University
stringclasses 19
values | Text
stringlengths 458
20.7k
|
|---|---|
Virginia Tech
|
The velocities for all the PVC pipe tests and concrete block tests were organized and
compared to identify if an overall consistent scratch velocity was maintained throughout the tests
that occurred with more than a foot in insertion. Most graphs seen in the experiments section that
did not suffer a massive dislodging of the scratch head, or a breakage of the tension wire will
show a region of constant velocity motion. The output of these velocities is available in Table 5-I
below:
Table 5-I: Average Velocity Values for PVC and Concrete Scratch Tests
Average
Concrete Average
Scratcher Block PVC Average
Test Diameter Velocity Velocity Velocity
Name (in.) (in./sec.) (in./sec.) (in./sec.)
1 Zone 1 0.045 0.54 0.54
1 Zone 2 0.045 0.49 0.49
2 Zone 1 0.045 0.43 0.43
2 Zone 2 0.045 0.6 0.6
3 Zone 1 0.045 1.36 1.36
3 Zone 2 0.045 0.66 0.66
1.1 0.045 0.5 0.5
1.2 0.045 0.6 0.6
2 0.045 0.768 0.768
3.1 0.051 0.44 0.44
3.2 0.051 0.52 0.52
Average 0.68 0.57 0.63
Standard
Deviation 0.31 0.11 0.25
The data indicate an average velocity of the movement of the scratcher head for the combined
PVC pipe and concrete block tests to be 0.63 in./sec. with a standard deviation for these values of
0.25 in./sec.. The movement in the concrete block had a higher average velocity (0.68 in./sec.)
than the average velocity in the PVC pipe (0.57 in./sec). Additionally the PVC pipe had a smaller
standard deviation of scratcher velocity, which means it had better velocity consistency than the
concrete block tests.
82
|
Virginia Tech
|
Force versus Displacement Analysis
Another useful type of plot can be developed for the collected information that portrays
how the force values change relative to the scratcher’s position in the hole. This makes the
buildup and release of stored energy from friction slippage on the rock become more evident.
This plot is generated in the “itchesplot” program and simply pairs the force and position values
for a given time value. A view of this plot for test two is seen in Figure 5.4 below, with a region
outlined in orange to be assessed further, the scratch direction, and a vertical red line denoting
the collar of the drilled hole:
Scratcher Head Movement Direction
Figure 5.4: Force vs. Displacement graph for Test One on Concrete Block with Orange Outline
of Study Area, Red Collar Line and Movement Direction Arrow
The outlined region corresponds to the constant velocity region from Figure 5.2. When
visualized up close, certain behaviors become evident. The highlighted region was selected for
an in-depth analysis because it is known that this area corresponds to constant velocity motion.
83
|
Virginia Tech
|
Sandstone Scratch Analysis
One final type of analysis was conducted on the data from the sandstone sample scratch
tests. Considering the only chosen change between the tests was the diameter of the scratchers, it
was deemed useful to consider the way that the forces may change between these tests. Some of
the original graphs for the sandstone tests can be seen in Figure 4.16, Figure 4.17, and Figure
4.18 where force and displacement are plotted against time, with the rest available in the
Appendix in Figures A.9 through A.20. An analysis of these images shows an increase in force
for the parts of the test where motion occurred followed by a flatline after the scratcher exited the
hole and hung until the test ceased. The mean force in the motion phase and was calculated and
organized in Table 5-1:
Table 5-II: Comparison of Force Averages for Different Scratch Diameters for Sandstone
Sample
Sandstone
Scratcher
Diameter Pull Force Averages (lbs) Overall Statistics
(Inches) Test 1 Test 2 Test 3 Test 4 Test 5 Test 6 Test 7 Average St. Dev.
0.045 8.28 6.19 5.34 4.48 n/a 6.59 6.55 6.24 1.29
0.051 12.75 7.76 10.39 4.509 5.44 11.63 4.87 8.65 3.49
0.055 6.49 15.19 12.22 9.17 6.71 n/a n/a 9.96 3.73
The magnitude of the force averages are visible in the blue column on the right. These results
indicate that increasing the scratch tip diameter will increase the force necessary to move it along
the rock surface. This is a result of the friction forces of the scratcher tips as they drag along the
rock surface and are subject to larger normal forces with increasing scratcher stiffness. As well
as the fact that if scratching occurred, the force would be applied over a wider area, which would
provide more resistance.
88
|
Virginia Tech
|
Conclusions
After conducting an analysis about the ability to quantify mine strata conditions, a
scratching device was developed to extract useful strength characteristics from legally mandated
one inch diameter test holes in mine roofs. The applied rock analysis technique is based off of
pre-existing methods of scratching core logs and determining their strength properties from
measured forces and cutting dimensions. A thorough review of mine strata conditions was
conducted in the literature review to identify what types of hazards exist in mine roofs as well as
how their geologic formation lead to some of the features that would be important for a strata
analysis device to target and identify.
Titled MRSAD, short for Mine Roof Strata Analysis Device, this rock analysis unit
consists of a mounting system and a series of instruments to gather properties from the rock. The
scratcher, pulled through the hole by a tension cable, has scratch tips contained within it that
interact with the rock to generate cutting forces. A winch attached to a stand winds this tension
cable and the generated forces are relayed by a custom-made tension transducer affixed in-line
with this cable. A position transducer, attached to the scratching head, travels the length of the
hole and relays position values to correspond to the force values. An installation protocol was
developed, as well as the modified pipe and special scratch tip retraction mechanism necessary
for installation.
The signal from the transducers connected into a data acquisition system that is paired
with a laptop computer. On the computer, a specially written LabVIEW program converts the
voltage signal into a spreadsheet. This spreadsheet can be imported into MATLAB where a
series of routines were written to aid in their processing. The results, when viewed as functions
of time and functions of position conveyed information about the motion of the scratch head in
89
|
Virginia Tech
|
the borehole. This device was first tested in a lab setting at a roof drill manufacturer facility
where a concrete block with various embedded strata was scratched to see if any changes in rock
could be observed. The data were too erratic to determine any changes in rock type along the
borehole. A second and third test were performed in a campus laboratory by conducting the same
scratch test in a PVC pipe and a sandstone block. This guaranteed proper anchorage and allowed
additional variables to be constrained such as scratcher diameter which were not controlled
during the tests at the roof drill facility.
Force measurements from the tension transducer relayed pulling forces in the sandstone
averaging 6.24 lbs, 8.65 lbs, and 9.96 lbs for the 0.045”, 0.051” and 0.055” diameter wire
scratchers, respectively. Additionally, tests performed on the PVC pipe show a similar increase
in applied force with increased scratcher diameter from 0.045” to 0.051” with a respective force
increase from 2.81 lbs. to 36.46 lbs. This indicates that increasing the scratcher wire diameter
increases the pull force and therefore, the force applied to the rock. Increasing the diameter of the
wire increases the stiffness of the scratcher and therefore increases the normal force exerted into
the surrounding media. Increasing wire diameter also reduced the deflection the scratchers may
undergo from scratch forces. The displacement transducer performed as expected and indicated
that the winch was able to control the velocity of the scratch head with an average of 0.63 in./sec.
velocity with a standard deviation of 0.25 in./sec.. The output from the graphs indicates a pattern
of sticking and slipping between the scratchers and the surrounding material as the head would
build up force and then travel a small amount. Although the numerical results didn’t provide an
ability to find rock strength values, it did indicate the promise of certain aspects of the device to
be used in later iterations of similar technology.
90
|
Virginia Tech
|
Future Work
In spite of the promise of several of the designed components, there were identified
aspects of this project where reevaluation could significantly simplify the system and increase
accuracy. The most glaring improvement would be to rework the force-sensing element to be
better suited for the forces that were encountered in this trial (many tests had valuable
information contained between 0-10 lbs but the load cell was designed for higher forces).
Additional efforts to improve filtering of the data as well as signal stability would also increase
precision and accuracy. A better configuration would be if the force sensor could be mounted to
the tension winch so that the forces would manifest themselves in the winch and the forces
applied could be derived from there and there bypass putting a force sensor in the hole.
Another improvement would be to integrate the position sensing properties of the
extensometer with the winch. Since the winch already has the cable in the hole, all the cable that
it pulls from the hole is moving at the same rate at the scratch head. With this knowledge, it
should be possible to configure a gear system to the pulley so that every N turns of the winch
will manifest themselves as a smaller number of turns on a potentiometer and therefore give a
position specific voltage output. This would bypass the expensive, delicate and cumbersome
extensometer with its accompanying cable and reduce the amount of things that need insertion
into, and could be damaged by, the borehole.
The scratch head also has areas of improvement. It is suggested here that the ideal means
of implementing the scratchers would be to have a head shaped much like the one from this
project, but have it sit within a sheath. This sheath would restrain the installed scratchers from
pushing into the rock mass and when placed in the correct position, the sheath could be slid off
the scratch head, with the scratcher arms springing into the surrounding strata. This would serve
91
|
Virginia Tech
|
several functions, as it would remove the amount of skill that the installer would have to have for
installing the device as well as increasing the stiffness of the scratchers that could be used. There
was no means for determining or controlling the contact force of the scratchers on the rock, a
method of doing that would provide other areas of improving scratcher design.
The use of thicker scratchers would increase the likelihood of gouging of the rock as the
PVC and sandstone experiment suggested that small changes in scratcher tip thickness can yield
higher pulling forces. A closer examination of the relationship between tolerance of the scratch
head and tolerance of the hole diameter would also likely provide more information about
circumstances that result in scratching versus simple friction sliding in the hole. Modifications to
the shapes of the scratching tips to better control their effect on the test should also be
considered. Lastly, a more portable (hopefully permissible), autonomous data acquisition system
that could bypass the cumbersome LabVIEW to Excel to MATLAB file transfer that had to
occur for this project would prove very convenient, especially if it could bypass a computer
altogether.
Several factors were identified that should be investigated further as sources of error in
the testing. The buildup of dust of certain amounts and compositions on the hole wall may
impact the effectiveness of the cutting system and its effect on this research is unknown. Also,
the presence of water in the rock may affect how the rock reacts to scratching. Controlled
dimensions of borehole diameter during scratching would be helpful for determining how out-of-
gauge hole profile affects scratcher behavior. Recommended safety practices include: keeping
only the most necessary people around the testing unit at any given time so that if an unlikely
failure occurs, there would be no risk to them, and only winding the spool from behind the
device so that nobody is in the way of the unit in the event of failure.
92
|
Virginia Tech
|
An Electrical Mine Monitoring System
Utilizing the IEC 61850 Standard
David Christopher Mazur
Abstract
Motor control assets are foundational elements in many industrial operations. In the
mining industry, these assets primarily consist of motor control centers and drives, which are
available with a comprehensive assortment of control and monitoring devices. Various
intelligent electronic devices (IEDs) are now used to prevent machine damage and downtime.
As motor control devices have advanced in technology, so too have the IEDs that protect them.
These advances have resulted in new standards, such as IEC 61850, that have embedded
intelligence and a standard set of communication schemes by which IEDs can share information
in a peer-to-peer or one-to-many fashion.
This dissertation investigated the steps involved in interfacing IEDs to a mining process
control network via the use of the IEC 61850 standard. As a result of this study, several key
technological advancements were made including the development of (i) vendor independent
system to communicate with IEDs in a mining environment over IEC 61850, (ii) command and
control methods for communication based assisted automation of IEDs for mining firms, (iii)
effective solutions to incorporate electrical distribution data in the process control system, (iv)
enhanced safety platforms through remote operation of IEDs, (v) standard visualization faceplate
graphics for HMI operators with enhanced security, and (vi) new methods for time stamped
dataflow to be correctly inserted into a process historian for “true” Sequence of Events Records.
|
Virginia Tech
|
Dedication
This work is dedicated to my family. Thank you for always being there and believing in
me. You have taught me over the years that I am capable of anything that I put my mind to.
This document is proof that anything is possible and that the North American engineering spirit
is truly alive and well. Thank you!
Ce travail est dédié à ma famille. Je vous remercie d'être toujours là et de croire en moi.
Vous m'avez appris au fil des ans que je suis capable de tout ce que je mets mon esprit à. Ce
document est la preuve que tout est possible et que l'Amérique du Nord d'ingénierie esprit est
bien vivant et bien. Je vous remercie!
Acknowledgements
I would like to take this moment in acknowledging those who have helped me along the
way in the development of this research. Special thanks to Mr. Erik Syme of ProSoft
Technologies for helping me along the way with the hardware and software development of the
gateway module. Thanks to Rob Entzminger of Rockwell Automation for helping me
understand the current problem with process historians and being there for guidance when
developing the historian event insertion tool. Thank you to Jim Elhert and Lee Ward of
Rockwell Automation for helping me obtain feedback from various mining and metals firms in
order to provide the community a meaningful and beneficial solution. Thanks to Blake Moret for
believing in this idea, by funding my research efforts.
Finally, special thanks to Joe Sottile and Tom Novak for agreeing to serve on my
committee and helping me develop this research into a beneficial result for the mining industry.
iii | P a ge
|
Virginia Tech
|
1. Chapter 1 - Introduction
Background
As new technology continues to drive innovative applications, the mining industry must
keep pace to remain competitive in an ever-changing marketplace. The trend requires replacing
outdated systems with high-performance, low-cost, option-rich devices that offer improved
flexibility while reducing operating costs. Many functions of new control systems are becoming
increasingly distributed to smart components capable of performing localized operations that
were once the responsibility of a central or master controller. The integration of intelligent
devices, device-level networks, and interrogation software into motor control centers
demonstrates improved diagnostics, permits early warnings for increased system reliability,
offers design flexibility, and provides for simplified wiring and an enhanced level of personnel
safety.
Energy consumption is an additional concern for mining operations. The ability to
integrate energy consumption and the electrical distribution network into the process control
system provides the mine process engineer many advantages. This dissertation focuses on
technology integration of intelligent devices via the IEC 61850 standard to traditional process
control networks utilizing programmable automation controllers (PACs) for both monitoring and
command and control of a mine electric distribution system.
Power system and industrial-based electrical protection devices have evolved over the
past three decades from electromechanical- to microprocessor-based relays. With the advent of
sophisticated microprocessor based relays, more advanced protection schemes that correctly
identify and clear faults in acceptable timeframes have been developed. As microprocessor-
based relay technology has evolved, so too have communication networks for electrical
protection. Electrical protection architectures have developed from hard-wired contacts to
communication networks using serial protocols such as Modbus. Serial communication has
evolved to communication protocols over the TCP/IP stack such as Modbus TCP and DNP3
LAN/WAN, which then allowed substation devices to communicate in a peer-to-peer manner to
share data.
The mining industry is the third largest North American vertical growth industry, behind
oil and gas, and power infrastructure. Multiple world market leaders are participating in the
1 | P a ge
|
Virginia Tech
|
power distribution system market place. Because of an increase in vendors, and the growing
complexity of interfaces for power system protection equipment, IEC 61850 was developed as a
global standard for substation communication. Now that technology in time synchronization,
protection, and fast-acting circuit breakers exists, a truly automated substation can now be
realized with the implementation of IEC 61850 across multiple equipment manufacturers.
The implementation of IEC 61850 also benefits automation and control companies as
they now can spend resources in developing SCADA systems. Additionally, other data
collection and historian systems that are inclusive of process and utility system can be controlled
by devices from multiple vendors. This will allow for better Human Machine Interfaces (HMIs)
as well as faceplates for operators to read and control relay operations in a substation
environment.
Problem Statement
The IEC 61850 standard provides a mechanism for interoperability, i.e., the ability for
information from Intelligent Electronic Devices (IEDs) from multiple manufacturers to be
exchanged with each other; however, the standard is limited to power system components, i.e.,
relays, meters, circuit breakers, etc. If this information could also be made available to the
control system, i.e., programmable logic controllers (PLCs) or programmable automation
controllers (PACs), the power system IEDs could be treated identically as control system
components, thus permitting a wide range of novel applications for these IEDs, e.g.,
optimization, demand-side load management, etc.
The electrical distribution and infrastructure systems of a mining operation are not very
well monitored. There needs to be an easier way to gather information to maximize process
yield and efficiency. Mining applications need to improve automation and control systems to
better understand the big picture of the overall operation. Current energy monitoring and
management systems are not vendor independent, thus making them difficult and expensive to
maintain. This dissertation solves all of these problems and presents a functional solution for
interfacing a mine distribution system to its process control system.
2 | P a ge
|
Virginia Tech
|
Scope of Work
Electrical SCADA (E-SCADA) platforms that interface with the process control system
of a mine were initially investigated. Research was conducted to evaluate the current E-SCADA
protocols and implementations to evaluate the pros and cons of traditional solutions. After the
evaluation of these traditional standards and protocols, the IEC 61850 standard and protocol suite
was evaluated for multi-vendor implementation of mining electric distribution systems.
After this initial research of current technologies, a novel solution was developed, tested,
and validated. The scope of this work was limited to the background research and construction
of a functional system to link a process control system to an electric distribution system.
This research produces a hardware and software solution that interfaces to the electrical
distribution system. A hardware gateway solution is defined and specified as well as software
developed to interface to this hardware solution. Visualization graphics are developed to allow
for an easy operator and engineer interface to various pieces of distribution equipment.
Additionally, a solution is specified and implemented to link traditional process historians with
E-SCADA systems for the benefit of long-term data storage and trending.
Thesis Format
This thesis is divided into nine chapters that describe the design and development of a
proposed interface to link an electric distribution system with a process control system via the
IEC 61850 standard. The following is an overview of the chapters follows.
Chapter 1 Introduction
This chapter defines the problem statement and provides an overview of the scope of
work that will be conducted and produced by the end of this document.
Chapter 2 Literature Review
Supervisory Control and Data Acquisition (SCADA) systems have been around since the
conception of automation and control systems. The first SCADA systems utilized data
acquisition by means of panels, meters, and lights. The operator manually exercised supervisory
control by adjusting control knobs. These devices still perform the supervisory control and data
acquisition of plants, factories, and power generating stations. As systems became more
3 | P a ge
|
Virginia Tech
|
distributed, SCADA systems played a larger role on process efficiency and yield. This chapter
discusses background information and evolution of traditional SCADA systems. Other topics
discussed include electrical distribution protocols and standards as well as time synchronization
standards.
Chapter 3 Conceptual Design
As previously described, the objective of this research is to design and develop a system
that links the protection and metering of the electrical distribution system at a mining operation
to a site-wide control system. The first steps in this process are to establish the system
capabilities, develop a preliminary conceptual design, and develop a working specification for
the system to achieve the research objective. This chapter discusses the development of the
conceptual design for the automation and control solution that was developed.
The design process of this solution begins with an initial conceptual design based upon
research, industry needs, and personal experience. In order to better fortify the design for an
industrial process network, the original conceptual design was presented to industry stakeholders
within the following disciplines: mining, metals, oil and gas, and engineering firms.
The original conceptual design was then compared against the engineering community
input in order to finalize the conceptual design presented at the end of the chapter. The chapter
begins with the original conceptual design followed by input provided by the engineering
community segments listed above. Subsequently, the final conceptual design is presented and
defined.
Chapter 4 Hardware Implementation
This chapter summarizes the hardware gateway module developed to implement the
conceptual design described in Chapter 3. A key component to this design is the multi-threaded
design to interface between the IED IEC 61850 network and the process EtherNet/IP network.
Each thread including the master control program is defined, and basic functionality is explained
in a functional block diagram. The various types of data that are used in this project are then
addressed followed by how data is packaged in packets to be sent to the process network via
EtherNet/IP.
4 | P a ge
|
Virginia Tech
|
The concept of the tag database is discussed, i.e., how it is utilized as a common space of
shared memory where both 61850 and EtherNet/IP drivers would read and write various tags.
The use of semaphore tags is also discussed in the database to avoid collisions in the tag
database. Finally, a practical example of this hardware module is used in a mine power system
application is provided.
Chapter 5 Software Implementation
In order for the hardware gateway to work properly with the process control system,
software needed to be developed that allowed for user interaction and configuration of the
gateway module. As described in Chapter 4, the configuration files are held by the Secure
Digital (SD) card. These files are downloaded to the SD card from the software tool that is
presented in this chapter. The procedure for how the hardware and software work together to
produce the final solution developed in this research is summarized in this chapter. A structured
example detailing the creation of a Configured IED Description, specifying the mapping of
information, and creating an Add-On instruction is explored in this chapter. Examples of
software are shown on the graphical user interface while additional XML code is provided to
show what files are created behind the scenes allowing the solution to function. Examples of
importing the Add-On instruction into code are shown as well as how the data are mapped as an
object into the controller data table.
In addition, a software tool was also developed to store information in the OSI PI
database on a precise timescale. The chapter discusses the traditional model of process
historians and how they do not adequately function for high speed SCADA and “smart” process
instrumentation. A new model is explored using off the shelf components to develop a bypass to
traditional historian data interface nodes. The tool creates a linkage between two databases, SQL
for alarms and events, and OSI PI via the OLE DB connector for historical data, thus allowing
for post processing of time critical data into the historian environment. The historian tool is also
discussed in this chapter.
Chapter 6 Visualization Benefits
This chapter defines the visualization developed for this research. Items discussed in this
chapter include: faceplate definition, human/machine interface discussion and definition, and
data management. The motivation for visualization is to represent an IED in logic and graphics
5 | P a ge
|
Virginia Tech
|
as close to the physical object itself. This provides the operator or engineer with the same look
and feel experience that they have with the physical device. At the same time, the graphics had
to be developed with various levels of security to allow only users with proper credentials access
to various command and control functions. Additionally, the research solution must adhere to
various graphics standards for both power and process control systems. This chapter discusses
the challenges in implementing a compliant visualization solution.
Chapter 7 Testing and Verification
This chapter discusses the testing and validation procedures conducted throughout this
research to validate and verify functionality of the proposed solution. As this research‟s goal
was to develop a mine monitoring solution utilizing the IEC 61850 standard to link the electrical
distribution system with process control system, it was determined that the best way to validate
the work and functionality of the proposed system would be to benchmark it against the
conceptual design. The chapter discusses each conceptual design milestone as well as the
procedure used to validate that this goal was met throughout the period of conducted research.
For milestones that required more than just visual inspection, the experimental setup is defined,
and testing procedure and experimental results are discussed.
Chapter 8 Practical Implementation
This chapter discusses the practical implementation of the mine monitoring system
presented in this dissertation.
Chapter 9 Conclusions and Future Work
This chapter summarizes the results of this research and discusses future work in this
area.
6 | P a ge
|
Virginia Tech
|
2. Chapter 2 - Literature Review
Introduction
Supervisory Control and Data Acquisition (SCADA) systems have been around since the
conception of automation and control systems. The first SCADA systems utilized data
acquisition by means of panels, meters, and lights. The operator manually exercised supervisory
control by adjusting control knobs. These devices still perform the supervisory control and data
acquisition of plants, factories, and power generating stations. As systems become more
distributed, the need and role of SCADA systems plays a larger role on process efficiency and
yield. The next few sections discuss the evolution of SCADA systems from simple sensor and
instrumentation panels to more complex systems and the justification for their use in mining
operations.
Mining Energy Usage
In June 2007, the U.S. Department of Energy (USDOE) conducted a study on energy
consumption in the mining industry. This study sampled 20 of the largest energy intensive
mineral extraction and processing facilities. The results of the study can be seen in Figures 2.1
and 2.2.
Figure 2.1 Mine Energy Usage
7 | P a ge
|
Virginia Tech
|
Figure 2.3 depicts energy usage by industry segment. It can be seen that mining, oil, and
gas extraction consume 32% of all industrial and commercial power usage in Canada [2]. Total
manufacturing, which consists of many subsectors as shown on the right of Figure 2.3, consumes
60% of all energy, but the largest individual energy consumer in the nation is the mining
industry. Figure 2.4 depicts the trend of energy usage. It can be seen that the mining industry
has shown continuous growth from 1990-2008, while other industries have declined. As
previously stated, much of this energy is consumed in crushing and grinding processes [2].
Figure 2.4 Energy Usage Per Segment
Taconite Mine Example
Taconite is an iron-bearing rock that is essential to the steel-making industry. Iron ore
mining and beneficiation are very energy intensive processes that require heavy crushing and
grinding procedures. To put this energy usage in perspective, in 2005 a Minnesota iron ore mine
used an average of 275 MW of power and is currently Minnesota‟s largest consumer of energy
[3]. The process begins with hauling taconite from an open pit mine to the central processing
facility. At the processing facility, raw taconite is delivered to the coarse ore crusher [4]. This
machine, which typically operates at 13.8 kV, reduces the mined rock into conveyor size
material. The coarse crusher is typically a 7850-horsepower synchronous machine with high-
torque, low-speed characteristics.
The output material of the coarse crusher is then conveyed to the fine ore crusher. The
belt conveyors are typically powered by dual 4160-V, 5000-hp induction motors, that continually
operate during production shifts [4]. These conveyors feed the fine crusher house, which
9 | P a ge
|
Virginia Tech
|
contains multiple mills that are also driven by 4.16-kV induction machines. The mills further
reduce the coarsely crushed material to 3 cm.
The fine material is then conveyed from the fine crusher house to the concentrator
building, which houses rod and ball mills for grinding. These mills are driven by 4.16-kV, 2600-
hp synchronous machines [4]. The output of these mills is a fine slurry mixture that is next sent
to magnetic separators.
The magnetic separators remove ferrous material from non-ferrous material. The non-
ferrous slurry material is pumped to the tailings pond, and the ferrous material is pumped in
slurry form to the agglomeration building [4]. This building contains vacuum discs and balling
drums to create taconite pellets. The pellets are then sent to a rotating kiln for drying and then to
storage for shipment. These processes are performed with low voltage equipment.
A typical one-line diagram of a processing facility is shown in Figure 2.5. These taconite
processing facilities are fed with 138 kV feeders from the utility.
From From
4.16KV BKR-A1 4.16KV BKR-A2
4 –3/C 500Kcmil 4 –3/C 500Kcmil
4160V MCC A1 4160V MCC A2
1200A
1200A
K N.C.
N.C.
MCC A1, 1200 AMP, 4.16 KV, 3 PHASE, 40 KA MCC A2, 1200 AMP, 4.16 KV, 3 PHASE, 40 KA
1200A 1200A
1200A 1200A 1200A 1200A N.O. N.O. 1200A 1200A 1200A 1200A
80E 80E 80E 24R 18R 18R 80E 80E
400A 400A 400A 400A 400A 400A 400A 400A
0 /2
#
C /3
0 /2
#
C /3
0 /1
#
C /3
0 /4
#
C /3
0 /1
#
C /3
0 /2
#
C /3
0 /2
#
C /3
0 /2
#
C /3
VFD VFD VFD VFD
0 /1
#
C
0 /4
#
C
0 /1
#
C
0 /2
#
C
To To /3 /3 /3 /3 To To
4160V-480V 4160V-480V 4160V-480V 4160V-480V
1000KVA 1000KVA 1250 2000 1500 1750 1000KVA 1000KVA
XFMR XFMR HP HP HP HP XFMR XFMR
Figure 2.5 Typical Taconite Mine Distribution One Line Diagram
10 | P a ge
|
Virginia Tech
|
From this point, every facility has two feeds of 13.8-kV in a main-tie-main configuration, each
with its own transformer. The 13.8-kV distribution voltage is then stepped down to 4.16-kV at
each building to feed individual loads. The 4.16-kV busses are also configurable in a secondary-
selective, main-tie-main scheme that can be implemented in case of bus failure. This
configuration is often preferred in large mining operations because it allows critical loads to be
transferred to healthy busses during times of an electrical system fault or insufficient spinning
reserve. Many relays are utilized for operating and protecting these facilities, including
instantaneous and time delay overcurrent (50/51 elements), differential (87 element), under-
voltage (27 element), and others.
Electrical protection has evolved over the past three decades from electromechanical to
microprocessor-based relays. Protection systems previously consisted of isolated, hard-wired
interlocked electro-mechanical and solid-state relays that had limited system visibility. These
devices only detected an electrical fault or overload and lacked a way to report additional
protection information. With the advent of microprocessor-based relays, more advanced
protection schemes, that correctly identify faults in acceptable timeframes, have been developed.
Communication schemes for providing electrical protection have also evolved with
microprocessor-based relays [5]. SCADA communications protocols and standards have been
developed for the purpose of collecting information from remote process locations for use in
central processing calculations. This dissertation proposes the use of the IEC 61850 standard to
serve as the backbone for communications for addressing the IED-interface issue. The IEC
61850 standard can also be used for controlling main-tie-main schemes through communications
rather than hardwire, in addition to its added SCADA benefits [6].
11 | P a ge
|
Virginia Tech
|
SCADA Evolution
Instrument Panels
Original SCADA systems consisted of sensor and instrumentation panels. These were
panels constructed of various discreet and analog I/O sensors that were hardwired back to a
reference point in a control panel [7]. A simple sensor panel can be seen in Figure 2.6.
Collection Point
Sensors
(Discreet or 4/20 mA)
Figure 2.6 Sensor Panel
Many process owners installed these sensor panels for the many advantages that they
offered, including simplicity with no CPU or programming software needed, sensors are directly
connected to meters, switches, and indicators, and the cost to construct these panels is cheap.
On the other hand, these panels had the following drawbacks:
The wire required to install a complex system could become unmanageable due to cable
quantity
The quantity and type of data received from the panel are rudimentary
Increasing capacity of the system becomes more difficult as the system expands
Reconfiguration becomes difficult as there is no way for data simulation using this
method
Data storage is minimal
12 | P a ge
|
Virginia Tech
|
There is not remote monitoring of data and alarms; thus requiring a person to be present
to monitor the system.
As a result of these drawbacks, the industry eventually progressed to the use of telemetry to
interconnect areas of a process network [8].
PLC / DCS Systems
Modern SCADA systems often utilize telemetry to move data across long distances and
connect remote nodes to their systems [7]. For example, in today‟s society, many heavy
manufacturing and industrial processes, such as metals, mining, utilities, and security, need to
connect equipment and systems often separated by large distances. These distances can range
from a few feet to hundreds of miles. Via the use of telemetry, process owners and control
engineers can send commands and programs and receive monitoring information over the
network from various locations [9].
A modern definition of SCADA is the combination of telemetry and data acquisition.
Furthermore, SCADA can be defined as the collecting of information, the process of transferring
this information to a central site, performing calculations to take corrective control action of a
process, and enabling actuators to take the corrective control action [7]. SCADA has also
evolved to displaying and trending data on operator and engineering workstations over the past
few decades.
At its conception, SCADA systems used relay logic to control production and plant
processes. With the evolution of CPUs and electronic devices, manufacturers, such as Modicon,
Allen-Bradley, and Siemens, began to implement this technology into relay logic equipment.
The result of this technology integration birthed/created the Programmable Logic Controller
(PLC). The PLC is still one of the most widely used control systems in industry today [10].
Instead of hardwiring a relay into a control panel, a sensor was simply wired back to an I/O
point. Computer logic, known as ladder logic due to its similar form of a relay ladder diagram,
was written to control the state of these devices [7]. As the need to monitor and control field
devices grew, PLCs were distributed, and control systems began to become more intelligent and
smaller in size. An example of a PLC or Distributed Control System (DCS) can be seen in
Figure 2.7.
13 | P a ge
|
Virginia Tech
|
Fieldbus
PLC
Sensors
(Discreet or 4/20 mA)
Figure 2.7 PLC System
This system had many advantages including the computer‟s ability to store and records a
large amount of data, the user‟s ability to customize data, connect sensors over a wide
geographic range, real time simulations for operators, and access to data from remote locations.
While DCS systems have many advantages, the system is much more complex than simple
panels; it requires programming knowledge to configure the system, there are wiring concerns
from the sensors to the PLCs, and the operator only has vision to as far as the PLC in the
network [7].
IED Systems over Fieldbus
As manufacturers and process owners began to push the envelope of SCADA and process
control systems, the need for smaller and smarter systems grew, and as a result, sensors were
designed with the intelligence of PLCs [7]. These new intelligent sensors became known as
Intelligent Electronic Devices (IEDs). As process engineers began to design systems from
process and instrumentation diagrams (P&IDs), a demand grew for these intelligent electronic
devices to perform both discreet and analog control, including analog output and PID control [8].
The IEDs were located on a fieldbus network, such as Profibus, DeviceNet, or Foundation
Fieldbus, where information could be exchanged from IEDs to PCs and other enterprise systems.
IEDs in and of themselves have enough intelligence to acquire data, communicate in a peer to
peer fashion, and contribute to the overall control system. These IEDs typically contained
14 | P a ge
|
Virginia Tech
|
multiple sensors for analogs and discreet inputs. An example of an IED system over a Fieldbus
network can be seen in Figure 2.8.
Fieldbus
IEDs
Figure 2.8 IED and Fieldbus SCADA System
Advantages of this system include the minimal need for wire, operator visibility to the sensor
level, the inclusion of sensor firmware information with received data , the plug and play ability
of devices for simple installation and replacement, and smaller physical footprint. The
disadvantages of this system include: the cost of training required for operators, higher sensor
prices, and IEDs that rely on a communication scheme for data transfer [7].
PLCs to PACs
Over the past 15 years PLCs have evolved to Programmable Automation Controllers.
These platforms are a rack based automation controller that accepts multiple modules for various
configurations. PACs were chosen because of its asynchronous capabilities. Most PLCs are
synchronous in nature in that they execute the following procedure: read inputs, execute a
program, and update output registers [11].
Asynchronous machines utilize two different sets of memory, program and I/O. These
two memory banks are in separate locations and can be written to independently. This means that
I/O information can get collected separately from the executing instructions in program memory.
I/O memory is populated via the use of a backplane circuit and a Requested Packet Interval
15 | P a ge
|
Virginia Tech
|
(RPI). Multiple data words can be simultaneously shared between modules connected to the
backplane as well as written to I/O mapped memory locations [10]. PAC processors are capable
of true multitasking, thus allowing the system to asynchronously gather information and provide
precise time tags while normally scheduled programs continue to operate. Many process owners
today are replacing PLC SCADA systems with PAC systems that work in conjunction with IEDs
on the same fieldbus for monitoring and process control.
SCADA Terminology
A SCADA system can be divided into two categories, hardware and software. It is
important that these two categories function equally in order for the SCADA system itself to be
self-sustaining and provide proper control for the process system.
SCADA Hardware
A traditional SCADA system consists of a number of remote terminal units (RTUs) that
collect field data and send it to a master station via a communication system. The master station
then displays the acquired data and allows the controls engineer to make process and control
decisions for control tasks. This acquisition of data in a timely fashion allows for process
optimization [7]. Other SCADA benefits include more efficient, reliable, and, most importantly,
safety operations. In all, this will result in lower cost of operation compared to non-automated
systems [8].
Complex SCADA systems consist of a five level hierarchy:
1. Field Level Instrumentation and Control Devices
2. Marshaling terminals and RTUs
3. Communication Systems
4. Master Station
5. Commercial data processing department computer system
The RTU provides the SCADA system interface to the field analog and digital sensors situated
on every remote node. The communications system provides the connection between master and
remote sites. This communication system can be defined by a number of media including: wire,
fiber optic, radio, telephone line, microwave, or even satellite. Specific SCADA protocols and
error detection philosophies are used for efficient and optimum data transfer between nodes [7].
16 | P a ge
|
Virginia Tech
|
Master Stations gather information from various RTUs and generally provide an operator
interface for information displays and control of remote nodes.
SCADA Software
There are two types of SCADA software: proprietary and open. Manufacturers engineer
proprietary software to communicate with their hardware. These are typically seen in “turnkey”
solutions where all engineering, integration, and startup are provided from a single vendor or
encompass group. There is one main failure with these SCADA systems: the process owner
heavily depends on the supplier of the system [7]. As a result, many process owners are
supporting open SCADA systems. Due to the interoperability these systems bring to the system,
multiple vendor manufacturer‟s equipment can be used in the same system.
Major vendors of SCADA system software include Citect and WonderWare, but there are
many manufacturers of SCADA software. Many packages today actually include asset
management in order to provide process owners more information about their systems. Key
software features for SCADA system software include user interfaces, graphical displays,
alarms, trends, RTU interface, scalability, access to data, database, networking, fault tolerance
and redundancy, and client/server distributed processing[8].
SCADA Protocols
This section discusses SCADA protocols. Since this dissertation addresses the electrical
distribution system in mines, this section focuses on the three main electrical SCADA protocols:
Modbus TCP, DNP3 LAN/WAN, IEC 60870-5-103, and IEC 61850. To be fair, there are many
electrical distribution protocols that can be implemented in any automation and control system.
Figure 2.9 depicts a listing of various electrical distribution SCADA protocols between the bay
and station level of a traditional electrical distribution substation.
In general, an electrical distribution substation can be broken into three levels of control:
station level, bay level, and process level. The process level of a substation houses the
instrumentation used to collect data, i.e. potential and current transformers. The bay level
consists of IEDs such as relays and meters that collect the process level data and make a
distributed control decision. In this case, control decisions may be network configuration,
energy usage, or load shedding, to name a few. Finally, the station level of a substation consists
17 | P a ge
|
Virginia Tech
|
of Human Machine Interfaces (HMIs), central supervisory control functions, and a station
gateway to send gathered status and data to further upstream substations or master controllers
[8]. Traditionally, this link is usually wireless or fiber optic.
Figure 2.9 Typical Electrical SCADA System
Modbus TCP
Modbus is a master-slave communication model that was developed by Modicon in 1979.
This communication protocol is well established with over seven million nodes in North
America and Europe alone. This protocol is very simple to implement and commission.
Although this protocol is very easy to implement, there are many drawbacks to this method.
Modbus TCP is very inefficient at managing data and network bandwidth. The protocol only
uses simple data types, such as integer and Boolean. Finally, it only can send static data.
Despite these drawbacks, Modbus is still the de facto standard in a multi-vendor integration [12].
Additional limitations to Modbus include no time stamp of data values for Sequence of Events
(SOE) Applications, no indication of a disturbance event, the need for the master to always ask
slave device for data, the inability of the slave to initiate communications to master, and the lack
of common data formats between devices. As a result, these limitations made it critical for the
power industry to find a communication protocol that provided data that more accurately
represented the events that occurred in a power application.
18 | P a ge
|
Virginia Tech
|
IEC 60870-5-101/103/104
As more services were being performed by IEDs in a substation, more intelligence was
required in the protocol that linked these devices. This need, combined with the limitations of
existing communication protocols, such as Modbus, led to the formation of new protocol
standards. In 1990, IEC created the IEC 870-5-1. This standard was then spun off into two
separate standards: IEC 60870-5-101 in 1995 and DNP 3.0 in 1993. This creation can be seen in
Figure 2.10.
Figure 2.10 Evolution of Electrical SCADA Protocols
Industrial communications protocols, such as IEC 60870-5 and DNP 3.0, provided unique
attributes that made data acquisition, data reporting, and overall communications much more
efficient and provided greater levels of detail to the devices in the network. DNP 3.0 is dominant
in North America, Latin America, South Africa, and Australia and is primarily used in the
water/wastewater, and oil and gas segments [13]. The IEC-60870-5 protocol suite is primarily
used in Europe, the Middle East, and Asia Pacific regions. Both protocols provide similar
application functionality and are popular in electrical SCADA applications for communication of
IEDs [14].
The IEC 60870-5-101/103/104 standards are used for power system monitoring and
control in the following: 101 defines serial communication, 103 defines protection relay data
formats, and 104 defines the Ethernet implementation of the protocol. IEC 60870-5-101 is the
first protocol to define a hierarchy of message structure. The protocol breaks data into two
formats, high priority messages (Class 1), and low priority messages (Class 2), and transfers data
19 | P a ge
|
Virginia Tech
|
using separate mechanisms. The 101 section of the standard also defines cyclic and spontaneous
updating schemes as well as the facility for time synchronization [15].
The 103 section of the standard defines power system control and associated
communications. It defines a mechanism that enables interoperability between protection
equipment and devices of a control system in a substation. The mechanism defines either two
methods of data transfer: the use of specified application service data units (ASDUs) or the use
of generic services for transmission of all possible information. This standard supports specific
protection functions and provides the vendor a facility to incorporate its own protective functions
on private data ranges [16].
The standard supports two modes of data transfer. The first mode of data transfer is
unbalanced which is defined by a master initiated message. The alternative to unbalanced
communication is balanced which is defined by a master/slave-initiated message. As previously
mentioned, data is classified into different information objects, each with its own specific
address. The process owner must then classify data into Class 1 and Class 2 data in order to
define transfer mechanisms [8].
The IEEE 802.3 Ethernet portion of the standard (104) classifies data into sixteen
separate data groups that can be individually interrogated. It allows for cyclic as well as
spontaneous data updating of data tags in addition to allowing for the benefit of time
synchronization [8]. These improvements from Modbus allowed for better data transfer with
more meaning to each data point. The ability to timestamp values, group non-similar data points
into custom groups, and interrogate data by two levels of priority provided large improvements
over traditional Modbus schemes [15]. The first implementations of DNP 3.0 developed from
the IEC 60870-5 specifications.
DNP3 LAN/WAN
DNP 3.0 was originally developed as a serial communications protocol for electrical
SCADA applications. This section will refer to the Ethernet based form of this protocol known
as DNP3 LAN/WAN. This protocol is primarily used for communications between a master
station and IEDs or RTUs [8]. This model can be seen in Figure 2.11.
20 | P a ge
|
Virginia Tech
|
Figure 2.13 Multi-Level DNP3.0 Network
In the multi-level drop configuration, there is one ultimate master within the system
(represented by black). All of the remaining nodes within the system are slaves to this master.
In addition, each level of the network designates a master to the lower levels of the network
(represented by gray). This device serves as a data concentrator to the devices below it to gather
data and push it to higher levels of the SCADA system [17].
DNP3 LAN/WAN builds upon IEC 60870-5 in that it supports complex data types,
including integers, double integers, reals, Booleans, dual point binaries, and counters with
quality flags. The protocol also allows for the time stamping of records based on event change
of state as well as analog deadband. The DNP3 LAN/WAN protocol also allows for unsolicited
reports by exception. Data is grouped into two types, dynamic and static. All dynamic data is
grouped into three levels of priority, Class 1, Class 2, and Class 3, while all static data is
referenced to Class 0 [8].
Process owners have adopted DNP3 LAN/WAN because it provides standardization and
interoperability. It is an open protocol and is optimized for SCADA communications. DNP3
provides interoperability between various vendors equipment and is supported by a substantial
number of SCADA equipment manufacturers [8]. The protocol is not static or unchanging; it
allows for extensions. Users can define complete data structures to pass industry specific
information in complete context [17].
22 | P a ge
|
Virginia Tech
|
Today industry is moving away from IEC 60870 and DNP 3.0 in the adoption of a new SCADA
protocol standard known as IEC 61850. This migration is occurring for a multitude of reasons
including: the advantages of high speed Ethernet, the high cost for data management, the loss of
information and functions during mapping of data, and the lack of consistency between vendor
implementations of similar devices [16].
IEC 61850
Introduction
Power system and industrial-based electrical protection has evolved over the past three
decades from electromechanical to microprocessor-based relays. With the advent of
microprocessor-based relays, more advanced protections schemes that correctly identify faults in
acceptable timeframes have been developed. As microprocessor-based relay technology has
evolved, so too have communication networks for electrical protection [5]. Electrical protection
architectures have developed from hard-wired contacts to communication networks over serial
communications such as Modbus. Serial communication has evolved to communication schemes
over the TCP/IP stack such as Modbus TCP and DNP3 LAN/WAN, which now allows
substations devices to communicate in a peer-to-peer fashion to share data.
With power systems being the second largest North American vertical growth industry
behind oil and gas, world market leaders are participating in the power system market place.
Furthermore, the mining industry is the largest consumer of electric power in the world [18].
Because of an increase in vendors and growing complexity of interfaces for power system
protection equipment, IEC 61850 was developed as a global standard for substation
communications. Now that technology in time synchronization, protection, and fast acting
circuit breakers currently exists, a truly automated substation can be realized with the
implementation of IEC 61850 across multiple equipment manufacturers [19].
The implementation of IEC 61850 also benefits automation and control companies as
they now can spend resources developing SCADA systems, as well as other data collection and
historian systems, that interface to multiple devices over one common protocol. This allows for
better Human Machine Interfaces (HMIs) as well as faceplates for operators to read and control
relay operations in a substation environment.
23 | P a ge
|
Virginia Tech
|
History
IEC 61850, “Communication Networks and Systems in Substations,” is an international
standard developed by Technical Committee 57 that focuses on substation automation [20]. This
standard was drafted and adopted in an effort to unify substation equipment and communications
on one common platform regardless of manufacturer. IEC 61850 is divided into ten sub-sections
ranging from 61850-1 to 61850-10 that define protocol terminology, communication
mechanisms, and conformance testing [20-35].
IEC 61850 was a combined international effort merging the Utility Communication
Architecture 2.0 (UCA), an EPRI project, with IEC standard 60870-5-101, -103, and -104 to
create one international standard for substation automation. UCA defined many of the protocols,
data models, and abstract service definitions, while IEC 60870-5 defined the basic
communication profile for sending control messages between two systems [36].
There are multiple benefits to the IEC 61850 standard. These benefits include the
support of comprehensive substation functions, ease of design, specification, setup, and
maintenance, strong functional support for substation communication, and its flexibility to
support system evolution [36]. Additionally, the standard was structured in such a way as to
accommodate current technology.
Protocol Definitions
IEC 61850 was designed as an object-oriented protocol in that data would be defined as
objects, and each object would have multiple attributes. This is very similar to protocols of high
level programming languages. Each physical Intelligent Electronic Device (IED) would be
accessed by a network address. The physical device would then be represented by a logical
device that contained all relevant, non-distributed, logical nodes (i.e. functions in the real
device). Each logical node would then contain data and data attributes which would describe
operations, positions, etc. of the IED. The data and data attributes are defined as dedicated data
values that are structured and well-defined. These values are exchanged according to defined
rules and communication mechanisms [24]. The way that this information is shared across the
network is defined by communications and mapping portions of the 61850 standard [24].
The traditional grid has evolved from analog electro-mechanical relays with one-way
communication that was human monitored to digital electronic relays with two-way
24 | P a ge
|
Virginia Tech
|
communication and managed by advanced control schemes. As previously described, with the
advent of digital IEDs, communication schemes became more advanced than the traditional
hardwired contact. Communication protocols such as Modbus TCP and DNP3 LAN/WAN were
developed by the power industry specifically for advanced control schemes. With the
development of the IEC 61850 standard, three main forms of communications are defined:
GOOSE, MMS, and SMV.
The IEC 61850 protocol is based on the Generic Substation Event Model (GSE). This
model is a fast and reliable system-wide distribution of input and output data values [36]. The
data transfer is based upon a publisher-subscriber mechanism. Additionally, simultaneous
delivery of the same generic substation event information to more than one IED can be achieved
via the use of multicast services. GSE describes the general structure for an event, and the
mechanism for transferring information between devices and locations are defined by GOOSE
and MMS.
The Generic Object-Oriented Substation Event (GOOSE) is one way to transmit
information under the IEC 61850 protocol. GOOSE is a single message sent by an IED, but it
can be received by multiple targets in a peer-to-peer fashion [32, 36]. It is used for fast, reliable
transmission of substation events such as alarms, commands, and indicators. In general GOOSE
is used for high speed, high priority applications such as protection interlocking and events when
event information needs to be delivered in a peer-to-peer fashion.
The Manufacturing Message Specification (MMS) is an alternative to transmitting
messages over GOOSE that is generally preferred for SCADA or non-time critical time data
acquisition. MMS is defined by ISO standard 9506 and is widely used in control networks. The
standard defines a reduced OSI stack with TCP/IP capability and Ethernet or RS232C as physical
media [37]. MMS defines communication messages transferred between controllers as well as
between engineering stations and controllers. Each IEC 61850 object is mapped to a
corresponding MMS object, and each IEC 61850 service is also mapped to a corresponding
MMS operation. All but GOOSE messages and the exchange of sampled values are mapped to
the MMS protocol stack, with the exceptions of time synchronization and file transfer. The
application described within this paper uses MMS to transmit report information from IED to an
automation controller.
25 | P a ge
|
Virginia Tech
|
The IEC 61850 object model is based on commonly used protection, control, and metering
functions in electrical substations. As shown in Figure 2.14, each physical IED can include one
or more IEC 61850 logical devices. Within each logical device, the manufacturer enables logical
nodes to represent the primary functions of the device. Each logical node encapsulates a
collection of data tags associated with the function in question. For example, in a feeder relay
commonly used for industrial power systems, the manufacturer defined logical devices for
protection, metering, control, and annunciation. Contained within the protection logical device,
are a number of logical nodes related to the operation of time overcurrent elements. Specifically,
in metering, the logical nodes are related to the variety of metering values in the relay [38].
Figure 2.14 Virtual IED
Most engineers are accustomed to SCADA protocols that reference tags by address or
index; one of the desirable attributes of IEC 61850 is that tags are referenced by a structured
name. As an example, A-phase electrical current is represented in a Polyphase Measurement
Unit (MMXU) logical node as MMXU$A$phsA$cVal. Table 2.1 describes the components of
this tag name. Additionally, each IEC 61850 enabled IED can self-describe all of the logical
devices and logical nodes it contains. A number of software manufacturers have IEC 61850
browsers that query devices on the network and display available data from each IED. Although
engineers commonly configure network communications using an offline Substation
Configuration Language (SCL) tool, which will be discussed later, self-description is a very
useful function for evaluating or verifying the configuration of a particular IED [39].
26 | P a ge
|
Virginia Tech
|
Table 2.1IEC 61850 Descriptor Example
Descriptor Component Type Description
MMXU Logical Node Polyphase measurement unit
A Data Object Phase-to-ground amperes
phsA Sub-Data Object Phase A
cVal Data Attribute Complex Value
The authors of the IEC 61850 standard defined many types of standard logical nodes that
are used by many manufacturers of protective relays. The common tag naming and structure of
the logical nodes help to simplify the integration of an HMI and other systems with power
system protection equipment. At the same time, it is important to understand limits to the
standard.
First, the standard does not dictate which logical nodes are implemented in a given relay.
Also, there is no way to define how the internal memory values of a given relay are mapped into
IEC 61850 tags. Each manufacturer chooses which logical nodes to implement and how
manufacturers or nodes associate their relay's functions to each logical node. Additionally, the
standard allows for custom logical nodes that can be created by any manufacturer. During system
design, the integration engineer needs to verify the existence and location of needed data from a
target relay.
Substation Configuration Language
In order to have a common method for describing and documenting the communications
network, IEC 61850-6-1 defines various SCL file types based on XML schemas. The specified
file types are System Specification Description (SSD), Substation Configuration Description
(SCD), IED Capability Description (ICD), and Configured IED Description (CID). This
dissertation only focuses on ICD and CID files. ICD files represent the default IEC 61850
configuration of an IED. CID files follow the same schema but represent the final configuration
of an IED in service.
27 | P a ge
|
Virginia Tech
|
If an ICD file were opened in a web browser or text editor, one would find the definition of all
logical devices and logical nodes for the IED. In addition, the ICD file can include definitions for
datasets, MMS reports, and GOOSE messages. Datasets are simply logical collections of tags
(not necessarily from the same logical node). Collecting the tags into datasets allows those tags
to be efficiently defined as part of a GOOSE message or MMS report.
Reports are unsolicited methods of sending datasets from an IED. The standard defines
two types, buffered and unbuffered. When using buffered reports, the IED keeps track of a client
message receipt so that any missed reports can be re-sent if there is a network problem.
Unbuffered reports do not fill in the gap if the link is lost. The MMS protocol can also provide
datasets via direct polling by the client. This method provides no buffering and eliminates
deadbands on analog datasets.
Automation Limitations
Even though SCL files define how an IED communicates on an IEC 61850 network,
many people do not realize that these files do not include configuration information for the
protection and control functions in a relay. Each manufacturer has software and proprietary
methods for enabling and configuring various protection elements and control strategies. The
IEC 61850 standard also provides no standard method for designing communication-assisted
automation [39].
Time Synchronization
Importance of Time in Applications
Everything in today‟s society is synchronized to some form of clock or source of time,
and it is important that in coordinated processes that these clocks be synchronized to provide the
most efficient process yield possible. Take the example of having a local controller with a
remote I/O device. Currently, measurements are typically taken by a controller-based I/O to
ensure that accurate timestamp information of events is recorded. Due to advancements in time
synchronization, this application can now be addressed by utilizing distributed I/O [40].
A sensor now can be connected to a remote I/O device. If each device contains a clock,
and they are time synchronized, discrete transitions from off-to-on and on-to-off can be captured
and timestamped as an input event. This timestamp now has meaning since both clocks are now
28 | P a ge
|
Virginia Tech
|
synchronized to each other, and this information can be sent across the network and used by the
controller or any other enterprise level process. Since the clocks are synchronized, the controller
can use the information to record sequences of events or timestamped data logs [41].
As a result, more than just traditional data can be sent across a time synchronized
network. In addition to timestamped inputs, scheduled outputs and synchronized actuation can
be achieved over time-synchronized networks. Four major areas of improvement can be realized
with the implementation of time-synchronized networks: sequence of events recording,
timestamped data logging, coordinated/synchronized operation, and motion control.
For power distribution applications, time synchronization plays a major role in everyday
process administration. Examples of time synchronization can be seen in Sequence of Events
recording (SER), protection schemes, power quality measurements, and data acquisition.
Microprocessor relays typically have the capability to synchronize to a common reference time
source such as GPS through IRIG-B [42]. Relatively newer technologies, including wide-area
measurement systems using synchronized phasor measurement units (PMUs), apply a GPS time
stamp to real-time data frames to capture a snapshot of the power system. More traditionally,
remote terminal units (RTUs) within a substation record events from devices throughout the
station and report statuses to a SCADA (Supervisory Control and Data Acquisition) Master,
commonly at a control center [9]. Although SCADA is not a real-time application due to its
synchronous design, a valid timestamp of events from substation equipment via either an RTU or
other intelligent electronic device is beneficial for both online and offline applications.
Mine applications are no exception to this need for time synchronization for both
electrical distribution systems and processes themselves. Most mines or processing facilities use
some form of backup or co-generation in order to power their processes. As a result, cascading
failures can occur that would cause massive system and process shutdowns, costing process
owners money for every moment they are not in production. Tracing the root failure of these
electromechanical systems can take a large amount of time and manpower. By having
synchronized SCADA systems, the downtime to locate the problem and take corrective action
becomes greatly shortened.
29 | P a ge
|
Virginia Tech
|
SCADA Timing Protocols
A point was made in each SCADA protocol definition to discuss the time
synchronization capabilities of each protocol. The timing protocols that will be discussed in this
section include GPS, IRIG, NTP, and PTP. Each protocol is compared to and contrasted with its
counterparts based on electrical SCADA conformance.
Global Positioning System
Global Positioning System (GPS) is not only a navigation system; it has evolved to
become the world‟s primary means of distributing precise time and frequency. GPS was
developed by the United States Department of Defense (DoD) in 1978 and is still maintained by
the organization today. Currently, 31 satellites orbit the earth and provide accurate time within
+/- 10 ns to GPS receivers [41]. The US DoD dictates the accuracy of the system and has the
right to limit this accuracy. Additionally, GPS systems can be costly to deploy over a distributed
I/O network since a GPS receiver is required at each node, and the antenna must have a clear
view of the sky to attain a locked signal.
Inter-Range Instrumentation Group Time Codes
In 1952, the commanders of the U.S. guided missile test ranges formed the Inter-Range
Instrumentation (IRIG) group as part of the Range Commanders Council (RCC) of the U.S.
Army in order to share information about range instrumentation [43]. Today, the steering
committee and ten technical working groups, including the Telecommunications and Timing
Group (TTG), control the IRIG time code standards. The IRIG standard was last updated in
September, 2004, and is titled IRIG Serial Time Code Formats [42]. The most well-known and
utilized code format is the IRIG-B time code from this standard. Despite multiple choices of
IRIG code formats, IRIG-B is the mostly widely used in the power industry. In fact, more than
ninety percent of substations use the IRIG-B format, which is typically better than +/- 1 ms [44].
There exist two different forms of the IRIG-B time code: modulated and un-modulated.
Modulated IRIG-B is transmitted on a carrier frequency sinusoid (similar to AM radio) and as a
result, must be demodulated at the receiving end in order to interpret the time data. With
advancements in phase locked loop (PLL) technology, modulated IRIG-B schemes can obtain
accuracies to within +/- 10 µs [45]. Installations with modulated IRIG-B must be isolated
30 | P a ge
|
Virginia Tech
|
through some form of transformer in order to prevent ground loops and signal degradation.
Unmodulated IRIG-B, or level shifted IRIG, is an alternative to the traditional modulated
approach. This form of time code transmission uses digital level shifting to achieve data
transfer. As no demodulation is necessary at the receiving end, time synchronization accuracy
improves to +/- 1µs [45].
Network Time Protocol
Network time protocol (NTP) is a protocol for synchronizing computer clocks using a
data network, such as the intranet or a wide area network (WAN). This protocol was developed
at the University of Delaware in 1980 and is the first protocol to address time synchronization
over variable latency packet switched networks [41]. This protocol provides accuracies that
depend on the setup of the network between each device and the performance of the computers‟
operating systems. Ideally, the connections in the network should be as short as possible, but
this protocol does include methods to estimate and account for round-trip path delay. Overall,
the accuracy of this protocol is in the low tens of milliseconds over wide area networks (WANs)
and better than a millisecond over local area networks (LANs) [46].
Precision Time Protocol
The formal title of IEEE 1588 is Precision Clock Synchronization for Networked
Measurement and Control Systems. This protocol is better known as Precision Time Protocol or
PTP. The standard specifies a protocol to synchronize independent clocks operating on separate
nodes of a distributed measurement and control network to a high degree of accuracy and
precision [47]. PTP is often used to synchronize distributed I/O devices over variable latency
packet switched networks such as Ethernet. This protocol was originally released in 2002
defining many of the regulations and specifications, and in 2008 it was revised to its current
version and released as IEEE 1588v2. The original version of this standard defined ordinary and
boundary clocks while version 2 defines transparent and hybrid clocks; all of these clock types
are addressed in this section [19].
In its simplest form, PTP was intended to be an administration-free protocol. The clocks
within a network communicate with each other across a network media and establish multiple
master-slave relationships. These master-slave relationships form what is known as a hierarchy
31 | P a ge
|
Virginia Tech
|
of clocks [47]. The overall intent of PTP is that distributed I/O devices manage the
synchronization of clocks automatically, thus requiring little if any network administrator input.
Not all devices within a network require the same level of time synchronization accuracy; as a
result, PTP allows for the support of a wide spectrum of clock accuracies to support the needs of
the end device or process [48]. For example, when dealing with protective relays, accuracy
within a millisecond is acceptable due to mechanical component tolerances. PTP can be
configured to meet the needs of both of these applications. Additionally, PTP can use multiple
sources of time as an ultimate time reference including, but not limited to, Global Positioning
System (GPS), IRIG, Network Time Protocol (NTP), or another PTP clock.
A system or network of clocks by definition consist of one or more devices capable of
becoming a master clock, while other devices within the network serve as slave clocks [47].
Normally one master clock is designated as the grandmaster clock. Figure 2.15 depicts the
combinations of clocks that PTP supports: ordinary, boundary, transparent, and hybrid.
Ordinary clocks consist of a single connection port, which industry commonly refers to as a PTP
port. This port can either be assigned as a master or a slave [41]. Examples of ordinary clocks
include GPS receivers and logic controllers and are typically located at the end nodes of a
network [49]. Boundary clocks contain multiple PTP ports that establish separate PTP domains
by segmenting the synchronization path between master and slave clocks. As their name
implies, boundary clocks form boundaries between PTP synchronization segments [47]. These
clocks are typically found in network switches. Transparent clocks are very different from
ordinary and boundary clocks in that they help compensate for the propagation delay through the
network rather than segment the network [41]. Finally, hybrid clocks are defined as a
combination of PTP type clocks in a device. The most common type of hybrid clock is a
transparent clock paired with an ordinary clock. Hybrid clocks are usually found within motion
devices and are used to perform synchronized actuation [41].
It was previously stated that PTP was designed to be an administration-free protocol in
that devices negotiate amongst themselves to determine a hierarchy of clocks. The algorithm
used to determine the hierarchy of these clocks is known as the Best Master Clock Algorithm
(BMCA). The BMCA, as defined by IEEE 1588, is the strict arbitration process employed to
determine the status of each network node, either master or slave [47]. As its name implies, the
32 | P a ge
|
Virginia Tech
|
BMCA determines the best master clock and names it the grandmaster of the PTP system. All
remaining clocks within the system are ultimately synchronized to the grandmaster.
Figure 2.15 System of Clocks
Announce messages are sent approximately every two seconds for any PTP device
claiming to be a master. These announce messages contain information about how well the clock
attributes compare to a scale [40]. When a node receives an announce message from another
device it compares the credentials that it receives to its own. The better of the two clocks the
serves as the master, while the lesser acts as the slave. This process continues until the status for
every clock within the network is determined.
The BMCA utilizes four criteria to determine the better of two clocks. These factors
include clock class, accuracy, variance, and priority. Clock class defines the relative measure of
clock quality. Accuracy defines how close the clock meters time to an absolute reference.
Variance is the measure of the clock‟s stability. Priority is a manual override that can be
established if a network administrator wants one clock to serve as grandmaster over another [47].
One main advantage to implementing PTP in a system is that it will dynamically update
to topology changes [41]. For example, if the current grandmaster is removed from the system,
the BMCA will attempt to designate a new grandmaster from the remaining clocks. On the other
hand, if a clock with better credentials is added to the system, the BMCA will designate this new
clock as grandmaster.
33 | P a ge
|
Virginia Tech
|
Figure 2.16 depicts the synchronization process for clocks utilizing the PTP protocol.
PTP utilizes four messages in order to synchronize two clocks: sync message, follow up
message, delay request message, and delay response message.
Figure 2.16 PTP Synchronization Messages
These messages are transmitted from master to slave and allow the clocks to make
frequency adjustments to change the rate at which the clocks meter time. Additionally, the
messages allow the clocks to measure the phase delay between master and slave and allow for a
value correction. The frequency adjustment is made by utilizing the sync and follow up
messages, while the phase adjustment is made by utilizing the delay request and delay response
messages [41]. Every time the synchronization process occurs, timestamps t1-t4 are collected to
determine the offset from master and frequency adjustments for the slave clocks. Clocks that are
PTP compliant that make both a frequency and value adjustment are known as tunable clocks
[41].
Tunable clocks are very important to use in any variable latency packet switched
network. Take the example with two clocks, one master and one slave, and start them at the
exact same time. No two clocks meter time at the exact same rate due to the natural frequencies
34 | P a ge
|
Virginia Tech
|
of individual clock crystals. Consequently, the clocks will begin to diverge from each other and
no longer remain synchronized.
One approach to correct this problem is to periodically reset the slave clock‟s value to
that of the master‟s. This solution does not address the issue of differing clock frequencies or the
rate at which the two clocks meter time. As a result, the clocks will only be brought into
alignment for a moment in time, then begin to diverge again. Tunable clocks, on the other hand,
allow for the proper synchronization of clocks. These clocks allow for the frequency of the slave
clock to be tuned to that of its master so they will meter time at the same rate [41]. In addition to
making a frequency adjustment, the tunable clock also applies a value or offset adjustment so
that master and slave clocks are truly synchronized. Both synchronization phenomena can be
seen in Figure 2.17.
Device
Time
Measurement
Master
Clock
Slave
Clock
Time
Slave Clock "Frequency
Disciplined" to Master Clock
Figure 2.17 Clock Synchronization Non-Tunable (Left) Tunable (Right)
It has been announced by the IEC that IEEE 1588 will be the protocol of choice to
synchronize clocks over the IEC 61850 protocol. PTP is only a time synchronization protocol
and must be implemented on a network that supports the protocol. IEEE 1588 was designed for
packet switched networks such as Ethernet. For experiments described later in this paper, the
EtherNet/IP (EtherNet/Industrial Protocol) was implemented across the network. EtherNet/IP is
a protocol that is part of the Common Industrial Protocol or CIP suite.
35 | P a ge
|
Virginia Tech
|
Common Industrial Protocol
The Common Industrial Protocol (CIP) is an industrial protocol suite that contains
message and service instructions for automation applications pertaining to control, safety,
synchronization, and motion [50]. This protocol is currently managed by the Open DeviceNet
Vendors Association (ODVA) and is an open development network supported by hundreds of
industrial vendors. CIP allows these applications to be implemented on enterprise-level Ethernet
networks. Benefits of implementing CIP networks include seamless integration of I/O control
and data collection, information flow across multiple networks, and implementable multilayer
networks without the need for the implementation of network bridges [51].
CIP can be defined as an object-oriented connection based protocol that supports both
explicit and implicit messaging. Explicit message connections provide generic, multipurpose
communications paths between devices. These messages provide typical request/response type
network communication [50]. Explicit messages are used by CIP for configurations, monitoring,
and troubleshooting. Implicit messages, also known as I/O connections, provide special purpose
communications paths between producing and consuming agents within a network. This I/O
data is often exchanged cyclically or at a requested packet interval (RPI) [50].
The principle behind how EtherNet/IP synchronizes time across a network of distributed
I/O can be attributed to the CIP Sync Object [40]. This object provides an interface to CIP Sync
that allows devices, such as logic controllers, to access the synchronization mechanism. CIP
Sync defines an offset clock model that addresses the requirements for various control
applications. This model is necessary since PTP defines a mechanism for distributing and
synchronizing time but fails to define a mechanism to compensate for step changes in time that
may occur at the grandmaster source [41].
Figure 2.18 defines the CIP Sync Object at a high level. This model shows a PTP master
represented by the circle and a PTP time slave represented by the top right side rectangle. In this
example PTP is used to discipline a local clock so that it ticks and meters time at the same rate of
the PTP master. The slave clock also maintains an offset between the local clock time and the
PTP system time. Any small delta, or step change in time, causes the slave device to make a
small adjustment to the “system to local clock offset” value [40]. In addition, the slave device
will continue to tune its clock. A large step change results in the device updating the offset
36 | P a ge
|
Virginia Tech
|
value, but not tune its clock. As a result, cyclic tasks such as SCADA reads, can be scheduled
based upon the local clock and are not affected by large step changes in time at the grandmaster
source.
Small
Time
PTP
Delta
Master
Local Clock Oscillator
Frequency Based
Any Time Delta
+
System To Local Clock Offset
-
Leap Seconds
=
Synchronized Time Between
Clocks
Figure 2.18 CIP Sync Object Diagram
CIP Sync represents time as a 64-bit long-integer (LINT) that can be expressed in either
nanoseconds or microseconds. The starting reference point in time for CIP Sync is January 1,
1970, starting at 12:00 AM. This time is represented in Universal Time Coordinated (UTC) and
adjusted to include leap seconds. In order to represent this 64-bit LINT in a readable format, an
algorithm must be developed to compute the current date and time in a readable, understandable
format for analysis. The algorithm consisted of several mathematical operations to convert the
LINT into two strings that are given in the following format: MM/DD/YYYY and
HH:MM:SS:µsµsµsµs.
Acceptance of Precision Time Protocol
Recent work has shown that PTP can be successfully used in various applications.
Precision time protocol was implemented successfully to measure the rotor angle of a
synchronous machine [52, 53]. Additionally, programmable automation controllers have been
used in conjunction with PTP and protective relays for a protection voting scheme [54]. PTP and
programmable automation controllers were also used together as Sequence of Events Recorders
37 | P a ge
|
Virginia Tech
|
3. Chapter 3 - Conceptual Design
Introduction
As previously described, the objective of this research is to specify and develop a system
that links the protection and metering of the electrical distribution system at a mining operation
to a site-wide control system. The first steps in this process are to establish the system
capabilities, develop a preliminary conceptual design, and develop a working specification for
the system to achieve the research objective. This chapter discusses the development of the
conceptual design for the automation and control solution that was developed.
The design process of this solution began with an initial conceptual design based upon
research, including industry needs and personal experience. In order to better fortify the design
for an industrial process network, the original conceptual design was presented to industry
stakeholders within the following disciplines:
1. Mining and Minerals
2. Metals
3. Oil and Gas
4. Engineering, Procurement, and Construction (EPC) Firms
The original conceptual design was then compared against engineering community input in
order to finalize the conceptual design presented at the end of this chapter. This chapter begins
with the original conceptual design followed by input provided by the engineering community
segments listed above. Subsequently, the final conceptual design is presented and defined.
Original Conceptual Design
After researching the various SCADA protocols that existed for electrical distribution
systems, as described in Chapter 2, it was determined that the IEC 61850 standard had a high
probability of being readily accepted into the industrial marketplace. The challenge now was not
to determine how to implement an IEC 61850 solution to control various intelligent electronic
devices, such as relays in a mine distribution setting, but how to interface it to a platform that
would be accepted by the engineering community.
39 | P a ge
|
Virginia Tech
|
Microprocessor Based or Automation Controller Based System
It was determined that developing a standalone microprocessor-based solution over PC
software would not be viable or serve the community‟s needs. This conclusion was based on the
fact that microprocessor-based solutions are difficult to maintain. Most microprocessor-based
solutions are written in some form of high level structured text language, e.g., C++, C#, and
Java. In addition, custom software for interfacing and database collecting would need to be
developed for interfacing and storing data. Consequently, specific message instructions over
TCP/IP would need to be developed in order to perform command and control of various
intelligent electronic devices.
Operations, maintenance, and engineering staff at mining and other industrial facilities
are not familiar with this type of system. Consequently, a microprocessor-based solution would
be very costly to maintain as well as make changes to. Operations personnel typically work with
ladder and function block programming in order to maintain, troubleshoot, and upgrade their
systems. As a result, it was determined that the automation and control solution should interface
directly with the facility‟s process control network. These networks are either controlled by a
DCS or PLC/PAC type of controller, which are widely used in the industry. By designing a
solution that interfaces to the electrical distribution system for these industrial facilities, the
following advantages can immediately be seen:
1. Power monitoring capabilities
2. Predictive maintenance capabilities
3. Auto Transfer Switch (ATS) load shedding capabilities
4. Command and control of rotating machinery
5. Command and control of breakers
6. Interfacing energy measurements into speed regulator process models
7. Ventilation On Demand (VOD)
8. Fast motor bus transfer
9. Main-Tie-Main Distribution Schemes
10. Data Historian Applications
11. On Demand Topology Transitions
40 | P a ge
|
Virginia Tech
|
In Chassis Solution or Standalone Gateway Model
The next issue that had to be addressed was determining the physical form of the
solution, in chassis or standalone solution. There are two practical approaches to address this
question:
1. To develop a module that resides in the PLC/PAC chassis, or
2. To develop a module that stands on its own and serves as a gateway or conversion box.
The first option has several advantages, including that the module would require no additional
network connections and would provide one less point of common coupling failure (PCCF). If
the module resides within the chassis itself, the throughput of the module to convert IEC 61850
to a usable industrial format, such as EtherNet/IP, is much greater than if the module sits external
to the local PAC rack. On the other hand, if the developed module is a standalone gateway
module, it would be much more flexible with the system acceptance. For example, not every
manufacturer creates the same rack-based PAC. Consequently, a module created for in-rack-only
would work for a specific model solution of controller. Meanwhile, if the solution is a
standalone module, multiple models of PAC and DCS systems can interface to the device. As a
result, the original design for the solution would be a standalone hardware gateway that would be
connected to an EtherNet/IP or equivalent industrial network. This solution provides more
flexibility and is much more likely to be accepted by industry compared with a module that
resides in the PLC/PAC chassis.
Quantity of Support
In order to have a solution that will be acceptable by the engineering community, it must
be able to handle an acceptable number of intelligent electronic devices, (IEDs) on the network.
After studying multiple drawings for both mining and metals applications, it was determined that
20 was an appropriate number of IEDs to support. Determining the maximum number of IED
support for each gateway module is necessary to permit multithreading on the hardware solution
side and to establish processor clock requirements. If the application required more than 20
IEDs, a second gateway module could be added to accommodate the additional devices. This
process could be completed until all devices could be accounted for within the distribution
system.
41 | P a ge
|
Virginia Tech
|
Communication Media
In order for the solution to be successful, a suitable communication media must be chosen
to move information from the gateway module to both the IED and process networks. The
following modes of communication were examined for this solution:
1. Copper 10/100/1000 Mbit/s communications
2. Fiber Single/Multimode
3. Wireless b/g/n communications
Copper media is the most common form of Ethernet based communication in North America.
Traditional 100 Mbit/s communications speed would be more than adequate to move measurands
and MMS messages throughout a large network of IEDs connected to the process network. The
disadvantage to this solution is that it is susceptible to electromagnetic interference (EMI). The
IEC 61850 standard defines various EMI specifications for devices on the network with
corresponding noise immunity requirements. According to the standard, copper communications
would not satisfy IEC 61850-10 for EMI if the module were within the substation environment.
As a result, if the solution were to include copper communications, the gateway would have to be
located in a control room in order to satisfy IEC 61850-10 for EMI.
Fiber communications would eliminate the EMI issue that a copper solution incurs, but
the disadvantage of this solution is the cost of fiber. Most automation and control networks are
traditionally copper, so adding a fiber interface may not be cost effective.
Wireless communications, while seemingly very attractive, would not be a viable choice
for the industry. With rising security issues and cyber-attacks at an all-time high, process
engineers are focused on maintaining network and system health. Wireless communications are
more vulnerable to attack than other solutions. In addition, wireless signals typically do not
travel well in most underground settings without the use of many repeaters and amplifiers.
Considering the advantages and disadvantages of copper, fiber, and wireless communications, it
was determined to select traditional copper, 10/100 Mbit/s communications for this application.
System Security
Security is a major consideration in the design and operation of industrial control
systems. Good security practices help reduce control product and system susceptibility to
accidental or unauthorized activities that affect safety, operational integrity, and data
42 | P a ge
|
Virginia Tech
|
confidentiality. Industrial control system security relies on layers of security using multiple
controls, methods and techniques that work together to protect a system‟s assets, operations, and
those who depend on its safe, reliable operation. Technical controls, including physical and
electronic mechanisms that compensate for risk, should be accompanied and balanced by non-
technical controls such as company policies, procedures and guidelines. To help protect key
assets, users should employ specific product-level security and protection features available
within a networked process system.
Cyber-attacks are at an all-time high [11]. With the use of intelligent devices, such as
managed switches, control engineers now have the ability to segment their network into zones,
thus mitigating cyber threats. Today, many Information Technology (IT) personnel advocate
constructing network infrastructures that are similar to the one shown in Figure 3.1. This model
is based upon ISA Standard 95 and the Purdue network model. The model defines levels down
the left-hand side of the figure and zones down the right. The model is separated into six
different levels and four zones. In order to keep a secure facility, managed switches and other
network equipment, such as firewalls, separate the plant network (levels 0-3) from the enterprise
network (levels 4-5). These two networks are isolated via a firewall which does not permit web
traffic, email, or process automation packets to pass directly through without passing through an
approved proxy. By segmenting these networks, the process owner can limit traffic and
authorized users who have virtual access to the plant floor production.
Figure 3.1 Network Security Model
43 | P a ge
|
Virginia Tech
|
The gateway module resides in Level 1, since it is part of the control solution for power
distribution systems. This solution would be accessed by Level 2 equipment, such as
engineering work stations (EWSs) and operator work stations (OWSs) for command and control
of various process components. System security would be maintained through best practices and
system architecture design by (1) keeping process control and instrumentation at the lower levels
of control, and (2) keeping enterprise control at the top of the network architecture.
Summary of Original Concept
Figure 3.2 depicts the original conceptual design for the solution to integrate the electrical
distribution network to the process control network of a mining and metals facility.
10/100 Mbit/s
Copper
Interface
Electrical Process
Power Enterprise
Distribution Gateway Control
Utility Network
Network SCADA Module Process Network
Comms. Network
Interface
Figure 3.2 Original Conceptual Design
As described, it was determined that the module should be integrated into the process control
system as a stand-alone, microprocessor-based solution that interfaces to a programmable
automation controller. This solution provides more flexibility with integration into more existing
process control systems compared with a module that resides in the PLC/PAC chassis. This
solution bridges the electrical distribution SCADA network with the process control network and
provides access from the enterprise network to the electrical distribution system as well as power
utility access, if granted, to the process control network. With the advent of smart grid
technology, this solution would provide bidirectional communications between a process owner
and the power utility, a major goal of smart grid technology.
44 | P a ge
|
Virginia Tech
|
The next step the engineering process was to take this conceptual design to various
personnel in the engineering community for additional input/feedback. The next portion of this
chapter discusses how this input was obtained as well as a summary of that input.
Engineering Community Input
After developing an original conceptual design for a SCADA interfacing solution to an
IED network, the concept was presented to the engineering community for input and feedback.
As mentioned, various industries, including mining, metals, oil and gas, and EPC firms were
surveyed for feedback. Firm names were removed to provide anonymous feedback Since this
.
solution bridges both the power and process control systems, survey questions were directed to
both power and instrumentation and control (I&C) engineers. When security concerns were
being addressed, these engineers were asked to consult with their IT Security office. As a result,
both I&C and power engineers needed to work together to provide a solution that would serve as
an acceptable compromise between them.
Each company was asked for their opinion and input on the following categories:
Packaging
Communications Media
Number of IEDs to Support
Safety Concerns
Security Concerns
User Interface Information
Any Additional Comments or Concerns
The results of the industry feedback are summarized in the following four tables.
45 | P a ge
|
Virginia Tech
|
Final Conceptual Design
After collecting industry input and comparing it to the original design, the final
conceptual design was specified and is presented in the table below. It can be seen than 17 of the
20 industry participants felt a stand-alone gateway would provide the best solution. As most
industries already had a copper Ethernet infrastructure installed, 70% of all companies surveyed
felt that developing an already existing copper infrastructure was more practical. Companies that
felt fiber communications were more beneficial cited EMI concerns for this opinion.
All of the industry segments surveyed felt that a minimum of 10 IEDs would need to be
supported by the solution. With regard to safety companies surveyed felt that the ability to
interlock between HMI screens and IED was needed to prevent unauthorized operation and mis-
operation of electrical devices in an industrial setting. Company feedback also stated that
security played a large role in process control. Feedback stated that HMI screens should have
different levels of operation and privileges based on role. They ultimately specified a minimum
of seven different levels of security settings.
Additionally, all industry segments surveyed agreed with the initial conceptual design
that alarms, targets, and measurands should be displayed. Also from the industry input, it was
determined that this solution should provide sequence of events reporting as well as a way to
interface to database software such as PI. Finally, it was determined from industry input that all
graphics must meet the current ISA standards for process control graphics. The final conceptual
design specifications can be seen in Table 3.5. Figure 3.3 depicts the final conceptual design for
this solution.
Table 3.5 Final Conceptual Design
Final Conceptual Design
Parameter Original Design Final Design
Packaging Stand-Alone Stand-Alone
Communications Media Copper Min. 100 Mbit/s Copper Min. 100 Mbit/s
Device Support Min. 20 device 20 devices
Interlocking, Select Before Operate,
Safety Interlocking
Classified Areas Considerations
Security Min. 5 levels Min. 7 levels
Alarms, Targets, Measurands, Timing /
Display Information Alarms, Targets, Measurands
SOE, Historian Interfacing Capability
ISA 5.5 Compliance, Provide
Visualization Information N/A Diagnostic Information, Command and
Control
50 | P a ge
|
Virginia Tech
|
4. Chapter 4 - Hardware Implementation
Introduction
As mentioned in Chapter 3, the hardware solution developed to link the electric
distribution and process control systems is a standalone hardware gateway. This chapter defines
the logic behind the gateway operation and how it bridges the IEC 61850 and process control
network. The gateway module consists of a main processor with various peripherals including a
network 10/100 Mbit/s copper interface and Secure Disk (SD) interface to store configurations
for the module as shown in Figure 4.1. It is noted here that there were several design iterations
before deciding on the components described in this chapter.
IEC 61850 Gateway
Network
Interface
Micrel Phy
Integrated MAC
KSZ8721BL
ARM Processor SD Card
10/100 Mbps Cirrus EP9302-CQZ 1 GB
Memory Block Flash Memory
256 MB SD RAM 4 GB
Figure 4.1 High Level Gateway Block Diagram
Figure 4.1 depicts a high-level functional block diagram defining the hardware
components. The most important component in this hardware configuration is the ARM
processor, shown by the largest block in the diagram. The processor, a single core 200 MHz,
ARM-9 processor, is responsible for all interfaces to memory (flash or SD RAM), as well as the
hardware level interface to the Micrel Phy for Ethernet based communications over a CAT-5
53 | P a ge
|
Virginia Tech
|
copper network. The processor additionally interfaces to the SD card, where configurations are
stored for various modes of operation. In addition, the processor provides a hardware level
interface to LEDs for both diagnostic and status indications during various modes of operation.
Three hardware blocks were considered for system communications: Network Interface,
Micrel Phy, and Integrated MAC, all depicted in the top left of the Figure 4.1. Figure 4.2
describes the data link path between physical media and the ARM-9 processor via the Phy. The
network interface consists of an RJ-45 receptacle to which a generic CAT-5 Ethernet cable can
be inserted to provide a link to the process control network switch, as seen on the right side of
Figure 4.2.
10/100 Mbps KSZ8721BL
MII MAC PHY
54 | P a ge
pihC-nO
srotsiseR
noitanimreT
MAC Comms.
RJ-45
Magnetics Connector
50 MHz Clock
Figure 4.2 High Level PHY Block Diagram
The magnetics block is a passive EMI filter which reduces electro-magnetic noise on the
wire. Additionally, an electro-static discharge (ESD) filter is built into the passive magnetic
circuit. The on-chip termination resistors were chosen to provide the highest level of signal
integrity, while further reducing EMI noise. The physical layer interface (PHY) is referenced by
the large block labeled KSZ8721BL. This transceiver provides the hardware level connections
between the processor MAC and physical media.
Figure 4.3 shows a high level block diagram of the physical layer interface chosen for
this solution. The benefit of this interface is that all low-level hardware drivers were already
implanted and functional. It can be seen from the block diagram level that the PHY has many
responsibilities and peripheral interfaces, including the auto-negotiation of network transfer
speed and the management of both a 10Mbit and 100Mbit connection. This choice of PHY
|
Virginia Tech
|
substantially reduced engineering time without having to code these low level communications
drivers.
4b/5B Encoder
NRZ
Scrambler
MLT3 Encoder
TX+ Parallel/Serial
Transmitter 10/100
TX- Pulse Shaper
Parallel/Serial
Manchester Encoder
MII/RMII
Registers
Adaptive EQ
RX+ And
Base Line
Controller
Wander Correction 4B/5B Decoder
Interface
RX- MLT3 Decoder Clock Recovery Descrambler
NRZI/NRZ Serial/Parallel
Auto
Negotiation
10BASE-T Manchester Decoder
Receiver Serial/Parallel
LED
Driver
Power Down
OR
PLL
Saving
Figure 4.3 PHY Block Diagram
In addition to holding the Linux operating system for this solution, the flash memory also
holds the firmware for the module as well as the web server for diagnostics and configuration.
The SD RAM is used to allocate and store memory for every client connection and IED, as well
as server socket connections and PAC. This memory is volatile, thus any data is lost on power
cycle. The SD card is responsible for storing the device configuration files. These files are
stored in XML format and read on startup, after which memory is then allocated within SD
RAM. Because Linux is chosen as the operating system, the SD Card peripheral is mounted as a
physical drive. Figure 4.4 depicts the startup sequence and memory allocation of the device
upon power-up.
55 | P a ge
|
Virginia Tech
|
BBoooott LLiinnuuxx 22..66 Spawn 61850
Mount SD Card Start Master Read
Power Cycle OOSS ffrroomm Client
as drive on Control Configuration
(+24 VDC) FFLLAASSHH Connections on
Linux OS Program on SD Card
MMeemmoorryy Configuration
Spawn Server
Allocate RAM Allocate RAM
Socket Initialize Tag Initialize Web
for each client for Server
Connection(s) Database Services
connection Connections
for EtherNet/IP
Figure 4.4 Power Cycle State Machine Sequence
The code for this solution was written in C++, in the windows environment. The code
was then compiled using the gcc tool to create a target file intended for the Linux 2.6
environment.
The remaining portion of this chapter focuses on the multi-threading of the processor and
interaction among IEDs, the gateway module, and the controller.
Solution Overview
The process of collecting data from various intelligent electronic devices (IEDs) and
converting it to an industry accepted process control protocol was performed through the concept
of multi-threading. Figure 4.5 depicts the various threads that are executing at any given time
during the operation of the hardware gateway.
Master Control Program (MCP)
Figure 4.5 shows six various types of threads that run parallel with communications to
the processor and each other. The processor contains the master control program (MCP). On
module startup, the MCP reads the configuration file from the SD card, as depicted in Figure 4.1.
Based on the configuration file, the MCP determines the number of individual threads needed to
be created for each driver and calls the corresponding tag database files to create and allocate
memory for moving parameters.
After startup, the MCP starts each driver individually. After each threaded drive has
returned a ready message to the MCP, the MCP changes the status of each individual (ready)
driver to run mode. The MCP also serves as a watchdog timer running on Linux that evaluates
56 | P a ge
|
Virginia Tech
|
Master
Control
Program
61850
EtherNet/IP
Driver
Ready
Ready
61850
EtherNet/IP
Driver
Running
Running
Figure 4.6 MCP Example
Should an error or fault occur during operation, the MCP will log this error in a log file
stored in the flash memory, as depicted in Figure 4.1. This file can then be retrieved for
debugging purposes by the configuration software that will be discussed in Chapter 5.
Hidden Threads
The telnet and ftp server threads are hidden from public use and are used exclusively for
debugging and pushing new system files to the processor. The web services thread hosts a
simple website whose web address is the IP address of the module. Information displayed on
this web page includes model and firmware revisions, as well as simple diagnostic information
about the module, such as error and fault status.
In order to enhance performance of the communication processes of the module, the web
services thread is tied to a jumper that can be removed to enable/disable this functionality. By
removing the web services jumper, the module, on power up, ignores the installation of the web
58 | P a ge
|
Virginia Tech
|
services thread, thus allowing more processing power to be dedicated to IEC 61850 and
EtherNet/IP communications. Additionally, module firmware is downloaded to the module via
the web services page.
Data Mapping
When the module is configured from the user interface software, report configurations
containing the parameters that are to be passed from the IED to the automation controller are
stored on the SD card. Additionally, when the user maps the IEC 61850 tags to EtherNet/IP tags,
each 61850 tag is allocated a certain number of bytes depending upon the data type, such as
boolean, real, double integer, etc. This configuration is also defined on the SD card.
Table 4.1 lists the data types that were supported for this research.
Table 4.1 Data Types
Data Type Definition Bits
BOOL Boolean 1
BYTE Byte 8
UBYTE Unsigned Byte 8
INT Integer 16
UINT Unsigned Integer 16
DINT Double Integer 32
UDINT Unsigned Double Integer 32
IEEE 754 Single Precision
REAL 32
Floating Point
IEEE 754 Double Precision
DREAL 64
Floating Point
STRING ASCII Character Array 32 + 8*length of text
UTC microsecond precision
DATETIME 64
date and time
SEMAPHORE Internal use semaphore 128
Table 4.2 gives a description of various data types supported by this research as well as their
corresponding range of values.
59 | P a ge
|
Virginia Tech
|
Table 4.2 Common Data Types Definitions
Type Description Range
BOOL Boolean data type 0 or 1
BYTE Signed 8 bit magnitude -128 to 127
UBYTE Unsigned 8 bit magnitude 0 to 255
INT Signed 16 bit magnitude -32768 to 32767
UINT Unsigned 16 bit magnitude 0 to 65535
DINT Signed 32 bit magnitude -2147483648 to 2147483647
UDINT Unsigned 32 bit magnitude 0 to 4294967295
Signed IEEE 754 32 bit
REAL ±10-44.85 to ±1038.53
floating point number
Signed IEEE 754 64 bit
DREAL ±10-323.3 to ±10308.3
floating point number
The STRING Data Type is the ASCII representation of a string of characters. The
STRING Data Type is composed of two fields shown in Table 4.3. The LENGTH field contains
the length of the string and DATA contains the ASCII characters. The size of the DATA array is
defined at run time by the configuration settings.
Table 4.3 Uncommon Data Types Definition
Field Data Type
Length UINT
Data UBYTE[]
A DATETIME data type represents a Coordinated Universal Time (UTC) date and time.
The time is expressed as seconds and microseconds since the Epoch and is stored as a 64-bit
structure. The epoch used is the standard UNIX epoch that corresponds to 00:00:00 UTC,
January 1 1970. The UTC time is stored with a one-microsecond resolution. A DATETIME data
type is composed by two fields, as shown in Table 4.4.
Table 4.4 Date/Time Definition
Field Data Type
Seconds UINT
Microseconds UINT
60 | P a ge
|
Virginia Tech
|
Mapping Example
The following example was created to illustrate how data, Alpha-Hotel, is packaged and
mapped in the gateway module. The data is shown in Table 4.5 and graphically mapped in Table
4.6.
Table 4.5 Data Mapping Example
Field Name Data Type Byte:Bit Size (Bits) Data Map
Color
Alpha (A) Bool 0:0 1
Bravo (B) Bool 0:1 1
Charlie (C) Byte 1:0 8
Delta (D) Dint 4:0 32
Echo (E) Byte 8:0 8
Foxtrot (F) Int 10:0 16
Gulf (G) Byte[3] 12:0 24
Hotel (H) Byte 15:0 8
Table 4.6 Data Mapping Example Graphical
Bit
3 3 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 9 8 7 6 5 4 3 2 1 0
1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0
Byte
3 2 1 0
C B A
D
F E
H G
SISCO Stack for IEC 61850 Driver
When the module is booted, the SD card information is loaded into the processor and SD
RAM for use. Every IED that is defined by the user in the reports has its own thread on the IEC
61850 client driver, which is based upon the SISCO stack. At any given time up to 20 threads or
physical devices can be interrogated by the gateway module. These threads read the
configuration of the report and request this information from the IEDs. After the information in
each report is populated, the information is transferred to the tag database where it is collected by
the EtherNet/IP driver.
61 | P a ge
|
Virginia Tech
|
EtherNet/IP Stack
As previously mentioned, when the user maps IEC 61850 tags to EtherNet/IP tags, a
certain number of bytes are allocated to each parameter or tag. The data is transferred from the
gateway to the automation controller via a Class 1 Common Industrial Protocol (CIP) message.
CIP messages are limited to 500 bytes of input data, 496 bytes of output data, and a
configuration header. The information in the tag database is grouped into chunks of 500-byte
packets or data table reads. A parallel thread is created for every group of 500 bytes that needs
to be sent to the controller. This thread is constructed using the Open Device Vendors
Association (ODVA) stack for EtherNet/IP. This organization manages the configuration for the
drivers for the CIP protocols.
Tag Database
The tag database is a shared piece of memory where data items can be located using a tag
name. The tag database is composed of two shared memory locations, as shown in Figure 4.7.
Symbols Shared Memory
Data Types Definition
Tag List Mapping
Data Shared
“INT” Data Address XYZ Memory
Figure 4.7 Generic Tag Database
62 | P a ge
|
Virginia Tech
|
The tag database is the central repository of process data and is the link between drivers.
The data is read by a driver and copied to the tag database where another driver can write it to a
device. This concept is shown in Figure 4.8.
IEC 61850 Tag EtherNet/IP
Driver Database
Tag Database
Read Command Write Command
Object
Tag Database
Write Command Read Command
Object
Figure 4.8 Interaction Between Drivers
The tag database thread can be thought of as a large array of data. The array begins at
index zero and ends with the index of the last byte of data that needs to be transferred. This
thread is the key to the operational success of the gateway module. The tag database is
populated by the IEC 61850 thread, and information is read from the tag database by the
EtherNet/IP thread to be sent to the process control system. Semaphore tags are used to signal
drivers that information has changed and a new, read/write cycle needs to be executed.
When the 61850 thread writes to the tag database, it writes to an index offset from the
base, i.e., the beginning location of where the parameter or electrical measurand lies within the
array of data to be captured. The same process is performed when the parameter is to be read by
EtherNet/IP driver. The process of writing to the tag database can be seen in Figures 4.9 and
4.10, respectively, for both the IEC 61850 and EtherNet/IP drivers. Figure 4.11 depicts the
interactions of the two drivers with the tag database. As described in this chapter, the IEC 61850
63 | P a ge
|
Virginia Tech
|
thread, shown in Figure 4.9, is instantiated upon power-up of the module. After the SD card
configuration is read by the module, the corresponding number of client connections (IED
threads/drivers) is instantiated. The IEC 61850 driver waits in an endless loop until the write
command is issued by the tag database for information to be passed from one driver to the next.
Also, on startup, the server-socket connection is created to transfer information from the tag
database to the controller via the EtherNet/IP driver. The EtherNet/IP driver, as shown in Figure
4.10, waits for a semaphore tag to update. This semaphore tag can be thought of as a collection
of boolean flags corresponding to various measurands and data that are designed to be passed
from the IED network to the process network. When the semaphore tag is updated, an interrupt
is triggered that allows the EtherNet/IP driver to read the tag database, construct data tables of
information to be sent to the controller, and issue a write command to the controller signaling
that it is ready to pass data from the gateway to the controller processor.
Write
Command
Scheduled?
Yes
Send Write Request
Lock Tag Database
Update Tags
Unlock Tag Database
Figure 4.9 IEC 61850 Driver Information Flow
64 | P a ge
|
Virginia Tech
|
reduce the distribution voltage to utilization levels, and trailing cables that connect the loads to
the power center.
Ordinarily, the switch house (as well as the substation and power centers) has devices
that protect each outgoing circuit. Current transformers (CTs) are typically used to measure
current while potential transformers (PTs) are used to measure voltage. A relay receives signals
from the PTs and CTs, processes the information, and the determines if there are any abnormal
conditions such as overloads, short circuits, ground faults, or under-voltage. If an abnormal
condition is detected by the relay, a trip signal is sent from the relay to the circuit breaker to de-
energize the offending circuit.
Mine Distribution
Trailing Cables
Cables
Load Loads
Utility Main Switchhouse Center
Substation
Load Loads
Center
Figure 4.12 Mine Power System Example
Figure 4.13 illustrates these protective devices in the switch house. Each power conductor
passes through a line current CT. In addition, all three power conductors pass through the zero-
sequence CT. The zero-sequence CT is used to detect ground faults through the application of
Kirchhoff‟s Current Law (and the theory of symmetrical components). In nearly all types of
normal and abnormal conditions, including line-to-line and three-phase faults, the sum of the
three line currents are zero, so there is no current induced in the zero-sequence CT. However,
during a ground fault, the sum of the three line currents is not zero, and the current measured by
the zero-sequence CT is equal to the ground fault current. Because many mining operations use
high resistance grounding to limit ground fault current to low levels to avoid arc and flash
hazards, zero sequence relaying can be used to detect low levels of ground current in a power
system having load currents many times higher than the ground fault current. For example, a
66 | P a ge
|
Virginia Tech
|
ground fault of 10 Amperes can easily be detected in the presence of load currents of several
hundred Amperes.
Relay elements shown in this diagram include 50-G and 50, and 51. Element 50 is an
instantaneous line overcurrent element typically used for detecting short circuits. Element 51 is
an inverse-time line overcurrent element typically used for detecting overloads. Element 50-G is
an instantaneous ground overcurrent element used to detect ground faults. It should/must be
noted that modern solid-state relays typically contain many additional elements, such as reverse
power flow, overvoltage, and under-voltage, that are not included in this example.
Vacuum
Circuit
Line Current
Disconnect Breaker Zero Sequence CT (1)
CTs (3)
Switch
50/51 50-G
Solid State Relay
Trip Signal
Vacuum
Disconnect Circuit Line Current
Switch Breaker CTs (3) Zero Sequence CT (1)
50/51 50-G
Solid State Relay
Trip Signal
Figure 4.13 Traditional Feeder Protection Scheme
Consider that a ground fault occurs on one of the branch circuits on the output of the
switchhouse, tripping the associated circuit breaker. All that would be known to downstream
personnel is that the power is off at their location, i.e., mine personnel would not know the extent
of, or reason for, the power outage. Consequently, they would begin calling other sections of the
mine, or outside, to inquire about the extent of the power outage. After many minutes, the
67 | P a ge
|
Virginia Tech
|
switch house eventually would be identified as the location where the circuit breaker tripped, and
an employee would be dispatched to determine the cause of the trip. Upon arrival at the
switchhouse, the individual would be able to determine from the relay panel that a ground fault
occurred and request that an electrician be sent to determine the cause of the ground fault and
initiate repairs.
Next, consider the situation depicted in Figure 4.14, in which the gateway module is
present. Immediately after a ground fault occurs, an individual at the monitoring-system
computer would know the location and reason for the power outage. As a result the correct
personnel could be dispatched immediately to conduct repairs, and while personnel in the
affected portions of the mine could be informed of the source and extent of the problem.
Although this is a relatively simple example of the application of the gateway module,
they show the potential behind the use of this module in other applications. Specifically, the
module provides the mine operator with the ability to not only monitor the status of the IED
(trip/not trip), but also to monitor the values or measurand of interest of the IED and make this
information available throughout the existing SCADA system. By having these electrical
parameters in addition to traditional process parameters, mines have the ability to expand the use
of these existing devices, such as implementing load shedding for energy management or
conducting root-cause analysis of failures through reconstructing sequence of events.
Vacuum
Dis Sc wo in tcn hect BC ri erc au keit r Lin Ce T C su (r 3r )ent Zero Sequence CT (1)
Process Process
Controller 1 Controller 2
Historian
50/51 50-G
Trip Signal Solid State Relay Gateway
Module
Vacuum
Dis Sc wo in tcn hect BC ri erc au keit r Lin Ce T C su (r 3r )ent Zero Sequence CT (1) Engineering
Work Station
Operator Client
Work Station
50/51 50-G
Trip Signal Solid State Relay
Figure 4.14 Gateway Module Implementation
68 | P a ge
|
Virginia Tech
|
Conclusions
This chapter has summarized the hardware gateway module which was developed to
implement the conceptual design described in Chapter 3. A key component to this concept is the
multi-threaded design to interface between the IED IEC 61850 network and the process
EtherNet/IP network. Each thread, including the master control program, was defined, and basic
functionality was explained in functional block diagram. The various types of data that were
used in this project were then addressed, followed by how data is packaged in packets to be sent
to the process network via EtherNet/IP.
The concept of the tag database was then discussed, i.e., how it is utilized as a common
space of shared memory where both 61850 and EtherNet/IP drivers read and write various tags.
Also discussed was the use of semaphore tags in the database to avoid collisions in the tag
database. Finally, a practical example of this hardware module could be used in a mine power
system application was provided. The next chapter focuses on how the user interacts with the
hardware via the use of a graphical user interface (GUI) software.
69 | P a ge
|
Virginia Tech
|
5. Chapter 5 - Software Implementation
Introduction
In order for the hardware gateway to work properly with the process control system,
software needed to be developed that allows the user to interact and configure the gateway
module. As described in Chapter 4, the configuration files are held by the Secure Digital (SD)
card. These files are downloaded to the SD card from the software tool that is presented in this
chapter. In addition, a software tool was developed to store information in the OSI PI database
on a precise timescale. This tool is also discussed in detail later in this chapter.
The procedure describing how the hardware and software work together to produce the
final solution developed in this research involves several steps. First, the user creates a
Configured IED Description (CID) file for each IED that is connected to the gateway module.
For example, if we consider a solid state relay, the user would typically configure it for various
parameters or measurands, such as active and reactive power, voltage magnitudes and angles,
current magnitudes and angles, and sequence components of current and voltage. This step is
conducted using the software provided by the IED vendor, following the vendor‟s instructions.
Next, the CID configuration is downloaded to the gateway module using the IEC 61850
standard. This step is accomplished by the software developed for the Gateway Module and is
described in this chapter. It is noted here that this software was developed to work with any IED
that follows the IEC 61850 standard; it is neither product nor device specific.
Next, the user selects specific information from the CID configuration that will be
transferred to the programmable automation controller (PAC). The software was developed to
allow the user to easily select any or all of the measurands or parameters included in the CID
configuration. With regard to the relay, for example, the user may be interested in sending the
voltage, current, and power quantities to the PAC, but not the sequence components of voltage
and current.
The software then maps these data to EtherNet/IP via the Open Device Vendor‟s
Association (ODVA) standard. The software also creates and exports an Add-On instruction
(AOI) for the PAC. Both of these steps are done automatically by the software developed for the
gateway module. For this information to be passed on to the PAC, the user only needs to add the
gateway module to the controller project and import the Add-On instruction into the controller
70 | P a ge
|
Virginia Tech
|
program. At this point, the IED appears as a device on a rung in the controller project and can be
used like any other automation and control device connected to the controller. (It is noted that
the software was developed specifically for Program Automation Controllers, manufactured by
Rockwell Automation.)
The remainder of this chapter describes how the software accomplishes these tasks and
also addresses process control historian issues.
Creating a Configured IED Description (CID) File
In order to have a common method for describing and documenting the communications
network, IEC 61850-6-1 defines various SCL file types based on XML schemas. The specified
file types are System Specification Description (SSD), Substation Configuration Description
(SCD), IED Capability Description (ICD), and Configured IED Description (CID). For the
purposes of this dissertation, only ICD and CID files will be considered. ICD files represent the
default IEC 61850 configuration of an IED. CID files follow the same schema but represent the
final configuration of an IED in service.
When opening an ICD file in a web browser or text editor, the definition of all logical
devices and logical nodes for the IED is listed. In addition, the ICD file can include definitions
for datasets, MMS reports, and GOOSE messages. Datasets are simply a logical collection of
tags (not necessarily from the same logical node). Collecting the parameters into datasets allows
the data to be efficiently used as part of a GOOSE message or MMS report.
Reports are unsolicited methods of sending datasets from an IED. The standard defines
two types: buffered and unbuffered. When using buffered reports, the IED keeps track of client
message receipts so that any missed reports can be re-sent if there is a network problem.
Unbuffered reports do not maintain the missing data if the link is lost. The MMS protocol can
also provide datasets via direct polling by the client. This method provides no buffering and
eliminates deadbands on analogs.
Figure 5.1 is a screenshot of an IEC 61850 configuration tool developed by a major IED
manufacturer. It can be seen from the figure that this project is configuring a 751, or feeder
protection relay. For the purposes of this research, the data sets and reports tab are of most
interest in this example. As previously defined, data sets are groups of measurands or variables
that are grouped together for reporting purposes via the 61850 standard. When designing
71 | P a ge
|
Virginia Tech
|
SCADA systems, as defined in Chapter 2, Manufacturing Message Specification (MMS) is the
mechanism for the transfer of information from server to client.
Figure 5.1 Manufacturer IEC 61850 Configuration Tool
Figure 5.2 shows how data sets are configured using the drag and drop feature of the
vendor software. As shown, all of the data is classified into various logical nodes as defined by
the standard. In this example, the data set contains fundamental measurands, since tags are being
chosen from the MX or Measurands Logical Node of the device. After the user is satisfied with
the information created in the data set, the data set is saved and assigned to a report.
After the data set is assigned to a corresponding report, the CID file is downloaded to the
IED via an FTP connection. In addition, the CID file is exported for use in this research. Figure
5.3 shows a CID file for the feeder relay configured in this example. The portion of the CID file
shown corresponds to the data set configured for the example with the name RA for Rockwell
72 | P a ge
|
Virginia Tech
|
Figure 5.3 CID File Example Feeder Relay
Importing the CID File into Software
As mentioned, the software tool was created to simplify the user‟s experience with the
gateway module and ease of implementation constructing complex SCADA systems. The
program was created in the Microsoft Visual Studio environment using Visual C++. The
software is responsible for reading the CID file from the relay, parsing the information, allowing
the user to configure final reports, mapping the data from IEC 61850 tags to EtherNet/IP tags,
writing a final configuration in XML for the gateway module, and downloading this
configuration to the SD card of the gateway.
Figure 5.4 depicts the workspace where the user interacts with the gateway module.
When the user opens a new project in the software, the gateway module is the only device that is
defined in the project by the gateway block. In this example the gateway is assigned an IP
address of 192.168.0.250. The left hand pane of the work space defines the devices that are
currently available to be configured; the project has the available EtherNet/IP connection for
communications to the controller. The bottom pane of the workspace defines a listing of all the
74 | P a ge
|
Virginia Tech
|
Figure 5.11 Non-Configured IEC 61850 Tags
After the configuration section is chosen for an individual IED, the user can browse
through the parsed CID file and has the freedom to browse through all configured logical nodes.
In this example the LLN0, or base logical node of the feeder relay, is selected. Furthermore, the
reports folder is chosen and buffered report one (BREP01) is selected to be mapped to the
controller. By dragging and dropping the desired information, i.e. the report folder, the
information is parsed from the report and displayed in the right workspace panel.
Once the information is mapped, the user sees a screen similar to that in Figure 5.12.
When the user wants to map remote bits for command and control, the user maps the control
value (ctval) of the remote bit (RBGGIO1) logical node. These bits are simple Boolean values
that can be mapped to actuate the IED, for instance with open/close or remote trip. The mapping
of these bits can be seen in Figure 5.13. The values mapped from reports are known as controller
inputs, while remote bits are known as controller outputs.
80 | P a ge
|
Virginia Tech
|
Figure 5.14 Adding EtherNet/IP Controller
Figures 5.15 and 5.16 show IEC 61850 tags are mapped to Class 1 CIP connections to
deliver them, via EtherNet/IP, to the automation controller. Prior to this screen being displayed,
the software reads the tag database, as defined by the IEC 61850 tag configurations. The tag
database is then parsed, and each tag is assigned a corresponding data type, each with the
corresponding number of bytes to be mapped to the controller. The user then has the option to
auto-assign all of the data to various CIP connections or assign each data tag point by point. For
purposes of this research and practical implementation, it is recommended that the auto-assign
feature be used to configure the system if no data priority is needed/required.
Once the data is mapped, as seen in Figure 5.16, various tabs in the right-hand workspace
window pane can be used to navigate between Class 1 inputs, Class 1 outputs, and Class 3
message instructions. It can be seen that the starting address of each data point is incremented by
the number of bytes that the data type corresponds. For example, the first item in Figure 5.16
defines the phase angle of the neutral current with a starting address at index zero of the first
connection with a length of four bytes. The second item in the list is the corresponding
magnitude to the neutral current, which in-turn starts at index four. The indicator down the right
side of the window in Figure 5.16 provides the user with a visual indication of the number of CIP
connections used, based on the data mapped from the gateway module to the control system.
82 | P a ge
|
Virginia Tech
|
After defining the mapping scheme from IEC 61850 to EtherNet/IP, the user can
download the project and configuration to the gateway as well as export Add-On Instructions
(AOIs) to the controller software so that the controller can correctly interpret and parse data
coming from the gateway module. This procedure is completed by right clicking the gateway
block in the workspace and selecting the export AOI option, as shown in Figure 5.17.
Figure 5.17 Exporting Add-On Instructions
Figure 5.18 shows the screens the user sees when he or she attempts to download the
project and configuration to the gateway module. When the user is ready to download the
configuration, the gateway block is selected, right clicked, and the download from PC to device
option is selected. The transfer file window appears and displays the IP address of the module.
The user should first select the test connection button to make sure that communication can be
established with the module. After successful communication is established, the download
button should be depressed. After the module downloads the configuration to the SD card,
“validate configuration” should be selected to ensure that the configuration was downloaded
successfully. At this point, the user has successfully configured both the IEDs and gateway
module to send and receive information over the IEC 61850 standard.
84 | P a ge
|
Virginia Tech
|
Figure 5.22, the object getting mapped to the automation controller is the 751A feeder relay,
defined in Line 3. The UDT name is defined as SEL_751A_1, as seen in Line 5. In Lines 7
through 35 (and beyond), the members of the object are defined. The UDT is an object with
multiple attributes, including members and modifying functions. Members are those attributes
associated with the 751A feeder relay. For this particular example, the UDT contains the same
mapped IEC 61850 information that have appeared in screens/or figures already presented in this
chapter.
Figure 5.22 User-Defined Data Type Definition
The later portion of the AOI file that is generated by the configuration software includes
a ladder routing that is imported into the automation controller code. The AOI defines how the
data is scanned by the controller and the data tables containing the information from the IEDs are
updated. The script can be seen in Figure 5.23. Lines 270-282 define the scope of the UDT and
what connections populate the data table. The <Routine> section of the code defines the ladder
logic which individually scans each object of the data stream, parse it, and copy its value to the
corresponding tag in the controller data table. The figure shows that each rung has its own rung
number, and the <text> field in each line defines the logic used to parse and copy the data.
89 | P a ge
|
Virginia Tech
|
Addressing Process Control Historian Issues
Traditional process historian consists of four major components: controller, data server,
interface node, and historian repository. Typically, the event timestamp is applied at the
interface node. Interface nodes collect, interrogate, and qualify information provided by the data
server. If the data values collected at the interface node exceed pre-defined dead band
thresholds, a timestamp is applied and the data is transferred to the historian repository. If the
data does not exceed the threshold, the data point is disregarded and not archived. This process
is better known as exception testing and can be seen in Figure 5.30.
Controller Data Server Interface Node Historian
1 2 3 4
Figure 5.30 Traditional Data Collection Scheme
This procedure for collecting and timestamping data does not work well for events that
are time stamped at the source device. In the above example, the event that is passed to the
interface node is value based, not time driven. After it has been determined that an exception has
occurred, a timestamp is generated and applied to the event. For example, if a process engineer
is trying to log and store events from IEDs, i.e., trips, alarms, etc., the timestamp of these events
is generated by the IED. Using value-based exception reporting will not suffice in transferring
meaningful timestamps from the IED to the historian repository. Referring back to Figure 5.30,
the controller does not provide a timestamp to the data server, but only a process value. For most
manufacturing environments, this schema is acceptable as time delay error is within the poll rate
of the data server.
In an IED system with IEC 61850 communications, the event drives both value and
timestamp changes in the data collection system. Both the event data and timestamp are created
at the IED and must travel together from the source of the event to the repository to be
94 | P a ge
|
Virginia Tech
|
meaningful. The previous two statements are the root cause of the problem in attempting to store
electrical distribution values in a process control data historian. The block diagram that
describes the solution to this problem can be seen in Figure 5.31.
Web Delivered
Trends / Reports
Relay Gateway Controller RSLinxEnterprise LiveDatF aT IH nterface FTH SE FTVantagePoint
1
2 3
RACE SOE -> FTH
4
Figure 5.31 Proposed Software Solution
Figure 5.31 shows the progression of data through the block diagram system. Block one
defines the physical IED. The IED(s) is (are) connected to the gateway module via an Ethernet
connection. Additionally, the gateway module is connected to the controller represented by the
Controller block. The timestamp that is generated on the event by the IED is passed through the
gateway to the controller, as shown by the progression from Boxes 1-3. Although the controller
data table contains both the event value and IED timestamp, it is not inherently passed to the
historian repository. In order to persist an event with an historian tag, the configuration of an
ALMD (Digital Alarm) instruction associates the event to the point ID of the historian tag. An
historian point ID is a unique identifier for a tag in the historian repository. When an alarm
occurs, the following parameters are known: timestamp of IED, unique point ID, timestamp of
entering alarm, value of alarm, and the Event Association ID. An example of this logic is shown
in Figure 5.32.
95 | P a ge
|
Virginia Tech
|
When the alarm is triggered, a message is sent to the Alarm and Events (A&E) data
server (RS Linx Enterprise) as depicted in 5.31. RS Linx Enterprise then populates the
ConditionEvent Table in the A&E database (red arrow). Figure 5.34 shows an example of this
table. The ConditionEvent Table shows the InputValue, EventAssociationID, Tag1Value (IED
timestamp), and Tag2Value (point ID). Figure 5.34 also shows the off/on and on/off transition
of an alarm event, represented by EventAssociationIDs being equal.
Figure 5.34 ConditionEvent Table (Partial)
In Figure 5.31, the top path in Block 4 represents the traditional data flow within the
process historian system. The solution developed for this research takes an alternative approach
to providing data to the historian repository, as represented by the lower path through the RACE
SOE / FTH tool. This GUI tool is shown in Figure 5.35. The GUI program was written using
the .NET framework. The Results workgroup represents the two collections: A&E records and
historian tags. Connections to both the A&E database, as well as the process historian, are
defined in OLEDB connection strings.
Figure 5.35 RACE GUI Tool
97 | P a ge
|
Virginia Tech
|
Now that an event with point ID 297 has been captured in the A&E database, the
information, including the associated IED timestamp, needs to be persisted to the historian tag.
This is accomplished by the user with an insert query to the process historian. After the insert is
successful, an artifact record is written to an additional table in the FT A&E database called
SOEtoFTH. Figure 5.36 defines the query to display information in the SOEtoFTH table. The
result of this query can be seen in Figure 5.37.
Figure 5.36 Query of SOEtoFTH
Figure 5.37 Query Results (Partial)
Note that Active is the input state of the alarm instruction (ALMD). Tag1Value is the
time stamp from the IED, Tag2Value is the point ID of the historian tag, and FTHTS is the
Factory Talk Historian Time Stamp. The FTHInsertTime is the time used in the insert to the
historian. The FTHTS and FTHInsertTime can be up to +/- 15 microseconds, as this is the
highest resolution that this timestamp object can represent.
98 | P a ge
|
Virginia Tech
|
Figure 5.40 shows the flow of information created by the development of the software discussed
in this chapter. As described, information moves from IEDs in the field through the gateway
module to the controller. From the controller, these values can then be inserted into historian and
trended in reports on view screens for visualization and reporting.
Figure 5.40 Practical Data Flow Diagram
Summary
In conclusion, this chapter has discussed software developments resulting from this
research that provide a functional, user-friendly solution to what program the gateway module to
convert IEC 61850 messages to EtherNet/IP messages. This chapter was broken down into two
sections: software created to interface with the gateway module, and software created to move
information to a process historian. Software developed to interface to the gateway module reads
CID files and allows the user to map information that is required by the automation controller.
After this information is defined as tags, these tags are mapped to corresponding EtherNet/IP
tags according to the ODVA standard. Once this mapping has occurred, the module
100 | P a ge
|
Virginia Tech
|
6. Chapter 6 - Visualization
Introduction
This chapter defines the visualization developed for this research. Items discussed in this
chapter include: faceplate definition, human machine interface discussion and definition, and
data management solution. The motivation for visualization is to represent an IED in logic and
graphics as close to the physical object itself. This provides the operator or engineer with the
same look and feel experience they would have if interacting with the physical device. At the
same time, the graphics were developed with various levels of security to allow only users with
proper credentials access to various command and control functions. Additionally, the research
solution must adhere to various graphics standards for both power and process control systems.
Faceplate Solution
Process control companies have developed solutions for process visualization. This
research takes process visualization one step further into the electrical distribution system. The
IEC 61850 standard allows for the visual representation of reports and alarms at the process
control level. With the development of more sophisticated Human Machine Interface (HMI)
screens, global objects have been introduced into the automation graphics. The use of these
global objects has allowed for the creation of faceplates.
A faceplate is defined as a reusable standard object. The advantage of the faceplate is
that it is a standard, prebuilt object that can be implemented repeatedly. Each faceplate has
security levels built into the objects themselves. These security features can be customized,
based upon user and application requirements. A prime example of the security benefit is the
command and control of IEDs.
The command and control portion of this research allows operators and engineers to
change settings and allow relay (or other IED) actuation from a remote site. Even though the
IEC 61850 standard defines how an IED communicates on an IEC 61850 network, many people
do not realize that these files do not include configuration information for the protection and
control functions in an IED. The IEC 61850 standard also does not provide any uniform method
for designing communication-assisted automation. Using faceplates as a solution addresses both
102 | P a ge
|
Virginia Tech
|
faceplates match National Electrical Manufactures Association (NEMA) standards. The color
schema adheres to that of the IEEE C37.1 definition of Supervisory Control and Data
Acquisition (SCADA) systems. The color schema differ from traditional process control
systems in that red usually references a stop or removal of power operation. However, in
distribution systems, red implies an application of power to terminals.
Figure 6.2 Relay Faceplate Global Objects
Additionally, to make the faceplates easily reusable, a standard base faceplate was
developed. Figure 6.3 shows the standard object and table formats used for the IED faceplates.
These objects are shown as visible when various tabs of the faceplate are displayed. Again, the
color scheme matches ISA 5.5 and complies with IEEE C37.1. Information tables, such as the
line to neutral voltages and currents table, are implemented in the engineering tab of each
faceplate. The “#” field indicates that the text field will be a numeric. The power conductors are
represented by red, the neutral conductor by white, and grounding conductor(s) by green. A logo
was also requested from various manufacturers for implementation in the faceplates. These
104 | P a ge
|
Virginia Tech
|
Figure 6.9 Physical and Graphical Comparison
Figure 6.9 shows a comparison of the physical IED to the virtualized representation of the
IED in the form of a faceplate. The graphic on the left is a photograph taken of an SEL 700G
generator protection relay. This IED is used to protect rotating synchronous generators that are
found in mining applications on sites that have their own co-generation. The graphic on the right
represents the IED in faceplate format. It can be seen that the faceplate is identified as a 700G,
or generator protection relay, from the title bar of the faceplate window as well as from the
manufacturer‟s image mark implemented as a global object. The target LED panels down the
left represent the same text that is represented on the physical IED, for example, Enabled, Trip,
Differential, etc. As all of these labels are configurable on the IED, the same text is also left up
to the user to define. For example, if the Differential LED was set to Reverse Power flow in the
IED, the faceplate could be altered to represent this change. The pushbutton LEDs, (right
column), are also graphically represented in the faceplate. Each button has a maximum of three
lines of text that could be assigned. The first and third lines of text refer to the operation of the
110 | P a ge
|
Virginia Tech
|
amber LEDs, while the second line of text references the operation of the push button.
Graphically this information is represented as follows:
1. The text on the second line with reference to the push buttons is graphically represented
as a faceplate pushbutton
2. The text on the first and third line reference the text with amber LEDs on the IEDs
themselves. These LEDs represent the internal state of the IED.
Additionally, each faceplate has an alarm and events banner built into the faceplate. This allows
the operator or engineer quick access on the home screen to determine any warning, diagnostics,
or trip information from the relay. This alarm banner is populated from logic written in the
controller.
The tabs across the top of the faceplate represent home, synchronization, engineering,
overall faults, X-side (Low Voltage) faults, and Y-side (High Voltage) faults. The next section
of this chapter presents the various screens for the IED faceplates developed in this research.
Faceplates were developed for four IEDs: generator, feeder, motor, and differential. These can
be seen in Figures 6.10-6.22.
111 | P a ge
|
Virginia Tech
|
Figure 6.24 shows a sample ladder program with two AOIs: one defining the UDT (rung
0) and the other defining the AOI for the faceplate (rung 1). The AOI in rung 1 defines 11
different input tags. The reference tags refer to the text of the target LEDs that the user wants to
show on the Faceplate when it is displayed. In this example, targets such as overvoltage, under
voltage, and overcurrent, can be specified depending upon system definition. The relay name
input parameter refers to the user-defined data type (UDT) as described in the software chapter.
In this example, the relay name is FeederRelayBusA. The alarm for this faceplate references
the name for the alarm that appears in the alarm and events database when this instruction is
executed.
Figure 6.24 Add-On Instruction Ladder
Figures 6.25 and 6.26 depict the steps to import the faceplate graphics into an existing
HMI project. From the file menu, the user selects the import wizard tool. When the wizard
loads, the user selects the import graphic displays into project option. After selecting this option
and backing up the project for revision control, the user then points to the faceplate(s) that are
intended to be imported into the project. These files are .gfx or graphics files that are designed
for FactoryTalk View Studio applications.
126 | P a ge
|
Virginia Tech
|
Conclusions
This chapter described the visualization developed for this research. Items discussed
included faceplate definition, human machine interface discussion and definition, and data
management solution. The goal of visualization in this research was to represent an IED in logic
and graphics which produced a virtual faceplate that would be familiar to users accustomed to a
particular IED. The virtual faceplate gives operators and engineers the same look and feel they
experience with the physical device. At the same time, the graphics were developed with various
levels of security to allow only users with proper credentials access to various command and
control functions.
Additionally, the research solution adhered to various graphics standards for both power
and process control systems. The visualization trending functionality of the solution was then
discussed by providing an example of a trend from a process historian repository. Finally, an
example of a secondary selective main-tie-main scheme was presented for a processing facility.
The one-line diagram was converted to an HMI screen, and multiple faceplates were tied to
various instrumentation blocks in the diagram to enable remote monitoring and control of the
system.
132 | P a ge
|
Virginia Tech
|
7. Chapter 7 - Verification and Acceptance Testing
Introduction
This chapter discusses the testing and verification procedures conducted to validate and
verify functionality of the proposed solution. Since a goal of this research is to develop a mine
monitoring solution utilizing the IEC 61850 standard to link the electrical distribution system
with the process control system, it was determined that the best way to validate the work and
functionality is to benchmark gateway performance tests against the conceptual design. The
chapter discusses each conceptual design milestone as well as the procedures used to validate
that this goal was met. For milestones requiring more than just visual inspection, the
experimental setup is defined, testing procedures are described, and the experimental results
discussed.
Conceptual Design Milestones
The original conceptual design of the solution proposed in this research was discussed in
Chapter 3. After surveying multiple industrial firms ranging from EPCs to heavy industry
segments, and comparing the feedback against information gathered in the literature review, this
research produced the results shown in Table 7.1. Table 7.1 shows seven major milestones
categories that were created for this research which include packaging, communications media,
device support, safety, security, display information, and visualization information.
Many of the physical milestones can be validated by simple inspection. For example,
packaging and media communications can be validated easily by examining a standalone
gateway module. In addition to inspection, further testing and validation needs to be performed
to benchmark the module‟s performance in various situations including being lightly loaded,
heavily loaded, and how it would react in dynamic situations.
The benchmarks tested in this chapter are the self-imposed “Final Design” specifications
shown in Table 7.1. Items that can be validated by inspection are qualified in paragraph form in
the next few sections, while testing procedures are thoroughly described throughout the
remainder of this chapter.
133 | P a ge
|
Virginia Tech
|
Table 7.1 Final Conceptual Design Milestones
Parameter Original Design Final Design
Packaging Stand-Alone Stand-Alone
Communications Media Copper Min. 100 Mbit/s Copper Min. 100 Mbit/s
Device Support Min. 20 device 20 devices
Interlocking, Select Before Operate,
Safety Interlocking
Classified Areas Considerations
Security Min. 5 levels Min. 7 levels
Alarms, Targets, Measurands, Timing /
Display Information Alarms, Targets, Measurands
SOE, Historian Interfacing Capability
ISA 5.5 Compliance, Provide
Visualization Information N/A Diagnostic Information, Command and
Control
Packaging & Communications Media
Once creating the final conceptual design specification, it was determined that a
standalone gateway module is the most versatile solution for linking the power distribution
system to the process control system. Additionally, in order to be functional with most of the
existing infrastructure, copper Ethernet media was determined to be the best form of data
transfer. Figure 7.1 shows a sample experimental system consisting of a controller, unmanaged
Ethernet switch, gateway module, 24 VDC power supply, and feeder protection relay. The
gateway module in Figure 7.1 is a standalone device (blue module) which is powered by the 24
VDC power supply.
Additionally, Figure 7.1 shows that the gateway module uses an RJ-45 copper media
Ethernet jack as the communications media. As defined in the hardware section, hardware for
the communications PHY was selected for 100 Mbit/s TCP/IP Ethernet communications, thus
satisfying the conception design milestone. As a result the packaging and media
communications milestones were successfully met.
134 | P a ge
|
Virginia Tech
|
Figure 7.1 Sample System Setup
Visualization Information
ISA 5.5 Compliance
After developing the final conceptual design specifications for visualization information,
it was important to ensure that any information displayed on process screens would be displayed
in a standard, acceptable fashion. The Instrumentation Society of America (ISA) defines proper
visualization in its 5.5 standard, entitled “Graphic Symbols for Process Displays.”
This standard was chosen since many mining facilities follow these guidelines as it is
similar to MSHA standards. One problem with some of the graphics developed for this research
is that power distribution and control often use color schemes opposite that of their process
counterparts. For example, the breaker closure is often represented as red, while the breaker
opening is often represented as green in power systems, whereas the breaker closure represents
the start of a process and breaker closure represents the stopping of a process. The IEEE
Industry Application Society Industrial and the Commercial Power Systems (IEEE IAS I&CPS)
color schemes were chosen for power since their safe, non-energized states are often referred to
with a green color, while their energized or possible danger states are symbolized by red.
All symbols used within the graphical faceplates for visualization follow ISA 5.5 guidelines,
with the following exceptions shown in Figure 7.2. The symbols shown in Figure 7.2 are not
135 | P a ge
|
Virginia Tech
|
defined in the ISA 5.5 symbols library and were added by exception to allow for full
functionality of the solution developed as part of this research. are all NEMA symbols that are
utilized in the power distribution realm and follow traditional power systems color schemes as
defined by the SCADA standard IEEE C37.1.
Figure 7.2 ISA Symbol Exceptions
Diagnostic Information
According to the IEC 61850 standard, diagnostic information is kept in the quality fields
of each relay. If the user decides to map the quality information from any parameters, these
values can be monitored for IED health. In addition, information about potential and current
transformers can also be brought back from the system. All of these quality flags from the IED
are transferred to the controller where they are constantly monitored for bad data. These quality
flags are then logically “ORed” together to determine if there was a diagnostic problem with the
relay. If there was a problem, an alarm was generated, and the corresponding faceplate was
136 | P a ge
|
Virginia Tech
|
Security
The original security specifications called for five distinct levels for the system, but after
industry consultation, the final conceptual design called for seven levels of security to be built
into the HMI graphics. The intent for levels of security is that those without proper credentials
cannot/are not able to perform command and control of the electric distribution system. The
solution developed in this research creates the possibility of safety risks if engineers or personnel
with improper credentials make unauthorized changes to the mine power system during
operation. As a result, seven levels of security were implemented to prevent unauthorized
actuation of power equipment. The seven levels of security implemented in the design solution
can be seen in Table 7.3. Tables 7.4-7.7 show the permissions for each security level.
Table 7.3 Security Levels
Security Level Description
1 Administrator
2 Engineer II
3 Engineer I
Senior
4 Operator
5 Operator
6 Maintenance
7 Remote Access
In order to test the integrity of the security built into the system, various trials were run
on each developed relay faceplate in order to ensure that no level could access data or command
and control actuation of a device for which it was not authorized. Tables 7.4-7.7 describe the
testing conducted on each relay faceplate with descriptions of each test case. A red „X‟ indicates
that access should not be and was not permitted, while a green check indicates that access was
granted for command and control or viewing of information. Each case was tested 50 times and
included closing the faceplate and opening it again for consistency. The results shown in Tables
7.4-7.7 represent a 100% successful testing of each test case.
139 | P a ge
|
Virginia Tech
|
Display
When reviewing the final conceptual design for the solution proposed in this research,
four distinct target milestones were established:
1. Display Alarms and Targets
2. Display Measurands
3. Display Sequence of Events Data
4. Provide Historian Interfacing Capability
Each of these milestones are discussed in this section as well as how the information is displayed
to the benefit of the operator or engineer.
Alarms and Targets
The first and most important functionality of IEDs is electrical protection of the devices
in the system. With this in mind, IEDs must be able to provide alarms and warnings to operators
and engineers to alert them that there is an electrical fault or other anomaly in their mining
system that could harm personnel or damage equipment. Prior to microprocessor-based relays,
SCADA systems were alerted by the closure of normally open and normally closed contacts on
the relays to make or break alarming circuits from the device.
Unfortunately, this method only provided information that the relay had picked up or
alarmed, but no diagnostic information as to why the relay had picked up or alarmed. As a
result, post mortem analysis of why faults occurred was time consuming, thus affecting process
yield due to downtime. With microprocessor based relays, more diagnostic information (such as
the reason for picking up) can be provided by the IED, but there is still no easy way in a
distributed system to retrieve this information without being directly plugged into the front port
of the relay. The solution provided in this research allows trips, alarms, and warnings to be
graphically displayed in a centralized location via the use of the graphical faceplates described in
Chapter 6.
The diagnostic tab in each relay faceplate provides all of the diagnostic information for
alarms or targets that may pick up during facility operation. This screen can be seen in Figure
7.4, while Figure 7.5 displays the generic targets on the home screen of the faceplate for easy
identification , in the event the relay picked up, and a summary view of why the relay picked up.
142 | P a ge
|
Virginia Tech
|
Each of the relay targets was tested via simulation software by forcing the target to
generate a trip signal based on computer forced command. Each target was tested five times in
conjunction with the relay faceplate and checked whether there was agreement between the
controller data table point and the faceplate graphic tag. The result of this testing indicates that
all relay pickup points were mapped correctly and operated correctly.
Measurands
With the advancement of microprocessor-based relays, individual measurands, such as
voltage and current, can now be gathered by the IED. Measurands for the solution proposed in
this research for mine monitoring and control play a major role in indicating the health of the
electric distribution system as well as the electromechanical equipment. These fundamental
measurands, both electrical and thermal, are important for establishing the long term health and
conducting the predictive maintenance of a system. As a result, these fundamental values are
included in the visualization portion of the relay faceplates, as this information, when provided at
a centralized facility from a distributed system, allows engineers to quickly asses the overall
performance of an operation. Figures 7.6 and 7.7 depict the measurands that are displayed
during the operation of the mine system.
In order to simulate values going to the relays, the IEDs generator, feeder, differential,
and motor were connected to a variable voltage source with a motor load. Voltages and loading
were then adjusted in order to provide for a range of electrical measurands, ensuring that the
relay updated these values accordingly. In addition, faceplate graphics were checked to ensure
they were mapped correctly to controller tags, thus appropriately updating from information in
the controller data table. It was determined after testing each relay and faceplate that the
information was correctly mapped and was being correctly transferred from IED to PAC.
144 | P a ge
|
Virginia Tech
|
Timing and Sequence of Events
With the advent of GPS timing, all IEDs from a distributed system can be referenced and
synchronized to the same time clock or reference to time. This allows for easier analysis of
cascading events in an electrical distribution system, and for reference to when events occurred
to an ultimate time source. Two places needed to be tested to ensure that timestamps were
properly passed throughout the system. The first place was between the IED and the automation
controller, and the second was between the automation controller and HMI software.
Timing between IED and PAC
The first step in validating proper time transfer was to establish a network of IEDs, a
gateway module, and controllers for receiving data from the IEDs. This setup is shown in Figure
7.8. The relays were configured to pick up every two seconds and reset automatically. When the
relays picked up, they were configured to send an IEC 61850 report, with corresponding
timestamp, to the gateway module, which in-turn published the information to the automation
controller.
First the timestamp between the IED sequence of events record (SER) and published IEC
61850 report was checked for having the same value. It was determined that timestamps from
trips were assigned in the same processing cycle, leaving a time difference of zero microseconds
between the actual protection element assertion time and the relay trip published time. This
point was later found as a part of the IEC 61850 annex defining that all IED manufactures should
follow this procedure to provide the most accurate information to SCADA systems.
The reported trip time was then compared against the data table LINT timestamp. It was
determined after 1000 tests of all four relays that the time difference between the IED timestamp
and controller data table timestamp was zero microseconds, thus meeting the milestone of
preserving the timestamp from the IEC 61850 network to the EtherNet/IP network.
146 | P a ge
|
Virginia Tech
|
Controller
Network
Managed Switch Gateway Module Managed Switch
Figure 7.8 Timing Test Setup
Timing between PAC and HMI
The intent of this solution is to provide a time scale that is as accurate as possible, given
current industry products and the capability of the system hardware. The ControlLogix controller
is capable of using a 64bit Integer (LINT) that represents the number of micro-seconds from Jan
1, 1970. This allows for a date timestamp to be held in either two registers (DINTS) or one
(LINT).
The FactoryTalk Alarms and Events server software is capable of storing, in a Microsoft
SQL Server Database, the event time at a resolution of one millisecond. This resolution is due to
the use of the DateTime data type. Unfortunately, at this time the DateTime field as specified by
Microsoft only accommodates a 32-bit timestamp, or one millisecond. This can be seen in
Figure 7.9. Microsoft is currently developing a DateTime2 field which represents the full 64-bit
range, or one microsecond, resolution.
Although there is no direct way to time a 64-bit timestamp to an alarm, it can still be
preserved within the alarm object. The object has four other tags storage locations with each
Alarm Event, stored in the FTAE database. As a result of this data object definition being able to
pass the LINT timestamp as an associated alarm tag, the timestamp can still be preserved in full
form and passed to a process historian for data vaulting. This feature can be seen in Figure 7.9.
147 | P a ge
|
Virginia Tech
|
The result of mapping the timestamp to Tag 1 also can be seen in the blue box in Figure
7.10, where the timestamp represents the same time microsecond format. It was determined after
1000 tests of each relay that the LINT timestamp, represented in the number of microseconds
located in Tag 1, matched exactly the timestamp in the data table of the PAC as well as the
timestamp transferred from the IED network. As a result, the preservation of the time milestone
was achieved for this solution.
Storing Values in Historian
As mentioned in Chapter 6, OSI PI historian does not accept foreign timestamps due to
the way traditional process historians function. Chapter 6 also discussed a solution for how to
address this issue with the creation of a software tool that bridges the Alarms and Events
database and OSI PI historian database. In order to test the functionality of this software tool,
2000 alarms were created and tested on IEDs in a system.
As long as a timestamp inserted into the historian was within 15 microseconds of the IED
published time, the tool was within tolerance and no further error propagated throughout the
system. The timestamps that were compared are shown in Figure 7.11. The red timestamp
corresponds to the timestamp published to the process historian, and the blue timestamp
corresponds timestamp published by the IED network. As long as the time difference between
the red and blue timestamps is less than or equal to 15 microseconds, the tool has no added delay
outside the tolerance of current process historian systems. The results of the testing can be seen
in Table 7.8 and Figure 7.12.
Figure 7.11 Time Comparisons
Table 7.8 Numerical Time Error Results
Total
Time Error [us] 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Number of results 172 176 137 116 97 164 170 152 151 161 115 112 92 85 63 37 2000
149 | P a ge
|
Virginia Tech
|
Time Error of Software Insert Tool
180
160
140
s
t 120
n
e
v
E 100
f
o
r e 80
b
m
u 60
N
40
20
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time Difference [microseconds]
Figure 7.12 Graphical Time Error Results
Table 7.8 and Figure 7.12 show that, after 2000 test trials, all resulting data points are
within the error bars of 15 microseconds. The data suggests that 74.7% of the results are within
9 microseconds of each other, well within the 15 microsecond tolerance provided by OSI PI.
In addition to the testing of time discrepancy, this test also checked whether fundamental
parameters could be trended against electrical faults in the system. To check this, a second
vendor‟s IED was tested as it had a built in simulator to test relay functionality and trip settings.
This software can be seen in Figure 7.13. By adjusting the slider bars in the simulator, the
currents, voltages, and harmonics of the corresponding waveforms were adjusted. The relay
tripped when the set thresholds of these parameters were exceeded, thus triggering an alarm to be
logged by the controller and historian system.
150 | P a ge
|
Virginia Tech
|
Figure 7.13 Software Simulator
Information for analog parameters can also be logged and trended in historian as seen in
Figure 7.14. In this figure, the maroon lines represent the trip status of the relay, which, in this
example, was set to trip on a ground current greater than 60 A (green), regardless of the currents
on phases A and B (blue and red). As shown in the figure, the maroon line is asserted any time
the process parameter, i.e., ground current (green), exceeds 60 Amps and resets when the relay
resets. In this example this occurs when the ground current is less than 60 Amps.
Analog waveforms can be depicted if the IED manufacturer allows this information to be
sent over the IEC 61850 standard. These waveforms can be seen in Figures 7.15 and 7.16,
respectively. Figure 7.15 shows waveforms over a four-second time interval, while Figure 7.16
shows those same waveforms zoomed to approximately eight electrical cycles.
151 | P a ge
|
Virginia Tech
|
Figure 7.16 Zoomed In Analogue Waveforms
Throughout the testing process design milestones were validated by successfully
transferring information from the IED network to the PAC and from the controller to a process
historian for storage and trending.
Device Support
In order to test quantity and throughput for device support, IEDs were instrumented with
a packet analyzer known as “wire shark” which captures and transfers data. Wire Shark is a
software network analyzer, recognized globally as an industry standard for determining
connectivity, data rates, and network transfers. In order to obtain information about the network,
a managed Ethernet switch was utilized. Managed switches have the ability to map an addition
port as a span port. The span port mimics or becomes an exact copy of the port to which it is
assigned to mirror. In this case, the switch port to which the gateway module was inserted was
mirrored to another port. so that a computer running wire shark could capture the data. This
setup is shown in Figure 7.17.
153 | P a ge
|
Virginia Tech
|
The blue plot in the diagram represents the IEC 61850 MMS data being transferred from
the IED to the gateway module. There is a distinct pattern of the blue trace consisting of three
distinct peaks, each representing a report being sent from an IED to the gateway module. The x-
axis represents time in seconds of the data capture, and the y-axis represents the number of bits
transferred. At approximately every 5 seconds the same pattern of three reports is generated and
sent to the gateway module, with each report marked by a blue peak. As IEDs are designed first
for protection and second for SCADA transfer, the timeframe is not exactly five seconds;
however, the error is less than 300 milliseconds on data transfer rate. The data flow to the
gateway module also spikes at these moments, as the green line spikes at the time of data
transfer.
Figures 7.19-7.21 show wire shark screen captures that provide information on packet
transfer between devices. Figure 7.19 shows information from the various frames captured by
wire shark with the corresponding protocols of each message. Figure 7.20 shows specific
information about an EtherNet/IP message, in this case between the gateway and the controller.
Figure 7.21 shows IEC MMS communications between the IED and gateway module.
Figure 7.19 Frame Capture
155 | P a ge
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.