question
stringlengths
19
6.88k
answer
stringlengths
38
33.3k
Problem Statement: This knowledge base article explains why the Aspen Batch Configuration Utility (BCU) scripts in the scheduler table may sometimes have the status waiting for data.
Solution: The status waiting for data means that a BCU script is about to fire, but needs additional data. For instance, the conditions for the trigger have been met, but the BCU script does not have all the data (yet) to write to the Aspen Batch.21 database. A common cause for this is that you specified to use interpolated data. In order to interpolate, the BCU needs two data points: one from just before the time the trigger fired and the next data point, which might not be available yet. Here is an example: The trigger fires, or becomes true, at 10:00. It has been configured to store a temperature characteristic, so looks at the history of its source tag, to see what its value was at 10:00. It finds that the last data value stored for that tag is 9:57. There is no data for that tag at 10:00. There has to be a data point at 10:00 or after 10:00, in order for Batch to determine what the temperature was at the time the trigger fired. Once it finds data for that tag at 10:00, it will finally store the characteristic, and the status of the script will change to OK. If you have a situation where this next data point might not occur in the near future, this might lead to unnecessary delays in your BCU script processing. The Extrapolate to current time feature would be a good alternative if this is the case. Knowledge base article #104737 describes this feature. Keywords: fire execute run process References: None
Problem Statement: Sometimes, a user may return a lot of rows of data with the AtGetTrend function in the Excel Add-ins. However, Excel 97 is limited to a 64k limit. As such if the data exceeds the 64k limit, then no data is returned to the Excel spreadsheet
Solution: We can overcome this with a macro. An example is shown below. Dim tmp Dim i As Integer tmp = ATGetTrend(atcai, New Server, , 9/1/2000 07:47:15 AM, 10/4/2000 09:47:15 AM, 9528, 5m, 0, 25, 0, 1) For i = 0 To UBound(tmp) Cells(i + 1, 11).Value = tmp(i, 0) Cells(i + 1, 12).Value = Format(tmp(i, 1), dd-mm-yyyy hh:mm:ss) Cells(i + 1, 13).Value = tmp(i, 2) Next i For Excel to recognize the AtGetTrend function, you must first enable the Add-In, by going to Tools->Add-Ins... and selecting AspenProcessDataAddin For VB to recognize the AtGetTrend function, you must first enable the Add-In, by going to Tools-> Keywords: Addin Excel 97 AtGetTrend References: s... and selecting AspenProcessDataAddin
Problem Statement: Error starting BCU - Failed to start unit scheduling: '(null)' 80004005
Solution: For Version 3.1.2 and 4.0.1: To resolve this error complete the following: Remove the bcu.areas file from the BCU directory. Then using the BCU Administrator - Tools | Options, click the Configure button. From this window, click the Add button and select the desired Area from the pulldown menu and click OK. Repeat as needed. Then use the BCU Server Manager, Start the BCU. For Version 4.1.2 and higher: With all Batch Services stopped run a query against your Batch.21 database using the SQL tool of your choice, removing the contents of these two tables in this order: DELETE bcu_triggers_status; COMMIT; DELETE bcu_units_table; COMMIT; Restart the BCU Server. Unit Scheduling should now start successfully. If the error persists, it may be related to the driver choice. This issue has been seen when using ODBC connection method. Try switching to the OLE Provider connection method. This is done in the ADSA component. Consult the Batch.21 Installation manual for specific steps. Keywords: BCU error References: None
Problem Statement: When a user tries to use the Aspen Process Data Add-In (Excel Add-In) to pull attributes from the InfoPlus.21 database by going to: Aspen/ ProcessData / GetData / AttributeValues...., an 'Attribute Values' dialog box appears. If the user then enters a tag name and cell reference location on the box, and clicks on the '...' button for attribute selection, the following error may be returned: #ERROR: -2147467259 [Microsoft][ODBC Driver Manager] Data Source name not found and no default driver specified
Solution: This error occurs because Desktop ODBC is not installed on the client PC. Desktop ODBC is only required to list the attributes. Please install the InfoPlus.21 ODBC driver on the PC where the error message occurs. This is usually accomplished by installing SQL+ client on the PC. Wordaround: If the user enters the attributes for IP_DESCRIPTION and IP_ENG_UNITS in Excel and then references the cells instead of clicking on the '...' button, the ATGetAttrVal function should return the attributes requested, even if Desktop ODBC is not installed on the client PC. Keywords: 2147467259 attribute ODBC References: None
Problem Statement: The Marginal Cost and Marginal Revenue curves indicate important information for a specific amount of feedstock purchased or product sold. The Marginal Value (DJ) calculated by Aspen PIMS relates both curves at a given amount of material and can be used to calculate one of them given the other. This documents describes the basic features of these curves and how they can be created using information created by an Aspen PIMS run.
Solution: Note: Please seeSolution 124710 for a Report Writer Template that automates the creation of these curves, the cumulative break even calculation and the Objective Function plot. Description of the Marginal Cost Curve The Marginal Cost curve indicates for any given unit of material the cost to purchase or produce that unit. The area under that curve is the total cost incurred to purchase or produce a given amount of material For feedstocks to the refinery, the Marginal Cost curve is typically a straight line with slope = 0, i.e. the cost for any amount of crude is the same, and is defined by the COST in table BUY for that material. If tiered pricing is used (through alternate tags defined in table ALTTAGS), the curve will be a series of straight lines with slope zero, as shown in the figure below. For products of the refinery, the Marginal Cost curve has to be calculated as indicated in the later sections. The typical shape is a curve with a positive slope, i.e. the Marginal Cost increases for additional units of material produced. Description of the Marginal Revenue Curve The Marginal Revenue curve indicates for any given unit of material the revenue obtained by processing that unit of purchased material or by selling that unit of a final product. The area under that curve is the total revenue obtained by processing a purchased material or by selling a given amount of material For feedstocks to the refinery, the Marginal Revenue curve has to be calculated as indicated in the later sections. The typical shape is a curve with a negative slope, i.e. the Marginal Revenue decreases for additional units of material processed. A sample Marginal Revenue Curve for a Crude purchase is shown below: For products of the refinery, the Marginal Revenue curve is typically a straight line with slope = 0, i.e. the revenue for selling any amount of product is the same, and is defined by the PRICE in table SELL for that material. If tiered pricing is used (through alternate tags defined in table ALTTAGS), the curve will be a series of straight lines with slope zero. Note: Due to the nonlinear nature of the Aspen PIMS refinery models used in these evaluations, the Marginal Revenue Curve for a Crude may show an increase when increasing the volume and then decrease again, or the Marginal Cost Curve for a product may show a decrease when decreasing the volume and then increase again. Marginal Value or DJ The profit or loss is calculated by Aspen PIMS for the last unit of material processed, for all purchases and sales. This number is called Marginal Value or DJ, and it represents the difference between the Marginal Revenue and the Marginal Cost, i.e.: (1) DJ = Marginal Value = Marginal Revenue - Marginal Cost If the Marginal Value is positive, it means that for the last unit you are making a profit equal to that value and there is an incentive to process more of this material. If the Marginal Value is negative, it means that the last unit you are making a loss equal to that value and there is an incentive to process less of this material. If the Marginal Value is zero, it means that the last unit produced does not make either a profit or a loss, so this is the optimum amount of material to process. From the above formula, a Marginal Value of zero means that the Marginal Revenue is equal to the Marginal Cost, i.e. they balance each other out. Net Profit /Loss calculation The Marginal Value, or the difference between the Marginal Revenue and the Marginal Cost for a given unit of material (e.g. at the 4100th barrel) is the profit or loss that you incur by processing that unit of material. The area between the Marginal Revenue curve and the Marginal Cost curve is the Profit or Loss of processing the total amount of that material. Maximum Profit The maximum profit is achieved when the Marginal Cost is equal to the Marginal Revenue, i.e. when the Marginal Cost Curve and the Marginal Revenue Curve cross, or what is equivalent, the Marginal Value is zero. For that specific unit of material, the cost of buying or processing is equal to the revenue of processing or selling it, therefore there is neither a profit nor a loss. How to build Marginal Revenue and Marginal Cost Curves using Marginal Values For a Feedstock (e.g. a crude) You need to set up a series of cases where you change the amount of the material that you want to buy. For example, analyzing a crude ANS, in each case you increment the amount of crude by 5 kbbl, starting from zero up to 45 kbbl (i.e. eleven cases). Once you have the results, identify the Marginal Value from the FullSolution for each case for this material. The Marginal Cost will be the COST defined in table BUY Based on formula (1), the Marginal Revenue will be calculated at each point as: Marginal Revenue = Marginal Cost + Marginal Value Below is an example of such a calculation For a Final Product (e.g. a gasoline) You need to set up a series of cases where you change the amount of the material that you want to sell. For example, analyzing a gasoline UPR, in each case you increment the amount of gasoline by 5 kbbl, starting from zero up to 45 kbbl. In the particular case for the data shown below, more granularity was required when running more than 40 kbbl, so a few cases were included where the increment was 1 kbbl. Once you have the results, identify the Marginal Value from the FullSolution for each case for this material. The Marginal Revenue will be the PRICE defined in table SELL. Based on formula (1), the Marginal Cost will be calculated at each point as: Marginal Cost = Marginal Revenue - Marginal Value Example The graph below, using the data from the product UPR from above illustrates this concept: The Marginal Revenue Curve is constant and equal to the PRICE in table SELL (107$/bbl). The Marginal Cost Curve is calculated as in the section before. The optimum amount of UPR to make is around 41kbbl, which is where the Marginal Cost and the Marginal Revenue curves cross Up to this amount, every additional bbl of UPR sold produces a profit, so the Objective Functions increases. The accumulated total profit is indicated by the shaded are between the Marginal Revenue and Cost curves From around 41 kbbl on, every additional unit of UPR is sold at a loss, so the Objective Function starts decreasing. It costs more and more to produce additional units of UPR, indicated by the steep slope in the Marginal Cost Curve. Keywords: Marginal Value DJ Objective Function Economics Marginal Cost Marginal Revenue References: None
Problem Statement: The Batch Query Tool (BQT) may give the following run time error immediately upon opening: Run-time error 2007. Invalid bands collection index. The above error message often appears after an Oracle application called ADI (Applications Desktop Integrator) has been installed on the computer. OKing this error shuts down the application, so it isn't possible to use the tool at all. Even if the BQT doesn't give a run time error, but is failing to start properly and the other Batch tools are working, the following
Solution: may help. Solution There are a few ways to reset the query tool and hopefully get it to work again. Make sure the default data source and area are valid and correct Delete the toolbar file Delete a bunch of Query tool run time information from the registry. Check the version, size and date on the following file: Actbar.ocx. About 1: Open the Batch Detail Display, choose Tools | Options, select the appropriate Data Source and area, click the Set As Default button. Then try to run the Query Tool again. (This should update the registry setting for default data source and area, which affects all Batch.21 client applications.) If the Batch Detail Display tool also isn't working, the user can modify their default Data Source and area directly in the registry. To do this, run Regedit and select: My Computer HKEY_CURRENT_USER Software AspenTech Batch.21 Settings Do a right mouse click on Settings and select Modify. About 2: The Query tool saves a .tb file which keeps information about the toolbar. On NT this file is in the C:\Temp directory, its name is: <user name>_BQTLayout.tb. Find this file and delete it, try running the BQT again. About 3: The query tool keeps information from the Options pages and other things like column widths, etc in the registry. Deleting this information from the registry forces the query tool to start up with default settings, and may fix the problem. To do this: Run Regedit and select My Computer HKEY_CURRENT_USER Software AspenTech Batch.21 QueryTool Preferences Do a right mouse click on this last item, Preferences, and select Delete. Then try running the Query Tool again. About 4: The version, size and date of the Actbar.ocx file (C:\Windows\System32) should be v1.0.6.1, 333kB, and 10/27/1999. If everything checks out then try re-registering the file. If your file has different attributes, then please rename it and import the file from another PC in your office, then reboot your PC and register the file. Keywords: 2007 invalid band collection index run-time References: None
Problem Statement: Can I change the maximum amount of velocity that a flare system can handle in Aspen Flare System Analyzer?
Solution: Yes it is possible from V7.2 version onwards. A new textbox is provided in the Calculation Options Editor to specify the maximum velocity a Flare System can handle. The default value is 500 m/s. In previous versions, this value was hard-coded and if the actual velocity exceeded this, Aspen Flare System Analyzer could not converge on swage pressure drop in some cases & gave a warning. In V7.2 version user can change the default value and the convergence of pressure drop in the swage will become more robust. This can eventually affect the static pressure values of all upstream operations. Keywords: maximum velocity, swage pressure drop calculations etc; References: None
Problem Statement: How do I change all the PSV into a conventional type valve?
Solution: There is not a easy way to change all the valve types to conventional type. You can change one by one in build-nodes. Another way will be to export your case to excel using Export Wizard. If you use the default definition file, you can change the value to 1 in 7th column (G column) on ReliefValves worksheet in Excel export file. Then you can import it back using Import wizard. In this way, all PSV should change as Conventional Valve. Type code for valve 0 = Balanced, 1 = Conventional, 2 = Pilot You can change to different type of value using above values. Keywords: PSV, conventional type, balanced type References: None
Problem Statement: How can I view the Pressure Profile in the Pipe Segment in Aspen Flare System Analyzer?
Solution: Generally, Aspen FLARENET does not report the pressure profile in the pipe segment. The program will only display the calculated upstream and downstream pressure values. Notwithstanding, the elements are still divided internally. The program just does not expose the calculated pressure values for each interval as Aspen HYSYS does. In order to generate segment wise; and expose the pressure profile similar to that available in Aspen HYSYS pipe segments, you will need to add the equivalent number of pipe segments in FLARENET to match the number of elements as divided in HYSYS. However, you cannot specify multiple segments in a pipe the way it is possible in HYSYS. Keywords: Pressure Profile, Pipe Segment, References: None
Problem Statement: What is the pressure drop correlation used for the inlet piping of the PSV?
Solution: Flarenet uses the same methods to calculate the pressure drop in the inlet piping as it does in the main piping, which depends on the correlation chosen. You can check the correlation in use from the Calculation Options. Keywords: Inlet piping, pressure drop References: None
Problem Statement: How can I choose a the printer I want to use in Aspen Flare System Analyzer?
Solution: Aspen Flare System Analyzer will use the default printer in your computer so in order to choose a different one, you need to select the desired printer directly in Control Panel and after that go ahead and print what you need from Aspen Flare System Analyzer Keywords: Printer References: None
Problem Statement: How to model a molecular seal in Aspen Flare System Analyzer?
Solution: There is no standard tool in Aspen Flare System Analyzer to model the molecular seals. The pressure drop by molecular seal can however be accounted for by using the Use Curve method for the flare tip. The pressure drop data can be obtained from the flare tip manufacturers. Following are the other options available to account for the pressure drop by the molecular seals: 1. Adding a fitting loss coefficient on the flare tip calculations tab. 2. Adding appropriate fittings to the pipe to the flare stack. Key Words: Molecular Seal, Flare tip, pressure drop Keywords: None References: None
Problem Statement: Why does not Aspen Flare System Analyzer take my composition when I import data from MS Excel?
Solution: In order to take the composition correctly from MS Excel, please check the following information: a) You have to have the exact components and order between your input data and Flarenet. This error especially occurs when you add more components to an existing case study. You might need import the existing components together with new components. b) You have to assign the source as composition basis instead of molecular basis. c) Your compositions for each source have been summed to 1. If your composition added up not equal to 1 (for example 0.9997), Flarenet might not get right values. You may want to normalize your data in MS excel before you import. Keywords: Import, MS Excel References: None
Problem Statement: How does one change the roughness for all pipes in a simulation? This option is not available from the File | Preferences menu.
Solution: The parameters for all pipes in a given simulation can be edited at the same time using the Pipe Manager. While the File | Preferences option is used to determine the parameters for new pipe that is added to the PFD, existing pipes can be edited using the Pipe Manager which can be accessed via the Build | Pipes menu item. If needed, multiple pipes can be modified simultaneously by pressing the Ctrl key while using the mouse to select the desired pipes. Keywords: edit, pipe, roughness, thermal, conductivity References: None
Problem Statement: Can I model dynamic depressuring in FLARENET? Does FLARENET have dynamic mode?
Solution: Flarenet does not have dynamic feature at this time, however this functionality is slated to be incorporated into a future release. If desired, one can import dynamic depressuring case results from HYSYS into Flarenet and create different scenarios to represent specific time steps.A To link FLARENET and HYSYS, please follow the procedure below: In HYSYS 1. Create a case file in HYSYS by adding components and fluid package 2. Add the stream in Simulation Environment fully defined 3. Press Ctrl+U to add ' Depressuring Dynamic' utility 4. Attach the stream to the utility and define the vessel information 5. Run the ' Depressuring Dynamic' utility and go to the performance | Strip chart | FLARENET Plot 6. Again run the utility and select the Flarenet plot to view the Historical data 7. Save the file as an MS Excel file (*.xls) and this should have all the Depressuring information In FLARENET 1. Add the information manually for the control / relief valve in FLARENET from the Historical data for FLARENET 2. Multiple scenarios may be created for depending on what flow rate/ pressure condition you want to rate in FLARENET file. Keywords: dynamic, depressuring, save as Excel References: None
Problem Statement: Why is the coefficient not deleted once the pipe fitting has been removed?
Solution: Click on the LINK button after you delete a fitting. If you have already left the fitting screen, go back to that form (i.e. double click on the Pipe | Fittings tab), select the fittings type, then press the LINK button. This should remove the extraneous coefficient. Keywords: pipe, fitting, delete, coefficient References: None
Problem Statement: In the Solver tab of the Calculation Options Editor, there is a section to input Damping Factors for the properties and for the Loop. What is a damping factor and how is it applied in the calculation? What are the correct damping factors to get most accurate results?
Solution: In general, a damping factor is a multiplication factor that controls the step size used in the loops. The damping factor is used in the iterativeSolution procedure. The damping factor in the Properties field is the damping factor applied to the pressure step when solving the inner properties / pressure loop. The damping factor in the Loop field is the damping factor applied to the steps taken when solving the outer pressure / flow loop when solving looped systems. By default, both of the damping factors are 1.0. This number is suitable for most of the cases. Values less than 1.0 may be specified to prevent oscillations in properties / pressure loop and in the pressure / flow loop to improve convergence. Keywords: Damping factor, Solver, Loop References: None
Problem Statement: Why does rated flow fluctuate in pure component system with superheat 0 when MAWP is changing a little bit?
Solution: The problem here is that for a single component, it has a fixed bubble point/dew point. So when you flash with a exact zero superheat or subercool, Flare System Analyzer has difficulty to know the vapor fraction(it can be either 100% vapor or 100% liquid). The flash usually calculates a value based on which side it is converging from (so vapor fraction can be either 0% or 100%). The workaround here is to use a small superheat/subcool temperature such as 0.0001 for 100% vapor/ or a small subcool for 100% liquid . Keywords: pure component, superheat, subercool, rated flow References: None
Problem Statement: Why I am getting warning message: Mach Number Violation at Flange for source, Chocked Flow at Flange for Source?
Solution: Flarenet solves for the pressure backwards, starting at the flare tip. If at any point the local velocity is higher than the sonic velocity, the pressure at the point is increased until the velocity is just equal to sonic so that the Mach number is equal to one. Flarenet then continues to solve upstream. User would need to increase the flange size or reduce the flow rate to reduce the local velocity. Keywords: Mach Number Violation, Chocked Flow at Flange References: None
Problem Statement: How to add an expander at the outlet of the Relief Valve in Aspen Flare System Analyzer
Solution: Aspen Flare System Analyzer performs expansion/contraction calculations from the orifice of the valve to the tailpipe. If you want to consider the flange connection to the tailpipe you only need to specify the diameter of the flange in the source tab. However, very frequently there are additional enlargers or pre-expansion pipes between the relief valve and the tailpipe. To address this configuration a connector can be added between the valve and the tailpipe, this adds also automatically a pipe segment between the relief device and the connector. The recommendation is to set the non-existing pipe with a 0 length; generally this causes a choked flow warning, which can be ignored. For more information and tips we suggest you take a look at the webinar: Rate, Design and Debottleneck Your Entire Flare Network with Accuracy and Efficiency Keywords: Flare System Analyzer, Expansion, Relief Valve, Flange, Tailpipe References: None
Problem Statement: How do I create data tables in Aspen HYSYS?
Solution: Data Table in Aspen HYSYS V8.3 and higher enhances the usability of HYSYS Simulation. This let the users to monitor key process variables in both steady state and dynamic modes. Data Tables can be placed in the PFD and can be exported to Excel. Adding Data Tables 1. Click Data Tables in the Simulation Navigation pane. 2. Click Add in the Data Tables Window. New data tables are added under the Data Table header in the Navigation pane. 3. Right click the default name of the data table to rename it. Adding Variables to the Data Table 1. With the target data table selected in the Data tables folder of the Navigation Pane, click Add in the data table’s setup window. 2. Use the Variable Navigator to select the variables to include in the data table. 3. Use the Add button to add variables to the table. 4. Use the OK button to close the navigator. In Aspen HYSYS V8.4 the table can be placed in the PFD or exported to Excel. Keywords: Data Table References: None
Problem Statement: Different fitting loss methods for TEE give different results; which is correct?
Solution: Different fitting loss methods available for TEE's are shown below: Equal Static Pressure (Ignored) - balances static pressure across the tee. The pressure drop across the TEE is not calculated. Simple - uses a constant flow ratio independent K factor for the loss through the branch and run. The Simple method uses fixed loss coefficients taken from API 521 RP and considers only two types of branch angles (90 degrees and <90 degrees). The loss coefficients are measured from experiments where the entire flow follows a single path (e.g Run to Tail or Run to Branch). In this method the loss coefficients are based on static pressures and the coefficients are always positive (meaning there can be no pressure recovery). Miller - uses a K factor which is interpolated using Miller Curves, which are functions of the flow and area ratios of the branch to the total flow as well as the branch angle. Miller loss coefficients are taken from the book Internal Flows Systems by D. S. Miller. In this approach the loss coefficients are treated as continuous function of the relative flow rates in each leg of the tee. Results are also presented for many different tee configurations (dividing, combining and different branch angles). The Miller loss coefficients are based on total pressures and for some configurations negative loss coefficients are possible, indicating pressure recovery. Miller (Area Ratio Limited) This is the same as the Miller method, but the area ratio used for calculation of the pressure drop is constrained to a lower value of 0.10. This is the lowest value for which data is presented for the curves. In terms of applicability, both Simple and Miller are equally applicable in a given situation but given that the Miller method is more refined (since the loss coefficients are treated as functions of flow ratios) this can be considered a more rigorous model. As to how much difference this would make, this is impossible to answer in general since it is very case specific and depends on how important the tee pressure losses are for the system being modelled. Both methods (Simple and Miller) were developed based on incompressible flows, but in the absence of additional data are often applied also to compressible flows. Under Swage Method field: Compressible assumes compressible flow, Incompressible assumes incompressible flow and Transition first calculates the pressure drop assuming incompressible flow, but if the pressure drop (expressed as percentage of inlet pressure) is greater than the value specified in the Compressible Transition box, it is recalculated using compressible flow. There is an explanation on the Aspen FLARENET Keywords: tee, fitting, loss, pressure, drop, miller, simple References: Manual.pdf page 70. The Incompressible method calculations are faster but will be less accurate at higher pressure drops. The Transition method can cause instabilities in some cases if the calculated pressure drop is close to the transition value. Users are not recommended to use incompressible method if pressure drop is higher than 10% of the inlet pressure.
Problem Statement: How can a Flare System be modelled as a simple open pipe without any engineered tip?
Solution: The correct theoretical approach for an open-ended pipe is to set a Fittings Loss Coefficient of unity and use a Fittings Loss Coefficient Basis of Total Pressure (on the Flare Tip|Calculations page). Note however that the overall effect of this is that a static pressure drop of zero would be observed at the end of the pipe (hence a pressure equal to atmospheric immediately upstream of the tip). The reason is that the loss coefficients are defined in terms of total pressure, which is the sum of static pressure and velocity head, resulting in the pressure change contributions from each one (non-recoverable losses associated with the diameter change and recoverable losses associated with the reduction of fluid velocity from the tip velocity to zero) cancel each other out:- Change in total pressure, delta P = K * 0.5 * rho1 * vel1^2 [1] Also: deltaP = (P1 + 0.5 * rho1 * vel1^2) - (P2 + 0.5 * rho2 + vel2^2) [2] Therefore, substituting [2] into [1]: (P1 + 0.5 * rho1 * vel1^2) - (P2 + 0. 5* rho 2+ vel2^2) = K * 0.5 * rho1 * vel1^2 [3] Finally, noting that K = 1 (from abrupt expansion formula) and vel2 = 0 (ambient fluid velocity) results in: P1 - P2 = 0 [4] This implies that there is no change in total pressure. In practice, the loss factor (even for an open pipe) might be slightly higher than unity since such factors such as exit turbulence etc. must produce some non-recoverable losses but these would be extremely hard to quantify. For true pipe flare tips, there is normally all sorts of extra equipment around the opening tip for stabilizing the flame etc. therefore these always have a loss coefficient greater than unity. Keywords: Open Ended Pipe, Fittings, Flare Tip References: None
Problem Statement: Why are my results not calculating for the case file in Aspen Flare System Analyzer?
Solution: Go to Calculation | Options | Scenario Tab | Select 'Current Scenario' or 'All Scenario'. At present you might selected, 'Selected Scenario' and no scenario is selected. Also make sure in PFD view, you are looking for the scenario results you are interested in. Keywords: None References: None
Problem Statement: When opening APS or MBO the next dialog box appears stopping the program execution: with the message Splash.CreateInstance Failed - - HRESULT: -2146959355
Solution: In order to fix the problem follow the next steps: 1. Open a command window using admin rights. 2. For a 64-bit OS go to the next folder: C:\Program Files (x86)\Common Files\AspenTech Shared For 32-bit OS the folder is: C:\Program Files\Common Files\AspenTech Shared 3. Unregister and register AspenSplash.exe using the command /unregserver and /regserver. 4. If the regserver command was successfully executed the problem should be solved. 5. If the next error is thrown when executing the command: Please contact support to have the required libraries sent to you. Keywords: Splash.CreateInstance, Failed, error, APS, MBO, MSVCP71.dll References: None
Problem Statement: What are the different print options available in the PFD Toolbar for printing a PFD?
Solution: In the PFD toolbar, there is a PRINT button in the top left hand corner; next to this is a PRINT PREVIEW button to view a copy of what you are going to print. To the right of the PRINT PREVIEW button is a drop down menu which determines what is printed. There are four options here: All - Single Page The complete PFD is printed on a single page. Selected Items selected on the PFD are printed. To select a certain area of the PFD, you can click the mouse in the top left hand corner of what you want to select, hold down the left mouse button, and drag to the bottom right hand corner. This will select multiple items which will be shown on the printout. Visible Print the section of the PFD visible on the screen. All - Tiled The complete PFD is printed, but on multiple pages. To set the level of detail on each page (and hence the total number of pages), go to File | Preferences | Printing and change the PFD Tiled Scale Factor. The smaller the value here, the more detail gets included on each page (and hence fewer pages are printed). Finally, using File | Printer SetUp, you can set whether to print PFD in Portrait or Landscape mode. Keywords: PFD, print, landscape, portrait, options, visible, tiled, PFD, toolbar References: None
Problem Statement: Is there any way in APS to set a default rate so that when a user creates an event the rate is automatically set, so there is one less thing for the user to do?
Solution: The default rate is specified in the TRANS_MODE table. The EVENT_TYPE column identifies the event type to which the transportation mode/rate applies. If an specific event type has not been defined in the TRANS_MODE table. The default rate to be used is the first rate with EVENT_TYPE set to NULL. E.g: Notice the next TRANS_MODES table. The first Event type that has not been specified is RAIL transportation mode. So the events that are not type 19 (Product receipt event), 16 (Pump Event) and 15 (Product Shipment) will have a default rate of 780. A Crude Transfer (not defined in TRANS_MODE) will have a default rate of 780: And a Product receipt event (defined in TRANS_MODES) will have a rate of 150: Keywords: Default rate, TRANS_MODES,EVENT_TYPE References: None
Problem Statement: How can I change default pipe segment icon color scheme on PFD? How can I change font type and size on PFD?
Solution: By default pipe segment color scheme is in grey color. But if user wish to change then it can be changed to white color icon by changing below settings File/Preferences/PFD/ Please select check box Use wire frame icon in display File/Preferences/PFD/ Please select font type from drop down menu box (by default it is in Arial) File/Preferences/PFD/Please select font size from drop down menu box Then click on OK Keywords: Pipe segment color, Font size, Font type etc; References: None
Problem Statement: When printing the MBO Blend Details dialog (using the PRINT button), it would be nice if it would print in color so that the hard copy would have all the great new color indicators that you see on screen.
Solution: 1. Turn on the printing of background color in Internet Explorer. From 'Internet Explore', Tool | Internet Options | Advanced tab, 2. Open your model in MBO 3. Go to Blend Details dialog 4. Click Print to File, verify the html file has table printed in the same colors as on the dialog 5. Click Print, select a color printer, verify the colors are printed out Keywords: References: None
Problem Statement: What is the consequence of changing the ‘Source inlet velocity basis’ option in the Calculation Options Editor, shown below?
Solution: This option influences the velocity used at the inlet of the source element (valve) for purposes of calculating the kinetic energy in the energy balance across the valve only. It is not a velocity specification. Thus, when the option ‘Zero Velocity’ is chosen, the static pressure upstream the valve does not equal the total pressure (i.e. the dynamic pressure is still calculated with the Inlet Pipe Velocity). In this case, the difference in enthalpy across the valve (hence, downstream pressure and temperature) will depend only on the velocity at the outlet, and not influenced by the inlet pipe size. If the ‘Inlet Pipe Velocity’ is chosen as basis, a dependence on downstream pressure and temperature on inlet pipe size appears (specified in Valve Editor, ‘Inlet piping’ menu), since the velocity upstream the valve now influences the energy balance across the valve (in the kinetic energy term). Keywords: Source, Inlet Velocity, Basis, Zero Velocity, Inlet Pipe, Kinetic energy References: None
Problem Statement: What does Beta=1 and Theta=0 mean in the fitting that is called Ball Valve (Beta=1, Theta=0)?
Solution: Beta (β) is the ratio of valve seat diameter divided by pipeline diameter. β = dvalve / dpipe Theta (θ) is called the reducer angle of approach. So, a value of Beta =1 and theta = 0 means that the valve has the same diameter as the pipe. These two parameters are used to calculate the resistance coefficient for the valve K Keywords: Ball valve, fitting, pipe, Fitting loss coefficient References: None
Problem Statement: How to obtain the whole system volume of Flarenet's network?
Solution: There is no direct way to get the whole volume of the network. The simple Excel Macro is written to calculate the volume. In order to run it, you have to keep the flarenet case study open. The attached version is only working for Flarenet 4.1 or lower version. If you have new version, you can change the reference in Macro editor and change flarenet declaration (such as Flarenet -> Flarenet2006_5). Keywords: Pipe, Macro, Excel References: None
Problem Statement: How do I add a constant superimposed back-pressure on a pressure safety valve in an Aspen Flare System Analyzer model?
Solution: The pressure at the flare tip is equal to the atmospheric pressure plus the pressure drop across the flare tip. This value is also the superimposed back pressure. There is no option to directly set the superimposed back pressure to a constant value. However, the user can use the following workarounds that models the same situation. 1. Using flare tip curves to manipulate DP so you can have the desired superimposed back pressure.  On the flare tip input form, go to Calculations and check the box for Use curves. On the Curves form, add points to the curve so that the pressure drop across the flare tip is zero for all flow rates. 2. Fixing the atmospheric pressure to be the desired superimposed back pressure. Go to Home | Options | General and set the atmospheric pressure to the desired constant value for superimposed back pressure. Keywords: Flarenet constant superimposed back pressure References: None
Problem Statement: How to add Mercaptan as a component, I can NOT find in the component database?
Solution: Select M-Mercaptan from the component list. Note: You can use * as wildcard character if you do NOT know the first few letter or last few letter of the component name. For example for Mercaptan type in *Mer, you will see the component name. Keywords: Mercaptan, component References: None
Problem Statement: In the File | Preferences... menu, what does the Display Velocity Properties Used By Pressure Drop Calculations option do?
Solution: In Aspen FLARENET, there is a checkbox on the General pagetab of the session preferences labeled Display Velocity Properties Used By Pressure Drop Calculations. When this option is activated, the velocity dependent properties (i.e Mach number, rho v^2 term, Reynolds number etc.) that are displayed for a given pipe are based upon rated flow rather than nominal flow if the Rated Flow For Tailpipes calculation option is enabled. Keywords: display, velocity, properties, rated, flow References: None
Problem Statement: Is there a way to set the system to bubble point or dewpoint?
Solution: If you enter a sub-cooling value, this is the temperature difference between the bubble point and the stream temperature. FlareNet calculates the bubble point, then the required temperature based on the specified sub-cooling. Similarly, for superheating, FN calculates the dew point, then the required temperature. To do this, you need to be using one of the VLE methods, since these are needed to get the bubble / dew points. In your model, rather than specifying stream temperatures, you could have specified the vapor composition, and a superheat value of zero. This would force the stream to be at its dew point. Keywords: bubble point , dewpoint References: None
Problem Statement: Which method should I use 'Compressible' or 'Incompressible' for Swage calculations of Relief Valve?
Solution: Under Swage Method field: Compressible assumes compressible flow, Incompressible assumes incompressible flow, and transition first calculates the pressure drop assuming incompressible flow, but if the pressure drop (expressed as percentage of inlet pressure) is greater than the value specified in the Compressible Transition box, it is recalculated using compressible flow. There is an explanation on the Aspen FLARENET Keywords: Compressible, incompressible, Swage, Relief Valve References: Manual.pdf page 70. 1. Compressible - pressure losses will be calculated assuming compressible flow through the relief valve at all times. 2. Incompressible - pressure losses will be calculated assuming incompressible flow through the relief valve at all times 3. Transition - pressure losses will be calculated initially using the assumption of incompressible flow. The Incompressible method calculations are faster but will be less accurate at higher pressure drops.The Transition method can cause instabilities in some cases if the calculated pressure drop is close to the transition value. Users are not recommended to use incompressible method if pressure drop is higher than 10% of the inlet pressure.
Problem Statement: Where can I see the flow map in Aspen Flare System Analyzer? What is the basis?
Solution: User Can able to see Flow Regime results from Toolbar/View ->Results-> Flow Map The flow map available in Flare System Analyzer displays the flow pattern correlation of Gregory Aziz and Mandhane, which is currently the most widely used method. It was based on almost 6,000 flow pattern observations, from a variety of systems, and many independent studies and it is strictly applicable only to horizontal flow. Typically, the superficial gas and liquid velocities in a horizontal pipe are the most important single parameters influencing the flow pattern. After running the case, you can view the Gregory Aziz and Mandhane flow map by selecting Results and then flow map from the View menu. You can display the flow map for each pipe segment by selecting the desired pipe segment from the drop down box on the top of the view. The upstream and downstream conditions are marked with a red dot and a label on the flow map. Unless the pipe segment has a single phase flow with a large pressure drop, both upstream and downstream pipe conditions will generally be close to each other. Keywords: Flow map, Flow Regime, Gregory Aziz and Mandhane method References: None
Problem Statement: Can I directly input fluid properties in Aspen Flare System Analyzer?
Solution: In Aspen Flare System Analyzer, you can not directly input the properties of the fluid. You have to set up the components or create hypos to represent your fluid. You have to input composition or molecular weight for each valve. The fluid properties will be calculated rigorously based on your input and thermodynamic model. Keywords: Fluid properties, PSV, control valve References: None
Problem Statement: How is the MABP approach calculated?
Solution: The MABP Approach is calculated as follows: As an example, please refer to PSV 1 located in Sample-S1 (C:\Program Files (x86)\AspenTech\Aspen Flare System Analyzer V7.3\Samples\Convergent). The atmospheric pressure in this case was set as 1.01325 bar. The Downstream static pressure is reported in the Summary tab in the Relief Valve Editor and the MABP was defined in the Conditions tab. Therefore the calculated MABP approach in this valve is: Keywords: MABP, approach, allowable backpressure, static pressure References: None
Problem Statement: How to add Single Source Scenarios for multiple relief and control valves?
Solution: There are two options available A. Go to tools on the menu bar and select Add Single Source Scenarios. If there are 100 relief valves, then this step will add 100 new scenarios with only one valve relieving to the network. B. Clone the base scenario on the Build menu and Scenario option. Edit each scenario and activate only one scenario on the sources tab. Keywords: Add, Scenario, Scenarios, Source, One References: None
Problem Statement: How to enter the vendor curve information for the flare tip
Solution: Aspen Flare System Analyzer allows the user to input a characteristic pressure drop curve. The program interpolates along the curve to calculate the pressure drop based on the flow rate. Aspen Flare System Analyzer also allows the user to supply multiple curves at different reference molecular weights to account for variations in the composition of the fluid being flared. The following procedure illustrates what is needed to input a single curve: 1. Open the property view for the flare tip and select the Calculations page. 2. To enable the use of curves, check the Use Curves box. 3. Select the Add button on the Curves tab to add a pressure drop vs. flow value. This will add another row to  the table which can then be filled in with the appropriate data. Keep adding rows until the requisite data has been input. If you wish to add multiple curves at varying molecular weights, do the following: 4. Select the Add Mol. Wt. Button to add an additional curve. (Note: a default molecular weight will be assigned. To change the default value, highlight the molecular weight and type in the new value). 5. Input a reference temperature for the newly created curve. 6. Select the desired curve using the Molecular Weight drop down menu. 7. Once the proper curve is selected, follow the same procedure as listed above to input the curve data. As a final note, the program will not extrapolate past the endpoints of the supplied data unless the Mol. Wt. Extrapolation or Flow Extrapolation boxes are checked. Keywords: Curve, Flare Tip References: None
Problem Statement: Is it possible to ignore a pipe in one scenario, but allow it in other scenarios?
Solution: Please note that in Aspen Flare System Analyzer the geometry and configuration of network are fixed. If you change something (i.e pipe diameter) in one scenario, it will be changed in all other scenarios too. What you can do that is only applicable per scenario is to change source conditions and ignore some unit operations. Keywords: scenario configuration, ignore unit operations References: None
Problem Statement: How to add multiple lines of Annotation on the PFD?
Solution: Use the letter A on the PFD toolbar to enter Annotation. At the end of first line enter CNTROL and ENTER key on the keyboard to type in next line of annotation. Press Enter key only on the keyboard once you finish entering the Annotation. To edit the annotation double click on the typed annotation on the PFD. Keywords: Annotation, PFD, Title, Text, Line, Multiple References: None
Problem Statement: What is the difference between the 'mixture velocity' and the 'superficial phase velocity' constraint options?
Solution: When check vel. constraint with is set to Mixture velocity in Scenario Editor | constraint tab the velocity violation occurs when the Mixture velocity is higher than lower of liquid velocity constraint & vapor velocity constraint, When check vel. constraint with is set to superficial phase velocity in Scenario Editor | constraint tab the velocity violation occurs when the superficial phase velocity is higher than corresponding phase velocity constraint. Keywords: Mixture velocity, superficial velocity References: None
Problem Statement: Why am I unable to connect objects on the Aspen FLARENET PFD?
Solution: Try pressing the Toggle Connect/Arrange Mode button. While in Arrange mode, the user can move items on the PFD, but is not given the ability to connect objects. Keywords: toggle, connect, arrange, PFD, object References: None
Problem Statement: How to read Aspen Flarenet file, saved in newer version, in older version?
Solution: User would need to export the File in MS Access / Excel / XML format and then the file can be imported back in older version of Flarenet to read it. For example, if file is saved in v 2006.5 and user wants to read it in version 3.51a, follow the steps below 1. Go to File | Export Wizard (Cntr+E) and export the file in version 2006.5 2. Open Aspen Flarenet v 3.51a and Go to File | Import Wizard (Cntr+I) and import it. Notes: 1. For using export import wizard refer to Aspen Flarenet reference manual orSolution id: 119474 2. To import all scenario at the same time, Go to Calculations | Options | Scenario and select All Scenario from the drop down menu 3. If user is trying to merge the file with old file, while importing remove the check box for 'Import as New' in the first step of import wizard process. Keywords: old, new, version, old version,new version, file, read References: None
Problem Statement: How does one include the file path and filename when printing a PFD?
Solution: Select File | Preferences from the main menu in Aspen FLARENET, then activate the Use Footer checkbox on the Formatting tab. Keywords: filename, file, path, name, PFD, print References: None
Problem Statement: What pressure drop correlations are available in Aspen FLARENET and when should they be used?
Solution: A brief description of the different pressure drop correlations available in Aspen FLARENET is provided below. For more information on these methods please refer to the Aspen FLARENET Keywords: Pressure, Drop, Correlations References: Manual. Isothermal/Adiabatic Gas These methods are for vapour phase only and as such should not be used to model fluids if a property method capable of predicting liquid is used, unless you are confident that you are in the single phase (vapour) region. In both cases acceleration effects are modeled, but changes in pressure due to elevation change are ignored. The adiabatic effect is strictly speaking the more accurate of the two methods since it takes into account the density changes due to JT effects as you progress along the pipe. It is however slower than the isothermal method and also unfortunately has a quadratic form that can give two numerically viable
Problem Statement: When running Aspen Flare System Analyzer in a Window Server 2003 you received the following error: .
Solution: This isSolution only applies to the Windows Server 2003 operating system. When running Aspen Flare System Analyzer on a Window Server machine you need to explicitly exclude it from the Data Excecution Prevention Services check. To do this, right click on My Computer and select Properties. On the Advanced tab, click on the Settings button found in the Performance group box. In the ensuing Performance Options window that opens, click on the Data Execution Prevention tab and activate the second option (i.e. Turn on DEP for all programs and services except those that I select:). Click the Add button browse for the Aspen Flare System Analyzer executable file (flarenet.exe). Click Open to select the file, then OK to close the Performance Options and System Properties windows. The problem will be resolved in the Aspen Engineering Suite V7.3 release. Keywords: V7.2 Flare System Analyzer, Windows server 2003, DEP References: None
Problem Statement: I have multiple Aspen HYSYS versions on my computer. How can I select which version to use when importing data? Aspen Flare System Analyzer is trying to open the Aspen HYSYS file with an older version.
Solution: Starting with Aspen HYSYS version 2006, select Programs | Aspentech | Aspen HYSYS | Restore File Associations from the Windows START menu for the desired version of HYSYS. For earlier versions of Aspen HYSYS (i.e. v2004.2 and older), simply open and close the version you wish to use, and this will automatically be assigned as the current version. Keywords: Aspen Hysys, Aspen Flare System Analyzer, import References: None
Problem Statement: What is the maximum allowed molecular weight of a hypocomponent in Aspen Flare System Analyzer?
Solution: In the field of molecular weight the program requires a numeric value not smaller than 2 and not larger than 500. This is because if you defining a hypo component based only on the molecular weight, a value higher than 500 can lead to a poor estimate of properties and a critical temperature above the limit. On the other hand, if you define a hypo component based only on the Normal Boiling Point, the molecular weight estimated can be larger than 500. In fact, the maximum molecular weight supported is 1000. Ideally to get the best estimate, all of the following parameters; molecular weight, Normal Boiling Point and Standard Liquid Density should be specified. Otherwise, the estimate may not be good and will not be accepted when you press OK. Normally, flare systems consist on low molecular weight hydrocarbons. If you have very heavy components with high molecular weight and high boiling point and the composition of these components is less that 1%. You may exclude these components from your simulation. Keywords: molecular weight, maximum, hypo component References: None
Problem Statement: Aspen Flare System Analyzer heat transfer solver is calculating unrealistic temperature at pipe outlet. What should I do?
Solution: 1. Choose Yes for Multiple Element Calculation in the Heat Transfer tab of the corresponding pipe. This will make sure that the heat transfer calculations are performed in as many elements as hydraulics calculations are performed. If heat transfer calculation are failing, using multiple elements will most likely solve the problem.Solution document 131691 has more information regarding number of elements. 2. Make sure that external medium velocity, insulation thickness etc are correct (both numbers and units). Keywords: Very high, very low, temperature References: None
Problem Statement: How to model a constant pressure drop in a horizontal separator.
Solution: In Flarenet, there is no direct way to set a pressure drop in KO drum. Since KO drum only considered inlet and outlet pressure changes, we can model a bleed for constant pressure drop between two KO drums to represent the horizontal separator. In the attached example, you will find that a bleed is used to set the constant pressure drop and zero pipe with the same diameters with KO drums to link two KO drums as a whole. Keywords: Horizontal separator, constant pressure drop, bleed References: None
Problem Statement: What is the meaning of fields on the calculation tab of flow bleed?
Solution: Flow bleed is the calculation block that allow you to specify a fixed pressure drop. Following input fields are available on the Calculations tab: Field Description Offtake Multiplier Specify the offtake multiplier. The default value is 0. Offtake Offset Specify the offset fo the offtake to compensate the changes in the inlet flow Offtake Minimum Specify the minimum value for the offtake Offtake Maximum Specify the maximum value for the offtake Pressure Drop Enter the pressure drop across the Flow Bleed Keywords: definitions, values, specifications References: None
Problem Statement: Why do I receive inlet piping pressure drop violation for ignored relief valve?
Solution: The warning message Inlet piping Non -recoverable pressure Drop Violation for ignored Sources is related to the inlet pipe to the relief valve and not related to the flare network in which the flow is ignored. It does not matter whether the valve is ignored or not, Aspen Flare System Analyzer calculates them for all valves. API recommends that the pressure drop in the inlet pipe should not be more 3.0% of the MAWP at rated flow of the relief valve. That's why this message is appearing for the valve. However user have an option to let the solver NOT display the warning message from version V7.1. Go to the calculations menu and select Options. On the Warning tab, check the option for Ignore Inlet Pipe Warnings for Ignored Sources. Keywords: Inlet, Violation, Inlet Piping pressure Drop Violation, Flarenet References: None
Problem Statement: I'm getting an Inner Loop Failure message when running my Aspen FLARENET case. What does this error mean and how do I correct it?
Solution: The Inner Loop Failure message usually indicates that Aspen FLARENET has just not been able to reach aSolution, and is often a result of flows that are too large and/or diameters that are too small. Increasing pipe diameters and/or reducing flowrates will usually rectify the problem. Other things to try/check are: 1. Go to Calculations | Options | Solver and set the following options: Unit Operations Pressure = 1.0e-05% Properties Pressure = 1.0e-05% Unit Operations Pressure = 1.0e-05% Properties Iteration Limit = 50 2. Ensure that an appropriate property package is being used. 3. Ensure that an appropriate pressure drop correlation is being used. 4. Check for choked flow, particularly around tees. If this message is the last message to appear before the Calculation time, then that indicates that Aspen FLARENET has failed to reach aSolution. If however it appears only as an intermediate error message and Aspen FLARENET then indicates that it has solved, the case has successfully converged (although the Problems tab should still be checked for any constraint violations). If you are unable to converge your flowsheet or you are unsure whether or not your flowsheet has converged please contact technical support for assistance at [email protected]. Keywords: convergence References: None
Problem Statement: Why does the component selection from a specific component family (for example, hydrocarbon, aromatics etc) affect the relieving capacity of the Relief Valve in Aspen Flare System Analyzer.
Solution: Depending on the component type selected on the relief valve editor, (i.e Relief Valve >> Compositions Tab >> Fluid Type Drop down option), the program calculates the different properties based on the components in which the composition is splitted internally. This is expected, and will therefore impact the rated capacity of the valve; this can be observed under the Rated flow parameters group of the relief valve editor. In the attached sample case (depending on the basis selected) the composition is split between Methane & Ethane for HC selection and H2O / N2 for Misc. fluid type selection. If the molecular weight is selected as the fluid composition basis, then the composition used by Aspen Flare System Analyzer will only be the two components with Molecular Weights either side of the specified value with the relative composition to give the specified Molecular Weight. Conversely, if the compositions of each individual source component is entered, then the actual compositions entered by the user will be used by Aspen Flare System Analyzer for the property calculations. Keywords: Molecular Weight, Fluid Type, Filter, Composition References: None
Problem Statement: How do I import source data using text file format?
Solution: Originally such feature was designed to import the HYSIM files in to Aspen Flare System Analyzer. Since HYSIM is no more released in the market, so this feature is not supported. However, you can use the following information. Refer to the attached text file. Text_Import.prn: File to be used for import. You need to copy exactly same format to be used for import purpose. Text_Import_ANNOTATED.prn: Not suitable for import - It helps you understand each line in the importable text file datastructure.text: Sample Data structure. It will help you if you trying to develop some automation to import the data to Aspen Flare System Analyzer Keywords: import, Export, text, format References: None
Problem Statement: When starting Aspen FLARENET, the following error message Pipe Schedule Database Missing appears. How does one eliminate this problem?
Solution: By default, the database files for components, pipe schedule and pipe fittings are not installed in the 'FLARE NET' folder; instead, they are usually found in the following location: C:\Documents and Settings\All Users\Application Data\Aspen Tech\Aspen FLARE NET(version)\Database. In this folder, there should be three database files: FITTINGS.MDB, SCHEDULE.MDB and COMPS.MDB. Within the Aspen FLARENET application, you can view the location of the database files via the Databases tab of the session preferences (select File | Preferences from the main menu). Ensure that the locations specified refer to the path name for the folder which contains the files FITTINGS.MDB, SCHEDULE.MDB and COMPS.MDB (as indicated above. If the error message Pipe Schedule Database Missing appears, the cause may be due to one of two problems: 1. The database files Comp.mdb, Fittings.mdb and Schedule.mdb are missing, and were not copied correctly during installation. 2. The path name indicated on the Databases tab of the session preferences is incorrect (i.e. the actual path name for the existing database files differs from what is shown). If database files are missing, there are two ways to resolve the problem: 1. Un-install then re-install Aspen FLARENET, ensuring that you select the Install this feature and all of its sub features option from the installation dialog so as to install all required components (instead of installing the components individually). 2. Unzip the attached files into the folder C:\Documents and Settings\All Users\Application Data\Aspen Tech\Aspen FLARE NET(version)\Database folder and ensure that the path indicated on the Database tab of the session preferences points to this same location. If the path name indicated on the Databases tab of the session preferences does not match the actual location of the database files, locate the three files described above and note the folder name. Next, open Aspen FLARENET and view the Databases tab of the session preferences. Ensure that the correct path is entered for each of the entries (or browse to the desired location by pressing the appropriate drop-down button). Please note that the entries need to include the file name so, assuming the default file locations, the three lines should read: C:\Documents and Settings\All Users\Application Data\Aspen Tech\Aspen FLARE NET (version) \Database\comps.mdb (for Components) C:\Documents and Settings\All Users\Application Data\Aspen Tech\Aspen FLARE NET (version)\Database\fittings.mdb (for Pipe Fittings) C:\Documents and Settings\All Users\Application Data\Aspen Tech\Aspen FLARE NET (version)\Database\Schedule.mdb (for Pipe Schedule) Keywords: pipe error, database file, components, schedule, fittings References: None
Problem Statement: Why the temperature increases through connector or tee instead of temperature decrease?
Solution: In the compressible flow, flarenet use equation Eq 2.57-2.63 in HTFS DR38 for swage calculation. Equation looks like H1 + (M1^2)/ )/(2 *rho1^2) = H2 + (M2^2)/ )/(2 *rho2^2). H1 and H2 are inlet and outlet enthalpy M1and M2 are inlet and outlet mass flowrate There are two types of calculation that you can do in Flarenet regarding Energy Balances. You can either include kinetic energy in the balance or ignore changes in kinetic energy (Calculation > Options > General tab). If you ignore kinetic energy then the change in temperature is much smaller. However, when you include kinetic energy, as there is an overall reduction in the velocity (it is dependent on density changes or Cp/Cv)), this means that there is an increase in enthalpy in order to maintain the energy balance. Therefore H2 = H1 + (M1^2)/(2* rho1^2)- (M2^2)/ (2* rho2^2). If H2 is significantly higher than H1, the temperature could be higher. Sometime, temperature also increases during the static pressure recovery. Static pressure in downstream could be higher than upstream. Keywords: References: None
Problem Statement: Why I don't get a choked flow error message for my orifice?
Solution: In Aspen FLARENET, the thin plate orifice equation is used for orifice calculations. A thin orifice won't choke, as the flow will continue to rise as the downstream pressure falls further because the vena contracta grows in diameter. Ref : Orifice Meters with Supercritical Compressible Flow, RG Cunningham; ASME, July 1951 Keywords: orifice, choke References: None
Problem Statement: What are the best practices to build a simple Aspen Flare System Analyzer model?
Solution: A. Components Basis 1. Use MW (Molecular Weight) as composition basis for only vapor phase systems. 2. Use Composition basis as mole fraction or mass fraction for multi phase systems. 3. Components with composition close to zero should be removed from the component list. Any additional component increases the calculation time by the square of total number of components present. 4. If you have multiple hypothetical components, calculation time may be reduced if the hypothetical components are combined together. For this step, choose the component in the selected components column and click on combine button. B. VLE, Enthalpy and Pressure drop method 1. For single phase vapor models, Compressible gas can be selected as the VLE method. For multiphase systems always select EOS models such as Peng Robinson. 2. Ideal gas is allowed as enthalpy method with compressible gas. Use EOS such as Peng Robinsons / Lee kesler as enthalpy method once VLE method is EOS such as Peng Robinson 3. Always check the pressure drop by selecting Compressible methods for swage for high Mach number (>0.3). It may take longer to converge the file but results are more accurate. Also for H2 rich system, use Compressible method for Swage. As incompressible method limit the Sonic velocity to 500 m/sec. Sonic velocity for H2 can be as high as 1500 m/sec. For Methane and other hydrocarbon 500 m/sec is a good assumption but not for H2. C. PFD Setup 1. Always connect the red dot as inlet stream and blue dot as outlet stream for pipe, TEE, orifice and connector. 2. Use connector where you do not need TEE. For example if the branch of connector is not connected to any pipe, use connector instead of using TEE. 3. Keep the PFD tidy. You can also regenerate the PFD to keep it tidy, using View | PFD | Regenerate option. Save your file, before doing it as if you do not like it, you can not go back. You need to start from backup copy. 4. Whenever modifying the network (such as adding pipe or changing pipe size), always modify in a base scenario (such as default scenario). Then clone a new scenario from base scenario to change scenario specific input. D. Convergence 1. For relief valves in multiphase systems, enter the estimates for the vapor fraction and MW on the method tab. 2. Since version 2006, for knock out drums, enter the mole fraction of the lightest component as 1 and delete the rest of the composition on the composition TAB of the KO drum. This helps improve stability and results in faster convergence. The value of 1 is only used when there is no vapor flow from KO drum. 3. Flarenet does not support multi-phase separators inside the loop in a looped system. If you have a case where separators are there inside the loop, and have multiphase feed, flarenet will not converge. When you try to solve, if you have trace window open, you will observe the relevant warning. Hence the way to approach this problem, is to identify the separators that are inside the loop & ignore them & try to converge the case. Once that converges, you can continue to unignore the ignored separators one by one & try to make it solve the case. If the separator that you unignore, gets multiphase feed, it will not solve. Then you move on to the next separator. 4. For closer boiling point mixtures, sometime convergence will fail at KO drum. If we introduce trace amounts of other chemicals in the composition of PSV, the case will solve easily as it widens the boiling range thus helping the KO drum to converge. 5. When in Design mode, provide reasonable initial pipe size estimates. This helps in solving the network in less time. 6. If the case file contains lot of loops and it fails to converge, then try to minimize the loop by ignoring some of the pipes. Also clear the estimates from Build | Scenario | Edit the scenario | Estimates tab | Press clear button. 7. For inner loop failure message, if it does not converge at default values then tighten the pressure tolerance (Properties | unit operation | loop) on the Calculation | Options | Solver Tab to 1 e-5 from the default value. If it does not resolve the issue, you might be choking the flow some where in the loop. Try to reduce the flow rate or increase pipe size. Select the sources that has maximum flow rate to reduce the flow rate. 8. For outer loop failure it is best to reduce the number of loops by ignoring the pipes. Once you get converge file, un-ignore the pipe one by one. You need to run the solver each time you ignore the pipe. Also ensure the option to keep estimate updated for convergedSolution is selected under Calculation | Options | Solver Tab. E. File backup 1. Under File | Preferences | General TAB, activate the autobackup option if you do not keep a manual backup, while making changes to the PFD. Once you are ready to start running the file, save one copy as backup. Remove the autobackup option before starting to run the file. 2. While closing Flarenet, first save the file and then close it from File | Exit. If you close the file from 'X' button, then Flarenet file may get corrupted as it fails saving all the information and closes the Aspen Flare System Analyzer first. 3. Using Export option, you can also keep your file backup in MS excel / MS Access format. For project work in final stage, I would recommend to use this option as well to have file backup. Once you loose your information, MS Access / MS excel file can also be imported in Aspen Flare System Analyzer. F. Export and Import feature 1. 119474 is aSolution I thought you may find useful: The topic is How do I use the Export \ Import feature in Aspen Flare System Analyzer?. It details you the step needed for export and import. 2. The import export feature can be used for following purpose Convert your file from 3rd party software to Aspen Flare System Analyzer Save your file in excel, XML, access format to be used later on. Merge multiple Aspen Flare System Analyzer files in to one single file Open the file (saved in the latest version) in a older version Cleanup the Aspen Flare System Analyzer file for any corrupt data. For example your Aspen Flare System Analyzer case file does NOT run, but you can open and close it. Keywords: best practices, best, practices, component, VLE, enthalpy, convergence, inner loop, outer loop, failure, backup, file, tips, Third Party, 3rd Party References: None
Problem Statement: Why is the flaretip velocity calculated as zero?
Solution: The Aspen FLARENET solver does not calculate flaretip velocity. This calculation should be done using third party software such as Flaresim?. For more information on this product, visit: http://www.softbits.co.uk/fs_overview.php Keywords: flaretip, velocity, Flaresim References: None
Problem Statement: The values for K(Cp/Cp-R) and Compressibility appear as empty values when reviewing the relief valve conditions.
Solution: One possible reason for this is that you directly specified the Rated Flow instead of allowing Aspen Flare System Analyzer to calculate it. As explained inSolution 135381, these values should only appear if the Auto checkbox has been activated. Note that the Rated flow parameters can also show up if you click on the Set button because this also allows Aspen Flare System Analyzer to perform the calculation of the rated flow. If these values do not show even after activating the automatic calculation of the Rated Flow, check the calculated vapor fraction at the relief valve conditions. If this is 0, Aspen Flare System Analyzer will show empty values because these parameters cannot be calculated for liquids. Keywords: Rated Flow Parameters, Compressibility, Specific Heat Ratio, K(Cp/Cp-R), Empty, Blank References: None
Problem Statement: Why do I receive the message “Sequence contains more than one matching element� when attempting to import data from Safety Analysis into Aspen Flare System Analyzer?
Solution: This message shows up if there are multiple valves on the Aspen HYSYS/Aspen Plus file that share the same name. When this happens, Aspen Flare System Analyzer (AFSA) will not be able to complete the import process, even if the user is not interested on importing this duplicated valve. In some cases, it has been reported that after deleting a valve in the Safety Analysis environment, this is not completely removed from the valve database, which can generate the problem if a newer valve has the same name as this deleted valve. From the import Relief Valve Sources tool of AFSA, the user is able to see the complete list of valves, and will be able to identify which valves are being found as duplicates: Once the duplicated valves have been identified, the user should try to rename the available valves inside Aspen HYYS/Aspen Plus and repeat the import process of the PSV information into AFSA. Keywords: Error, Import, Safety Analysis, Duplicate, Matching, PSV, Relief Devices References: None
Problem Statement: Are tank inventories used during each period optimization in Aspen Refinery Multi Blend Optimizer?
Solution: Yes. MBO balances all tanks at every period ensuring that the inventory constraints are enforced. If there is no way to make a tank feasible in a given period then penalty structure will be activated. The penalty value used is based on the interpolation between the time of the particular period and the beginning and end of the campaign. Keywords: Periods Tank inventories Multi-Blend Optimizer References: None
Problem Statement: Why does the 'wall temperature violation' warning appear for a pipe segment?
Solution: The Wall Temperature reported in the P/F Summary is the internal wall temperature. The reported wall temperature in the P/F summary is an average over the length of the pipe. The wall temperature violation may appear sometime even if you do not expect the average temperature to go beyond specified temperature range for the pipe. If you get wall temperature violation warning message, also check the heat transfer calculation error message on View ll Results ll Messages ll Solver tab. If the heat transfer calculation failed then you will see the message here. Please check the following 1. Select the Calculations ll Options menu item, and make sure Enable Heat Transfer option is checked (on the General page tab). Also under Energy Balance, make sure Include Kinetic Energy is checked. 2. On each pipe, go to the Heat transfer tab and select Yes for the Heat Transfer Enabled option 3. Make sure you are using the correct VLE method under calculation ll options ll methods. For example if the source is mostly methane, you are releasing it at ambient temperature, most likely all the stream will be vapor phase only. In that case use compressible gas as VLE method and ideal gas as enthalpy method. 4. Run the solver. If still you are getting the same error message, you need to increase the number of elements on calculation ll options ll methods tab. the default is 10. Increase to 25 and then 50. Note: For two-phase calculations, the pipe segment is divided into a specified number of elements. On each element, energy and material balances are solved along with the pressure drop correlation. In simulations involving high heat transfer rates, many increments may be necessary, due to the non-linearity of the temperature profile. Obviously, as the number of increments increases, so does the calculation time; therefore, you should try to select a number of increments which reflects the required accuracy. Keywords: temperature, pipe wall, pipe wall temperature, wall, wall temperature, pipe temperature References: None
Problem Statement: The PFD is always printing in conner of page and not in middle of page.
Solution: There may be empty space on the top left side of the PFD corner. Zoom in or out your whole flowsheet and make the whole network visible. Then make sure the bottom and right window scroll bars of PFD are at the end. Select whole network and drag it to the left-top coner of PFD. Click print review and check whether it is in middle of page or not. Keywords: Printer, PFD References: None
Problem Statement: How will AspenTech handle the support activity for the old HTFS Research Network (web version) after we release the new HTFS Rresearch Network (aspenONE Exchange) starting in V8.8?
Solution: Starting in V8.8, all HTFS Research Network resources can be accessed from aspenONE Exchange, which is accessible in Aspen Exchanger Design and Rating (EDR), Aspen Plus and Aspen HYSYS programs. AspenONE Exchange offers a convenient way to search research literature and documentation about EDR and no longer needs a login account. We encourage all users to upgrade to V8.8 or later versions. In case you cannot upgrade but still want to use the old HTFS Research Network website, our support policy is documented below: 1. If you are an existing user with an HTFS Research Network account (note that this is a seperate account from Aspentech support account), you can still access the website (//htfs.aspentech.com/ ) with your current account and password till end of you current contract. 2. If you are an existing user but you have problems accessing the website (such as forgotten account name or password or the server goes down), you can contact [email protected] to request our technical support. 3. If you don't have an account in the old HTFS Research Network website, we no longer provide any new accounts and you will need to upgrade to V8.8 or later to access HTFS Research Network documents. Keywords: HTFS, Web Page, Support Policy, V 8.8, Account References: None
Problem Statement: I am getting error messages for Inlet Piping pressure Drop Violation even though the Pressure Safety Valve(PSV) is ignored and I have ticked the option for calculate ignored sources with zero flow.
Solution: American Petroleum Institute(API) recommends that the pressure drop in the inlet pipe should not be more 3.0% of the Maximum Allowable Back Pressure (MAWP) at rated flow of the relief valve. This is a standard engineering practice for designing inlet pipes (that is the pipe connecting the source and the relief valve). It is not part of the network but an additional useful calculation and in no way corresponds to Aspen Flarenet Solver. So it is not affected by whether a relief valve is ignored or not from the solver point of view. Anyway there is already an option for calculating it with nominal flow. Go to Calculations -> Options, Uncheck Rated Flow for Inlet pipes. This will still however calculate this warning message with nominal flow even for ignored sources. An option to Ignore Inlet pipe warnings in Calculation | Options | Warnings tab is available in Aspen Flare System Analyzer V7.1. Keywords: Inlet pipes , pressure drop, rated flow ,relief valves References: None
Problem Statement: How to get heat exchanger, or cooler/heater duty in user variable?
Solution: The two sided heat exchanger, and heater/cooler store the duty differently. The attached model contains a cooler unit and a heat exchanger unit. Each has a user variable and its code to retrieve duty variable. Keywords: Duty, heat exchanger, cooler, heater References: None
Problem Statement: Why the temperature increases through connector or tee instead of temperature decrease?
Solution: In the compressible flow, flarenet use equation Eq 2.57-2.63 in HTFS DR38 for swage calculation. Equation looks like H1 + (M1^2)/ )/(2 *rho1^2) = H2 + (M2^2)/ )/(2 *rho2^2). H1 and H2 are inlet and outlet enthalpy M1and M2 are inlet and outlet mass flowrate There are two types of calculation that you can do in Flarenet regarding Energy Balances. You can either include kinetic energy in the balance or ignore changes in kinetic energy (Calculation > Options > General tab). If you ignore kinetic energy then the change in temperature is much smaller. However, when you include kinetic energy, as there is an overall reduction in the velocity (it is dependent on density changes or Cp/Cv)), this means that there is an increase in enthalpy in order to maintain the energy balance. Therefore H2 = H1 + (M1^2)/(2* rho1^2)- (M2^2)/ (2* rho2^2). If H2 is significantly higher than H1, the temperature could be higher. Sometime, temperature also increases during the static pressure recovery. Static pressure in downstream could be higher than upstream. Keywords: References: None
Problem Statement: What is meant by the error message: Molecular weight is not bound by composition?
Solution: If Mol. Wt. is selected as the source basis on the composition tab and there are multiple components added to the model, then Aspen FLARENET will distribute the composition according to the closest two molecular weight components. For example; if methane (MW=16), ethane (MW=30 and propane (MW=44) are the available components and 27 is entered as the Mol. Wt., then Aspen FLARENET will distribute the composition in between methane (0.214) and ethane (0.786). Note that the entered molecular weight must fall between highest and lowest molecular weight components available in the component list. If the entered molecular weight doesn't fall somewhere between highest and lowest molecular weight components, the error message Molecular weight is not bound by composition will appear. Keywords: molecular, weight, MW, bound, composition References: None
Problem Statement: How can I display variables with object names throughout the PFD?
Solution: By default, each object on the PFD has a label that displays its name. If desired, you can enhance all object labels to include the current value of a key variable. To select the label type for all pipe segments and nodes shown on the PFD, navigate through the PFD tool bar and click on the drop down list to select the variable of interest, as in the image below: Key words display, variables, PFD, label Keywords: None References: None
Problem Statement: How does the combine option works for components in Aspen FLARENET?
Solution: The combine option is used to replace multiple components by a single component. The process is used to reduce the calculation time. The step is recommended once the the user has multiple components with zero or close to zero composition. Please Note: Once you have done this, you cannot undo the grouping. 1. In order to access the component manager, go to Build | Component in the menu bar. Select the components and click on combine button. 2. You will then be asked about component base selection and basis of combine option. 3. In the Component Manager you can select a group of components (eg say nBBenzene - oM Styrene) and group them together as if they were a single component by highlighting a number of components and pressing Combine. You can then select the component you want from this group and how you want to represent the group. Thus, for instance with nBBenzene - oM Styrene you might use n-Octane if the majority was n-Octane. FLARENET will then effectively delete the rest of the components and use the properties of the chosen component but with the full flowrate of the individual components in the group. Keywords: Combine, Component, Combine component References: None
Problem Statement: Where can I find the complete name of a component in Aspen FLARENET? In the component manager I can only see the component abbreviation.
Solution: To see the name of a abbreviated component, follow the steps below. 1. From the main menu, select Database | Components. 2. In the component editor window, move the scroll bar towards right until you see Aspen Name. 3. Expand the Aspen Name cell to view the full name of the component, as shown in the figure below: For example, if you are not sure of what CC6= represents, you can follow the above steps and see it as cyclohexene. Keywords: component, full, name, abbreviation, aspen References: None
Problem Statement: In scenario of one source, I want all other valves not lifted but all other sources not completely ignored. How can I do that?
Solution: You need to include calculation of ignored sources with zero flow. To enable this option, select Calculations | Options, then activate the Calculate Ignored Sources with Zero Flow checkbox. Keywords: one source, scenario, back pressure References: None
Problem Statement: How do I change PSV and control valve's value for all scenarios?
Solution: There is not a direct way to change the value of sources for all scenarios in Aspen Flare System Analyzer. However, workaround will be to export your case to MS Excel file, change the values in Excel and import back to Aspen Flare System Analyzer again. Keywords: All scenario, relief valve, control valve References: None
Problem Statement: How does Flarenet calculate rated flow for two-phase flow?
Solution: For API 1976 method and API 1993 method, Flarenet calculation is based on the method described in API 2000 section 3.6 with some minor differences. This calculation is applicable for single phase gas/vapor. In API 2000, Flarenet can also calculate rated flow for the following cases: 1. Two phase system with saturated/sub cooled liquid & saturated vapor/ non condensable gases (section D.2.1) 2. Sub cooled liquid (section D.2.2) 3. Two phase system with non-condensable gases; condensable vapor & sub cooled/ saturated liquid (section D.2.3) There are cases where discrepancies have been reported by the user falls in regimes 1& 2 described above. These methods (1, 2 & 3) are not available in API 1976 and even in API 1993. In these situations Flarenet still calculates, using the method described in API 2000, section 3.6 with some minor modifications depending on the year. This is not the most accurate thing to do and we are planning to add a warning message saying that API 1976 and 1993 are not applicable in these situations (CQ00296279). Keywords: Two phase flow, rated flow, relief valve, valve, rated References: None
Problem Statement: Why is the allowable backpressure used instead of the actual downstream pressure to calculate the rated flow?
Solution: Aspen Flare System Analyzer will try to use all the available information that has been specified to a valve in order to calculate the rated flow. The specified allowable backpressure on the Conditions tab will be used to calculate the rated flow in cases where the simulator fails to detect a specification for a known backpressure. Since it is common practice to only specify the allowable backpressure, the program ignores the actual backpressure. If you have data for a known backpressure, you can enter it in the Methods tab under the sizing section. When a value has been specified in the sizing section, this will be the value taken into account to calculate the rated flow: Note: To apply the changes, go back to the Conditions tab and click on “Set� to make the valve recalculate the flow. Keywords: Backpressure, Allowable, Actual, Rated Flow, Relief Valve References: None
Problem Statement: What does Char. Volume mean when creating Hypothetical component?
Solution: Char. Volume is characteristic volume for the COSTALD model. It is different with critical volume. It can be found in most of pure component databank. It should be available in Aspen plus databank (VSTCTD) and Aspen Hysys for most of component. Keywords: Flarenet, pure component, Char volume. References: None
Problem Statement: In the Calculation Options Editor / Warnings, users can specify a minimum and a maximum temperature for carbon steel pipes and for stainless steel pipes. Is this temperature checked against the wall pipe temperature or against the fluid temperature?
Solution: The minimum temperature specified in the Calculation Options Editor (see below) is based on the fluid temperature. This is because the fluid temperature is smaller than the pipe wall temperature as shown below. Hence, checking against the fluid temperature will be a more conservative approach as low temperatures in flare networks can lead to embrittlement of equipment and pipework metal walls. Keywords: fluid temperature, wall temperature, minimum temperature References: None
Problem Statement: What is the difference between link and paste options in Pipe Editor | Fittings?
Solution: Both link and paste transfer the coefficients of the fittings for the selected equation into the Fittings Loss box. Link also maintains the list of fittings and you will not be able to change the fitting loss value by typing because the value is linked and shown as gray out. Whereas Paste will not retain the list of fittings and you will be able to change the fitting loss value by typing, the value will be pasted and shown in black. Keywords: paste, link, fittings, pipe editor References: None
Problem Statement: How to define the import and export definition file?
Solution: In the attached report, the detailed procedure is given. More information can also be found in Aspen Flarenet reference manual. Keywords: Import, export, definition file, Excel. References: None
Problem Statement: Is it possible to see Total Pressure in the plot?
Solution: Currently, Aspen Flare System Analyzer only displays the Static Pressure in the plot, but in the Pressure/Flow Summary table you can include total pressure and export the table to Excel. To have the Total Pressure in the Pressure/Flow Summary, check Display Total Pressure under: File | Preferences | General. Keywords: Plot, Total Pressure, Static Pressure References: None
Problem Statement: How do I customize the labels shown on the Flowsheet in Aspen Flare System Analyzer?
Solution: By default, a label with the name of each Object is displayed in the Flowsheet as shown below: You have the option to include additional labels to show the results of your AFSA model. This option is available under the Process Flowsheet tab. Go to the Modify tab and locate the Labels section. When you click on the drop down menu, a couple of variable combinations are available in order to show the results on your PFD. Note: For detailed instructions for AFSA versions prior to V9, please seeSolution 146649. After V9 you can customize the groups of labels displayed in your flowsheet. For that purpose, the Manage Labels button was included. Follow the instructions below in order to create a new group of labels with the desired variables: 1. Click on Manage Labels. The Labels Editor is displayed. 2. Click the + button to create a new group (by default the name of this group is Custom1, right click to rename it). 3. Once located on the Custom group, display the drop down menu to see the available variables. 4. Select the variable you want. 5. Click the + next to the variables drop down menu. 6. Repeat steps 4 and 5 to select all the desired variables (only 3 items are allowed per group). 6. Click on OK to apply the changes. The new group is created, and the variables will be shown in the Flowsheet. Keywords: Labels, Flowsheet, Customize, Manage Labels, Modify, Variables References: None
Problem Statement: Why is the pressure calculated for all the upstream operations from my orifice plate displayed as NaN?
Solution: A NaN value stands for Not-a-Number and is often found when noSolution was found for a certain variable. In the particular case of objects upstream of an orifice plate, the problem appears if the specified Orifice diameter is bigger that the Internal Diameter of the upstream pipe. Because it is impossible to set an orifice that is bigger than the actual pipe, Aspen Flare System Analyzer is unable to determine the pressure drop across this orifice and the NaN value is returned after failing to converge. This can be accidentally caused if the orifice plate is specified with the same diameter as the Nominal Diameter of the upstream pipe, and in reality the Internal Diameter of this pipe is smaller than the Nominal size. (i. e. Schedule 40 Carbon Steel 2.5 Nom. Diam. has an Internal Diam. of 2.469, which is smaller) To avoid this problem simply verify that the specified Orifice diameter is equal or smaller than the Internal diameter of the upstream pipe. Keywords: NaN, Orifice Plate, Pressure, Result References: None
Problem Statement: Why un-checking or checking rated flow of the tail pipe does not make a difference in back pressure?
Solution: Rated flow calculations in tail pipes require two steps: (1) Choose Tail Pipe = Yes in the connections page of the tailpipes and (2) check rated flow of the tail pipe under Calculations || Options. If you have un-checked (or checked) rated flow of the tail pipe and it did not make a difference in back pressure, then you missed the first step. Keywords: Tail pipe, rated flow References: None
Problem Statement: The total pressure in the system is below atmospheric pressure. Is this an operating concern?
Solution: The total pressure in the line can drop below atmospheric pressure if there are downhill pipes in the flare network and especially at lower flow rates. Total pressure only takes into account static pressure and kinetic head. It ignores the pressure change due to elevation. At commonly-encountered flow rates, the elevation pressure drop is negligible when compared to total pressure. However, at lower flow rates, the elevation pressure drop (which is negative for decreases in elevation) becomes more significant and dominates over total pressure. This is not an operating concern and will not cause reversed flow since at this point, the flow becomes gravity-driven and not pressure-driven. Keywords: Total static dynamic pressure negative below atmospheric References: None
Problem Statement: Can I export the PFD to some other program?
Solution: You can save the PFD as .emf or PDF file. The picture below shows where the option is. Exporting the PFD as PDFformat requires the icons to be wire frame icons. To change the icons to wire frame icons go to File || Preferences || PFD || check for Use Wire Frame Icons. You can copy the PFD to paste it in other applications. The same picture file shows this option as well. (I'll put a better picture later on) You can print the PFD. The various options related to it are described in theSolution document below. http://support.aspentech.com/webteamasp/KB.asp?ID=121054 Keywords: .emf, pdf, export, pfd, copy References: None
Problem Statement: How does one change the orientation of PFD icons in Aspen FLARENET?
Solution: 1. Choose the target object on the PFD using your left mouse button 2. Click on the Rotate Selected PFD Objects icon found on the PFD toolbar ( ) 3. Select the desired option (i.e. Rotate 90, Rotate 180, Rotate 270, Flip X or Flip Y) Keywords: transform, rotate, PFD, icon, flip, mirror References: None
Problem Statement: Why is the calculated pressure in the PSV downstream piping higher than the pressure at PSV outlet?
Solution: The static pressure in the inlet of the downstream pipe or tailpipe can be higher than the pressure of the PSV outlet if there is a flange in the PSV and diameter of the flange smaller than the diameter of the pipe. Static pressure recovery usually occurs in sections of the flare system where there is an increase in diameter. This is highlighted in the screenshot below. By default the Aspen Flare System Analyzer displays the static pressure in the PFD. This allows the users instances of static pressure recovery. You can display total pressure by selecting this option in the Preference Editor. The total pressure accounts for both the dynamic and static pressure components in the system. If you display the total pressure then the downstream pressure does not show higher than the PSV outlet pressure as highlighted below. Keywords: Static Pressure, Pressure Recovery References: None
Problem Statement: Why is there a difference in the heat transfer coefficients calculated in Aspen Flarenet compared to the Aspen HYSYS pipe-segment?
Solution: The differences comes from two sources: 1. Aspen HYSYS does not support the Radiation Heat Transfer. In the Aspen Flarenet model, you can disable the radiation, in the heat transfer page of the Pipe editor. This will reduce the difference between Aspen HYSYS & Aspen Flarenet calculation. 2. Although lesser than before, this still leaves considerable differences. The reason for that is Aspen HYSYS calculates the convective Heat Transfer Coefficient(HTC) at outer layer using the forced convection model over a flat plate (Simple model). Aspen Flarenet uses a convection model of flow over a circular tube with cross flow. Please refer to the Aspen Flarenet and Aspen HYSYS attached files and screen shots for sample cases. Keywords: Pipesegment , Heat Transfer References: None
Problem Statement: Is it possible to do heat transfer calculations under sea water?
Solution: Starting with Aspen Flare System Analyzer V7.0, the capability to do heat transfer calculations under sea water has been implemented. The calculated heat transfer coefficient is a function of the following parameters: a) External medium Velocity (user input) b) External medium Temperature (user input) c) Wall Temperature (calculated internally) d) External medium Density as a function of temperature (calculated from correlations) e) External medium Heat Capacity as a function of temperature (calculated from correlations) f) External medium Thermal Conductivity as a function of temperature (calculated from correlations) g) External medium Viscosity as a function of temperature (calculated from correlations) Keywords: HTC, Sea water, heat transfer, heat, pipe References: None
Problem Statement: When using Rated Flow for tailpipes calculations, by default the velocity results reported are the ones calculated with Mass Flow. Is it possible to have the velocity results based on the Rated flow?
Solution: When the option Rated Flow for tailpipes is checked, the rated flow is used to calculate the pressure drop. Other properties are calculating using mass flow. Nevertheless you can change this for velocity. To report velocity using rated flow rather mass flow, user needs to: 1. Go to File>Preferences and check the option Display velocity properties used by pressure drop calculations. Once this options is checked, the velocity results reported for Tailpipes will be the ones calculated with Rated flow. NOTE-Rated flow for Tailpipes has to be checked, otherwise Mass flow will be used for all calculations. For more information please refer toSolution 127933 Keywords: Rated flow, velocity, Mach number, Tailpipes, Flare System Analyzer, Preferences References: None
Problem Statement: How do I link to an application using VBA in Aspen Flare System Analyzer V7.3?
Solution: Aspen Flare System Analyzer V7.3 could not use CreateObject or GetObject to link the application. You will have to use new key word. Dim FlarenetApp As Object Set FlarenetApp = New AspenTech_FlareSystemAnalyzer_Interfaces.Application FlarenetApp.Visible = True FlarenetApp.OpenModel YOURFILEFOLDER_And_FILENAME Keywords: Automation, VBA, CreateObject, GetObject References: None
Problem Statement: I have enabled heat transfer calculations in the Calculation Options Editor, but Aspen FLARENET is still not performing heat transfer calculations. What should I do?
Solution: Heat transfer calculations need to be enabled in both the Calculation Options Editor and individual pipe segments (Heat Transfer tab) in order for Aspen FLARENET to do them. Note that local settings within a pipe will override the values set in the Calculation Options Editor. For information on enabling heat transfer calculations for multiple pipes simultaneously, please refer toSolution 109959. Keywords: heat transfer References: None
Problem Statement: Rated flow for relief valve in V7.0.
Solution: Rated flow: Rated flow is the manufacturer guaranteed flow and it is stamped on the relief valve nameplate. The Aspen Flare System Analyzer (Aspen FLARENET) calculation for rated flow is based on API 520. Aspen FLARENET 2006.5 and earlier Single Phase System: API-520 -1976 and API-520 -1993 are used for single phase flow. For API-520 -1976 method and API-520 -1993 method, Aspen FLARENET calculation is based on the method described in API-520 -2000 section 3.6 with some minor differences. This calculation is applicable for single phase gas/vapor. Multiphase System: Aspen FLARENET calculates rated flow for the following cases using API-520 -2000: 1. Two phase system with saturated/sub cooled liquid & saturated vapor/ non condensable gases (section D.2.1) 2. Sub cooled liquid (section D.2.2) 3. Two phase system with non-condensable gases; condensable vapor & sub cooled/ saturated liquid (section D.2.3) There are cases where discrepancies have been reported by the user falls in regimes 1& 2 described above. These methods (1, 2 & 3) are not available in API 1976 and even in API 1993. In these situations Aspen FLARENET still calculates, using the method described in API-520 -2000, section 3.6 with some minor modifications depending on the year. Aspen Flare System V7.0 and later Several improvements were done in the rated flow calculation of pressure relief valves. - For liquid relief rated flow calculation the procedure described in section 3.8 & 3.9 for API 520 -2000 / section 4.5 & 4.6 for API 520 -1976 was implemented. - For steam relief rated flow calculation the procedure described in section 3.7 for API 520 -2000 / section 4.4 for API 520 -1976 was implemented. In previous versions (Aspen FLARENET 2006.5 and earlier) these procedures were not implemented and the procedures described in section 3.6 for API 520 -2000 / section 4.3 were used for API 520 ? 1976, which are applicable for compressible gas/vapor (other than steam). The main issue in implementing these methods will be to get the necessary technical data from figures given in API 520 RP by regression analysis. - For two-phase relief rated flow calculation for sub-cooled liquid flashing inside the relief valve, Aspen Flare System Analyzer uses a regression formula (for critical pressure ratio calculation) from the data of figure D-3, API 520 RP (2000). The regression formula holds only in a limited region. To extend the range of applicability the equation that was used to generate the original plot was implemented. (Ref: http://committees.api.org/standards/tech/docs/520ati.xls) - Methods of API 520 (1976) & API 520 (1993) do not apply for two-phase flow. Aspen Flare System Analyzer now gives a warning when these methods are selected for conditions where two phase flows exist. Keywords: Rated flow, relief valve, PSV, Valve, Flow References: None
Problem Statement: Why do I get some inner loop failure warning messages in design mode?
Solution: In design mode, the diameters of pipes will start from initial values. For some of pipes, those initial diameters could be too small for your relief loads. In this case, Aspen Flare System Analyzer will give you these warning messages. However, those warning message most likely are intermediate messages. You may ignore them after you run your design results and verify it in rating Mode without getting the same messages. If you still get the same warning messages or you do not want to get these messages, you may follow the nextSolution I'm getting an Inner Loop Failure message when running my Aspen FLARENET case. What does this error mean? Keywords: Design Mode, inner loop failure, rating mode. References: None