Sponsored by the U. S. Nuclear Regulatory Commission Volume 13 / Number 1 / December 2013
Download 326.54 Kb.
|
TRACE User Problems …continued from page 1 TRACE. Inclusion of a pending update into TRACE typically requires additional review and testing by the NRC staff. The most current developmental version of TRACE is Version 5.831 as of December 31, 2013. Updates included in Version 5.831 can be found on the TRACE user support web site (https://www.nrccodes.com). New Trouble Reports Still Open The following new trouble reports are still open and have not been resolved or closed. Trouble Reports 653 – Power weighting of reactivity feedback parameters. A developer while reviewing the FxBoronReact update noted some inconsistencies in the power weighting for the core reactivity feedback parameters. In general, the intent of the power weighting is to increase the importance for a given reactivity feedback parameters in the region where the neutron flux or local power is relatively large. The specific problem identified was the power weighting factors were not normalized correctly when the core bypass and water rod regions of the core were included. In addition, the power weighting factor for the water rods and core bypass for BWRs was set to 1.0 and the power weighting factor should be approximated with the lower fuel power factor. This trouble report is still under investigation. Trouble Reports 658 – PARCS bug for ADF rotation when there is more than 1 mesh in 1 fuel assemble. A user reported that when PARCS V3.1m07 was used to simulate a PWR core in steady-state the code crashed. The neutronic mesh for each fuel assembly is 2x2 and the ADF option was set to true in the tree_xs input. The ADF rotation option in ROT_ADF was also set to true. User reported that outside reflector lattices are rotated according to their positions and macro cross-sections were input a PMAX files. This trouble report is still under investigation. Trouble Reports Closed During the time period from September 26, 2012 to December 18, 2012 the following trouble reports have been closed with updates included into the NRC developmental version of TRACE. In some instances, these trouble reports have been closed by rerunning the test problems of interest with the most current version of TRACE and/or by modifying TRACE documentation to address the issue. Trouble Report 287 – Inconsistencies in reactivity feedback models in TRACE. Update R5fdbkFix went into version 5.793 of TRACE. Update R5fdbkFix added RELAP5 reactivity feedback model capability into the BWR reactivity feedback models in TRACE. This closes this trouble report. Trouble Report 299 – Error in mass flow rates used in HTC calculations when 1D PIPE is deadend. The MIT pressurizer test problems were rerun with version 5.765 and the 1D and 3D models of these tests gave consistent results. The Clean1Ddrag update went into version 5.750 and closed this trouble report. The Clean1Ddrag update moved the 1D interfacial and wall drag from a component based method to a global data structure method. With this change the stratified flow weighting factor (i.e., wfhf) was calculated consistently for 1D and 3D deadend fluid cells (i.e., inlet but no outlet). Trouble Report 325 – Typical velocity error for restart with components replaced with ASCII input. Update FxRstVels went into version 5.721. With this update if there is a relatively small inconsistency for a given cell edge between the fluid velocity in the ASCII as compared to the fluid velocity in the restart file a user warning is written and restart file velocity is used for that cell edge in the ASCII input file. If the velocity difference is significant, then input error is thrown the calculation is stopped. This trouble report is closed. Trouble Report 354 – Excessive warning messages. Update newErr went into version 5.780. This update changed the TRACE input error processing logic so that the first input error encountered results in the calculation stopping after writing an input error message for the users. In addition, an effort was made to improve the error messages and attempt to provide the user with more information concerning why the calculation could not continue and what needed to be done to correct the error. This approach reduces the number of warning and error messages written before TRACE would normally stop at the end of input processing. The user had a difficult time determining which messages were fatal input errors and which messages were warnings or information messages. With the new methodology in the newErr update input error messages are easily identified, since the code stops for the first input error encountered and all other user messages are clearly labeled as warnings, information, etc. This update closes this trouble report. Trouble Report 355 – Recovery from a Time-Warp. Update AUTO went into version 5.770 and closed this trouble report. Update AUTO adds to TRACE the capability to backup more than one time step if TRACE fails with time step cannot reduce below user input for DTMIN. When TRACE repeats the last time step over and over with smaller and smaller time step sizes, this indicates that TRACE cannot resolve this numerical problem with a smaller time step. For earlier versions of TRACE this required a rerun or restarting with a smaller DTMAX in attempt to solve through this numerical problem. With AUTO TRACE has the capability to backup more than one time step and rerun through the problem time period with smaller time step size. With this update the robustness of the code is significantly improved. Note this AUTO backup logic does not currently work with PARCS coupled or CONTAN coupled TRACE calculations. Trouble Report 398 – Accumulation of round-off error in the shifting of tables relative to a trip. Update unShiftTables when into version 5.763 and closed this trouble report. The original logic in TRACE shifted tables each time step relative to when a controlling trip status had last changed. This continuous shifting and adjustment of tables relative to trip status changes could result in significant accumulation of round-off error in a given table depending upon how many time steps the table was shifted. Update unShiftTable replaced this logic so that the time when a controlling trip status changed could be saved with the table and used to determine the correct independent variable values to be used to evaluate the table. This eliminated the accumulation of round-off error, since the tables are no longer shifted. Trouble Report 400 – Getting a -0 for Diff->statName. When two TRACE output files are the same the TRACE perl testing scripts compare them using the diff program and result should be a diff file with 0 length. However, sometimes on the RedHat LINUX operating system the file length would be -0. This would indicate that results for this test problem changed or were different, when in fact they were the same. Using the most current perl scripts and with Lahey compiler on the LINUX 64 bit OS with version 5.820 this behavior was no longer observed and this trouble report was closed. Trouble Report 403 – Signal variables for CONTAN component. Update CONTANCTRL went into version 5.670 and update CONTANSV went into version 5.651 and closed this trouble report. These two updates added signal variables and control block capability to the TRACE CONTAN component. Trouble Report 468 – APEX1000 small break modeling guidelines. Application of NRC small break modeling guidelines to APEX1000 indicate that when there is an orifice in the cold leg pipe wall critical flow model should be turned on there and if controlling choking is not expected downstream from this location, then this can be the only critical flow location simulated. For example, SJC connected from cold leg to an atmospheric break. However, long runs of relative small piping from the cold leg to the break flow tank with valves between the cold leg and the break flow tank may result in the controlling critical flow moving from the cold leg break orifice to the VALVE component. Therefore, the complete critical flow discharge line including VALVE components may need to be simulated for these types of tests. In a plant small break accident with a hole or split in the side of a cold leg or hot leg, there is not discharge piping to be simulated. An SJC connected to a BREAK component is the appropriate model. This trouble report is closed. Trouble Report 536 – Test problem manopipe02.i results in erroneous amplitude decay. Update FxLevelBackup went into version 5.610 and closed this trouble report. This update changed the logic when the level tracking model would force a backup. With this update and with the initial conditions in equilibrium and dtmin = 0.001 the manometer oscillations amplitude does not decay. As the time step is increased, the manometer oscillations amplitude will decay. Initial conditions in equilibrium are important to minimize the impact of phase change as the level moves up and down in the vertical flow paths. If phase change is turned off, then initial conditions in equilibrium are not important, however current versions of TRACE do not include a user input to turn off phase change. Trouble Report 584 – Enable side junctions for PRIZERS. Update FxPRIZERSJ went into version 5.680 and closed this trouble report. Update FxPRIZERSJ added pipeType option 9 to the PIPE component. PIPEs with pipeType option 9 have all of the PRIZER component modeling capability included into the PIPE component. This allows side junctions to be included into a PIPE component with pipeType equal to 9. Trouble Report 585 – Enhancements to built-in heater model in PRIZER component. Update FxPRIZERSJ went into version 5.680 and closed this trouble report. Update FxPRIZERSJ added pipeType equal to 9 to the PIPE component, which implies PIPE components can be used to simulate pressurizers. PIPE component power to fluid table and trips input can be used to trip off pressurizer heaters when pipeType is 9. Trouble Report 591 – Potential error with direct moderator heating. Update FxDirModHEat went into version 5.783 and closed this trouble report. This update adds the capability to TRACE for the user to specify as functions of time the fraction of the fission heat and/or decay heat that is direct moderator heating. Direct moderator heating associated with the fission power is due mainly to neutron slowing down or interacting with the moderator water. Direct moderator heating associated with decay heat power is due mainly to gamma ray interaction with the moderator water. Fraction of these two sources of power that is direct moderator heating would not normally be the same value. Note that fission and decay heat power levels are only calculated by TRACE when the TRACE point kinetics model is turned on. If the point kinetics model is turned off and the power is specified by the user, then the fission power direct moderator heating input fraction is used for the total power specified and the decay power direct moderator heating fraction input is ignored. Trouble Report 596 – Behavior of ISNOTB in point kinetics model turns off all boron reactivity feedback. Update FxDirModHeat went into version 5.783 and closed trouble report. This update removed all of the ISNOTB logic from TRACE. The ISNOTB input is still read in so that old input decks still run, but the ISNOTB input is ignored. The ISNOTB input was associated with using burnable poisons in then boron reactivity feedback. If at a future date a user need is identified for reactivity feedback due to changes in burnable poison concentrations this model can be revisited. Trouble Report 597 – Suggestion to add core avg parameters in reactivity feedback model to XTV graphics. The FxDirModHeat update went into version 5.783 added the core average parmeters used in the reactivity feedback model to the XTV graphics variables. There would some SI to English units conversion problems for when the reactivity feedback coefficient for Doppler was based on the square root of the core average fuel temperature. These last concerns associated with this trouble report were addressed with the FxNameListUnits update that went into version 5.821. This trouble report is closed. Trouble Report 600 – Error in energy source term for water coming from TRACE BREAK component and flashing in CONTAN compartment. The energy source term error was fixed in version 5.541. This fix was also included in an update FxContanFlash along with two other fixes for the CONTAN component. The modified version of the FxContanFlash update went into version 5.701 and closes this trouble report. The two other fixes were associated with mass and energy source term for FILL components connected to a CONTAN compartment and logic associated with flashing in the liquid pool in an CONTAN compartment. Trouble Report 601 – SJC pump has small flow when it should be zero. Update FxErrMess went into version 5.690 and closed this trouble report. This update allows velocities in SJC pump cell edges to go to zero and stay at zero if boundary condition velocities are zero. All other active cell edges if at zero velocity are reset to small velocities (i.e., 1.2e-10 m/s) in the direction of the pressure gradient so that donor cell logic can anticipate when the flow starts to move again. For SJC PUMPs the flow are boundary conditions are not solved for from the TRACE momentum equations and therefore should not be reset. Trouble Report 603 – AMNCSS and AMXCSS for CSS input applies to what parameters? Modified documentation was included into the FxErrMess update and documentation in the NRC TRACE repository was changed and closed this trouble report. The constrained steady-state (CSS) input documentation was modified so that it is clear which parameters have a minimum and maximum specified by the user input AMNCSS and AMXCSS. In addition, the units for these parameters was clearly identified. Trouble Report 605 and 606 – Hoop stress and axial stress in gap models The FxErrMess update went into version 5.690 and closed this trouble report. This update includes logic changes so that heat structure gap model hoop and axial stresses can be plotted. Trouble report 606 is a duplicate of this trouble report 605. Trouble Report 607 – Negative gap width reported in output. Update GapEdits went into verison 5.690 and closed this trouble report. With the GapEdits negative gap width are not allowed. A negative gap width could be calculated when the fuel pin outer radius expands beyond the inner radius of the cladding. A negative gap width maybe used in the future to estimate the contact pressure between the fuel pin and cladding. With version 5.690 and larger gap widths are not allowed to be smaller than the roughness of the fuel pin and cladding surfaces. Trouble Report 608 – Add more gap heat transfer parameters to the output/graphics files. Update GapEdits went into verison 5.690 and closed this trouble report. The GapEdits update adds a number of gap heat transfer parameters to the output and graphics files. Trouble Report 609 – Implement a fuel swelling model. Update GapEdits went into version 5.690 and implemented FRAPCON 3.4 fuel swelling model into the TRACE gap conductance model. FRAPCON has been compared favorable to experimental data for the steady-state fuel rod center line temperature and is used as a benchmark for the TRACE fuel rod models. Update FxFCModels went into verison 5.761 and closed this trouble report. This update implemented the following fuel rod modeling recommendations based on comparison to FRAPCON models: Consistent with FRAPCON gap conductance model, surface roughness factors are not included in open gap HTC calculations. Consistent with FRAPCON gap conductance model, closed gap modes are applied when gap thickness < (roughness of the fuel and cladding). When the gap is closed then the roughness of the fuel and cladding is included in the closed gap HTC calculation. The factor of 1.8 in the temperature jump correlation in the FRAPCON documentation was not supported by other references and was removed from the TRACE temperature jump model. Consistent with the FRAPCON conductance model coding that b factor is only included in the closed gap conductance model. The minimum allowed for the characteristic dimension across the gap was set to 0.025 mils. Trouble Report 610 – Possible bug in contact resistance model. Update GapEdits went into version 5.690 and closed this trouble report. With the GapEdits update, user can input separate roughness for the fuel outer surface and the cladding inner surface. With separate roughness, then the contact resistance model can be evaluated correctly. In addition, the fuel gap model documentation was updated to reflect that the contact pressure is estimated from the cladding yield stress. The implied assumptions is as the gap closes the cladding stress is sufficient to deform the cladding to the expanded radius of the fuel and that cladding stress is at or close to the cladding yield stress. This appears to be a reasonable assumption and relatively easy to implement. If this assumption is found to be significantly inaccurate, then a detailed multi-dimensional mechanical deformation model may be required. Trouble Report 611 – Investigate use of 1.27e-07 m as default for RFCLAD. With update FxFCModels into version 5.761 and FxNFCI into version 5.762 this trouble report is closed. In general the gap conductance goes as Kg / drg + contact conductance + radiation heat transfer. Kg is the conductivity of the gas in the gap and drg is the characteristic gap width dimension. drg will typically include any temperature jump corrections due to the dimension of the gap approaching molecular mean free path distance and surface roughness if the gap is considered closed. In addition, the FRAPCON closed gap model includes a b factor that is correlation factor based on fuel rod experimental data. The minimum that drg is allowed to go will have a direct impact on the closed gap conductance. The initial minimum allowed for drg was based on the FRAPTRAN coding (i.e., 1.27e-07 m). As the TRACE models moved toward to FRAPCON (i.e., including the b factor in the closed gap conductance model in TRACE), then the minimum as implemented in the updates listed above is consistent with FRAPCON (i.e., 0.025 mils = 6.35e-07 m). If TRACE moves toward the FRAPTRAN models, then this minimum allowed for the drg and for the fuel and cladding roughness should be revisited. Based on implementation of the FRAPCON models into TRACE, then this minimum should be consistent with the FRAPCON models. Trouble Report 612 – Possible gap model improvements. Update FxNFCI went into version 5.762 and closed this trouble report. This update removes the old TRAC-P dynamic gap model and the RELAP5 fuel slumping model. The TRACE dynamic gap model currently in TRACE is considered a best estimate model and is based on FRAPCON 3.4. The available input options for NFCI for the TRACE heat structure component has been reduced from a large number of possibilities to three values (i.e., = 0 implies constant gap conductance, =1 implies best estimate models for cladding thermal and mechanical deformation but not including plastic deformation and/or rupture, = 2 same as 1, except includes plastic deformation and cladding rupture). Trouble Report 613 – Implement actual FRAPCON fuel relocation model. Update FxFCModels went into version 5.761 and closed this trouble report. This update implemented the FRAPCON 3.4 fuel relocation model. Trouble Report 614 – Determination of interface pressure in contact resistance model. With update GapEdits going into version 5.690 this trouble report is closed. The interface pressure for a closed gap is calculated based on assuming that the cladding stress is at the cladding yield stress. If this assumption proves to be inaccurate, then a more mechanistic model can be implemented at a future date. Trouble Report 619 – Large component number fails for CHAN component grouped view factors and path length files when TRACE attempts to open these files. The logic in generation of file names for grouped view factor and path length files used to store data as CHAN component input is processed assumed CHAN component numbers no larger than 9999. With the GapEdits update this logic was modified to handle CHAN component numbers as larger as 999999. These grouped view factor and path length files are written out for user convenience since significant cpu time may be spent calculating these grouped view factors and path lengths. Once calculated if the geometry does not change then these input files can be pasted into CHAN component inputs and TRACE calculation for these view factors and path lengths can be skipped. For relatively simple geometry analytical formulas are used to determine the view factors and typically do not require significant cpu time. However as the geometry becomes more complicated the view factors may be calculated using Monte Carlo methods, which may require significant cpu time depending upon the number of CHAN components, the number fuel rods and water rods in each CHAN component, whether the geometry varies axially or is constant and how many view traces are used per surface segment within the fuel assembly. Under these conditions the user may only want to calculate the grouped view factors and path lengths once. Trouble Report 620 – Slight difference in PCT for restart for CHAN nfciTest-13.Rev.FE. Update GapEdits went into version 5.690 and closed this trouble report. Determined that heat structures radiation heat fluxes were not included in the TRACE TPR library dump/restart files. Update GapEdits include heat structure radiation heat fluxes in the TRACE dump/restart files and resolves this difference in PCT at restart. Trouble Report 627 – RELAP5 reactivity feedback models in TRACE not working for transient restart. For earlier versions of TRACE the RELAP5 reactivity feedback models were not available for TRACE models of BWR. Update R5fdbkFix went into version 5.793 and closed this trouble report. Trouble report 287 is similar to this trouble report and with update R5fdbkFix in version 5.793 both of these trouble reports are closed. Trouble Report 628 – Choking problem with V5.0 Patch 3. TRACE calculations were compared to a ROSA-2 intermediate side break test. TRACE results were not consistent with expectations until choking was turned at both the break orifice in the cold leg and on at the VALVE components in the discharge line. There are periods during the transient where choking may occur at both the break orifice and at the VALVE components. If the discharge is long enough and has sufficient flow loss and relatively small areas, then as the two phase mixture expands in the discharge line the minimum break flow may occur at the two-phase choking at VALVE rather than at the break orifice, even though the flow area at the break orifice is smaller than the flow area through the valve. With the choking model turned at the correct multiple locations the results look reasonable and consistent and this trouble report is closed. Trouble Report 629 – Warning message about 1D level tracking with SEPD components no longer needed. Update R5fdbkFix went into version 5.793 and this trouble report is closed. Logic was added to update R5fdbkFix to remove this user warning message that is no longer needed or useful. Trouble Report 630 – DirectCoolantHeating in the Regression test problem suite for PARCS does not with TRACE since version 5.540. Update ExMappingFix went into version 5.673 and this trouble report is closed. Logic errors in heat structure input processing was causing these test problems to fail and testing scripts were not picking up that these test problems were all failing. The ExMappingFix resolves this input processing error Trouble Report 631 – Heat transfer rates from HS to fluid cells not backed up correctly. Update PowerBackupFix went into version 5.764 and closed this trouble report. This update fixes the HS backup logic so that heat transfer rates are backed up correctly. Trouble Report 632 – Input error messages confusing. Update FxErrMess went into version 5.890 and update newErr went into version 5.780 and closed this trouble report. With the newErr update it is clear whether or not a user message is an input error, warning, information, etc. In addition, any fatal input error results in TRACE stopping immediately. Multiple fatal input error messages are no longer possible. These updates significantly reduce the confusion level for input error messages. Trouble Report 633 – TRACE V5.0 Patch 3 does not read data curve. Update ContBrkFix went into version 5.820 and closed this trouble report. This update fixes the logic error that was writing over the BREAK component solute table. Trouble Report 637 – Power component’s maximum rate of change programmed reactivity (i.e., RRPWMX) affects also the maximum programmed reactivity. Update FxCladRupEdit went into version 5.672 and closed this trouble report. This update fixes a logic error in the application of the maximum allowed change in reactivity in the POWER component. With this fix the user input for RRPWMX is used correctly with the programmed reactivity. Trouble Report 638 – Initialization errors flagged by Regression Test Problem Set on LINUX 64 bit OS with Lahey compiler. Update LaheyFix went into version 5.772 and closed this trouble report. This update fixes the run time errors and initialization errors identified by the Lahey compiler on the LINUX operating system. Trouble Report 639 – Cladding rupture fine mesh index may be off by one. Update FxCladRupEdit went into version 5.672 and closed this trouble report. This update improves logic for determination of the fine mesh node index at the cladding rupture location. Trouble Report 642 – Version 5.80 results in compiler errors for Lahey compiler on LINUX OS. Update FxFCModels went into version 5.761 and closed this trouble report. This update includes fixes for the compiler errors identified by Lahey compiler on LINUX OS. Trouble Report 643 – lenNamelist in NamlistDat module should be 123 for version 5.690. The fix for this trouble report went into version 5.701 and closed this trouble report. The variable lenNamelist needs to be consistent with the number of name list variables initialized in the NamlistDat module. lenNamelist was 124 and it should have been 123. When lenNamelist is too large then there will undefined entries in the namelist data structure. Trouble Report 645 – Potential error in indiex used in NCG mixture property evaluation fluids model. Update FxMixNCGTP went into version 5.773 and closed this trouble report. This update fixes some index errors in the non-condensable gas (NCG) mixture thermal and fluid properties model. The NCG mixture properties model is based on the FRAPCON 3.4 model and is used in the gap conductance model and in the fluids component models when mixtures of NCG are present. Trouble Report 646 – Restart problem for models with iGas > 10 and VESSEL component. The FxMixNCGTP update went into version 5.773 and closed this trouble report. This update fixes restart/dump problems for the VESSEL component when there are more than one NCG present. Trouble Report 647 – Error in the 3” axial renode logic when fine mesh, cladding rupture, and metal-water reaction model are turned on. FxMWRXReNode update went into version 5.784 and FxMWRXAppK update went into version 5.800 and these updates close this trouble report. With these updates the user has the option to run with the inside surface of 3 inches zone at the cladding rupture location for the inside surface metal-water reaction to be forced or relaxed. If the 3 inch zone is forced and the axial noding is too small for 3 inch renode around the cladding rupture location, then the calculation stops. If the 3 inch zone is not forced, then the calculation will continue if the axial coarse mesh noding is too small for a 3 inch renode around the cladding rupture location. Trouble Report 648 – Critical flow at SJC edges may not be correct. Based on the resolution of trouble report 628 and investigation of the results for this trouble report, this trouble report was also resolved and closed. Results are consistent with expectations for the models used. For a split in the side of a reactor pressure vessel (RPV) an SJC PIPE connected to an atmospheric break is the best representation of this geometry. If the break is a small pipe connected to the RPV and there is non-zero length of pipe between the break and the RPV, then a PIPE component connected to the RPV and then to the atmospheric break may be the best approach. The intent is to provide the best estimate for the fluid conditions upstream from the choking plane. If the pipe stub connected to the RPV is short, then going back to an SJC PIPE connected to the BREAK is a reasonable approach. With the SJC PIPE the upstream choked flow plane fluid conditions are in the RPV fluid cell and will typically result in larger critical flow rates as compared to a short pipe where flashing and phase change may occur upstream from the choking plane. For very short pipe stub, no significant flashing will be expected. If larger critical flow is conservative for the transient of interest then lean toward to the SJC PIPE and BREAK combination. Trouble Report 649 – Not clear which form of the reactivity feedback coefficients and units used when R5fdbk = 1. Update FxBoronReact went into version 5.791 and FxMWRXAppK into version 5.800 and closed this trouble report. Documentation was update and logic errors corrected so that it should be obvious to the user which reactivity feedback coefficients and the units are used with the TRACE point kinetics reactivity feedback models. Trouble Report 650 – Table driven POWER component does not work during steady-state. The default behavior for a TRACE table or tripped table is during steady-state a constant initial value is used. The typical TRACE table is not evaluated until the transient starts. In this case the user wanted to use control systems during steady-state to determine pressurizer power and pump power as the steady-state is approached. Using PIPE with pipeType 9 could use the existing pressurizer models to control the pressurizer heater during the approach to steady-state and then at the start of the transient trip off the TRACE internal pressurizer models and trip on powered HS components that will be used to simulate the transient pressurizer behavior. PUMP models will turn on the pump friction heating models. If a user need for steady-state controls for POWER component is identified it can added at a future date. This trouble report is closed. Trouble Report 651 – Version 5.780 has a USE ErrorInterface, ONLY: error missing in TestForBackup. Update FxMWRXReNode went into version 5.784 and closed this trouble report. The newErr update implemented optional arguments into the error routine. This implies that an error routine interface must be defined in any modules where error is called. In a few cases the newErr update missed where USE ErrorInterface, ONLY: error needed to be added to TRACE source files. Trouble Report 652 – Input deck failures for version 5.780. With the newErr update some test problems that did run were now failing. The failure was due to the change in some input error testing logic that changed e character missing from the ASCII input from a warning to a fatal input error. The e character in the TRACE array input implies that the array input is finished. A few input models have the e character in column 81 and therefore does not get included into the fixed format input in the .echo file. When the array input is read from the .echo file it finds the e character missing and throws a fatal input error. This is the expected behavior for TRACE input processing and this trouble report is closed. Trouble Report 654 – HS fatal input errors called from CountingCoup. Update FxMWRXAppK went into version 5.800 and closed this trouble report. The CountingCoup module in TRACE scans the ASCII and restart file input to determine the total number of fluid, heat structure, etc. components in the TRACE input model and the total for each type of component. This information is needed to allocate arrays and data structures. The HS component is scanned using the HS input routines, but during the scanning the component number, etc is not saved. So if an input error is detected then with the newErr update that went into 5.780 will cause the calculation to stop, but the error message itself will not be complete since it was called from the input scan initiated by CountingCoup. Logic was added to the HS input processing routines to skip input error checks during an input scan initiated by CountingCoup. Trouble Report 655 – Appears to be backup problem with version 5.784. Update FxBackupMB went into version 5.811 and closed this trouble report. Certain time integrated mass and energy flow rate parameters were not included into the TVar data structures used to save solution variables at selected potential backup times as the calculation proceeds. This time integrated parameters are part the TRACE edited mass and energy balances. Without these parameters in the TVar data structures a backup would appear to result in a mass error. With these parameters added to the TVar data structures, then a backup would not add to the mass error. Trouble Report 657 – Version 5.792 compiled on LINUX 64 bit with Lahey compiler uncovered coding errors. Update FxErrCalledBy went into version 5.822 and closed this trouble report. This update fixes compiler errors uncovered with the Lahey compiler on a LINUX 65 bit OS. It also adds calledBy argument to the error routine so that calls to error can pass the name of the subroutine that is calling error. This information will be useful to developers when users provide them with the error message of interest. Trouble Report 659 – Negative pressure calculated in some CONTAN test problems. Update NegPresFix went into version 5.812 and closed this trouble report. This update modified logic in PredictVoid routine so that the estimated new time NCG partial pressure did not go negative even though the total pressure was approaching the triple point pressure. Trouble Report 660 – Divide by zero error for CONTAN test problem CVTR_T3.rev7sv. Update NegPresFix went into version 5.812 and update ContBrkFix went into version 5.820 and closed this trouble report. The divide by zero occurred during initialization of the CONTAN component and was associated with certain condensation film parameters not correctly initialized. The updates listed above modifies logic so that the CONTAN component parameters are correct initialized during initialization. Trouble Report 661 – Divide by zero when non-ideal helium is used. Non-ideal helium does not appear to be used anymore in TRACE. This model was removed from version 5.801 of TRACE. This closes this trouble report. Trouble Report 662 – Floating point overflow when TRACE attempts to write to XTV file. The steady-state pressurizer models in TRACE use a scaling factor of 1.0e+35 times the pressurizer volume and flow area to ensure that the pressurizer pressure stays constant during a steady-state calculation. This implies that the interfacial heat transfer coefficients written to the graphics file (i.e., XTV file) will include this scaling factor. Given that the graphics file data structures are mostly single precision this results in a floating point overflow. The scaling factor was changed to 1.0e+15 in version 5.801. This closes this trouble report. Trouble Report 663 – Divide by zero error for higher order numerics test problem. The NegPresFix update went into version 5.812 and closed this trouble report. This update changed in the logic in StbVel1D routine so that for a mixed numerics calculation with a PUMP component, SETS is turned off and diagonal terms in the vvt and vlt coefficient matrix are set to 1.0 and the off-diagonals are set to zero. This eliminates the divide by zero in the TriSolve routine for the vvt and vlt velocities. In addition, logic was added to catch zero diagonal terms and throw a run time error. Trouble Report 664 – Divide by zero error in heat structure logic. The fprop routine was not getting called for PLENUM components during initialization. The fprop routines calculates fluid viscosities and other fluid properties. The fluid viscosities are transferred to any HS component connected to these PLENUM components. However, since viscosities were calculated during initialization the first time heat transfer coefficient calculations need a Reynolds number a divide by zero error occurs. The fix is to all a call to fprop for the PLENUM component during initialization. This fix was added to version 5.801 and closed this trouble report. Trouble Report 665 – “floating invalid” error reported in the HtStrShared module. Run time error occurred in the heat structure fuel pin dish volume calculation when the fuel pin was not dished. Logic was added to skip the fuel pin dish volume calculation if the fuel pin is flat (i.e., no dish). This logic change was made in version 5.801 and closed this trouble report. Trouble Report 667 – Input description for PGAPT and GMLES are confusing and needs to be reworked. Update ChanGMLES went into version 5.824 and closed this trouble report. New input was added to the CHAN component so that when name list inpu legacyFRM is FALSE then GMLES input for the CHAN is expected. In general, if the fuel rods simulated by the HTSTR or CHAN input is fresh fuel, then PGAPT can be used to calculate the number gram-moles of fill gas in the fuel rod gap and plenum. The user input for PGAPT is for cold conditions and the cold condition gas volume will be used to determine the number of gram-moles of fill gas. If the fuel rods have burnup, then a code like FRAPCON or something similar would be run to determine the number of gram-moles of gas in the gap and fuel rod plenum and the composition of this gas, since it will contain not only fill gas but also fission product gases. For this case GMLES input should be used for the HTSTR or CHAN input. If the user inputs non-zero input for GMLES, then PGAPT input must be zero. If the PGAPT input is non-zero, then GMLES must be input as zero. Input descriptions for PGAPT and GMLES were modified to be consist with the ChanGMLES update and describe when which input should be used. Trouble Report 668 – Found a case where the choked flow critical liquid phase velocity was zero. Update FxNegChoke went into version 5.823 and closed this trouble report. The lower limit for the pressure at the choking plane was 1.0e-06 Pa and under certain conditions as the correct critical flow search is going it was possible pressure at the choking plane to go so low that the fluid properties were negative. This resulted in negative critical flow velocities that were limited to zero. This fix for this problem was change the minimum pressure at the choking plane to be the triple point pressure. Properties at the triple point pressure are reasonable and consistent and eliminates the possibility for zero velocities at the choking plane. Trouble Reports Resolved with Updates Pending During the time period from January 1, 2013 to December 31, 2013 the following Trouble Reports have been resolved with updates or documentation modifications pending. These Trouble Reports are awaiting NRC review and a decision to implement or not to implement the associated code updates or document modifications into TRACE. Trouble Report 402 – Incorrect vapor internal energy fit between 1 bar and 20 bars. During external independent review of TRACE, Marv Thurgood identified the saturated vapor internal energy fits from the TRAC-P computer code as inadequate in the range between 1 and 20 bars. The derivative of internal energy with respect to pressure and temperature are of the wrong sign. These errors for the saturated vapor phase also impact the superheated steam internal energy calculation. The TRAC-P EOS fits are not recommended for use and are currently only maintained for old input decks that use these EOS fits. Starting with version 5.600 the default EOS model for TRACE is the IAPWS steam tables. This trouble report is marked as resolved. There will be no attempt to fix these fits since they are no longer recommended and are maintained only for backward compatibility. Trouble Report 604 – Inaccurate interfacial heat/mass transfer as 3D vessel level crosses nodal boundary using 3D level tracking. A user reported that TRACE 3D model of the MASLWR test facility predicted a pressure spike when the mixture level in the pressurizer region of the reactor pressure vessel dropped from level 27 to 26. These results were observed with verison 5.540. These calculations were repeated with version 5.680 and the pressure spike went away. There were a number of level tracking fixes between versions 5.540 and 5.680 that may have been the reason for these improved results. This trouble report is marked as resolved. Trouble Report 618 – Removal of unnecessary units definitions in the VarLabels array. The FxCleanupUnits update is pending for NRC review. This update removes 217 unused units definitions from the VarLabels array. The VarLabels array defines the units for a given TRACE variable or parameter that would be in TRACE input or output or graphics. This units definition is used for converting the parameter between SI and English units. As models changes and input and output changes and graphics variables change units definitions that were needed in older versions of TRACE are no longer needed. For version 5.831 the length of the VarLabels array is 1465 and removal of 217 elements from this array would reduce it’s size by ~15%. Trouble Report 636 – Motor torque curves no accurately used by TRACE. The motor torque curves in TRACE were not correctly converted from English to SI units. This problem was fixed with the FxCladRupEdit update that went into version 5.672. The final issue for this trouble report was resolved with the FxIPMPTY3 update which contains documentation updates to address the name list input PUMPFRICQ input associated with pump friction heat loss model. The update is pending. Trouble Report 644 – Logical function calls in an IF (…) or DO WHILE (…) relative to Fortran 90/95 standard. There were a number of places in the TRACE source that included within IF (…) tests and DO WHILE (…) coding where a logical function used in the IF or DO WHILE that may have secondary effects. For example, IF (FUN1(x, y) .AND. FUN2(w, z)) THEN there is no requirement that both FUN1 and FUN2 will be called or which one is called first. Therefore if FUN1 and/or FUN2 have secondary results, then there is no guarantee that the function will be evaluated. The PURE update was developed to address this trouble report and resolves this trouble report. The PURE update forces all logical functions in TRACE to be PURE Fortran 95 functions (i.e., no side effects). Functions that could not be made to be pure were modified to be subroutine and the logical result returned through the argument list. This update is pending. Trouble Report 656 – Description of NFAX is insufficient in Volume 1 of the TRACE Users Guide. Documentation modifications in the FxNameListUnits update identify and explain the input requirements for the NFAX array for both the CHAN and HTSTR components. The NFAX input requirements are enforced by TRACE but were not documented for the user. This update is pending. Trouble Report 672 – Differences between 5.0 Patch 3 and 5.791. TRACE version 5.791 was not restarting from the restart that user expected. Given this difference in restarts between 5.0 Patch 3 and 5.791 the calculated results were significantly different. For the 5.0 Patch 3 restart the user use ASCII input so that the calculation started from the last restart dump for the steady-state calculation. For the 5.791 restart the user included name list input closest = .TRUE., which turns on TRACE logic to restart to the dump closest to the user requested dump given by user input DSTEP. The steady-state wrote two restarts dump (i.e., one at time step 0 and one at time step ~25520). For the 5.791 restart with closest = .TRUE. and DSTEP = 2552, TRACE restarted from time step 0 dump. TRACE reported which dump it restarted from, but this was not immediately noticed in the TRACE output files. With DSTEP changed to 24000 TRACE restart from the final steady-state dump and calculations were consistent with the version 5.0 Patch 3 results. This trouble report is resolved. Directory: Newsletters Newsletters -> Last year four of the best Afrikaans Christian singers joined forces to tour Down Under with the production ‘Manne wat Glo’ Newsletters -> An Extension Cord to mars Newsletters -> In this issue Newsletters -> - Newsletters -> Acoustics Shock Vibration Signal Processing September 2011 Newsletter Newsletters -> In this issue Newsletters -> Engaging Families Supporting Students from Cradle to Career Newsletters -> Thank You Spring Retirees Newsletters -> Market direction Download 326.54 Kb. Share with your friends:
|