question
stringlengths 19
6.88k
| answer
stringlengths 38
33.3k
|
|---|---|
Problem Statement: The DMCplus controller turns off but the TDC controllers don't shed. | Solution: If the alarms status of the timer and on/off flag points are disabled, then the CL code will not shed the local PID controllers to the specified shed mode. They will time-out on their own at the time specified for each PID. This time is normally 2.5 times the DMCplus controller execution cycle. If it is less than 2.5 times, then the PID may shed on a single missed cycle.
Keywords:
References: None |
Problem Statement: Under certain conditions, the DMCplus controller could accept a -9999 (Cimio BAD flag) and be used as a valid value for an intermittent variable and base its online calculations on this. This could occur when the analyzer update logic erroneously sets the NEWPV=1 trigger. | Solution: In previous DMCplus versions AspenTechnology stressed the use of vetted logic calculations to trigger an analyzer update (i.e. setting NEWPV=1). Under certain conditions, it was difficult for some users to always guarantee that the correct logic is used. Thus, and additional validation check has been added to the online controller to reject BAD values as valid numbers.
Install Engineering Release 2 for AMS6.01 (seeSolution 114336).
Keywords: DMCplus
References: None |
Problem Statement: A user is unable to load a product receipt and product shipment event through EIU to an oracle database. The exact same test using Oracle export to Access, and the product event (shipment, receipt) loaded from EIU without a problem. | Solution: If we enter a valid TRAN_MODE in SHIP Via field, then the Product event can be audit successfully. See below the red circle area.
For Oracle, sqlldr is used to import data to database, and we set TRAN_MODE as the last column for importing ORION_MGR_EVENT_IMPORT. If the final field in a given row of your data file is unspecified (NULL), sqlldr will reject that records.
Keywords: EIU
Oracle
SQL
Receipt
Shipment
TRAN_MODE
References: None |
Problem Statement: Build v1.0x.xx lets me build a CCF calculation that exceeds 180 characters (RHS) but it won't load the file again. It crashes and reports this program has performed an illegal operation. What is the problem? | Solution: updated September 2009.
Applicable Release(s):
Multiple
Keywords:
References: None |
Problem Statement: DMCplus Implementation, IP-X / IP.21 | Solution: Summary of recent discussions for implementing DMCplus with IP-X.
When DMCplus is run against the IP-X database, DMCplus internal scheduling can be used. The database reads and writes will be asynchronous, but as long as they occur as fast as or faster than the controller runs there should be no problems.
When there are external calculations that must be synchronized with the controller execution, it will be necessary to schedule the DMCplus controller via the DMCplus API call coded in a task scheduled by IP-X. This task can also perform the required calculations.
Here's the recommendation for InfoPlus-X systems with regard to the windup and controller mode:
Create an individual IP_DVDef (integer) record for each of the AWSCOD and LOOPST parameters for each MV in the InfoPlus-X database (2 points per MV). Populate these DVs via a Cim-IO get record (IoGetDef) using smart data types CBAW and CBST. Point the DMCplus controller to those points in the database (and use regular DBVL source types, i.e., don't convert them again).
There is a template file that you should start with to do a DMCplus project communicating with InfoPlus-X and using DMCplus View as the operator interface.
The decision to make points local or build points in the database is based on what you want to be able to trend in the database and whether the operator will use displays that see data in the DCS or in the database or the controller directly.
The suggested way to handle the outputting of setpoints is to use a single IoPutDef record with an occurrence for each setpoint. Schedule an SQL script to set the individual processing switch for each of the setpoints in that one record based on the controller overall status (ONREQ and ONSTS) and the individual MV status (INDSTA).
IP.21 is going to write everything in the list, every time ... whether or not the value was updated.
This means every setpoint (VINDSP) gets written every time, whether or not the MV is in service or the controller is on.
However, there is hope.
Recommended :
Associated with each tag in an IoPutDef record is a field called IO_OUTPUT_TYPE. If this field has the value STORE the that tag will be written to the DCS. If this field has the value DO NOT STORE then it will not be written to the DCS.
An SQL script can be triggered to examine the status of each MV (and the controller) and decide whether to set their IO_OUTPUT_TYPEs to STORE or DO NOT STORE. After the script completes then the IoPutDef record can be triggered.
However the MV status information is written to IP.21 after the setpoint values are written. This means that you would have to wait until after all the writes are completed before running the SQL script and triggering the IoPutDef records.
In turn, that means that you probably want to use the scheme of having an AWRITE update a dummy tag which triggers the script to be scheduled after a 3 or 5 second delay (the delay gives the controller time to complete all I/O to IP.21.)
Yes it's a kludge, but the only other way to do this properly is to write a custom CIM-IO client for IP.21 that handles things in a manner that is compatible with the way the controller does I/O. The standard IP.21 CIM-IO client and IO records just don't offer any cleanSolution.
Note: Associated with each tag in the IoPutDef record there is also a field called IO_DATA_PROCESSING which can be ON or OFF. If OFF then it that tag is not sent. HOWEVER, you should not use this field to control whether or not to write because toggling that field for any tag causes a new CIM-IO list to be created, causing extra burden on the CIM-IO server and slowing I/O.
Not recommended, doesn't work:
Another alternative is to put each setpoint in it's own IoPutOnCosDef record. The IoPutOnCosDef record is very similar to IoPutDef, but it gets triggered when any of the tags in its list are changed.
Using this method, with one IoPutOnCosDef for each VINDSP, you would write the setpoint as soon as it was written by DMCplus, and only if it was written by DMCplus.
* * If the controller is riding a constraint and not changing the value, then the writes stop. * * *
If and only if you tie the IoPutOnCosDef to an Asynchronous CIM-IO client then there should be no significant delay in I/O .... figure a conservative 0.1 seconds per I/O.
Doing this for a few medium controllers is aok, but large numbers of outputs done this w
Keywords: DMCplus, IP21, IP-X, recommendations
References: None |
Problem Statement: In Windows 7, when I tried to register a .dll file using the command in DOS window 'regsvr32', sometimes I get this error message,
The module C:\!SUPPORT\US\OrionVb.dll was loaded but the call to DllRegisterServer failed with the error code 0x80004005. | Solution: The reason is you have to register as an administrator. You need to open a DOS window by choosing 'Run as Administrator'. The DOS window should opened as 'Administrator: Command Prompt', instead of 'C:\Windows\System32\cmd.exe',
Keywords: register
registering
DLLRegisterServer
error code
0x80004005
References: None |
Problem Statement: I have setup a HYSYS Dynamic Model with a compressor which has an electric motor. Why is there a negative duty warning when I lower the synchronous speed of the compressor? | Solution: In the attached case as the user can lower the synchronous speed and it causes the compressor to give a ‘compressor with Negative Duty.’ This error message is shown below:
The governing energy balance equation that models the startup and shutdown of a compressor is provided within the attached HYSYS Dynamics Documentation as equation 1.39. The energy balance around a compressor can be broken down into 4 main terms:
torque =duty loss due to friction + duty loss due to inertia + power
The torque relates to the moment of force exerted by the motor. The power relates to the energy exerted on the passing fluid through the compressor. Consider the case for shutdown where the compressors synchronous speed is gradually lowered to zero. As the synchronous speed approaches zero the inertia, friction and torque terms also decreases to zero. The resulting power term can become negative due to the resulting enthalpy difference between the input and exiting streams of the compressor.
Keywords: Compressor, HYSYS Dynamics, Motor
References: None |
Problem Statement: What differentiates a Unit Parameter (type 1) event and a Unit Mode (type 2) event? | Solution: If the Value of all the Operating Parameters of the events can be any number, then it is type 1. However, if the Value of ANY parameter can be discreet values e.g. Severity of FCC can be High or Low (definitely not a number), this event is of type 2. In other words, when you create a Unit Operation event that contains a dropdown list (e.g. select mode operation), APS recognizes this event as type 2. If you only define numbers, APS recognizes it as type 1.
Using the Demo model, Unit Operation events for the IS4 - Butane Isomerization will be of type 2, because there's a dropdown list - Charge, as shown below:
On the other hand, Unit Operation events for the DLC - Delayed Coker will be of type 1, because there isn't a drop down box, as shown below:
Keywords: drop down menu
References: None |
Problem Statement: I need to change the control scheme during the dynamic simulation, say for example that initially one control loop is in action (say for example a pressure controller), but later another controller (e.g. a flow controller) takes control of the same valve. How can I do this? | Solution: You have two options. The first option is to run the dynamic simulation with the first PID control loop (possibly using the PIDIncr block), then pause the simulation when you need to change the scheme, and reconnect the control signal (PV in this case). Note that any change of the simulation structure during a dynamic run will cause the Restart function to stop rewinding the simulation to time 0, and rather it will rewind to the time at which you applied the change. While ACM/AD allows this by design, this approach is useful only if you know you will not need to go back to time zero. In other words, this is probably not the preferred option.
A better method is to create the two controllers at time 0, configure them correctly. You need also to create a SignalSelector block, which allows any number of input signals to be connected (in this case the outputs of the PID controllers). You can then connect the OP of each controllers to the inputs of the SignalSelector block, and connect the output of the SignalSelector block to the valve position. You can then select which control loop is active by setting the SelSig variable of the SignalSelector block.
If you also need a bumpless transfer when switching the controllers, you should consider using a task (as you will have to change several variables simultaneously to achieve this).
The example attached illustrates the method.
Keywords:
References: None |
Problem Statement: Sometimes it is useful to bring data from the APS database (Acces or SQL/Oracle) that it is not available through other means (E.g. #INPUT from the user workbook). A way to achieve this is by creating a connection to the database using the units workbook. In this | Solution: a VBA example is shown that brings the API gravity of the crudes to the spreadsheet INIT (Units.xlsm attached).Solution
1. Open the Units workbook and hit Alt+F11 to open Excel’s VBA.
2. Create a new Module called GetAPI
3. In order to make the connection to Access or SQL (via .dsn) the next code is being used:
4. Then the next code is used to execute a query to that database and paste the values in the specific cells:
5. All the code is pasted as sub function GetAPI and to try it you can use the demo model with the attached units.xlsm (remember to refer to this unit file)
6. In INIT we have defined GetAPI in #MACRO, in order that when the simulator is reloaded the values appear in the specified cells
Keywords: Database, connection, query, Access, Oracle, SQL
References: None |
Problem Statement: CIM-IO connections to more than 1 remote OPC servers cannot be started automatically from CIM-IO for OPC version 2.0.1 according to any of the documented procedures. Also, multiple connections to a single OPC server cannot be started automatically from CIM-IO for OPC version 2.0.1 according to any of the documented procedures. The following approach using CIM-IO Kernel 4.9.4's cimio_autostart.bat file provides this desired functionality. | Solution: This document explains how multiple CIM-IO asyncdlgp processes can be automatically started up to communicate with multiple remote OPC servers.
(This procedure can also be used to automatically start up multiple CIM-IO asyncdlgp processes to communicate with a single OPC server if desired This could be useful for example when you are running DMCplus Collect processes against an OPC server while also running IP.21 against it, since the DMCplus Collect will be using synchronous CIM-IO communications to the asyncdlgp while IP.21 is free to use asynchronous requests to the asyncdlgp. Note that the OPC server itself should be free to receive asynch or synch requests from the IP.21 CIM-IO server even while processing the synchronous requests from the DMCplus Collect CIM-IO server since these synchronous CIM-IO requests will become asynchronous OPC requests at the CIM-IO server level.)
This procedure, when used for DMCplus or AspenIQ alone, depends on the cimio_autostart.bat file of CIM-IO kernel 4.9.4. The cimio_autostart.bat file of CIM-IO kernel 5.0 does not apparently work at the moment, but once it is fixed, it should also be possible using CIM-IO kernel 5.0.
The following six files should be copied into the %CIMIOROOT%\Commands directory after renaming the current cimio_autostart.bat and cimio_autostop.bat files out of the way. (To find out the directory name corresponding to %CIMIOROOT%, you can type SET C at a DOS prompt.)
=====cimio_autostart.bat=====
# add any batch file to be run by the CIM-IO Manager service at
# startup
%CIMIOROOT%\commands\startopcservers.bat
=====end cimio_autostart.bat=====
=====cimio_autostop.bat=====
# add any batch file to be run by the CIM-IO Manager service when it stops.
%CIMIOROOT%\commands\stopopcservers.bat
=====end cimio_autostop.bat=====
=====startopcservers.bat=====
call %CIMIOROOT%\commands\stopopcservers.bat
start OPCSERV1 Cim-IO Server /min %CIMIOROOT%\commands\cioopcserv1.bat
start OPCSERV2 Cim-IO Server /min %CIMIOROOT%\commands\cioopcserv2.bat
=====end startopcservers.bat=====
=====stopopcservers.bat=====
%CIMIOROOT%\code\cimio_shutdown cioopcserv1
%CIMIOROOT%\code\cimio_shutdown cioopcserv2
=====end stopopcservers.bat=====
=====cioopcserv1.bat=====
set SVCNAM=CIOOPCSERV1
set SERVNAM=OPC.APS
set NODNAM=A1_15
%CIMIOROOT%\io\cio_opc_api\asyncdlgp -tcpservice%SVCNAM%-opcserver%SERVNAM%-nodename%NODNAM%
exit
=====end cioopcserv1.bat=====
=====cioopcserv2.bat=====
set SVCNAM=CIOOPCSERV2
set SERVNAM=OPC.APS
set NODNAM=A2_50
%CIMIOROOT%\io\cio_opc_api\asyncdlgp -tcpservice%SVCNAM%-opcserver%SERVNAM%-nodename%NODNAM%
exit
=====end cioopcserv2.bat=====
After copying the files into place, verify that the cio_opc_api directory including the asyncdlgp.exe is located under %CIMIOROOT%\io. If it is not, you will need to replace the %CIMIOROOT% symbol used in the cioopcserv1.bat and cioopcserv2.bat files accordingly.
You will also need to edit the cioopcserv1.bat and cioopcserv2.bat files to change the OPC server names and nodenames appropriately. The example given here, OPC.APS, is the OPC server name for a Yamatake IOGS. For a Honeywell APP node, the OPC server name would be Hci.TPNServer. For Yokogawa ExaOPC, the OPC server name might be something like Yokogawa.CS1000_OPC. You will need to verify the name of the OPC server process on your machine by checking the registry. (If you are attempting to start multiple processes communicating to the same OPC server on the local machine, you may want to go ahead and delete the environment variable NODNAM and the nodename portion of the asyncdlgp command line.)
These batch files should start up two asyncdlgp processes talking to two different remote OPC servers under process names CIOOPCSERV1 and CIOOPCSERV2. The processes will be started by the CIM-IO Manager service under the user name and password you have set for that account. You should set the CIMIO Manager service to start automatically under the user name and password specified by your vendor for communication to your OPC servers.
To set the user name and password for your CIM-IO Manager service, go to Control Panel and double-click on the Services icon. When the services window appears, highlight CIM-IO Manager select the Stop button to stop the service if it is already running. Then, with CIM-IO Manager still highlighted, select the Startup button. Select the checkbox next to Account and type in the required user name and password. Then select the radio button next to Automatic startup and select the OK button to finish entering the information. (At some point, you may receive a message indicating that NT is automatically changing the specified account to be enabled to be started as a service. You can make sure the account is enabled to be started as a service by going to the User Manager program under Administrative Tools on your start menu.)
After configuring the CIM-IO Manager service and setting it up to use the required user account, restart the CIM-IO Manager service by selecting the Start button on the Service window with CIM-IO Manager highlighted.
At this point, there may be issues with the CIM-IO Manager not actually managing to execute the cimio_autostart.bat file. You should be able to verify if the asyncdlgps were started successfully by looking at the Task Manager process list. If you need to troubleshoot, you may want to verify that the individual cioopcservn.bat file commands are set up correctly by bringing up a DOS window and typing in the batch file commands directly to the prompt to check if they execute correctly without error messages. This is your chance to re-verify that the path to the command is entered correctly, as well as the nodename, service name and OPC server name. Another likely problem is that the cimiomanager process is simply failing to execute the cimio_autostart.bat file. You can verify this by putting a command like the following at the beginning of the startopcservers.bat file: echo This is the startopcservers.bat file being executed. > c:\startopcservers.log. You can then check to see if this file gets created when you start up the CIM-IO Manager service.
For systems where IP.21 is installed, an alternative to using the CIM-IO Manager's cimio_autostart file to start these asyncdlgps is to set up the startopcservers.bat and stopopcservers.bat files as tasks in IP.21 with the task associated with startopcservers.bat being set to execute at startup while the task associated with stopopcservers.bat is set to be skipped at startup. The customer should then make a point of manually executing the stop task prior to shutting down IP.21.
Miscellaneous Notes:
These batch files need to be started as a task or service rather than being executed from the command line, because, when run from the command line, they 1) leave minimized windows for each CIM-IO task launched and 2) launch the tasks under the current user name and password. Normally, the current user name and password will not correspond to the user name and password that will be accepted by the OPC server process.
So far, we have been unable to get the CIM-IO 5.0 kernel's CIM-IO Manager process to successfully execute the cimio_autostart.bat file, possibly because of the included space in the directory Program Files within %CIMIOROOT%.
Version 3.0 of IP.21 should allow for configuring CIM-IO server processes to run as services from within IP.21 configuration. This may be one of the betterSolutions in the future.
The command line asyncdlgp commands being used in these batch files correspond to the OPC Start/Stop utility, but we discovered that the current version (2.0.1) of CIM-IO for OPC has a bug in the Start/Stop utility which makes it impossible to start a connection to a remote OPC server from Start/Stop. Using the command line version is both a way around that problem and an automated method of starting these remote OPC server connections. This bug in Start/Stop is also expected to be corrected in an engineering release soon.
NOTE: Version 2004 and newer provide the OPCProperties utility for configuring additional opc servers.
OPCProperties can be found in Program Files / AspenTech / Cimio / IO / cio_opc_api.
Refer to the Cimio to OPC Users Guide for additional information.
Keywords: OPC, CIM-IO, DMCplus
References: None |
Problem Statement: How to set up a PC to run a TPN Server client remotely. | Solution: When the Cimio Server is installed and running on the DMCplus machine, thisSolution can be used to use the TPN Server Client.
Note: The account and password of the TPN Server machine must be known so they can be set up identically on the PC.
Summary:
The following reSolution describes how to set up any PC to run a TPN Server Client.
Solution
How to set up a node to run a TPN Server client remotely:
On the client node:
Mount the TPS System Software CD (i.e. the CD with APP software)
Install the Microsoft Redistributable Libraries. Reboot.
Check which version of the App node you have:
If you have a release prior to APP 120 then: Select either Hi-Spec Base Software or APPSolution Package. Click Install Package. DESELECT EVERYTHING but TPN Server Connectivity Client (or TPN Server if using APPSolution Package) and HCI Runtime.
If you have APP 120 or later then: Select HCI Connectivity Client and click on Install Package.
Install. You may see a number of errors starting with Unable to set identity to TPSComServer. Errors related to unable to set permissions will follow. These can be ignored. They occur when the client is not part of the same NT domain as the TPN server.
If you are using a Honeywell application, it may make use of an HCI Component name. This does not apply if you are using a third party client. To create the HCI Component name, the following steps must be taken. Go to PROGRAMS->HONEYWELL TPS->Non TPS Client Configuration Utility. Select Configure an HciComponent Manually. For Base ProgID: select Hci.TPNServer. For Hci Alias Name: specify the HCI name you used for the TPN Server when you added it to the TPS Domain. For Host Name: specify the hostname of the computer with the TPN Server installed. Leave the TPS Domain Name: blank. This will allow usage of the Hci component name.
If the client and TPN Server are in different NT domains, then create a local TPSComServer account with the same passowrd as the domain account on the client machine.
Go to User Manager ->Policies->User Rights. Grant the Right, Access this computer from network to the TPSComServer account (either the local account if they are in different domains, or to the domain account if they are in the same domain).
On the TPN Server Node:
If the client and TPN Server are in different NT domains, then create a local account on the TPN Server node, that matches the account that the client will run under. Use the same password.
Go to User Manager ->Policies->User Rights->. Grant the Right, Access this computer from network to the client account (either the local account if they are in different domains, or to a domain account if they are in the same domain).
On the node where TPN Server resides, enter RUN->DCOMNCFG. Select the line HCI_TPNServer - exe Server. Under Properties, check Security. Check the access permissions and make sure the client account is allowed to access the TPN server.
DCOMNFG>
Default Authenticatin level
From
connect -> none
Impeumake -> Anonymous
Identity from -- to interactive user.
Keywords: DMCplus, Aspen IQ, ACO Base, TPN, TPN Server, Server, client, remotely
References: None |
Problem Statement: When Aspen HYSYS Dynamics is linked with Excel using Aspen Simulation Workbook, you can create “Profile Tables” to display outputs from the simulation as a function of time. How can I change an input variable in Excel? | Solution: 1. Add a user defined variable (e.g. P/F Spec on a stream or Set Point of a controller) into the Model Variables page in the Organizer.
2. Click “Create table”. This will be a static table.
3. Create a “Profile Table” and include the same variable.
4. Change the variable in the static table.
5. The correspondent value in the Profile Table will be the same as in the static table.
In the following screenshot, the set point of a controller was included in the static (top) and Profile (bottom) tables.
Keywords: Table, variable, value, Excel
References: None |
Problem Statement: Why do we provide a Kp tuning parameter instead of Kc for a Surge Controller? | Solution: Both Kc and Kp are essentially the same and follow the HYSYS PID algorithm. The surge controller acts like PID controller operations but uses the extensions of the PID algorithm to prevent the Centrifugal Compressors from surging. For instance, if the process variable (PV) is operating above a certain margin over the surge flow limit, the surge controller operates exactly as a PID Controller. However, more aggressive action is taken by the surge controller if the Centrifugal Compressor is close to surging.
Keywords: Compressor, Surge Controller, PID controller, Tuning Parameters
References: None |
Problem Statement: Dais Trader will not run.
The following log file tells us the TRADER is not working on a VMS node named ALICE.
The log file is:
DAIS320$ROOT:[TMP.ASPENACO_ALICE_LOGS]ASPENACO_ALICE_TRADER.LOG
$ Set NoOn
$ IF F$MODE() .EQS. NETWORK THEN EXIT
$ if f$mode() .eqs.OTHER then set noverify
%SET-W-NOTSET, error modifying ALICE$DKB201:
-SET-E-INVDEV, device is invalid for requested operation
2000/12/18 14:31:24 $5$dkb201:[sys1.syscommon.dais320rtl.][bin]trader.exe;1:
default checkpoint file: ''DAIS320$ROOT:[TMP.ASPENACO_ALICE_LOGS]traderlog''
2000/12/18 14:31:24 $5$dkb201:[sys1.syscommon.dais320rtl.][bin]trader.exe;1:
current checkpoint file:
''DAIS320$ROOT:[TMP.ASPENACO_ALICE_LOGS]alice_sarnia_sunoco_caTrLog.''
($5$dkb201:[sys1.syscommon.dais320rtl.][bin]trader.exe;1) :: capsule
564141720 Terminate: TRADER: Failed to ForceRegister Offer: 17560
(UserException)
DMCPLUS job terminated at 18-DEC-2000 14:31:26.76 | Solution: We recommend that the Dais Trader log files get deleted.
Sometimes it doesn't get shut down properly or completely and a log file gets corrupted and interferes with registration of the applications.
Shut down the applications.
Delete the log files at DAIS$ROOTTEMP:[ACO_hostname.location of dais log files)
Restart the Trader.
Startup the applications.
A good test to ensure all is well is to do an application shutdown and an application startup. If the apps startup as expected, things are back to good.
Keywords: Dais, Trader, Dais Trader, ForcedRegister, Offer, DMCplus, AspenIQ
References: None |
Problem Statement: A recommendation for handling the Blue Screen of Death (BSOD) - on a DMCplus ( or any NT ) Server | Solution: [quote from Microsoft(r) Windows(r) Platform News - 1998, Issue No. 30]
FORCE WINDOWS NT TO REBOOT AFTER A CRASH
If you spend any time administering Windows NT, you're far too familiar with the Blue Screen of Death (BSOD) that displays the cause of the crash and gives some information about the state of the system when it crashed. The BSOD will sit on the screen until someone reboots the system, which could be very bad for a system that should be running 24 hours a day, like an Exchange server. You can force NT to automatically reboot after a crash by setting the value of HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\CrashControl\AutoReboot to 1. Once you've changed this value, NT will reboot after writing the crash log file.
[end quote]
Keywords: Reboot, Crash, Windows, BSOD
References: None |
Problem Statement: %%%%%%%%%OPCOM 26-SEP-1997 04:04:01.16 %%%%%%%%%
Message from user SYSTEM on CORVXG
Event: Carrier Check Failure from : Node LOCAL :.CORVXG CSMA-CD Station CSMACD-1
========================================================================= | Solution: CSMA-CD is another name for Ethernet hardware protocol. A Carrier Check Failure event would suggest hardware failure detected in mid- packet. Like maybe an intermittent coax or a flakey 10BaseT port.
Client changed the Vax transceiver then tested the VAX transfer across the network, this seemed to eliminate the CSMA-CD error.
Keywords: Carrier Check, CSMA, Failure, CORVXG
References: None |
Problem Statement: How can I change the time unit in a strip chart? | Solution: Go to the strip chart of interest, right mouse click on it and select Graph Control then select Time Axis and choose the Units you want to use.
Keywords: Time scale, Units, strip chart
References: None |
Problem Statement: How do I fix the error 'User-defined type not defined' while running automation in APS? | Solution: This error can be seen if orion.tlb is not referenced correctly.
In the VBA interface, select Tools |
Keywords: APS automation
User-defined object not defined
orion.tlb
References: s. In the reference window, browse to C:\Program Files (x86)\Aspentech\Aspen Petroleum Scheduler and then point to the file orion.tlb |
Problem Statement: How can I show the pressure profile of a pipe in Aspen HYSYS Dynamics? | Solution: When running an Aspen HYSYS Dynamics case we can do time profile of a variable so X axis will always be Time, you can access this trough Pipe Segment> Dynamics tab> Stripchart (see below). We cannot see other variable as a function of time and pipe length simultaneously.
The changes of a specific variable at each length point of pipe can only be seen in steady state (Pipe> Performance >Profile).
Keywords: profile, pipe, length
References: None |
Problem Statement: I have an atmospheric pressure tank or CSTR in my Hysys Dynamics simulation. The feed to the tank is liquid only. Why do the pressure and temperature decrease when I decrease the liquid level (either manually or via liquid level control)?
Background
Aspen HYSYS considers the tank or Constantly Stirred Tank Reactor (CSTR), a fixed volume vessel that is not open to atmosphere (or external pressure source) unless it is connected to a feed stream to represent these conditions.
If the feed to this tank is liquid only, the only way to increase the vapour space within the tank (i.e. reduce the liquid level) is to flash (i.e. reduce the pressure) of the vessel contents. In addition to the vapour production, the temperature will decrease due to the evaporative effect.
Furthermore, if the liquid level reduction is significant, a very low pressure will be required to create sufficient vapour to fill the vapour space in the vessel. This may be accompanied by a very low temperature. This effect is very prominent in liquid only feeds of low volatility. In such cases, liquid level control will be very noisy and the system can behave erratically.
If in the reality, the tank or CTSR is open to atmosphere (or an external pressure source, e.g. nitrogen (N2) header), it will not accurately modeled by this simulation configuration. | Solution: Add a gas or vapour feed stream to the tank or CSTR. This stream should be representative of the real operation conditions.
For instance, if the vessel is open to atmosphere, a feed stream of 100% mole fraction Air component can be used. This stream can be connected to a valve (of very large Cg - so as to simulate negligible pressure drop for the flow) and have a pressure specification of 101.3kPa (typical atmospheric pressure). A pressure-flow correlation can be used (with a check-valve - to prevent backflow)
If the vessel is connected to an external pressure source (to maintain vessel pressure), e.g. N2 header, a pressure controller connected to a valve can be used to modulate the flow (and maintain pressure) of N2 to the system. If the dynamics of this N2 injection is not required, a similar valve configuration / stream dynamic specification to the atmospheric air injection (as stated above) can be used. The only difference will be the composition and process conditions of this injection gas.
Keywords: Tank, CSTR, Open to atmosphere, Liquid Level, Control, Vapour space, N2 injection, blanket,
References: None |
Problem Statement: I'm opening a Flowsheet in Edit mode. When I first open it, the font size is small:
However, when I switch to Simulate mode, the font size gets larger:
After this, I go back to Edit mode, and the flowsheet looks scrambled. | Solution: To fix this problem, go to View | Settings | User Settings | Font Size Factor for Flowsheet Screen and change it to a fraction (0.666 is standard).
This way, the same font size will be kept throughout the modes.
Keywords:
References: None |
Problem Statement: Collect - The end of file option for circular files frequently leads to a corrupted binary file. When eofchoice=1, the binary input file becomes a circular file. When the maximum number of samples is reached, space used to store collected data is re-used (thus limiting the size of the binary file). Several sites have reported that the collection is corrupted and no data can be recovered after the file wraps around. | Solution: Use eofchoice=0 for a fixed binary file size. If the chosen size becomes inadequate to hold the necessary number of samples for each tagname, start a differently named collection before the first one reaches the size limit, at which time it stops.
If a fixed binary file size is not a requirement, then you should use eofchoice=-1 for a file that will continue to grow as long as the collect runs. You should monitor your available disk space.
Complete information regarding the EOFCHOICE can be found in DMCplus Online Help.
Keywords: EOF, dmcplus, collect, EOFCHOICE
References: None |
Problem Statement: Finding Non-Default Values with Variable Find | Solution: The Variable Find dialog box has two new options to find:
Variables and parameters whose values are not at their default
Variables and parameters whose specifications are not at their default
These options make it easy for you to find all changes to the default values and specifications that you have made for review, reporting, or other purposes.
In fact, the graphical layout has changed since v7.1. See details in the following screen shot.
Left side shows prior to v7.1 and right side is for v7.1 onward
Keywords: default, finding, variable, find
References: None |
Problem Statement: *** 21-Aug-1998 01:04PM *** PATTERSON, WAYNE : 13108 ***
When not using
the Bridge to start up the FOXIA CimIO and instead are starting it up manually, they may need this.
There are two things that can cause the large output files. One is the environment variable, CIMIO_FOXIA_NOLOGFILES, that causes the logging of the files, and the other is the setting of the DEBUG flag in the data file read by the FOXIA Cim-IO server upon startup. This flag is located in /opt/cimio/etc/cimio_foxia_config.dat.
A lot of the Cim-IO debug prints would not occur if DEBUG were 0 instead of 1 so you may just want to modify that file. Make DEBUG a 1 since I only ever write to the output files if I'm trying to debug something. | Solution:
Keywords:
References: None |
Problem Statement: Using DMCplus Build, a program runtime error can occur
when using the save from the File, Close, Save before Exiting prompt. Modifications since last save will be lost. | Solution: The program logic for the File Close functionality has been found to have errors. Some modifications in the CCF will trigger a runtime error when saving during the file close, ending the session without saving the file. Do not use the File Close menu option to exit the CCF. Instead, simply click on the CCF window close icon (the x in the upper right corner of the CCF window) to close the CCF. This method works as expected.
Keywords: DMCplus, Build, CCF, Runtime, error, exiting
References: None |
Problem Statement: UCX 4.1 ECO 6 has a problem running out of buffers to use.
System Error, errno=55 no buffer space available
Sat May 30 07:14:38 1998
dmc_u20fr2 Error writing list 5: IDB: IDB_Write failure: PDB: CIMIO_USR_PUT_CONNECT, Error connecting to device CIMIO_MSG_CONN_SOCK_CREATE, Error creating an outbound socket CIMIO_SOCK_OUT_CONN_FAIL, Error connecting to the server System Error, errno=55 no buffer space available | Solution: *** 05-Dec-1997 02:13PM *** PATTERSON, WAYNE :13108 ***
$ set output=00:00:10
$ set verify
$ set noon
$ ucx show comm
$ ucx show comm/memory
$ ucx set comm/large_buffer=max:1600
$ ucx set config comm/large_buffer=max:1600
$ ucx show comm
$ ucx show comm/memory
$ exit
-
Keywords:
References: None |
Problem Statement: How can I use the ACM2004 calling convention in procedures? | Solution: We will use the example discussed inSolution 120202. Please refer to it for the beginning of the discussion.
The new procedure with ACM2004 option will be modified as below:
Procedure pSizeLimit2
Library: SizeLimit.dll;
Call: sizelimit2;
Compatibility: ACM2004;
implementation: subroutine sizelimit2.f;
language : fortran;
inputs:
sizes as real(*),
frac as real(*),
limit as real;
outputs:
smed as realvariable;
End
This allows you to declare the name of the arguments (inputs and outputs). The syntax is different of the variable declarations in models as you put the array indication after the argument type (e.g. you write sizes as real(*), while in a model you would declare sizes([1:n]) as realparameter;). It may be a bit confusing, but after all these are not actual variable declarations, but just indications for ACM when it generates the template code of the subroutine to be written to use the names selected by the user, instead of making up some names that may be hard to follow.
The template that is generated by ACM starts as follow:
SUBROUTINE SIZELIMIT2(ICALL,
& sizes, sizesDim1,
& frac, fracDim1,
& limit,
& smed,
& IFAIL)
The other effect of the ACM2004 option is that it changes the order of the arguments. The call type is now in front of all other arguments and is always present. (With the previous spec, it was passed after the variables, and only when needed). We hope the new convention is more simple (but of course it won't be if you were used to the old one). Finally, the way the derivatives are returned (in case the procedure returns analytical derivatives) is dramatically different. This example does not return derivatives, so let's not worry about this.
The example attached contains the code of the procedure with version 12.1 and 2004 compatibility options. Note that the ACM2004 compatibility option may only be used with version 2004 and above. The template shows ACM13 instead of ACM2004, but they are equivalent (ACM13 may be deprecated in the future).
The example was created with version 2004.1.
Keywords:
References: None |
Problem Statement: Why does my vessel outlet flow decrease when I increase the outlet nozzle diameter? | Solution: The flow rates of the vessel are determined by the Dynamic pressure-flow solver. So if there is no pressure change around the vessel, the outlet flow should not be changed.
The reason why the nozzle diameter has an impact on the outlet flow rate is as follows:
When the liquid product nozzle diameter changes, each phase fractions in products will be recalculated based on the current liquid and vapour level. An enlarged nozzle would result in an increase in vapor fraction and decrease of liquid fraction on product fluid, which would cause the overall molar density to decrease. Since holdup volume is conserved, and V = M/rho, the decreased density results in a decrease in the molar flow rate. The disturbed system will eventually return back to the original state because the flow is governed by the pressure drop which hasn't changed.
Keywords: Nozzle, Vessel, HYSYS Dynamic
References: None |
Problem Statement: Here's a summary of which parameters are on each of the CIMIO lists and what types of points they are. | Solution: List 3 READ List - All reads (AM, PM or HW points)
List 4 WRITE List - ONSTS, CURMOVE, CONSTRNT, DMCSTAT, LPTARG (AM points) List 5 LWRITE List - Controller Messages to the MAIN controller point (AM Point - Main) List 6 PWRITE List - Controller Setpoints (Regulatory control could be AM, PM or HW) List 7 AWRITE List - RESETCMD, STRTSTOP (AM Timer point)
List 8 RDWRT List - DMCON, CLPREQ, PRDSWC, ULIMIT,LLIMIT (AM Points - Main, sub and associated)
Keywords: cim-io, lists
References: None |
Problem Statement: User is running Cimio T_API to confirm connect to the data source through cimio. When running the test, he discovers that the first time through the values and retrievel are good. The rest of the values appear to be the same as the first. | Solution: Set list id = -1
This clears out the previously retrieved information and replaces it with current fresh information.
Keywords:
References: None |
Problem Statement: When a Petroleum Supply Chain application crashes, it can be useful to send the dump file to AspenTech after the crash for further investigation. How do I create this dump file? | Solution: 1. For Aspen Petroleum Scheduler (APS), Aspen Refinery Multi-Blend Optimizer (MBO), Aspen Olefins Scheduler (AOS) and Aspen PIMS, you may see a dialog showing the reason why the application is crashing. Press the PrtScn key to caputure a screen shot to record this information. E.g. (For some other applications, you do not see the below window).
Click on OK, and you will see the following window:
Â
2. Start Task Manager located at C:\Windows\SysWOW64\taskmgr.exe. This is a 32-bit program better for dumping our 32-bit applications. Do not use the one from Taskbar or Ctrl+Alt+Del.
3. Right-mouse-click your application and then Create dump file.
4. Send the screen shot, and the dump file to [email protected] to analyze the crash.
If you are able to identify the cause of the crash, please include that information. If you are not able to identify it, but the problem happens again, please capture the following information:
· Event Screen you were working with.
· How many screens were open simultaneously?
· Events you modified and how modified.
· Did you save the model when you closed it?
· Was the Excel file still open after the crash?
· What other programs were open during the crash?
· Anything else you consider relevant
Keywords: Crash, dump file, troubleshoot, PIMS, APS, MBO, AOS
References: None |
Problem Statement: This knowledge base article discusses LP Errors and Co-Linearity. | Solution: LP Error Code / Explanations
General Note 1:
DMCplus v1.0x.xx didn't allow any LP parameter changes.
General Note 2:
Problem: Changing relative costs between two MVs causes the
CV constraint to change from riding the limit to violating the limit.
Answer: Two possibilities are that either there is a very high ECE on that CV or there is a scaling inconsistency, i.e., rpm from the instrument but 1E+3 rpm in the model.
Another possibility is that there is a CV with a constraint rank of 1000. CVs with constraint ranks = 1000 have their constraints treated in an economic sense; that is, the constraints are violated if it makes economic sense to do so. The violation can be removed by decreasing the ECE for the constraint.
LP Parameters
LP matrix conditioning tolerance (EPSCND) (range 0.0 - 0.01, default 0.001)
The description in the entry dictionary is wrong. This value is used as a check on the condition of the dynamic control matrix.
LP matrix cost tolerance (EPSCST) (range 0.0 - 0.001, default 1.0E-08)
This tolerance is used by the LP to differentiate a significant change to the LP objective function when choosing a basis variable. If the reduced cost changes by less than this value, then the change is considered insignificant.
EPSCST : Do not change the value by more than one order of magnitude.
LP matrix divide by zero tolerance (EPSDV0) (range 0.0 - 0.001, default 1.0E-08)
This tolerance is used by the LP to ensure that division by zero does not occur during the calculation of the variable that must leave the basis. Do not change the value by more than one order of magnitude.
LPSolution feasibility tolerance (EPSFEA) (range 0.0 - 0.01, default 0.00001)
The LP uses this value to check that the LP constraints are feasible. The linear constraints are considered feasible if the difference between the calculated constraint and the actual constraint is less than this value.
Matrix inversion tolerance (EPSINV) (range 0.0 - 0.001, default 1.0E-15)
This value has a role similar to EPSDV0. It is used to check for zero pivoting and divide by zero in the dynamic controller calculations.
Solution reSolution tolerance (EPSRES) (range 0.0 - 0.00001, default 1.0E-08)
The entry dictionary value should be changed to ?LPSolution reSolution tolerance? This value is used in the LP as a check onSolution accuracy, and for determining the variable that must leave the basis.
Matrix sparseness tolerance (EPSSPR) (range 0.0 - 0.001, default 1.0E-15)
This entry is used to evaluate non-zero values. If the absolute value of a double precision floating point number is larger than this tolerance, then the value is considered significant. If not, then the value is considered to be zero.
LPSolution tolerance (LPSTOL) (range 0.0 - 0.01, default 0.01)
LPSTOL is used to ensure that the dynamic moves calculated by the controller will place the MVs at the desired steady state values calculated by the LP (within a tolerance of LPSTOL).
LP consistency weight (WTLP) (range 10.0 - 1.0E+05, default 5.00E+02)
Weighting factor for the dynamic matrix to ensure that the dynamic moves calculated by the controller will place the MVs at the desired steady state values calculated by the LP.
LP Post Processing Warning
2811, 2812
This does not turn off the controller.
This results from confusion in counting / comparing the number of constraints and the number of MVs. It is really a warning for an external targeting agent or some other application watching the number of constraints. It is not related toSolution validity, just a confusion in counting.
It can occur as a result of the external target resting at the variable high or low limit such that the ET constraint is counted again as a high or low variable constraint.
The message can occur when many MV's are at their limits which is the case when several MV's are OFF.
LP Internal Error
2995
The 2995 error specifically results from too many LP iterations and increasing MXLPIT above its default may alleviate the problem.
The default value for MXLPIT should be number of MVs X 100. Therefore, if you have 45 MVs, then MXLPIT should default to 4500.
MXLPIT should be increased above the default if there are many ranked sets of constraints, or several ETs are used, since each unique rank set and ET will result in a separate call to the LP. A rule of thumb is to increase MXLPIT by 50 percent. For the example below, increase MXLPIT from 4500 to 6750. If the LP requires more iterations still, there is most likely a problem with the controller gains. However, if the client has lots of CPU power, they may be able to get away with setting MXLPIT to a larger value. For some really badly conditioned problems, the LP will cycle forever, so changing MXLPIT won't solve the problem.
Note that there is currently a BUG in the way that MXLPIT is written to the CCF that prevents changing it in DMCplus BUILD - the format in the CCF is incorrect, as shown below: The variable type should be I4, and the number formatted as an integer, remove the decimal. You will have to do this after each edit in Build.
<incorrect as written by BUILD>
.MXLPIT~~~CONSTANT~~~R4~~~4500.~~~
<corrected, manually>
.MDLNAM~~~CONSTANT~~~CH(132)~~~Butane09.mdl~~~
.MXLPIT~~~CONSTANT~~~I4~~~4500~~~
.NORMOV~~~CONSTANT~~~I4~~~0~~~
LP Internal Error
2901, 2902, 2903
2921
2931 - 2937
2941 - 2947
2951 - 2957
2961, 2962
2971, 2972
2991, 2992
CV co-linearity is the likely cause of these errors.
Co-linearity results when one CV's steady-state gains are calculated from those of another CV, directly or indirectly. Two LP parameters can be adjusted to alleviate the problem. These are EPSDV0 and EPSCST with default values of 1.0E-08. Change EPSCST to E-07 or E-06. Change EPSDV0 to E-07. The co-linearity can be identified and the gains adjusted to eliminate the problem. See the notes below for the specific version of DMC or DMCplus.
DMC v5.02
Users should plan to upgrade to DMCplus V1.10. In the meantime, the model, the number of MVs and a sim file can be evaluated for short-termSolutions. Get a .eng file that shows the LP errors.
DMCplus v1.0x.xx
Users should plan to upgrade to DMCplus v1.10.
DMCplus v1.1x.xx
LP errors can be evaluated from .mdl and .dbg files. Users of previous versions should send .mdl file and number of MVs. Get .eng file or valid.err portions that show the LP errors. Send the files to [email protected]. Changes in CCF parameters EPSDV0 and EPSCST from the default of 1.0E-08 to either E-07 or E-06 may help with LP internal errors.
There is an LP tolerance factor called EPSRES that can be adjusted to reduce the occurance of type 2971 LP errors. This factor defaults to 1.0E-8, and can be loosened up to 1.0E-6. (similar to EPSCST, EPSDV0). The factor will not normally be in the CCF file; it has to be added by BUILD; or you could duplicate the line for adjusting EPSDV0, change the name to EPSRES and reload - the formats are the same.
LPSolution Fails Feasibility Check
2981, 2982, 2983
2999
Paramater EPSFEA is the feasibility check tolerance. It defaults to 1E-06 in DMC and 1E-05 in DMCplus. Change this to 1E-05 or 1E-04, respectively. The error may result from the use of high ECE (1,000 - 100,000).
(Version 1.1 added paragraph below.)
Check the parameter EPSDV0 as well. It has been noted that errors originally starting with near zero problems later show up as feasibility errors. If EPSDV0 is at its default of 1.0E-08, you may as well change that parmeter at the same time (increase it to 1.0E-07).
LPSolution Warning
2938
2948
2958
This has been removed from the v1.10 engine.
LP ExtendedSolution Warning
2928 (added v1.10)
2968
2978
2998
This does not turn off the controller.
It is related to re-tries. Initially, the LP uses some matrixSolution shortcuts. If the shortcuts don't give aSolution, the LP gives up on the shortcuts and inverts the matrix, etc, more frequently.
Keywords: LP Internal, Errors, LP Extended, DMCplus, Post Processing, EPSCST, MXLPIT, LPSTOL, EPSSPR
References: None |
Problem Statement: How to use a custom name for OrionVB.dll | Solution: While working with multiple models, we have to be sure that we have unregistered all other orionvb dlls and register the right one required for the model.
To avoid confusion, we could rename the orionvb.dll files for the different models. But we have to make sure that this dll is recompiled with the new name.
After the dll is recompiled with the new name, add a CONFIG table entry called VBDLL_NAME and set it the value of the renamed dll file.
The example below is the entry when the dll name is OrionVB1.dll
Note: this approach will work only if the DLL is recompiled with the new name.
Keywords: CONFIG, OrionVB, VBDLL_NAME
References: None |
Problem Statement: Explanation on how ranking affects the LP | Solution: .
LPSolution
In short, the LPSolution is in two sequential parts:
1.a. Feasible by CV rank order (1 - 999). This means that lower ranks get solved first then the higher ranks are added in. At each rank, if theSolution is feasible, then the limits become hard for the next rank. If not feasible, then the violated limit is moved to theSolution for that CV for use in the next rank. If multiple CVs are violated, then the ECEs are used.
Note: When I say that the limit for an infeasible CV is moved when that rank is solved, I mean that the new limit is a hard limit in the next rank to be solved. For example, a CV in rank 100 had limits of 1 and 2 but could only be solved at 3. The new limits for that CV become 1 and 3 when passing down to the next rank.
Regarding the effects of ramps in ranks, please note that the LP is solving for the ramp to balance, not be within limits. Balance may occur at strange ramp values in the LP. You should probably have ramps in their own rank and that should be the most important rank. The LP must solve a ramp or the controller aborts, with exceptions based on allowed imbalances.
The steady-state target shown for a ramp variable has nothing to do with the LPSolution. Since the LP is solving for the steady-state rate of change, the true LP Target for a ramp variable is always zero. This is not useful for the operator, so what the controller shows as the SS target is the end of the closed-loop prediction vector. This value may or may not be within the limits, largely depending on the dynamic ECEs for the ramp variable, and the move suppressions. For a ramp CV, a SS target outside the limits DOES NOT mean that the LP has failed to balance the ramp.
1.b. Feasible by ET rank order (1 - 999). Each CV (ranks 1 - 999) is feasible. Solve sequentially for each ET rank (i.e., each CV).Solution handling should be similar to that for the CV ranks.Solution is now feasible or has minimum infeasibility. The constraints are hard.
EconomicalSolution by introducing MV costs and rank 1000 CVs. TheSolution may or may not satisfy rank 1000 variables, based on LP equal concern errors and the feasible space. If any ETs are in use, then the entire set of LP costs and criteria (CST, LPCRIT) is replaced by the external target costs and criteria (ETMCST, ETMCRT).
LPSolution with CLP
If you have a CLP, then the CV ranks and ET ranks are replaced by CLP equivalents, but the process is analogous to that described above. CLP implementation is under development and is not supported for stand-alone implementation without close expert consultation from AspenTech.
Shadow Prices
DMC 5.x and DMCplus without the use of CV ranks always give shadow prices. When CV ranks are introduced into the problem, DMCplus provides shadow prices only when theSolution is completely feasible; any violated constraints inhibit shadow prices.
LP Targets
What is the meaning of the LP target value for an MV when:
Controller is OFF and all MV switches are ON, but not cascaded to DCS
Controller is ON and all MV siwtches are ON, but not cascaded to DCS
Controller is ON and MV's cascaded to DCS, but one subcontroller is OFF with MV switches ON (but not cascaded to DCS) for those MV's in the subcontroller that is OFF.
Controller is ON, MV switch is ON and cascaded to DCS, but one MV is ON but not cascaded to DCS.
Controller is ON, MV switch is ON and cascaded to DCS, but one MV is OFF and not cascaded to DCS.
Is the LP target value tracking, optimizing, or...?
First, some preliminaries. You can interpret the LP target for an MV by looking at the variable status for that MV.
If the status is...
the LP target is...
Good
calculated to minimize the objective function (optimized as you called it)
Ready
handled the same as Good
FFwd
set to the current input value (tracking as you called it)
Bad
handled the same as FFwd
When the controller is OFF, theSolution the controller calculates is what it would do if you turned it on. This applies not just to the LP targets, but to the move plans as well.
Ready is a special status which is used only when the controller is OFF. It means that all the parameters for that MV are valid except that LOOPST is off (i.e., the MV is not cascaded to the DCS). If you turn a controller on with an MV in this condition, its status will become FFwd because the controller cannot actually move it. Note that in this case theSolution will change when you turn it on, so you would not normally want to do this. If you're not planning to put an MV into cascade when you turn the controller on on you should set that MV's service switch to FFWD. Then theSolution you see before you turn on the controller will be consistent with the one you actually get after you turn it on.
With this information, all you need to know for your five scenarios are what status the MV will have in each:
Ready
FFwd
FFwd
the MV which is not cascaded will be FFwd, the others will be Good
Bad
Solution
See explanations above.
Keywords: DMCplus
References: None |
Problem Statement: How can I create the dll for my procedures with Visual Studio .NET 2003? | Solution: The most simple way is to let ACM do the work. You can have ACM generate a makefile with commands to compile and link the procedure code into a dll. When Visual Studio .NET 2003 is installed correctly (accessible from the command path, or detected by ACM), it will be used to compile and link.
In some cases, you may want to use the Visual Studio environment to develop the code, compile and link. This is also possible.
The important settings for the project are:
- in General, Configuration Type: select Dynamic library (.dll)
- in C/C++, General, add under Additional Include Directories the folder c:\program files\aspentech\AMSystem 2004.1\procedures
- in C/C++, Preprocessor, set the Preprocessor Definitions to WIN32;INTEL_X86;_NO_PROTO
- in Linker, General, Additional Library Directories, add c:\program files\aspentech\AMsystem 2004.1\procedures
- in Linker, Input, Additional Dependencies, add in front libsax.lib and libatdll.lib
An example is attached.
Keywords:
References: None |
Problem Statement: When exporting a model from ACM, you might want to invoke nmake manually over the intermediate files generated by ACM in order to troubleshoot any possible errors. If the following error appears:
nmake -f MakeUntitled command failed (return code 2) Is C++ (required to export compiled flowsheet) installed - Can''t find nmake.exe Are PATH and other environment variables set correctly? To see all errors, invoke the command manually
This | Solution: illustrates how to do this.Solution
1. In ACM go to Tools > Settings > Export Models and uncheck Delete intermediate files
2. Export your model and even if the process fails, you will notice that in the working folder some files will be generated, these contain the c++ source code and nmake script among others.
3. Open a cmd window (Start Menu and type cmd.exe ) and run the following command:
>C:\Program Files (x86)\Microsoft Visual Studio XX.X\Common7\Tools\vsvars32.bat (with quotes)
This sets the environment variables for Visual studio components including compiler and linker.
In the line above replace XX.X with the Visual Studio version installed in your machine
4. Navigate to the working folder where intermediate files are stored, in the line below use your own path:
>cd C:\SomePath
5. The nmake script will be a file with the name starting by Make followed by the ACM model name and with no extension, for example MakeHeater. Invoke nmake over the script file:
>nmake makemodel
Substitute the quoted name in the line above with the one for your model.
nmake program should return information on the assemble process of your dll including any possible error.
Invoking nmake manually can also be used to solved a compatibility problem with the version of ACM and Visual Studio installed, please seeSolution 147161. To register the extension manually seeSolution 141281.
Keywords: ACM, nmake
References: None |
Problem Statement: What open VMS mailbox is used in v1.7 and 1.8 of the interface. | Solution: The processes use the first available mailbox when they connect. Ie. it makes the Init_for_chip() call with blank parameters, and it is assigned the first available mailbox
Keywords: CIM-IO for Fisher-Rosemount CHIP
mailbox
References: None |
Problem Statement: How do Aspen Custom Modeler and Aspen Dynamics locate custom binary files such as dynamic linked library (DLL) for procedures or simulation access extension (SAX)? | Solution: Aspen Custom Molder (ACM) and Aspen Dynamics (AD) use the same approach as all other Windows program to locate a dynamic link library (dll), details please see http://msdn2.microsoft.com/en-US/library/ms682586.aspx
ACM/AD also will search in:
- the folder where the dynf/acmf file is located (GUI working folder)
- the simulation working folder (typically AM_simulation name created under GUI working folder
Finally, you can also specify the full path to the dll in the Simulation access extension window or in the procedure definition (LIBRARY: specified with full path).
The preferred option is to put the dll in the same folder as the dynf/acmf. The next best is to update the PATH environment variable. If you want to deploy smoothly the dll to various users, then the best is to put it in bin folder of ACM/AD (for example, c:\program files\aspentech\AMSystem 2006\bin).
Note that this is applicable only to dll loaded directly by ACM/AD. When the user code is dynamically loading other dlls, then it is the user's responsibility to work out how to load those dlls.
Keywords: DLL, dynamic link library, search, locate, binary, custom code, fortran
References: None |
Problem Statement: How to configure Aspen CimIO with Store & Forward when there are multiple Chip databases and one Aspen Cim-IO client. | Solution: Example configuration using multiple chip databases with one Aspen Cim-IO client.
CLIENT
Services
cimiochip
33000
cimiochipu1
33003
cimiochip_sc
33004
cimiochip_st
33005
cimiochip_fw
33006
According to the Aspen CimIO for Fisher Rosemount Chip user manual p.15-4 the DIOP's do not need to be defined on the client side except for the unsolicited.
Logical devices
iochip1 node1 cimiochip
iochip2 node2 cimiochip
iochip3 node3 cimiochip
SERVERS
node1
Services
cimiochip
33000
cimiochipr1
33001
cimiochipw1
33002
cimiochipu1
33003
cimiochip_sc
33004
cimiochip_st
33005
cimiochip_fw
33006
Logical devices
iochip1 node1 cimiochip
The logical devices file is not necessary on the server node unless you plan to run the cimio_t_api program. However, I always recommend it just in case.
node2
Services
cimiochip
33000
cimiochipr1
33001
cimiochipw1
33002
cimiochipu1
33003
cimiochip_sc
33004
cimiochip_st
33005
cimiochip_fw
33006
Logical devices
iochip2 node2 cimiochip
The logical devices file is not necessary on the server node unless you plan to run the Aspen Cimio_T_API program. However, I always recommend it just in case.
node3
Services
cimiochip
33000
cimiochipr1
33001
cimiochipw1
33002
cimiochipu1
33003
cimiochip_sc
33004
cimiochip_st
33005
cimiochip_fw
33006
Logical devices
iochip3 node3 cimiochip
The logical devices file is not necessary on the server node unless you plan to run the cimio_t_api program. However, I always recommend it just in case.
Keywords: store & forward
References: None |
Problem Statement: Installation of the DMCplus 1.10.04 offlines - when we start Model we've got the message: Security module not found. Most likely, the module has been plugged in the laptop's parallel port before the installation. | Solution: Try running the keysetup utility in C:\program files\DMCplus\support.
You may also need to reconfigure the port from EPP or Bidirectional to standard. This is usually done by entering the BIOS setup utility. (On some machines you press F1 or ESC during a reboot.)
Keywords: DMCplus, Model, Security, Dongle, Sentinel
References: None |
Problem Statement: After installing CIMIO for Chip on a VMS box the following error might occur when the DLGP process and DIOPS are starting up:
error=11 no more processes.
The write and unsol process that are configured for DIOP 1 refuse to start. The error message: Error = 11: no more processes is in the DLGP output file. | Solution: This error indicates that the VMS setup parameter for page file quotas was not large enough, which prevents CIM-IO processes from starting successfully. Since the read DIOPs start first they will use the available space, which would prevent write DIOPs from starting. To resolve this problem, in the MCR AUTHORIZE interface modify the PGFLQUO (page file quota) parameter and increase it to be large enough to allow the processes to start.
Another approach to solving the problem is to disable the DLGP spawning DIOPs, so that the cimio_chip_startup.com starts the DIOPs.
Also, make sure you are running the latest version of the interface.
Also make sure to have commands to setup env vars for CIM-IO and to set the default directory to the interface directory at the sys$login:login.com file of the user that starts the interface:
$ @dra3:[aspentech.cimio.commands]cimio_k_set_env.com
$ @cimioroot:[io.cio_fch_chp]cimio_chip_set_env.com
$ SET DEF CIMIO_CHIP
Asserting the directories in this manner is required only for DLGP spawning DIOPs. Otherwise the executables could not be found.
Following is an example of a good VMS account set-up. Note that PRCLM should be the number of processes that could be spawn by a process:
Primary days: Mon Tue Wed Thu Fri
Secondary days: Sat Sun
No access restrictions
Expiration: (none) Pwdminimum: 6 Login Fails: 0
Pwdlifetime: (none) Pwdchange: (pre-expired)
Last Login: 29-NOV-2001 09:55 (interactive), 28-NOV-2001 18:00 (non-interactive)
Maxjobs: 0 Fillm: 100 Bytlm: 512000
Maxacctjobs: 0 Shrfillm: 0 Pbytlm: 0
Maxdetach: 0 BIOlm: 150 JTquota: 4096
Prclm: 8 DIOlm: 150 WSdef: 16500
Prio: 4 ASTlm: 250 WSquo: 60000
Queprio: 4 TQElm: 10 WSextent: 64000
CPU: (none) Enqlm: 2000 Pgflquo: 200000
Authorized Privileges:
CMKRNL GROUP GRPNAM GRPPRV IMPERSONATE NETMBX
OPER PHY_IO PRMGBL PRMMBX PSWAPM READALL
SETPRV SHMEM SYSGBL SYSLCK SYSNAM SYSPRV
TMPMBX
Default Privileges:
CMKRNL GROUP GRPNAM GRPPRV IMPERSONATE NETMBX
OPER PHY_IO PRMGBL PRMMBX PSWAPM READALL
SETPRV SHMEM SYSGBL SYSLCK SYSNAM SYSPRV
TMPMBX
Keywords: no more processes
error=11
quota
References: None |
Problem Statement: What is Friction loss factor field for a compressor/Pump? | Solution: Friction loss factor is a factor parameter from the rotation mechanical energy and this is related to the rotation speed.It is used to model the rate of energy lost to mechanical efficiencies in Dynamic mode.
Duty Loss to Friction = friction loss factor * speed ^2, where the speed is with the unit radians/s.
This value is convert from speed in rps, speed = spd in rps * 2 *pi.
The friction loss factor in HYSYS should be possibly provided by vendor.
Keywords: Pump,Compressor,Friction Loss Factor,Inertia
References: None |
Problem Statement: How to access the 32-bit ODBC Data Source Administrator from the Start menu or Control Panel in 64-bit Windows, to create ODBC / file DSN Connection to work with Aspen Petroleum Schedular? | Solution: In 64-bit Windows, there are two ODBC Data Source Administrators: a 32-bit and a 64-bit variant.
The ODBC Data Source Administrator that is directly accessible in the Control Panel, under Administrative Tools (which is located in System and Security if your Control Panel is set to view by category), is always the same bitness as the installed Windows platform.
That is, on 32-bit Windows - it is the 32-bit administrator;
on 64-bit Windows - it is the 64-bit administrator.
On 64-bit Windows, you can execute 64-bit and 32-bit applications transparently.
Starting with 2010, Microsoft Excel is available in both 64-bit and 32-bit versions.
Versions priors to Excel 2010 are 32-bit only.
It is important to understand that 64-bit applications can only use 64-bit drivers and that 32-bit applications can only use 32-bit drivers.
64-bit or 32-bit. Data Source Names (DSN) are similarly divided by bitness, since a DSN is defined for a specific driver.
The 32-bit version of the ODBC Data Source Administrator must be used to configure data sources for 32-bit applications. This is the source of many confusing problems where what appears to be a perfectly configured ODBC DSN does not work because it is loading the wrong kind of drive
ANS: In order to create or modify 32-bit DSNs on a 64-bit Windows, you must run “C:\WINDOWS\SysWOW64\odbcad32.exe�
You may find it useful to create a shortcut to this on your desktop or Start menu if you access it frequently.
Keywords: DSN, DATA-SOURCE,32 BIT DSN, 64 BIT WINDOWS, ODBC, APS
References: None |
Problem Statement: How can I generate the temperature profile along the separator as a function of time? | Solution: In Aspen HYSYS Dynamics you can add the variables and create a strip chart. The strip chart can be used to track historical data from a simulation and observe trends in the system behavior. If you want to know more about strip chart in HYSYS please refer the knowledge baseSolution below.
How do I create a strip chart?
http://support.aspentech.com/webteamasp/KB.asp?ID=108911
You can include heat transfer in a vessel in the Heat Loss page via the Rating tab as shown below.
The recycle efficiency can be set less than 100% to represent non-equilibrium flash in the vessel. See knowledge baseSolution 111644 for more details on recycle efficiency.
How are phase holdup properties calculated when Recycle/Nozzle Efficiencies are < 100%?
http://support.aspentech.com/webteamasp/KB.asp?ID=111644
A strip chart showing temperature profiles around the separator has been created in the attached case. The user can change the temperature in the feed stream and observe the transient effects of temperature distribution around the separator.
Keywords: Temperature Profile, Heat Loss
References: None |
Problem Statement: How can I model equilibrium reactions in a dynamic model? | Solution: The issue with equilibrium reactions is that if you attempt to model this naively, you will end up with an index>1 formulation.
We can suggest three (3) ways to address this issue:
One is to conceptually replace the equilibrium reactions with fast forward and reverse reactions.
Another approach is based on the paper Dynamic model of a system with phase- and reaction equilibrium by Moe&al (Computers and chem. Engng vol 19, pp S513-S518, 1995). The mathematical details are a bit involved, but essentially the idea is to reduce the number of state variables by the number of equilibrium reactions. This can be done by applying a rectangular transformation matrix to the component holdups. This matrix is calculated from the singular value decomposition (SVD) of the stoichiometric coefficients.
The third approach is based on the one used in Aspen Dynamics version 2006.5 to support true electrolytes.
The attached simulation requires Aspen Dynamics and a custom modeling license. It compares four models:
RCSTR of Aspen Dynamics
a model (reactor_rate) using similar approach as in ART
a model (reactor) using the SVD approach
a model (reactor_rate2) using the AD True Electrolytes approach (of course this model contains no electrolytes)
You can see the results are very similar.
Rate approach (see reactor_rate model)
The material balance is:
// material balance
for c in componentlist do
$Mc(c) = in_f.F * in_f.z(c) - out_p.F * out_p.z(c) + sigma(foreach(rxn in nr_) SM(rxn,c) * rate(rxn));
endfor
The equilibrium relation would be written as:
Keq(rxn) = ConcProd(rxn);
where Keq is the equilibrium constant, and ConcProd the product of the concentrations of the components taking part in the reaction with the appropriate exponents. This is correct for a steady state model, but not for a dynamic model. It is precisely this equation that leads to the index>1. The index>1 comes from the fact we have some algebraic variables (rate) that are present only in dynamic equations.
The trick is to calculate the equilibrium reaction rates from:
rate(rxn) = K * (Keq(rxn) - concProd(rxn)) / Keq(rxn);
where K is a reasonably large constant (e.g. 10000). The larger K, the less we deviate from equilibrium. If we setting K to an extremely large value, Keq(rxn) would be equal to concProd(rxn) for any finite value of rate(rxn). What happens is that the rate takes the appropriate value to remain close to the equilibrium. It is a bit like a proportional controller, where ConcProd is the process variable, Keq the setpoint, rate the manipulated variable and K/Keq the gain. You can also see why this cures the index problem, as now we use the variable rate in algebraic equations.
SVD approach (see reactor model)
The mathematical details are a bit involved and will not be explained here. We recommend you read the paper referred above. For copyright reasons we don't distribute this paper.
We can only write nk = nc - nr differential material balances, where nc is the number of components and nr the number of equilibrium reactions. We call these nk state variables (Nc) the holdups of key components. Note that if we had no equilibrium reactions, we would be back to the simple material balance.
// material balance
for i in nk_ do
$Nc(i) = sigma(foreach(k in componentlist) TM(i,k)*(in_f.F*in_f.z(k) - out_p.F*out_p.z(k)));
endfor
The component holdups Mc are related to the key component holdups by the transformation matrix TM, which is made of the singular vectors of SM, the stoichiometric matrix.
for i in nk_ do
Nc(i) = sigma(foreach(k in componentlist) TM(i,k) * Mc(k));
endfor
The transformation matrix TM is calculated by the SVD subroutine.
call (TM, rank) = pCalcSVD_rightvectors(SM, 1, nk);
AD True Elec approach (reactor_rate2 model)
It is using a kinetic approach to model the equilibrium, although done in hopefully a smart way.
The idea is to introduce a new array of holdups, Mcp (p stands for pseudo). The actual component holdups are calculated from those Mcp with rateelec which is another array of reaction rates that essentially drive the composition to the equilibrium. The parameter Ktc is controlling how close we force theSolution to be near the true componentSolution (in AD, we set Ktc to 100). If you make Ktc small, you will observe larger deviations from equilibrium.
We also found it necessary to use Sigma(Mcp) in the energy balance, i.e.
E = Sigma(Mcp) * h;
instead of:
E = M * h;
In the equation:
Mc(i) = Mcp(i) * (1.0 + RateEquil(i)) + Tol_Mcp * RateEquil(i);
the first term is converting the Mcp to actual component holdups. The 2nd term on the right Tol_Mcp*RateEquil(component) was added to improve robustness when Mcp = 0. without this, the solver pushes the RateEquil value very high. This additional term eliminates the singularity. (In AD, we set Tol_Mcp to 1e-10).
Finally, the equilibrium reaction rates are constrained by the reaction extends (see at the end of the model reactor_rate2).
These models are difficult to solve the first time they are placed on the flowsheet. For the rate model, one trick is to set the rates to zero (e.g. set K to zero). You can then change back the value of K to 10000. For the SVD model, you can copy the values of the rate model.
One important point to note is that with the rate model, you may specify initial conditions that are far away from the equilibrium. This will force the integrator to take very small steps. This may be a disadvantage of thisSolution. With the SVD model, it is simply impossible to specify out of equilibrium initial conditions. You may find that sometimes the SVD based model takes fewer integration steps, but practically speaking the difference is not always very large.
The simulation shows that all models give very similar results.
Keywords: None
References: None |
Problem Statement: I would like to set up the data reconciliation case (special type of estimation case) using scripting or automation. | Solution: Please note that the scripting for the estimation has changed for version 2004.1 and 2006, so this method cannot be applied with version 12.1.
The attached example illustrates how this can be done.
Set Est = Simulation
Est.ResetEstimation
' Steady State Experiments
SET steadystateexperiment1=Est.newexperiment(SteadyStateExp_1,Steady State)
steadystateexperiment1.Weight=1
steadystateexperiment1.AddSSMeasurePoint Streams(cold-in).F,760,0.1
steadystateexperiment1.AddSSMeasurePoint Streams(cold-in).T,21,1
steadystateexperiment1.AddSSMeasurePoint Streams(cold-out).F,740,0.1
steadystateexperiment1.AddSSMeasurePoint Streams(cold-out).T,44,1
steadystateexperiment1.AddSSMeasurePoint Streams(hot-in).F,1050,0.1
steadystateexperiment1.AddSSMeasurePoint Streams(hot-in).T,59,1
steadystateexperiment1.AddSSMeasurePoint Streams(hot-out).T,23,1
Add experiments to the simulation
Est.AddExperiment(steadystateexperiment1)
In version 2004.1/2006, only the AddSSMeasurePoint on the experiment should be invoked. You no longer need the AddEstimateVariable (for a data reconciliation case) - this would cause an error message. If you try to drag & drop on the Estimation, Estimated Variable some fixed variable you will get the same error message. This is because now all variables are on the Measured Variables sheet of the experiment.
Keywords: None
References: None |
Problem Statement: When exporting model to msi format, ACM reports compiler not found error. | Solution: Please check the following items on your system:
1. You have the environment variable VS80COMNTOOLS defined.
Note, to check, go from Windows start menu,settings, control panel, system, under advanced tab, click on environment variables, you should see a list of them under system variables. Click on new button to add.
2. If it is defined then check that file vsvars32.bat exists in the directory defined by VS80COMNTOOLS
Note that these settings should be setup automatically if you had C++ compiler installed first and ACM later. Note also that the official compiler for ACM 2006.5 is Microsoft Visual Studio 2005, and if you use other versions of compilers or other C++ compilers all together, there could be unexpected and unsupported issues, which would be beyond scope of Aspen Tech support.
Keywords: VS80COMNTOOLS, vsvars32.bat, vcvars32.bat, cl, nmake, C compiler, C++, C++ compiler
References: None |
Problem Statement: How can I remove a variable selected on the OnLine Links feature of ACM using automation? | Solution: The documentation for the methods RemoveInputVariable and RemoveOutputVariable gives an example which is incorrect. Note that you need to know the name of the variable you want to remove. To do this, you need to keep a record of which variable is added. The attached example contains two scripts: one to setup the online links and another one to remove the variables. For ease of reference, the scripts are copied below.
Script to setup the on-line links
' OLL scripting
dim name_input(3)
dim tag_input(3)
name_input(1) = b1.theta
tag_input(1) = Simulated Card.Simulated Node.COS
name_input(2) = b1.beta
tag_input(2) = Simulated Card.Simulated Node.FLOAT
name_input(3) = b1.invalid
tag_input(3) = Simulated Card.Simulated Node.INVALID_I4
dim name_output(1)
dim tag_output(1)
name_output(1) = b1.alpha
tag_output(1) = Simulated Card.Simulated Node.FLOAT
Application.Simulation.RunMode = Dynamic
Set OLL = ActiveDocument.OnLineLinks
'Set up the server
OLL.ServerName = SST.SimulationOpcSvr.1
'Modify Period
OLL.UpdatePeriod = 200
'Set up the inputs
for i = 1 to 3
OLL.AddInputVariable name_input(i), tag_input(i)
next
'Set up the outputs
for i = 1 to 1
OLL.AddOutputVariable name_output(i), tag_output(i)
next
OLL.Enable = ON
Script to remove the variables
' OLL scripting
dim name_input(3)
dim tag_input(3)
name_input(1) = b1.theta
tag_input(1) = Simulated Card.Simulated Node.COS
name_input(2) = b1.beta
tag_input(2) = Simulated Card.Simulated Node.FLOAT
name_input(3) = b1.invalid
tag_input(3) = Simulated Card.Simulated Node.INVALID_I4
dim name_output(1)
dim tag_output(1)
name_output(1) = b1.alpha
tag_output(1) = Simulated Card.Simulated Node.FLOAT
Set OLL = ActiveDocument.OnLineLinks
'Set up the inputs
for i = 1 to 3
Set var1 = OLL.FindInputVariable(name_input(i))
OLL.RemoveInputVariable var1
next
'Set up the outputs
for i = 1 to 1
set var1 = OLL.FindOutputVariable(name_output(i))
OLL.RemoveOutputVariable var1
next
Keywords:
References: None |
Problem Statement: Why does performing a FILE SAVE-AS to a new file name cause recently modified parameters to revert to their former values? | Solution: The changes would be overwritten if a customer:
1. rewound to a time before the changes were made. e.g. running at t=5. make tuning changes. save-as. restart
2. Load latest snapshot on opening the simulation is checked (Snapshot Manager) when the simulation is opened - the latest snapshot will overwrite the changes.
Keywords: None
References: None |
Problem Statement: How to use structures with automation (VBscript or Excel macros)? | Solution: The attached example illustrates a few methods available in Aspen Custom Modeler with flowsheet scripts.
- how to create a structure instance
' shows how to create a structure
on error resume next
application.simulation.CreateStructure AntoineData, ADtest
if err.number <> 0 then
application.msg err.description
err.clear
' application.msg structure already exists
end if
The documentation of the DeleteStructure method is not clear enough at this stage to provide an example. You will have to delete the structure instance manually.
- how to access the value of variables and parameters in a structure instance
With the first method, you build the path to the structure object, starting from the Structure container (folder).
With the second method, you get the path to the structure object using the Resolve method.
' access data in a structure
' method 1
set s = Application.Simulation.Flowsheet _ .StructureContainers(Structures).Structures(plxData)
application.msg A = & s.A.value
vars = s.FindMatchingVariables(~)
for each v in vars
application.msg v.name & & v.value
next
application.msg ==============
' method 2 : using Resolve
set s1 = Application.Simulation.Flowsheet.Resolve(Structures.plxData)
vars = s1.FindMatchingVariables(~)
for each v in vars
application.msg v.name & & v.value
next
- how to specify the external structure in a block
This is very simple, since you just need to specify the full name of the structure instance in the external structure pointer, e.g. the declaration is plx as external AntoineData in the model used by the block B1.
' specify structure in block
B1.plx.value = Structures.ADtest
Keywords: None
References: None |
Problem Statement: How is it possible to improve the execution speed of scripts or automation calls? | Solution: Using the approach of ACMObj.Application.Simulation.Flowsheet.Resolve(S1.F) as the base, accessing with ACMFlo.S1.F takes 70% of the time, and accessing with ACMFlo.Resolve(S1.F) takes just 40% of the time (where ACMFlo = ACMObj.Application.Simulation.Flowsheet). The former is an expected and documented feature of VB, but surprisingly the Resolve method appears to be much quicker than the direct access of the variable.
Creating a new object that defines the path to what you are interested in saves time. You can even go further than ACMFlo.S1.F and create ACMFloS1 = ACMFlo.S1 if you are accessing multiple variables in S1.
Resolve is generally faster because ACM is resolving the path more efficiently than VB. VB needs to do it one step at a time talking across processes via COM at each step. With resolve there is only one such call. For each of your examples the number of COM calls is as follows:
4 calls for ACMObj.Application.Simulation.Flowsheet.Resolve(S1.F)
2 calls for ACMFlo.S1.F
1 calls for ACMFlo.Resolve(S1.F)
Keywords: None
References: None |
Problem Statement: How to import depressuring utility peak flow results into Aspen Flare System Analyzer V7.3 | Solution: The composition, temperature, pressure and mass flow results of the depressuring utility peak flow can be imported directly as the data for the sources in the flare system networks built in Aspen Flare System Analyzer.
1. Go to AspenOne and Select Import Sources.
2. In Import HYSYS Sources View, select Dynamic Depressuring Utilities when importing the Aspen HYSYS.
3. Click the Browse button to find the Aspen HYSYS file with your depressuring utility.
4. Click Open button to import the Aspen HYSYS data into Aspen Flare System Analyzer.
Keywords: depressuring utility, peak flow, import source
References: None |
Problem Statement: How are zwitterions defined with ENRTL-RK? Could you provide an example? | Solution: The ENRTL-RK model supports zwitterions, compounds with both positive and negative charges but net charge of zero.
Zwitterions are defined as:
Components of type Conventional, similar to solvents;
Parameter ZWITTER set to 1; all other components in the Aspen Physical Property System default to zero for ZWITTER;
Parameter PLXANT/1 less than -1.0E20 so that they are non-volatile.
Zwitterions are handled as follows:
A zwitterion interacts with other molecular species through NRTL parameters only, excluding any interactions through Henry constants and pair parameters;
The activity coefficient is calculated as a solvent;
The contribution from zwitterions to theSolution enthalpy and Gibbs free energy are calculated as solutes using the infinite dilution heat capacity model CPAQ0, DGAQFM, and DHAQFM.
Then, a zwitterion will be declared as a Conventional similar to solvents but is distinguished from solvents with a value of unity for the parameter ZWITTER; all other components in Aspen Plus default to zero for ZWITTER. The zwitterion interactions with other molecular species are accounted for with NRTL binary interaction parameters.
The activity coefficient for the zwitterions is calculated as a molecular solute. The contributions from the zwitterion to theSolution enthalpy and Gibbs free energy are calculated as molecular solutes using the aqueous phase Gibbs free energy and heat of formation at infinite dilution and 25A?C, DGAQFM and DHAQFM, and the infinite dilution heat capacity, CPAQ0. PLXANT/1 is set to -1E20 for zwitterions so that they are non-volatile.
Please also see the Aspen Plus Help at the topic Zwitterions in Symmetric and Unsymmetric Electrolyte NRTL.
A useful example on how to define zwitterions in Aspen Plus V7.3.2 is the 'Rate_Based_PZ+MDEA_Model_v732.bkp' attached. In this example you can find a Rate-Based Model of the CO2 Capture Process by Mixed PZ and MDEA.A This file is part of the 'Aspen Plus Carbon Capture Models August 2012 Update'.
Please open the example 'Rate_Based_PZ+MDEA_Model_v732.bkp' and for the definition of the zwitterion HPZCOO, please see the USRDEF tab from Properties Environment | Methods | Parameters | Pure Components (in the print-screen below):
As one can see, for the component HPZCOO the ZWITTER parameter is set to 1 for specifying that this is a zwitterion and then it has both positive and negative charges, while the CHARGE parameter is set to 0 because the net charge of the zwitterion is zero. The DGAQFM and DHAQFM parameters should also be defined.
On the other hand, for the other components one should set either the positive or negative charges through the CHARGE parameter, and set the ZWITTER parameter to 0 (default value for ZWITTER parameter).
For more details on this simulation please download the 'Rate_Based_PZ+MDEA_Model_v732.pdf' guide attached.
Keywords: zwitterions
References: None |
Problem Statement: How do I import Aspen HYSYS Depressuring Utility peak flow results into Aspen Flare System Analyzer V7.3? | Solution: From Aspen Flare System Analyzer V7.3, the composition, temperature, pressure and mass flow results of the depressuring utility peak flow can be imported directly as the data for the sources in the flare system networks.
1) go to AspenOne button->Import Sources->HYSYS Sources
2) Instead of choosing Streams, you should use option: Dynamic Depressuring Utilities before you load your case.
Make sure that your Aspen HYSYS depressuring utility already solved and had the peak results.
Keywords: Depressuring utility, peak conditions, Flare System Network.
References: None |
Problem Statement: What is the Fix Qualities of Zero Flow Pools to Initial Values option? | Solution: This is found on the Advanced 2 tab of the Non-Linear Solver (XSLP) Settings. It is only applicable to Aspen PIMS Advanced Optimization (PIMS-AO).
This setting is used to control the value of stream qualities saved in theSolution file when the stream to which they apply has a zero flow. Since there is no stream flow there are really no qualities. But in order to use theSolution file to initialize a future run - in which the stream may not have a zero flow it is important to have a reasonable value for this quality. If this box is not checked the values will commonly end up being initialized to the minimum or maximum of the quality. If the box is checked it will be set to the last value it had when this stream had a non-zero flow.
While the default for this option is unchecked, it is suggested that this setting be checked.
Keywords: None
References: None |
Problem Statement: Why does the column not converge after adding water in component list even though the fraction is zero? | Solution: The reason column doesn't converge after adding water in the component list is, because HYSYS considers water is present, even though the fraction is zero. HYSYS does it because column solves iteratively. In one iteration water flow can be zero, and in the next iteration water can be present. To keep the iterations simple column solver(s) considers presence of water right after addition of it, into the component list.
So, if you add water in HYSYS component list later and the column doesn't converge, you may either turn on the Auto Water Draw under 2/3 Phase Or you may add water draw in stages or in the condenser (using 3 phase condenser).
Keywords: Column water draw, Column Convergence,
References: None |
Problem Statement: What is the tag name addressing format for DDP (Detail Display Parameters) access? | Solution: As stated in the CIM-IO for Fisher-Rosemount CHIP Interface User's Manual, the tag name addressing format is:
tagname.DDP#<DDP number>
The example in the manual is correct, however some customers have had trouble because they didn't realize that the string is case sensitive. Therefore DDP must be uppercase, just as it appears.
Keywords: ddp
case sensitive
usage
detail display parameters
References: None |
Problem Statement: The DLGP fails to spawn the DIOPs
If the current directory is not CIMIOROOT:[io.cio_fch_chp] the startup fails to start the DIOPs. | Solution: There are twoSolution to this issue.
#1
Add the following statement at the sys$login:login.com file for the account that starts the cio-fch-chp interface.
$!--- sys$login:login.com
$! .... put the following line close to the end so that it executes last
$ @dkaxxx:[cimio.commands]cimio_k_set_env.com
$ set def cimioroot:[io.cio_fch_chp]
$!----end ----
#2
Disable autostart of diops by the dlgp,
Edit the cio_fch_chp.cfg and set the autostart diops option to 0.
Keywords: CIM-IO for Fisher-Rosemount CHIP
diop
cio_fch_chp.cfg
References: None |
Problem Statement: Should the values for NB_TAGS and NB_UNSTAGS be set to reflect the number of CHIP points or IP.21 points? | Solution: The config file setting for NB_TAGS and NB_UNSTAGS is for IP.21. It creates memory to be shared by the DLGP and DIOP processes. This number is the size of shared memory the bridge allocates for tags and represents the amount of tags you wll be using from the CHIP system. The closer to the actual amount of tags used the less space is wasted in memory.
The CHIP points are limited to 10,000 by DBI, but creating a number larger than 10 000 will cause no problems as it does not reflect the limit in the CHIP database.
Keywords: Fisher
Rosemount
cio_fch_chp.cfg
References: None |
Problem Statement: How to implement load balancing with the Aspen Cim-IO for Fisher-Rosemount Chip interface. | Solution: There are three different approaches to managing the Read DIOPs of this interface.
1. A single DIOP (no load balancing)
2. 2 to 5 DIOPs (load balancing off)
3. 2 to 5 DIOPs (load balancing on)
Scenario 1: A single DIOP (by definition, no Load Balancing):
If this is a small system (i.e. few tags, slower scan rates) one read DIOP might handle the tag load just fine. In that case, configure one Read DIOP as follows:
cio_fch_chp.cfg:
LOAD_BALANCE = 0;
DIOP Unit Section (bottom of file)
UNIT#1 set READ=Y and all other DIOPS to N; UNIT#2 set all DIOPS to N.
UNIT# = 1
READ = Y
WRITE = Y
UNSOL = N
HISTORY = N
Services file:
Note that the Unit example above has corresponding service entries for parameters set to Y:
CIMIOCHIPR1
3301/tcp
CIMIOCHIPW1
3302/tcp
Transfer Record(s) in Aspen Infoplus.21:
IO_DEVICE_UNIT: 1
By specifying IO_DEVICE_UNIT to 1, the Get Record will depend by default on the one configured Read DIOP.
Scenario 2: 2 to 5 DIOPs (Load Balancing Off)
To share the scan load between several DIOPs, specify which transfer records use which DIOP. Up to 5 Units per DLGP service can be configured. Here are some example parameters for that scenario using two Units, one with a Read/Write DIOP, and a second with a Read DIOP:
cio_fch_chp.cfg:
LOAD_BALANCE = 0;
By setting LOAD_BALANCE to 0, you need to manually configure each Get Record to use a configured Unit.
DIOP Unit Section (bottom of file);
UNIT#1 set READ=Y, WRITE=Y and all other DIOPS to N; UNIT#2 set READ=Y and all other DIOPS to N.
UNIT# = 1
READ = Y
WRITE = Y
UNSOL = N
HISTORY = N
UNIT# = 2
READ = Y
WRITE = N
UNSOL = N
HISTORY = N
Services file:
Service entries for every configured Unit:
CIMIOCHIPR1
3301/tcp
CIMIOCHIPW1
3302/tcp
CIMIOCHIPR2
3303/tcp
Transfer Record(s) in Aspen InfoPlus.21:
Set IO_DEVICE_UNIT to 1 or 2 for all Get Records to specify which DIOP to use for Read Requests.
GETREC1, IO_DEVICE_UNIT = 1
GETREC2, IO_DEVICE_UNIT = 2
GETREC3, IO_DEVICE_UNIT = 2
Scenario 3: 2 to 5 DIOPs (Load Balancing On)
If this is a busy system then configuring up to 5 units is supported, and by turning Load Balancing on, the system can dynamically allocate transfer record requests to the DIOP of its choice, giving the newest request to the least busy DIOP.
Here are some example parameters for that scenario. Configuring three Units, with a Read/Write DIOP on Unit 1, and Unit 2 and 3 with Read DIOPs.
cio_fch_chp.cfg:
LOAD_BALANCE = 1;
By setting LOAD_BALANCE to 1, you will allow your Get Records to use any configured unit.
DIOP Unit Section (bottom of file);
In the DIOP Unit section (bottom of file), UNIT#1 set READ=Y, WRITE=Y and all othe DIOPS to N; UNIT#2 set READ=Y and all other DIOPS to N; UNIT#3 set READ=Y and all other DIOPS to N.
UNIT# = 1
READ = Y
WRITE = Y
UNSOL = N
HISTORY = N
UNIT# = 2
READ = Y
WRITE = N
UNSOL = N
HISTORY = N
UNIT# = 3
READ = Y
WRITE = N
UNSOL = N
HISTORY = N
Services file:
Service entries for every configured Unit:
CIMIOCHIPR1
3301/tcp
CIMIOCHIPW1
3302/tcp
CIMIOCHIPR2
3303/tcp
CIMIOCHIPR3
3304/tcp
Transfer Record(s) in Aspen InfoPlus.21:
Set IO_DEVICE_UNIT to 0 for all Get Records. This allows the system to control which DIOP will get each Read Request.
GETREC1, GETREC2, GETREC3, etc:
IO_DEVICE_UNIT = 0
Keywords:
References: None |
Problem Statement: The documentation of software requirements for ACM model exports states that:
You must have a C++ compiler?Microsoft® Visual C++ Studio Net is recommended for exporting models on machines running Microsoft Windows. Your C++ compiler must be configured so that it can be used from the command line.
Can Microsoft Visual C++ .NET 2003 Standard be used instead of Visual Studio .NET (or .NET 2003) be used instead? Microsoft Visual C++ .NET 2003 Standard is much cheaper. | Solution: We now have successfully tested Microsoft Visual C++ .NET 2003 Standard, on a Windows 2000 system, using AES 12.1.1 products. No problem occurred.
Therefore, Microsoft Visual C++ .NET 2003 Standard is suitable to export models from ACM 12.1 to be used in Aspen Plus 12.1 or HYSYS 3.2.
Note The C++ compiler is only needed to create the DLL when the model is exported from your Aspen Modeler product. A DLL can be used in Aspen Plus or HYSYS without having a C++ compiler.
Keywords:
References: None |
Problem Statement: This knowledge base article describes how to interpret the data which is in a normal diop startup message. Diop messages are written to the cimio_msg.log file. | Solution: This message will be interpreted as an example.
Thu Feb 22 00:34:18 2007, Logged by CIMIOCHIPW2 on node HKLAP.ASPENTECH.COM:
Starting Up CIMIO CHIP WRITE DIOP Unit #2 Service:CIMIOCHIPW2 DbgLvl:0(0) Pid:4158 (0x103e) Version:@(#)REL_CIOFCHCHP_V2.3.1_EA Build Date:Nov 9 2001 22:08:18 CIM-IO Kernel Version: Linked:4.9.8.1, Compiled:4.9.2
This message means the following:
- The message was logged by the dlgp process named CIMIOCHIPW2
- It is running on the machine with the fully qualified domain name of HKLAP.ASPENTECH.COM
- The write diop is in the process of Starting Up
- There is no additional Aspen Cim-IO for Fisher Chip interface-level debugging enabled (DbgLvl:0)
- The process ID of the write diop which is starting is 4158. The hexadecimal equivalent of 4158 is 0x103e.
- The version of the Cim-IO for Fisher Chip interface is V2.3.1_EA Build Date:Nov 9 2001 22:08:18
- The version of the Cim-IO kernel is 4.9.2
Keywords: Dissect
understand
log
shutdown
start up
References: None |
Problem Statement: Which discretization method to use for counter current flows in partial differential equation models? | Solution: The attached example is a simple plug flow reactor with counter current cooling and shows how this can be done. The trick is to use two domains, one for each flow direction. As the discretization has to be against the flow direction, one domain will use the BFD1 method (for the process side), and the other the FFD1 (forward finite difference for the counter current coolant flow).
The example also contains the same model implemented without using the PDE features of ACM. You may find useful to study that model to understand better the way the nodes are indexed and how to use successfully the FOR...ENDFOR loops to handle equations coupling distributions declared over different domains.
Keywords: None
References: None |
Problem Statement: You can select the units of measurement for display in Tables in Aspen Custom Modeler (ACM) that are different than the global UOM set. However, when you create a Profile Plot of a variable that has the units of measure changed, the plot reverts back to the global units for the variable. How do you change a Profile Plot's units of measurement for the variable displayed. | Solution: Unlike other types of plot changes to the display, the UOM for individual variables within the profile is not reflected in the Profile Plot. Note: If you display the Profile Plot as a Profile Table, the change in the variables UOM is displayed. You can only change the UOM for a profile variable in the plot by changing the display UOM set.
Keywords: profile plot
units of measure
UOM
References: None |
Problem Statement: Is there a library of icons for Aspen Custom Modeler models? | Solution: No, there is no official library. Attached is a zip file with a library of icons created from the icons available in Aspen Dynamics library for heat exchangers, pumps, compressors, columns, reactors, etc. You can use those icons in your simulations. To use those icons, go to the File menu, Import Types, and select the file icons.acmf. You can then copy and paste the icons from the icons model into your custom models. You can then adjust the universal port location, or use the port browser to locate the ports of your model to the appropriate locations.
Keywords: None
References: None |
Problem Statement: I have installed Visual Studio C++ .NET 2003 and Intel Fortran 9.0 for Aspen Custom Modeler 2006. I have also installed Visual Studio C++ 2005 for ACM 2006.5. However, when I try to compile a fortran procedure in Aspen Custom Modeler 2006.5, Aspen Custom Modeler sets up to use Visual Studio C++ 2003. Why? | Solution: This occurs only if one of the procedures you are compiling is implemented in FORTRAN.
For 2006.0, Aspen Custom Modeler invokes the batch file supplied with the intel fortran tool under c:\program files\Intel\Fortran\9.0\IA32\bin\ifortvars,bat. It is this batch file that invokes the setup of Visual Studio 2003 with the statements:
@call C:\Program Files\Microsoft Visual Studio .NET 2003\Vc7\Bin\Vcvars32.bat
title Intel(R) Visual Fortran Compiler 9.0.028 Build Environment for 32-bit applications
Thus you end up with the correct configuration.
If Intel Fortran 8.0 is installed, Aspen Custom Modeler will call the batch file c:\program files\Intel\Fortran\8.0\IA32\bin\ifortvars.bat.
Finally, if no procedure is implemented in fortran, then ACM calls the batch file c:\program files\Microsoft Visual Studio .Net 2003\Vc7\Bin\Vcvars32.bat.
With ACM 2006.5, when all procedures are in C/C++, then ACM invokes c:\program files\Microsoft Visual Studio 8\Vc\Bin\Vcvars32.bat (which is Visual Studio 2005). If some procedures are in fortran, then ACM calls the batch file c:\program files\intel\Fortan\9.1\IA32\bin\ifortvars.bat (intel 9.1) or c:\program files\intel\Fortran\9.0\IA32\\bin\ifortvars,bat. Those batch files typically set up the Visual Studio environment 2005.
However, there is a problem with the following installation history:
install Aspen Custom Modeler 2006
install Visual Studio 2003
install Intel Fortran 9.0
install Aspen Custom Modeler 2006.5
install Visual Studio 2005
The problem is that the C++ 2003 is not supported by Aspen Custom Modeler 2006.5 (and C++ 2005 is not supported by Aspen Custom Modeler 2006), so we end up with a link failure if we try to compile a fortran procedure in Aspen Custom Modeler 2006.5. We get the following message:
Building DLL...
Intel(R) Visual Fortran Compiler 9.0.028 Build Environment for 32-bit applications Copyright (C) 1985-2005 Intel Corporation. All rights reserved.
Setting environment for using Microsoft Visual Studio .NET 2003 tools.
(If you have another version of Visual Studio or Visual C++ installed and wish to use its tools from the command line, run vcvars32.bat for that version.) ..
Stop.
nmake -f MakeUserCode command failed (return code 2) Are required languages installed and environment variables set correctly?
To see all errors, invoke the command manually
1 error , 0 warnings generating procedure code. See above for more details.
The problem is not so much that ACM 2006.5 is calling the wrong batch file (it calls the fortran setup), but that this fortran batch file itself calls the wrong Visual Studio setup. The work around is to modify the file c:\program files\Intel\Fortran\9.0\IA32\bin\ifortvars,bat and change the Statement
@call C:\Program Files\Microsoft Visual Studio .NET 2003\Vc7\Bin\Vcvars32.bat
to
@call C:\Program Files\Microsoft Visual Studio 8\Vc\Bin\Vcvars32.bat
Note that you will have to undo the change to create procedures in Aspen Custom Modeler 2006.
Keywords: coexistence
References: None |
Problem Statement: Sometimes user needs to populate a TimeData block from Excel. | Solution: It is quite straightforward to populate a TimeData block from Excel into ACM/AD.
Here is sample code:
Private Sub CommandButton1_Click()
Dim acmobj As Object
Set acmobj = CreateObject(ACM Application)
acmobj.Application.Visible = True
acmobj.OpenDocument (c:\temp\acm\timedata.acmf)
Set flowsheet = acmobj.Application.simulation.flowsheet
For i = 0 To 99 'number of points minus 1
'36 is the sample time
flowsheet.B1.x(i * 36).Value = Sheet1.Cells(i + 1, 2).Value
Next i
End Sub
Where your data would reside in 2nd column of sheet1 in Excel.
You would of course adjust number of points and sample_time for your case.
Keywords: TimeData, Excel, Fill in
References: None |
Problem Statement: This | Solution: shows a usage example of new parameter with new units used in an ACM modelSolution
The sample ACM model has a new quantity Permeability and its unit of measures defined and a parameter is created to use the newly defined quantity.
Open up the model; look at the code in script tts under flow sheet folder. It defines the new quantity Permeability and its unit of measures. Look at the new parameter ttsp under custom modeling folder/parameter types; it defines the ttsp parameter referring to Permeability quantity. Look at model ttsm under custom modeling/models where the new identities are used there.
Notice the all variable table of block B1, the unit of its mypermeability parameter only shows literal Permeability since script tts has not been run yet upon loading the model, and as such the new quantity has not come to life yet.
Now invoke script tts, close the all variable form of block B1 and open it up again, observe the unit of mypermeability is punitm now.
Go to menu tools l Units of Measurements and select ENG, notice its unit is now punitk.
Please see associatedSolutions for UOM system default settings and other details.
Keywords: UOM, quantity, physical type, units
References: None |
Problem Statement: Installing PIMS 2006 (v17.1) components breaks solver in Custom Modeler 2004.1 (v13.2) or other modeler based products (Aspen Dynamics, Aspen BatchSep, Aspen Chromatography, Aspen ADSIM).
Installing version 2006 of Aspen Custom Modeler, Aspen Dynamics, Aspen BatchSep, Aspen Chromatography, Aspen may also cause version 2004.1 of those products to fail in the same way.
Installing version 2006.5 of Aspen Plus or its patches may also cause Aspen Custom Modeler 2004.1 or other modeler based products to fail. | Solution: To fix the 2004.1 version of the product, follow the procedure outlined after installing the product:
1. Click on the Windows START button
2. Click on the RUN option
3. Type in cmd in the Open text box
4. Within the DOS window type: cd c:\Program Files\AspenTech\Aspen Open Solvers 12.3\Bin
5. Enter the DOS command Regsvr32 /u Decompserver.dll
6. Re-enter the command without the /u as in: Regsvr32 Decompserver.dll.
Keywords: AspeOne 2004 Custom Modeler
AspenOne 2006 PIMS
ACM
References: None |
Problem Statement: Sometimes it is desirable to call up a sub model in a loop, do some calculation and save only a subset of the results from the sub model. | Solution: Sub models such as property models are instantiated, rather than called. For example,
nstage as integerparameter (3); //the length
stageflashes([1:nstage]) as props_flash2; // you are declaring an array of sub models actually
The instances of a sub model have their own memory allocated respectively, thus all the variables used in the instances spend memory space and it would not be possible to call upon a sub model in a loop like the common practice in a procedural language such as C or Fortran.
In another word, if procedural run is desired, one would need to create procedures rather than sub models.
To see the differences in terms of the number of equations and variables required by sub model usage and procedure usage, please see attached ACM sample model where only one of the last two lines of code should be active.
If the procedure code is used (the default), by looking at the flow sheet properties, one would find out the number of equations is under 10. Now comment out the procedure line and make the sub model live active by removing the leading literals //, recompile and rerun the case, check the flow sheet property, one would notice the number of equations and variables were increased considerably. Notice the 2 approaches are carrying out similar calculations, but the sub model approach requires more equations and variables.
The underneath reason is that the procedure actually hides the logic in lower level language and reduces the number of variables required to be explicitly in the ACM model.
Also, for an example of defining a procedure step by step, please seeSolution 120202.
Keywords: Procedure, sub model, instance
References: None |
Problem Statement: Are there any limitations in passing string parameters to procedures? | Solution: String parameters are not limited in length in ACM. However, the length is limited when you pass a string parameter to a procedure. Only the first 39 characters will be accessible as an internal buffer of 40 characters is used internally and string uses a null character to mark its end.
If you need to pass a longer string to a procedure, you have to split it into a string parameter array.
Note also that it is possible to return a string parameter from a procedure. However, the parameter must be an input argument (because only variable types are allowed as outputs of a procedure). The procedure must be declared with the options precall and changesinputs as explained in the on-line help. You can modify the value of the parameter during the precall call. You must make sure the string you return is 39 characters or less, otherwise you will cause an access violation (crash) to ACM simulation server.
Keywords: None
References: None |
Problem Statement: I need to evaluate the sensitivities of variables with respect to some fixed or initial variables, as preparation work for estimation or optimization. How can I do that? | Solution: This example shows how this can be done using the ACM automation interface. We use Excel as a user interface for those calculations.
To use the example, you need to open the spreadsheet EvalSensitivities.xls (Excel 2003). You can then set the name of the independent and dependent variables, the time step to use and the end time. When you click the Open button, you can select an ACM or an Aspen Dynamics simulation. Note the code is set to start version 2004.1. Modify the code if you want to use another version.
When you click the Run button, the simulation is started. First we add the variables to the sensitivity calculation (with AddSensitivityParameter and AddSensitivityVariable), then we do an initialization run. We also repeat as many dynamic steps as required. The results are recorded on the sheets Time data and the sensitivities returned by GetSensitivityValue are written to the Results sheet. Two plots have been created for the ACM example, but if you modify the variables or other settings you will have to recreate the plots.
The sensitivity shows how a variable depends on another. For a steady state case (not handled by this example), the sensitivity says that if you change the fixed variable by 1, the dependent variable will change by the value of the sensitivity. In other words, a large sensitivity means the variable will change a lot.
The same applies for dynamic simulation, although in this case we have a value of the sensitivity which is function of time instead of just a single value. Moreover, the independent variable may also be an initial variable.
The sensitivity will show how the calculated variables depend on the estimated parameters. For example, it can help you detecting the case where an estimated parameters has little influence over the measured data.
Keywords:
References: None |
Problem Statement: Is it possible to retrieve the predicted values after an estimation run? | Solution: The predicted values for the measured variables are displayed on the Estimation graphical interface on the Experiments. The predicted values are also printed to the simulation messages window.
It is also possible to retrieve those data using the automation interface. The zip file attached contains two examples illustrating how to retrieve the predicted values after an estimation run. The first example (test-estimation.acmf) is a steady state estimation case. To run the example, open the simulation file in ACM 2006.5, then run the estimation run, then invoke the flowsheet script show. This should write to the simulation messages window the following lines:
exp1 0.75
exp2 1.7
exp3 2.65
This is the script:
' show the predicted values
dim expname(3)
expname(1) = exp1
expname(2) = exp2
expname(3) = exp3
for i = 1 to 3
y = Application.Simulation.GetEstimationPredictedValues(expname(i), B1.y)
application.msg expname(i) & & Cstr(y(0))
next
Note the GetEstimationPredictedValues is available only in ACM 2006.5 and above. Note also that this method returns an array with the predicted values, even if there's only one value (as in the example above). This is why we have to use y(0) to get the numerical value (y is an array in this script).
The second example is a dynamic estimation. To run the example, open the spreadsheet Reactor-data.xls, go on the Sheet2. Click the Open button to open the simulation (reactor-ans.acmf, a simple batch reactor example), then run the estimation in ACM. When the estimation run is complete, click the Demo button on the spreadsheet. This will retrieve the predicted values for the first dynamic experiment.
Private Sub cmdDemo_Click()
On Error GoTo ErrorHandler
If ACMDocument Is Nothing Then
MsgBox You need to start ACM with the Open Simulation button first, _
vbInformation, cmdRun
Exit Sub
End If
' get dynamic predicted values
v = ACMSimulation.GetEstimationPredictedValues(DynamicExp_1, Rct.CA)
For i = 0 To UBound(v)
Sheet2.Cells(i + 2, 6) = v(i)
Next
Exit Sub
ErrorHandler:
Application.Cursor = xlDefault
MsgBox Error & CInt(Err.Number) & : & Err.Description, vbCritical, _
cmdRun_Click
End Sub
There are some limitations with the automation interface, as there is no method to retrieve the name of the experiments, nor to retrieve the name of the variables that were measured in each experiment. Finally, there is no method to retrieve the time corresponding to the dynamic measure. You will therefore have to design a way to get that information, e.g. ask the user of your script. If you set up the estimation case completely with automation/scripting, you will need to record the name of the measured variables as well as the name of the experiments used.
Keywords: None
References: None |
Problem Statement: I am exporting a model from Aspen Custom Modeler to be used in Aspen Plus, using the model export feature. However, I am confused as the modifications made in the Aspen Custom Modeler model don't seem to be picked up correctly in Aspen Plus. | Solution: When you export a model from Aspen Custom Modeler, the installation script is automatically launched and prompts you if you want to install the model. This installation script detects if there was a previous installation of the same model. In that case, it gives you two options:
Repair
Remove
The repair option will not install the new model. It will only restore the old model, in a similar way to the restore option in the operating system (e.g. to repair a corrupted file, registry, etc). Most likely this is not what you want.
You should use the remove option instead. This option will remove the previous installation of the model. You should then locate the msi file in the Explorer and double click it to proceed with the installation of the updated model.
Keywords: None
References: None |
Problem Statement: Where can I define my own units of measurement for Aspen Custom Modeler, Aspen Dynamics, etc? | Solution: See under Automation reference in ACM on-line help for detailled information. See alsoSolution 113152 for basic information. The purpose of this note is to explore the alternative locations where the units can be defined.
The units are defined in the file OnNewDocumentScript.vb located in c:\program files\aspentech\AMSystem x.x\bin for versions up to 2004.1 (included) and in OnSystemNewDocumentScript.vb for versions 2006 and upward. These scripts are executed automatically when a file is opened or a new simulation is created. We have introduced this OnSystemNewDocumentScript.vb in version 2006 to allow a user to separate the system changes to the custom user changes. We recommend custom user changes should be implemented only in OnNewDocumentScript.vb and the file OnSystemNewDocumentScript.vb to not be modified. Note that since OnNewDocumentScript.vb is invoked before OnSystemNewDocumentScript.vb, if you want to create a new physical quantity in OnNewDocumentScript.vb in version 2006, you need to put the statements to declare the UOMSets.
dim UOM
set UOM = ActiveDocument.UOM
' Define the UOM sets
bOK = UOM.AddUOMSet (Metric)
bOK = UOM.AddUOMSet (US)
bOK = UOM.AddUOMSet (ENG)
bOK = UOM.AddUOMSet (ENGPETRO)
bOK = UOM.AddUOMSet (MET)
bOK = UOM.AddUOMSet (SI)
bOK = UOM.AddUOMSet (METCBAR)
bOK = UOM.AddUOMSet (SI-CBAR)
bOK = UOM.AddPhysicalQuantity(MyQuantity, m3/hr)
bOK = UOM.DefaultDisplayUOM (Metric, MyQuantity, m3/hr)
If the units of measurements are very specific to your simulation file, another option is to create them in the OnLoadSimulation script in the Flowsheet folder. This script is invoked after loading the simulation file (e.g. after OnNewDocumentScript and OnSystemNewDocumentScript.vb have been executed). You can use the same code as in the OnNewDocumentScript.vb (although you don't need to define the UOMSets as at this stage they are already created.
Keywords: unit of measure, units, engineering units
References: None |
Problem Statement: How can I return analytical derivatives from a procedure? | Solution: You need to do three things.
The very first one is to figure out if it is worth evaluating the derivatives in the procedure. If you see your simulation having troubles to converge (residuals getting small, but never getting lower than the tolerance, with the solver apparently running in circles) and some procedures inputs are calculated, then it may be worth doing it. By default, ACM evaluates the derivatives of output variables with respect to input variables using numerical perturbation. These derivatives are required for the Newton method (non-linear solver) when input variables of procedures are being calculated.
You simply need to add DERIVATIVES option in the procedure description. You need also to regenerate the wrapper (that's done automatically by default by ACM).
The more difficult part is to actually write the code to evaluate the derivatives. You need to modify the source code of your procedure to return the derivatives of each output variable with respect to each input variable.
The example attached shows with a simple example how a procedure can return derivatives for the two compatibility options.
The procedure takes two input variables (x and a) and return two output variables (y = 3*x^2 and z = x - a). You can see in the source code (example.f and example2.f) how the derivatives are returned.
Note that returning incorrect derivatives is highly likely to cause convergence problems, especially if the sign is wrong! You can use the solver option to check the derivatives (under Run, Solver Options, Diagnostics, change the option Check procedure derivs to Active Derivatives).
Keywords:
References: None |
Problem Statement: How can we limit the number of components used for a blend? | Solution: To place a limit on the count of something in an optimization model requires us to use mixed-integer techniques. In older versions of Aspen PIMS, the modeler could build mixed-integer constraints using Table ROWS. More recently, we have added Table BLNMIP, which automates the use of mixed-integer constraints.
The figure below demonstrates how to use Table BLNMIP to limit the number of components for a blend.
The row name UPR is the name of a blended product. In column MAXCOMP the user can specify the maximum number of components to be allowed into the blend. Column MAXVOL holds a number larger than the maximum volume of the blend. MAXVOL is not intended to be an upper bound on the blend volume; it is used in building the mixed-integer constraints.
Based on the limit in Table BLNMIP, Aspen PIMS builds constraints as shown in the figure below. Columns BALKUPR, BCCGUPR, etc. represent the volumes of components going to blend UPR. Columns MALKUPR, MCCGUPR, etc. are binary variables which take on a value of 1 if the component is fed to blend UPR, and 0 otherwise. Constraint LMIPUPR limits the number of active components in the blend to the MAXCOMP value.
The figure below shows the specification blend report for UPR in the model without BLNMIP. Note that the blend uses 5 components.
The figure below shows the specification blend report for UPR in the model that uses BLNMIP. Note that the blend uses four components.
Keywords: None
References: None |
Problem Statement: UCX 4.1 ECO 6 has a problem running out of buffers to use.
System Error, errno=55 no buffer space available
Sat May 30 07:14:38 1998
dmc_u20fr2 Error writing list 5: IDB: IDB_Write failure: PDB: CIMIO_USR_PUT_CONNECT, Error connecting to device CIMIO_MSG_CONN_SOCK_CREATE, Error creating an outbound socket CIMIO_SOCK_OUT_CONN_FAIL, Error connecting to the server System Error, errno=55 no buffer space available | Solution: *** 05-Dec-1997 02:13PM *** PATTERSON, WAYNE :13108 ***
$ set output=00:00:10
$ set verify
$ set noon
$ ucx show comm
$ ucx show comm/memory
$ ucx set comm/large_buffer=max:1600
$ ucx set config comm/large_buffer=max:1600
$ ucx show comm
$ ucx show comm/memory
$ exit
-
Keywords:
References: None |
Problem Statement: How can I use the ACM2004 calling convention in procedures? | Solution: We will use the example discussed inSolution 120202. Please refer to it for the beginning of the discussion.
The new procedure with ACM2004 option will be modified as below:
Procedure pSizeLimit2
Library: SizeLimit.dll;
Call: sizelimit2;
Compatibility: ACM2004;
implementation: subroutine sizelimit2.f;
language : fortran;
inputs:
sizes as real(*),
frac as real(*),
limit as real;
outputs:
smed as realvariable;
End
This allows you to declare the name of the arguments (inputs and outputs). The syntax is different of the variable declarations in models as you put the array indication after the argument type (e.g. you write sizes as real(*), while in a model you would declare sizes([1:n]) as realparameter;). It may be a bit confusing, but after all these are not actual variable declarations, but just indications for ACM when it generates the template code of the subroutine to be written to use the names selected by the user, instead of making up some names that may be hard to follow.
The template that is generated by ACM starts as follow:
SUBROUTINE SIZELIMIT2(ICALL,
& sizes, sizesDim1,
& frac, fracDim1,
& limit,
& smed,
& IFAIL)
The other effect of the ACM2004 option is that it changes the order of the arguments. The call type is now in front of all other arguments and is always present. (With the previous spec, it was passed after the variables, and only when needed). We hope the new convention is more simple (but of course it won't be if you were used to the old one). Finally, the way the derivatives are returned (in case the procedure returns analytical derivatives) is dramatically different. This example does not return derivatives, so let's not worry about this.
The example attached contains the code of the procedure with version 12.1 and 2004 compatibility options. Note that the ACM2004 compatibility option may only be used with version 2004 and above. The template shows ACM13 instead of ACM2004, but they are equivalent (ACM13 may be deprecated in the future).
The example was created with version 2004.1.
Keywords:
References: None |
Problem Statement: How to configure Aspen CimIO with Store & Forward when there are multiple Chip databases and one Aspen Cim-IO client. | Solution: Example configuration using multiple chip databases with one Aspen Cim-IO client.
CLIENT
Services
cimiochip
33000
cimiochipu1
33003
cimiochip_sc
33004
cimiochip_st
33005
cimiochip_fw
33006
According to the Aspen CimIO for Fisher Rosemount Chip user manual p.15-4 the DIOP's do not need to be defined on the client side except for the unsolicited.
Logical devices
iochip1 node1 cimiochip
iochip2 node2 cimiochip
iochip3 node3 cimiochip
SERVERS
node1
Services
cimiochip
33000
cimiochipr1
33001
cimiochipw1
33002
cimiochipu1
33003
cimiochip_sc
33004
cimiochip_st
33005
cimiochip_fw
33006
Logical devices
iochip1 node1 cimiochip
The logical devices file is not necessary on the server node unless you plan to run the cimio_t_api program. However, I always recommend it just in case.
node2
Services
cimiochip
33000
cimiochipr1
33001
cimiochipw1
33002
cimiochipu1
33003
cimiochip_sc
33004
cimiochip_st
33005
cimiochip_fw
33006
Logical devices
iochip2 node2 cimiochip
The logical devices file is not necessary on the server node unless you plan to run the Aspen Cimio_T_API program. However, I always recommend it just in case.
node3
Services
cimiochip
33000
cimiochipr1
33001
cimiochipw1
33002
cimiochipu1
33003
cimiochip_sc
33004
cimiochip_st
33005
cimiochip_fw
33006
Logical devices
iochip3 node3 cimiochip
The logical devices file is not necessary on the server node unless you plan to run the cimio_t_api program. However, I always recommend it just in case.
Keywords: store & forward
References: None |
Problem Statement: Installation of the DMCplus 1.10.04 offlines - when we start Model we've got the message: Security module not found. Most likely, the module has been plugged in the laptop's parallel port before the installation. | Solution: Try running the keysetup utility in C:\program files\DMCplus\support.
You may also need to reconfigure the port from EPP or Bidirectional to standard. This is usually done by entering the BIOS setup utility. (On some machines you press F1 or ESC during a reboot.)
Keywords: DMCplus, Model, Security, Dongle, Sentinel
References: None |
Problem Statement: After installing CIMIO for Chip on a VMS box the following error might occur when the DLGP process and DIOPS are starting up:
error=11 no more processes.
The write and unsol process that are configured for DIOP 1 refuse to start. The error message: Error = 11: no more processes is in the DLGP output file. | Solution: This error indicates that the VMS setup parameter for page file quotas was not large enough, which prevents CIM-IO processes from starting successfully. Since the read DIOPs start first they will use the available space, which would prevent write DIOPs from starting. To resolve this problem, in the MCR AUTHORIZE interface modify the PGFLQUO (page file quota) parameter and increase it to be large enough to allow the processes to start.
Another approach to solving the problem is to disable the DLGP spawning DIOPs, so that the cimio_chip_startup.com starts the DIOPs.
Also, make sure you are running the latest version of the interface.
Also make sure to have commands to setup env vars for CIM-IO and to set the default directory to the interface directory at the sys$login:login.com file of the user that starts the interface:
$ @dra3:[aspentech.cimio.commands]cimio_k_set_env.com
$ @cimioroot:[io.cio_fch_chp]cimio_chip_set_env.com
$ SET DEF CIMIO_CHIP
Asserting the directories in this manner is required only for DLGP spawning DIOPs. Otherwise the executables could not be found.
Following is an example of a good VMS account set-up. Note that PRCLM should be the number of processes that could be spawn by a process:
Primary days: Mon Tue Wed Thu Fri
Secondary days: Sat Sun
No access restrictions
Expiration: (none) Pwdminimum: 6 Login Fails: 0
Pwdlifetime: (none) Pwdchange: (pre-expired)
Last Login: 29-NOV-2001 09:55 (interactive), 28-NOV-2001 18:00 (non-interactive)
Maxjobs: 0 Fillm: 100 Bytlm: 512000
Maxacctjobs: 0 Shrfillm: 0 Pbytlm: 0
Maxdetach: 0 BIOlm: 150 JTquota: 4096
Prclm: 8 DIOlm: 150 WSdef: 16500
Prio: 4 ASTlm: 250 WSquo: 60000
Queprio: 4 TQElm: 10 WSextent: 64000
CPU: (none) Enqlm: 2000 Pgflquo: 200000
Authorized Privileges:
CMKRNL GROUP GRPNAM GRPPRV IMPERSONATE NETMBX
OPER PHY_IO PRMGBL PRMMBX PSWAPM READALL
SETPRV SHMEM SYSGBL SYSLCK SYSNAM SYSPRV
TMPMBX
Default Privileges:
CMKRNL GROUP GRPNAM GRPPRV IMPERSONATE NETMBX
OPER PHY_IO PRMGBL PRMMBX PSWAPM READALL
SETPRV SHMEM SYSGBL SYSLCK SYSNAM SYSPRV
TMPMBX
Keywords: no more processes
error=11
quota
References: None |
Problem Statement: When using group decomposed solvers, are there Improvements made for performance of Exported Models? | Solution: Since version 2006.5, improvements have been made to optimize the performance for Aspen Custom Modeler models exported for use in Aspen Plus or HYSYS when using group decomposed solvers.
The solvers are selected on the Block Options form, Simulation Options tab. The group decomposed solvers are DSPARSE, DLSSWP and DDMO.Solution of the block should be significantly faster when using one of these solvers.
To take advantage of this improvement, you must export the model from Aspen Custom Modeler 2006.5 or later. Models exported from earlier versions continue to run in Aspen Plus and HYSYS 2006.5 onward, but do not show the performance improvement.
To export models, one should refer to ACM online manual, Exporting Models to Aspen Plus and HYSYS.
Keywords: Group, Decomposed, Solver, DSPARSE, DLSSWP, DDMO, Export, Performance
References: None |
Problem Statement: How to convert mole fractions to mass fractions? | Solution: User would to retrieve mole weights of the component list used by the model by calling built-in procedure pMolWeights. Then relate mole fraction and mass fraction with the constitutional material balance equation. Refer to the example code below as well the attached sample file, which is for V2006.5 format.
MODEL ttsm
x(componentlist) as molefraction(fixed);
xm(componentlist) as massfraction;
Mws(componentlist) as Molweight;
//To convert moel fraction to mass fraction, first get mole weights
call (Mws) = pMolWeights();
// Then relate mass fraction xm to mole fraction x
xm*sigma(x*Mws) = x*Mws;
End
Keywords: Fractions, convert, mole weights
References: None |
Problem Statement: What is the meaning of a red triangle, Simulated column in an Audit Inventory dialog box? | Solution: The red triangle in the Simulated Column indicates the Plant Value and Simulated value are different.Bolded blue text indicates values you have changed.
In this example, Plant value is 5000 and Sim value is 3000. If you put same values in ORION_MGR_TNKINV_IMPORT table and try again, the red triangle goes away, as shown below:
About the red color tank , it should be in red when the property(different than VOL) is outside of Min and max value defined in Orion_MGR_TNKINV_RANGE table and also if there is NOT a red triangle in Simulated value column, as shown.
ORION_MGR_TNKINV_IMPORT:
Use this table to store imported opening-inventory data for tanks. This table acts as a staging location, where data from an inventory system can be audited (validated) before it is imported into the model.
ORION_MGR_TNLINV_RANGE:
Use this table to set minimum and maximum value limits for opening inventory data imported from a refinery information system (e.g., ATOMS, ADVISOR, etc).
Keywords: ORION_MGR_TNKINV_RANGE
ORION_MGR_TNKINV_IMPORT
Audit Inventories
References: None |
Problem Statement: Cannot run data regression with polymers. Simulation crashes. | Solution: Data regression runs cannot handle polymer components, because they do not handle polymer component attributes (e.g. SFLOW, DPN, MWN, etc.). TheSolution is to declare your polymer component as type Oligomer rather than type Polymer (this is done on the Components / Specifications form). Oligomers are similar to polymers in that they are composed of segments. However, unlike polymers, they have a fixed structure, and do not possess component attributes.
Once you change the component to type Oligomer, the Components | Polymers | Characterization form will be incomplete, so you must also enter the segment composition of the oligomer, i.e. you tell Aspen Plus how many of each segment is contained in the oligomer.
Data should be entered in terms of the solvents and the Oligomer with the Constraint that only the solvents exist in both the liquid and vapor phase.
For the Regression, the parameters used in the model should be regressed. For example, for the Polymer NRTL (POLYNRTL) model, there are NRTL parameters for the solvent-segment interactions and for the segment-segment interactions that can be regressed.
In the attached example, the NRTL parameters for the ethylene repeat unit and the ethylene solvent are regressed from polyethylene oligomer and molecular ethylene TPXY data.
After performing the data regression run and obtaining the parameter results, you can switch the run type back to Flowsheet. However, you must remember to switch the component type from Oligomer back to Polymer.
Keywords: Data regression, polymer, oligomer
References: None |
Problem Statement: On a PC with multiple instances of SQL Server installed, the second instance installed will have a slash \ character added automatically to differentiate it from the primary install. The Aspen Database Wizard cannot handle the slash character. | Solution: This is a partial work-around, that still needs developer intervention to help complete the process:
This problem usually occurs when there is more than one SQL Server instance on the same box. For example, SQL Server 97 coexists with SQL Server 2000. In that case the SQL Server 2000 instance will typically be named like
<MyComputerName>\<MYSQLServerInstanceName
It is necessary to specify the SQL Server instance without using a slash character. During the Database Wizard execution, when asked for the SQL Server instance, this can be done by specifying the port the node and the port for SQL Server, instead of using the default name from the drop-down list. See the section below titled Determining my SQL Server Port, for instructions on obtaining that information.
Changing The Default Driver
1. Stop the Apache Tomcat Service.
2. Download the attached zip file. This zip contains three other zips: (1) msbase.jar; (2) mssqlserver.jar; and (3) msutil.jar.
3. Copy these three zip files to:
C:\Program Files\Apache Tomcat 4.0\webapps\aebrsserver\web-inf\lib
4. Change the name of the apps.jar file in this directory to apps.zip.
5. Create a directory in the root directory named meta-inf.
6. Extract Manifest.mf to this directory.
7. Edit the file by typing the names of the three files added to the \lib directory, separated by spaces, at the end of the existing list. Save your changes.
8. Delete the existing Manifest.mf file from the zip.
9. Now drag and drop the meta-inf directory (not just the Manifest.mf file!) from the root back into the zip file, and confirm the change to the zip.
10. Rename apps.zip back to apps.jar. As clean-up, you can also delete the c:\meta-inf directory and its contents.
11. Next, go to the cfg_source directory, located at:
C:\Program Files\Aspentech\AEBRS\cfg_source.
12. First, edit the config.m2r_cfg file. Change the INCLUDE_FILES line so that instead of referencing db_mssql.cfg it references db_mcsql.cfg instead. Save your changes to that file and close it.
13. Next, edit the codify_all.cmd file using Notepad. To view the file correctly, make sure to turn off Word Wrap. Examine the file to see if a reference to db_mcsql.cfg already exists. If it does, close the file without saving changes. If it does not, copy the existing db_mssql.cfg line and paste it as the following line. Change db_mssql.cfg to db_mcsql.cfg in the new line, in both relevant places in the line.
14. Save the changes to codify_all.cmd
15. Run codify_all.cmd
16. Restart the Apache Web Server.
17. In the path file, located in \AEBRS\path.m2r_cfg, change the Database Connection Options as follows:
DB_URL_SERVER=<servernameonly>
DB_PORT=<port being used by SQL Server>
NOTE: To determine the DB_PORT, choose Start | Programs | Microsoft SQL Server | Server Network Utility. TCP/IP will be one of your Enabled Protocols. Highlight it, and choose Properties to obtain the port number. This port number was chosen dynamically at installation time, by SQL Server, and remains static after that point.
18. Go to cfg_source and edit config.m2r_cfg. Change the INCLUDE_FILES line reference from db_mssql.cfg to db_mcsql.cfg. Save the changes.
19. Next, edit the codify_all.cmd file, and search for mcsql.cfg. If no reference exists, copy the existing line which refers to the mssql.cfg file, paste the line below the existing line, and change the new line from mssql.cfg to mcsql.cfg.
20. Delete the existing AeBRS database which was not successfully created before, along with the AEBRS user.
Re-run the database Wizard. The Wizard will still fail, but before the failure point it will now create the entire correct AeBRS database. The only failure point left is that the database needs to be manually signed. To manually sign the database, developer intervention is required. An Aspen developer can do the final steps by referring to an internal procedure documented in Support call 738198. Please contact Aspentech Support and reference thisSolution number if seeking that assistance.
Keywords:
References: None |
Problem Statement: Change the maximum point count (total number of data records) for Aspen Infoplus.21 database is described in KB000078151.
If your Infoplus.21 is in a cluster, how to change the maximum point count? | Solution: Note: The Aspen IP.21 Manager is still used to configure tasks but it’s no longer used to start and stop InfoPlus.21 in a cluster. The Aspen MES Cluster Administrator should be used to start and stop InfoPlus.21 instead.
1. Go to the active Infoplus.21 server and use Aspen MES Cluster Administrator to stop IP.21
a) Right-click the Owner and select Stop application group.
b) Click on Yes
c) Ensure that the operation completed successfully.
2. Go to Microsoft Failover Cluster Manager | Bring Aspen Data Source Directory and Aspen Infoplus.21 Task Service online.
3. Launch IP.21 administrator | expand Infoplus.21 | Right-click the data source | select Set Point Count | Type the number of maximum point count | click on OK.
Note: Please also check if another node has the same Maximum Point Count.
4. Start IP21 using Failover Cluster Manager | Right-click Roles | Select Start Role
5. Check the license state of your IP.21. Launch IP.21 administrator | expand Infoplus.21 | Right-click the data source | select License Status | The license State should be License Granted
Key Words:
point
count
license
cluster
Keywords: None
References: None |
Problem Statement: If a particular phase is in a Not Ready state, and it is preceded by multiple conditions, how can one determine the state of those conditions, to determine why execution flow has not reached phase? | Solution: There are no API functions to do this. And as part of the design of AeBRS, all direct queries to AeBRS tables (i.e. tables that have ebr_ as part of the name) are controlled: If the query includes terms like Insert, Update, Delete, etc. they are not allowed. However Select statements are allowed. Since Condition blocks are unique elements in the database, it is possible to query the Condition block directly, and return the value of the flow_state parameter to gather useful information about the status of the Condition. Here are two examples:
If you know the Order Code and Repetition number, you can return an integer value for flow_state as follows:
select e.flow_state from ebr_order_rpl_element e, ebr_order o where e.id_order = o.id_order and o.code = 'order code' and o.repetition_no = repetition and e.tag = 'condition name'
If you know the internal Order ID, a simpler query will return the same information:
select flow_state from ebr_order_rpl_element where id_order = <myorderid> and tag = 'condition name'
Here are the detailed definitions of the integer values returned by these queries:
Table - Integer Status Definitions
Integer
Name
Definition
0
NOT_READY
Execution flow has not reached the Condition.
1
BLOCKED
Execution flow has reached the Condition, but it is blocked by a synchronization link (synchronization links are not implemented in 6.0.)
2
READY
Condition is eligible for evaluation because all previous Conditions/Phases are completed or skipped.
3
COMPLETED
At the time of the check, Execution flow has reached and passed this element (see Note 1.)
4
SKIPPED
Evaluation of previous conditions has caused Execution flow to bypass the condition.
5
CHECKING
Execution flow has reached this element and its condition is being periodically evaluated (see Note 2.)
6
SLEEPING
Execution flow has reached this element but its condition is not being periodically evaluated (see Note 3.)
7
ZOMBIE
Execution flow has not yet reached this element - i.e.: Not ready - but its corresponding phase is not Enabled - i.e.: Disabled or Cancelled (see Note 4.)
8
CANT_EXEC
Phase is Completed but its requirement condition evaluated to NO, so it cannot be executed
Note 1 - further information re COMPLETED status: Completed elements may receive Ready events due to Repeatable phases left Enabled/Completed finishing - state persistence/notification takes care of not taking into account these state non-changes.
Note 2 - further information re CHECKING status: It can receive Fulfilled and Not Fulfilled external events depending on the condition evaluating to YES or NO - it can also receive a Sleep event from corresponding phase to disable this evaluation.
Note 3 - further information re SLEEPING status: A Condition returns sleeping status because its corresponding phase is not Enabled - it can receive an Awake event from the corresponding phase to enable this evaluation again.
Note 4 - further information re ZOMBIE status: A condition with zombie status can receive an Awake event to go back to Not ready or get a Ready from incoming flow control and goes to Sleeping passing the event down to the non-Enabled phase to process.
Keywords:
References: None |
Problem Statement: By design, the TRANSLATE function only pulls translations from specific language files that ship with the product. It is not possible to add your own terms to the language file without the risk of it being overwritten on a recompile or upgrade. | Solution: To add custom variables (for instance language-specific or company-specific terminology) to your AeBRS installation, there are at least two different ways to do it, either via customizing the TRANSLATE function (Method One, below), or understanding how to use the GET_CONFIG function (Method Two, below.)
This has been tested by Customer Support for Version 6.0.1
Method One: The first approach is to configure your system so that the TRANSLATE function will transparently pick up your default language file and also any custom terminology you want to add.
This method requires modifications to configuration files. The second method does not. However, if you have extensive eBRS coding already done using the TRANSLATE function, this might still be your preferred approach.
By default, the LANGUAGE key near the top of the path.m2r_cfg file is blank. Edit this file and give the LANGUAGE key a value, like the following:
LANGUAGE = lngext
lngext is the name of the new user resource file we will define (the name itself is not important.)
Close this file and save your changes.
Now use Notepad to define your custom strings, in the following format. The number sign can always be used to provide remarks:
# This is my custom resource file
CustomTermBefore1 = CustomTermAfter1
CustomTermBefore2 = CustomTermAfter2
CustomTermBefore3 = CustomTermAfter3
Save the file as lngext.m2r_cfg using this path:
C:\Program Files\AspenTech\AEBRS\cfg_source\lngext.m2r_cfg
Next, make sure every time eBRS files are codified (turning text files into compiled objects), your new user resource file is codified and placed in the right directory. To do this, you will use notepad to edit the codify_all.cmd file. Add the following line (this is all one line, no hard returns allowed!) You might find it easiest to cut-and-paste the last codify line already in the file, and just edit it:
call C:\Program Files\AspenTech\AEBRS\cfg_source\codify C:\Program Files\AspenTech\AEBRS\cfg_source\lngext.m2r_cfg C:\Program Files\AspenTech\AEBRS\resource\lngext_en.lng
IMPORTANT NOTE: The TRANSLATE function only sees the language resource file that matches your Windows language settings. So if your language is set to French, for example, instead of en, the filename you specify would have to be lngext_fr.lng.
ADDITIONAL NOTE: In step 1 we defined the LANGUAGE key to look for the user resource file lngext. When AeBRS starts, it will automatically append an extension based on the language settings. So by defining the LANGUAGE key lngext on an English/American workstation, the file AeBRS will look for wil be lngext_en.lng. (If your language setting is French, it will automatically look for lngext_fr.lng.)
Close the file and save your changes.
Now doubleclick on codify_all.cmd. After codify_all executes, verify that your new compiled user resource file exists:
C:\Program Files\AspenTech\AEBRS\resource\lngext_en.lng
Your custom language extensions are now ready for use. For the language file used in the above example, a simple BP is defined with the following code in the Previous Actions window:
TestTranslate1 := TRANSLATE(CustomTermBefore1) TestTranslate2 := TRANSLATE(CustomTermBefore2) TestTranslate3 := TRANSLATE(CustomTermBefore3)
When the BP loads and the Previous Action code executes, the translated value for each key will be assigned to the variable. For example, in the first line, the variable TestTranslate1 will be assigned the translated value for key CustomTermBefore1, which is CustomTermAfter1.
Add expressions to your BP form with TestTranslate1, TestTranslate2, and TestTranslate3 as the values for the Expression property of the expressions. Here is an example of a BP created using the steps outlined in this Knowledgebase Article (the .CHK file for the BP is attached to thisSolution) :
Method Two: Use GET_CONFIG
GET_CONFIG works similarly to the TRANSLATE function discussed above. A Previous Action code block like the one we used above would be virtually identical, and would function the same way:
TestTranslate1 := GET_CONFIG(CustomTermBefore1) TestTranslate2 := GET_CONFIG(CustomTermBefore2) TestTranslate3 := GET_CONFIG(CustomTermBefore3)
The differences are as follows:
GET_CONFIG is language independent. It can provide a common library of translated terms available to your eBRS system regardless of the language settings.
It requires no additional customization -- right out-of-the-box GET_CONFIG can be used to quickly add your own custom terms without editing config files, as was outlined in Method One, above.
So if you are starting a new project where you may want to have custom terminology translated, it might be a good idea to start off using GET_CONFIG. Steps to implement are as follows:
Use Notepad to edit the local.m2r_cfg file, located along this path:
C:\Program Files\AspenTech\AEBRS\cfg_source\
The syntax is the same as the custom file we created in Method One, above:
# All defined key here would update server values
CustomTermBefore1 = CustomTermAfter1
CustomTermBefore2 = CustomTermAfter2
CustomTermBefore3 = CustomTermAfter3
After adding your custom terms and their translated values, close the file, saving your changes.
The codify_all.cmd file already has a line in it to codify your changes. So doubleclick on codify_all.cmd to create the compiled object containing your translation table.
Test the results by creating a BP (or modifying the one you used in Method One above!) to use the GET_CONFIG function.
Final note -- after defining a User Resource file, you may want to make sure it can be accessed by every AeBRS workstation. SeeSolution #113679 for an explanation of how to make your resource directory shared.
Keywords:
References: None |
Problem Statement: When a unit is set to idle, the expert zeroes all the pipe flows. The reconciliation results are then saved. If the unit should not have been set to idle, you may need to go back in time to un-idle the unit. The problem is that when the Expert system re-runs the reconciliation, the pipe flows continue to be zero. | Solution: The system was not designed to pick up instances where the unit went from idle to active, so the values in the pipes do not change. Normally the values would change. If a user wants to have the system update and change the values, then they must change the raw value, run expert, then change the raw value back to the original value. This will cause that single value to change without upsetting all of the other values.
Keywords:
References: None |
Problem Statement: Performance of the EBRS MOC modules (i.e. Orders, RPL Design, Workstation, etc.) is unacceptably slow. For example, it takes up to 10 minutes for a list of orders ready for execution to appear. | Solution: This is a problem more likely to occur when the underlying database is Oracle, for one of the following reasons:
AUTO-EXTEND
If the performance decrease happens gradually over time, one possibility is that the initial table size parameters in Oracle were too small. Every time an Oracle table fills its existing allocation of space, it auto-extends. Oracle recommends allowing no more than 10 auto-extends per table. After 10 auto-extends, performance will start to degrade. However, if necessary, tables will auto-extend over 100 times.
Consult with your Oracle DBA to verify whether or not your tables have too many auto-extends.
[do we have any code that shows the number of auto-extends for a particular table?]
MISSING INDEXES
To maintain optimal performance, AeBRS depends on a set of indexes. If these indexes are not present, the application will continue to function, but with poor performance.
These two queries provide a list of the indexes in your AeBRS database, and their sizes. Make sure to change the owner to mach the schema/username (this is the user used to connect to the database listed in your AeBRS config files):
select index_name, initial_extent, next_extent from dba_indexes where owner = 'AEBRS'
select segment_name Index, max (extents) Extents , sum(bytes) Total Size, sum(blocks) Blocks from sys.dba_segments
where owner ='AEBRS' and segment_type ='INDEX'
group by segment_name, segment_type
having segment_type not in ('TEMP', 'CLUSTER')
order by 1
If the indexes are missing, a likely cause is that the database was exported from one system and imported on another. There are flags related to the indexes that are part of both the export and import operation. If the index flag for either operation was set to NO, then the indexes will not be present in the new database.
Attached is a SQL script that will recreate the indexes for a Version 6.0.1 database.
Keywords: slow
freeze
References: None |
Problem Statement: This knowledge base article explains what an Entity Relationship Diagram is and how it relates to Aspen Advisor. | Solution: An Entity Relationship Diagram (ERD) is a diagram that shows tables, columns, primary (PK) or foreign keys (FK), and relationships between tables. Abbreviated ERDs for various tables can be found in the Aspen Advisor Database online .HLP file.
NOTE: Aspen Advisor has no PK, or FK's defined in a database - the foreign relations are inferred and enforced in software, but not in the native database due to the different referential integrity models used by various database vendors.
Keywords:
References: None |
Problem Statement: I have a block of code that requests the identity of $CURRENT_USER in my Basic Phase that executes fine. However the same code when inserted into a Condition block, returns no value for $CURRENT_USER. Why? | Solution: Code that exists in condition blocks is not associated with any particular user, but instead the RPL (and is executed by the Apache Web Server on the AeBRS server, not on a workstation.) So by design, this condition-related code belong to no user. The code is evaluated for all users on the system (i.e. after a condition evaluates to YES, the following step/phase will be available for execution by any user.)
A request for the $CURRENT_USER are only scoped correctly when executed from code within a Basic Phase.
Keywords:
References: None |
Problem Statement: Because the implementation of an AeBRS system requires development specific to the customer, often it is first installed in a development environment before moving to a production machine. For this reason, a set of steps must be followed to properly migrate the system. | Solution: Backup the AeBRS relational database
Backup the Working and Project folders - including PROC_DESIGN and SCREENSHOT (if the screen shot feature is used)
Keep a copy of the .m2r_cfg files for reference
Import the original AeBRS database into the new machine
Replace the Working folders with the original ones - only PROC_DESIGN and SCREENSHOT
Install AeBRS pointing to PROC_DESIGN, SCREENSHOT, the existing relational database
Customize config files from previous copy, if there are any changes on the new machine. In this step you should modify the new .m2r_cfg config files based on the previous copy and run codifyall.bat.
If also upgrading to a new version, after moving the installation, run SQL upgrade scripts as indicated in the upgrade section of the AeBRS installation manual
For further reference on the various files and dependencies of AeBRS, please refer to the AeBRS_install.txt file which is loaded during an AeBRS installation.
Keywords: maintenance
backups
References: None |
Problem Statement: Help files may fail to appear in the AeBRS client applications. | Solution: The HELP_DIRECTORY_NAME key now points to the url for the help files on the server, yet the client's config.m2r_cfg file may still have the localhost name in the key, rather than the server name. To resolve the problem, move the HELP_DIRECTORY_NAME key and its value from the config.m2r_cfg file to the path.m2r_cfg file. After making this change, run codify_all so that it is recognized. This will allow the help files on the client to be called correctly.
Keywords: help
path
url
References: None |
Problem Statement: AeBRS v6, service pack 1 makes use of a new environment variable, SERVER_PROC_DESIGN, in order to determine the location of the Proc_Design directory.
Unfortunately, this environment variable is not created by the service pack installation. As a result, users may see errors from the MOC, such as:
E1100: Aspen eBRS database unavailable or not configured correctly | Solution: There are two options for resolving issues related to the missing environment variable:
Create the system environment variable SERVER_PROC_DESIGN, specifying the path to the Proc_Design folder. Note: The machine must be rebooted in order to recognize the new environment variable.
-or-
Open the config.m2r_cfg file and locate the line: DESIGN_ROOT_PATH = %SERVER_PROC_DESIGN%\\ Remove the environment variable and hard-code the path to the Proc_Design folder. Note: In order for AeBRS to recognize the change, you will need to save the file, then run codify_all.cmd.
Keywords: upgrade
installation
startup
sp1
database error
References: None |
Problem Statement: MOC fails to start up with this error:
Could not find the main class | Solution: Main class not found errors reflect a problem with one of the necessary resources referenced in the AeBRS.cmd directory not being available.
Open up AeBRS.cmd, located in this directory:
C:\Program Files\AspenTech\AEBRS
Verify that each path, for L, J and D actually exist and are mapped on the node:
SET L=C:\Program Files\Apache Tomcat 4.0\webapps\AeBRSserver\WEB-INF\lib
SET J=C:\j2sdk1.4.0_01\BIN\JAVAW.EXE
SET D=C:\Program Files\Apache Tomcat 4.0\lib\DriverJDBC.jar
START= /i %J% -cp %L%\Apps.jar;%D% %1
Keywords:
References: None |
Problem Statement: Get the following error message when trying to open a custom report made in AORA using Microsoft Visual Basic. | Solution: 1) Close all Microsoft Office applications
2) Navigate to C:\users\<username>\AppData\Local\Temp\Excel8.0 and delete MSForms.EXD
3) Navigate to C:\users\<username>\AppData\Local\Temp\VBE\ and delete MSForms.EXD (if it exists)
4) Microsoft visual basic should now be able to access the AORA report.
Keywords: Error 438
Microsoft Visual Basic
AORA Reporter
References: None |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.