ML082380268: Difference between revisions
StriderTol (talk | contribs) (Created page by program invented by StriderTol) |
StriderTol (talk | contribs) (Created page by program invented by StriderTol) |
||
Line 16: | Line 16: | ||
=Text= | =Text= | ||
{{#Wiki_filter:WASHINGTON STATE%;UNIVERSITY Nuclear Radiation Center August 6, 2008 United States Nuclear Regulatory Commission Document Control Desk Washington D.C. 20555-0001 Docket Number 50-27 Facility License Number R-76 This purpose of this letter is to transmit a communication from General Atomics declaring that a document submitted by Washington State University, "BLOOST Code Validation Report" on June 13, 2008, may be made available for public release.A copy of the General Atomics letter authorizing release of the BLOOST Code Validation Report is included with this letter.I declare under penalty of perjury that the foregoing is true to the best of my knowledge. | {{#Wiki_filter:WASHINGTON STATE | ||
Respectfully Submitted Donald Wall, Ph.D.Director Nuclear Radiation Center Washington State University P.O. Box641 300, Pullman, WA 99164-1 300 509-335-8641 | %;UNIVERSITY Nuclear Radiation Center August 6, 2008 United States Nuclear Regulatory Commission Document Control Desk Washington D.C. 20555-0001 Docket Number 50-27 Facility License Number R-76 This purpose of this letter is to transmit a communication from General Atomics declaring that a document submitted by Washington State University, "BLOOST Code Validation Report" on June 13, 2008, may be made available for public release. | ||
-Fax: 509-335-4433 | A copy of the General Atomics letter authorizing release of the BLOOST Code Validation Report is included with this letter. | ||
I declare under penalty of perjury that the foregoing is true to the best of my knowledge. | |||
Respectfully Submitted Donald Wall, Ph.D. | |||
Director Nuclear Radiation Center Washington State University P.O. Box641 300, Pullman, WA 99164-1 300 509-335-8641 - Fax: 509-335-4433 * *ww.wsu.edu/nrc | |||
Q GENERAL ATOMICS From: Ken Mushinski Refer To: kjm08:018 Project: WSU To: Don Wall, Washington State University Date: 5 Aug. 2008 | |||
==SUBJECT:== | ==SUBJECT:== | ||
Bloost Code Validation Report, 21C025 Revision 0 Please use this memo as official notification that | Bloost Code Validation Report, 21C025 Revision 0 Please use this memo as official notification that GeneralAtomics approves the unrestricted release of General Atomics document 21C025 Rev. 0, Bloost Code Validation Report, regardless of the proprietary denotation on the Issue/Release Summary page contained therein. | ||
ONGHARAK NUCLEAR RESEARCH CENTER PROJECT BLOOST Code Validation Report Prepared for the Government of Thailand Ministry of Science, Technology and Environment Office of Atomic Energy for Peace Under Contract No. 56/2540 April, 2000 Prepared By TRIGA Technologies Under Contract to General Atomics | 1 of I Kjm08-018 Bloost Validation Report | ||
+. TRIGA TECH ISSUE/RELEASE | |||
21C025 Revision 0 ONGKHARAK NUCLEAR RESEARCH CENTER PROJECT BLOOST Code Validation Report Prepared for the Government of Thailand Ministry of Science, Technology and Environment Office of Atomic Energy for Peace Under Contract No. 56/2540 April, 2000 | |||
+ GENERAL ATOMICS, Marubeni HITACHI Qnsto zdfnT +TRIVA TECH! | |||
ONGHARAK NUCLEAR RESEARCH CENTER PROJECT BLOOST Code Validation Report Prepared for the Government of Thailand Ministry of Science, Technology and Environment Office of Atomic Energy for Peace Under Contract No. 56/2540 April, 2000 Prepared By TRIGA Technologies Under Contract to General Atomics | |||
+. TRIGA TECH ISSUE/RELEASE | |||
==SUMMARY== | ==SUMMARY== | ||
T-1495 (REV. 0/99) | |||
REVISION APPROVAL/ | T-1495 (REV. 0/99) | ||
REV. PREPARED DESCRIPTION DATE BY ENGINEERING QA -. PROJECT | SAFE.CLS QA LVL/CLS DOC. TYPE PROJECT DOCUMENT NO. REV. | ||
EXCEPT WITH THE WRITTEN CONSENT OF GA, (I) THIS DOCUMENT MAY NOT BE COPIED IN WHOLE OR IN PART AND WILL BE RETURNED UPON REQUEST OR WHEN NO LONGER NEEDED BY RECIPIENT AND (2) INFORMATION CONTAINED HEREIN MAY NOT BE COMMUNICATED TO OTHERS AND MAY BE USED BY RECIPIENT ONLY FOR THE PURPOSE FOR WHICH IT WAS TRANSMITTED. | EJ R&D APPVL LI DV&S LEVEL SR-A IIA R 2641 2641 21 C025 21 C025 00 0 DESIGN Q N/A 2 SYSTEM SEIS. CLS. SEIS. CAT. E. CLASS 21 N/A N/A N/A II TITLE: | ||
L] NO GA PROPRIETARY INFORMATION PAGE iOF* | BLOOST Code Validation Report CM APPROVAL(S) REVISION APPROVAL/ REV. PREPARED DESCRIPTION DATE BY ENGINEERING QA -. PROJECT IRELEASED-I 1 R. Sherman R. Sherman K. Maxwell AR ca Initial Release 74.?,0VO gA~- ,qjqeo1 Independent Director of Elor: E gi*'rig USE CONTINUATION SHEET NEXT INDENTURED DOCUMENT (S) | ||
LIST OF EFFECTIVE PAGES Page Number Issue/Release Summary List of Effective Pages Table of Contents Pages 1-5 | *See List of Effective Pages COMPUTER PROGRAM PINs PG000258 APPROVAL BY GENERAL ATOMICS :ý N/A APPROVAL BY CUSTOMER N/A Z GA PROPRIETARY INFORMATION THIS DOCUMENT IS THE PROPERTY OF GENERAL ATOMICS. ANY TRANSMITTAL OF THIS DOCUMENT OUTSIDE OF GA WILL BE IN CONFIDENCE. EXCEPT WITH THE WRITTEN CONSENT OF GA, (I) THIS DOCUMENT MAY NOT BE COPIED IN WHOLE OR IN PART AND WILL BE RETURNED UPON REQUEST OR WHEN NO LONGER NEEDED BY RECIPIENT AND (2) INFORMATION CONTAINED HEREIN MAY NOT BE COMMUNICATED TO OTHERS AND MAY BE USED BY RECIPIENT ONLY FOR THE PURPOSE FOR WHICH IT WAS TRANSMITTED. | ||
/TABLE OF CONTENTS 1.0 CODE IDENTIFICATION | L] NO GA PROPRIETARY INFORMATION PAGE iOF* | ||
............................................................................. | |||
1 2.0 PROGRAM DESCRIPTION | LIST OF EFFECTIVE PAGES Page Number Page Count Revision Issue/Release Summary 5 0 List of Effective Pages 1 0 Table of Contents 1 0 Pages 1-5 5 0 Attachment A 12 0 Attachment B 10 0 Total Pages 30 ii 21C025, Rev.O | ||
.................................. | |||
1 3.0 VALIDATION TESTS ................................................................................... | / | ||
2 | TABLE OF CONTENTS 1.0 CODE IDENTIFICATION ............................................................................. 1 2.0 PROGRAM DESCRIPTION .................................. 1 3.0 VALIDATION TESTS ................................................................................... 2 | ||
==4.0 CONCLUSION== | ==4.0 CONCLUSION== | ||
S | S ............................................................................................ 4 | ||
............................................................................................ | |||
4 | |||
==5.0 REFERENCES== | ==5.0 REFERENCES== | ||
............................................................................................. | ............................................................................................. 4 6.0 INDEPENDENT EVALUATION ................................................................. 5 7.0 ATTACHMENT A-BLOOST (TRIGA VERSION) USERS MANUAL ...A-1 8.0 ATTACHMENT B - SANDIA ACPR CORE PERFORMANCE (PULSE OPERATION) ............................................................................................ B-1 iii 21C025, Rev.0 | ||
4 6.0 INDEPENDENT EVALUATION | |||
................................................................. | 1.0 CODE IDENTIFICATION Program ID No. (PIN) - PG000258 Program Name - The BLOOST Code (TRIGA Version), referred to as TRIGA-BLOOST Computer Platform - Dec Alpha with UNIX operating system Executable Location - /gaprod/decalpha/tbloost/bin/tbloost3 Sample Test Problems - /gaprod/decalpha/tbloost/examples 2.0 PROGRAM DESCRIPTION BLOOST is a combined reactor kinetics-heat transfer computer code developed by General Atomics. A special version, utilizing a space-independent heat transfer model, was developed for pulsing TRIGA reactor design work. A short description of the TRIGA-BLOOST code's purpose and capabilities is given in the user manual: | ||
5 7.0 ATTACHMENT A-BLOOST (TRIGA VERSION) USERS MANUAL ... A-1 8.0 ATTACHMENT B -SANDIA ACPR CORE PERFORMANCE (PULSE OPERATION) | "The kinetics model employs the space-independent form of the reactor kinetics equations with very general forms of reactivity input and feedback. | ||
............................................................................................ | The first basic assumption of the method is that the shape of the neutron flux does not vary with time regardless of how large the change in amplitude may be. While this criterion is rarely satisfied in an exact sense, it is reasonably adequate for a large number of practical cases, including the small cores used in TRIGA reactors. | ||
B-1 iii 21C025, Rev.0 1.0 CODE IDENTIFICATION Program ID No. (PIN) -PG000258 Program Name -The BLOOST Code (TRIGA Version), referred to as TRIGA-BLOOST Computer Platform -Dec Alpha with UNIX operating system Executable Location -/gaprod/decalpha/tbloost/bin/tbloost3 Sample Test Problems -/gaprod/decalpha/tbloost/examples 2.0 PROGRAM DESCRIPTION BLOOST is a combined reactor kinetics-heat transfer computer code developed by General Atomics. A special version, utilizing a space-independent heat transfer model, was developed for pulsing TRIGA reactor design work. A short description of the TRIGA-BLOOST code's purpose and capabilities is given in the user manual: "The kinetics model employs the space-independent form of the reactor kinetics equations with very general forms of reactivity input and feedback.The first basic assumption of the method is that the shape of the neutron flux does not vary with time regardless of how large the change in amplitude may be. While this criterion is rarely satisfied in an exact sense, it is reasonably adequate for a large number of practical cases, including the small cores used in TRIGA reactors.The heat transfer calculation uses simple lumped parameter models to compute the average temperatures in the UZrH fuel material and the water coolant. These average temperatures are used in combination with temperature coefficient data appropriate to the whole core to determine the reactivity feedback terms in the kinetics equations. | The heat transfer calculation uses simple lumped parameter models to compute the average temperatures in the UZrH fuel material and the water coolant. These average temperatures are used in combination with temperature coefficient data appropriate to the whole core to determine the reactivity feedback terms in the kinetics equations. The second basic assumption of the method is thus that a single average fuel element can be used to determine reactivity feedback from temperature changes occurring over the entire core." | ||
The second basic assumption of the method is thus that a single average fuel element can be used to determine reactivity feedback from temperature changes occurring over the entire core." With respect to TRIGA reactor core design, TRIGA-BLOOST is used for calculating the maximum fuel temperature in the core as a function of time reached during a fast transient (similar to a short power pulse). These types of transients are generated by rapid positive reactivity insertions into the core under cold startup or full power operating conditions, followed very quickly by a reactor control rod system scram. Please see the User's Manual, 1 21C025, Rev.0 included in this document as Attachment. | With respect to TRIGA reactor core design, TRIGA-BLOOST is used for calculating the maximum fuel temperature in the core as a function of time reached during a fast transient (similar to a short power pulse). These types of transients are generated by rapid positive reactivity insertions into the core under cold startup or full power operating conditions, followed very quickly by a reactor control rod system scram. Please see the User's Manual, 1 21C025, Rev.0 | ||
A, for more detailed information on the solution algorithms used, as well as the input instructions. | |||
3.0 VALIDATION TESTS Many of the TRIGA reactors built by General Atomics were designed to operate safely in a pulsing mode of operation. | included in this document as Attachment. A, for more detailed information on the solution algorithms used, as well as the input instructions. | ||
With such designs, it was imperative that an accurate and reliable means of calculating core performance during rapid transients (power pulses), be available. | 3.0 VALIDATION TESTS Many of the TRIGA reactors built by General Atomics were designed to operate safely in a pulsing mode of operation. With such designs, it was imperative that an accurate and reliable means of calculating core performance during rapid transients (power pulses), be available. The TRIGA-BLOOST code was developed for just that purpose. TRIGA-BLOOST has been used at General Atomics for over thirty years to calculate maximum fuel temperatures and total energy release during rapid reactivity transients. The code has been applied to both pulsing and non-pulsing TRIGA reactors. | ||
The TRIGA-BLOOST code was developed for just that purpose. TRIGA-BLOOST has been used at General Atomics for over thirty years to calculate maximum fuel temperatures and total energy release during rapid reactivity transients. | A useful benchmark problem for TRIGA-BLOOST has been documented (Ref.1) as part of the safety analysis report for the TRIGA Annular Core Pulse Reactor (ACPR). A comparison is made of the measured versus calculated core performance characteristics of the Sandia Laboratories ACPR for a transient pulse of $4.40 reactivity insertion. Conditions at the onset of the $4.40 transient were as follows: | ||
The code has been applied to both pulsing and non-pulsing TRIGA reactors.A useful benchmark problem for TRIGA-BLOOST has been documented (Ref. 1) as part of the safety analysis report for the TRIGA Annular Core Pulse Reactor (ACPR). A comparison is made of the measured versus calculated core performance characteristics of the Sandia Laboratories ACPR for a transient pulse of $4.40 reactivity insertion. | * To =25C SPo 10watts | ||
Conditions at the onset of the $4.40 transient were as follows:* To =25C SPo 10watts* 141 Fuel elements + 6 fueled followers* $4.40 pulse with transient rods fully scrammed after I second Using this basic data, as well as important calculated input core parameters such as the effective delayed neutron fraction, | * 141 Fuel elements + 6 fueled followers | ||
The results, taken from Table 7.1-15 of Ref.1, are summarized in Table 3.1.2 21C025, Rev.0 Table 3.1 Sandia ACPR Pulsing Operation, TRIGA-BLOOST versus Experimental GA Calculated Performance Parameter Sandia Experimental Sandia GA Parameters Parameters Peak Power (MW) 12,100 12,500 14,200 Period, T (msec) 1.31 1.38 1.31 Energy Release (MW-sec)Prompt Burst 76 77 80 I sec -83 86 5 sec 88 88 92 10 sec 90 90 94 Peak Fuel Temp (*C) --860 Peak Temp at Thermocouple | * $4.40 pulse with transient rods fully scrammed after I second Using this basic data, as well as important calculated input core parameters such as the effective delayed neutron fraction, P3 lf, the prompt neutron lifetime, e, and the prompt negative temperature coefficient, a, TRIGA-BLOOST can be used to model core performance during the transient. The results, taken from Table 7.1-15 of Ref.1, are summarized in Table 3.1. | ||
(*C) 735 (7 sec) 745 (15 sec)Parameters Effective Delayed Neutron 0.0073 0.0073 Fraction, Peff Neutron Lifetime, £ (psec) 33 34 32 Power Peak/Avg. (total) 2.54 2.20 Prompt Negative Temp 9.67x 10-' 9.6x 10-'Coefficient, ct (8k/'C)3 21C025, Rev.0 Note the very excellent agreement between the measured performance parameters and those calculated with TRIGA-BLOOST. | 2 21C025, Rev.0 | ||
For TRIGA reactor calculations, the experience at General Atomics has been that the TRIGA-BLOOST code generates results that are somewhat conservative (slightly higher peak powers, energy release, and maximum fuel temperatures) relative to experimental values. The results shown in Table 3.1 confirm this observation. | |||
The entire section from Ref.1 concerning the Sandia ACPR transient is included in this document as Attachment B.This benchmark case shows that TRIGA-BLOOST provides accurate core performance results for TRIGA reactors undergoing rapid reactivity insertion transients. | Table 3.1 Sandia ACPR Pulsing Operation, TRIGA-BLOOST versus Experimental GA Calculated Performance Parameter Sandia Experimental Sandia GA Parameters Parameters Peak Power (MW) 12,100 12,500 14,200 Period, T (msec) 1.31 1.38 1.31 Energy Release (MW-sec) | ||
It also shows the adequacy of the GA methods for generating important input to TRIGA-BLOOST, such as | Prompt Burst 76 77 80 I sec - 83 86 5 sec 88 88 92 10 sec 90 90 94 Peak Fuel Temp (*C) - - 860 Peak Temp at Thermocouple (*C) 735 (7 sec) 745 (15 sec) | ||
Parameters Effective Delayed Neutron 0.0073 0.0073 Fraction, Peff Neutron Lifetime, £ (psec) 33 34 32 Power Peak/Avg. (total) 2.54 2.20 Prompt Negative Temp 9.67x 10-' 9.6x 10-' | |||
Coefficient, ct (8k/'C) 3 21C025, Rev.0 | |||
Note the very excellent agreement between the measured performance parameters and those calculated with TRIGA-BLOOST. For TRIGA reactor calculations, the experience at General Atomics has been that the TRIGA-BLOOST code generates results that are somewhat conservative (slightly higher peak powers, energy release, and maximum fuel temperatures) relative to experimental values. The results shown in Table 3.1 confirm this observation. The entire section from Ref.1 concerning the Sandia ACPR transient is included in this document as Attachment B. | |||
This benchmark case shows that TRIGA-BLOOST provides accurate core performance results for TRIGA reactors undergoing rapid reactivity insertion transients. It also shows the adequacy of the GA methods for generating important input to TRIGA-BLOOST, such as i3eff, £, and a. | |||
==4.0 CONCLUSION== | ==4.0 CONCLUSION== | ||
Line 69: | Line 82: | ||
==5.0 REFERENCES== | ==5.0 REFERENCES== | ||
: 1. "TRIGA Annular Core Pulse'Reactor Safety Analyses and Preliminary Design for the Nuclear Safety Research Reactor," TRIGA Reactors, Gulf Energy and Environmental Systems, E- 117-174, July 1972.2. "Control of Scientific and Engineering Computer Programs," TRIGA Technologies Engineering Procedure EP-3.7, latest revision 4 21C025, Rev.0 6.0 INDEPENDENT EVALUATION This Validation Report has been reviewed and evaluated and it is concluded that the validation procedures met the requirements of Ref.2. The sample test problems were demonstrated to be available, and the validation results were adequately evaluated and documented. | : 1. "TRIGA Annular Core Pulse'Reactor Safety Analyses and Preliminary Design for the Nuclear Safety Research Reactor," TRIGA Reactors, Gulf Energy and Environmental Systems, E- 117-174, July 1972. | ||
The evaluator is very familiar with the use of the code for most all past TRIGA design calculations and the comparisons with operating reactor results.This evaluation consisted of reviewing the benchmark calculations and results, as well as their comparison with operating reactor results. Input was reviewed by sampling random values, checking volumes, and engaging in question and answer sessions with the nuclear designer.The validation effort was direct and straightforward; comparing calculated and experimentally determined results for a reactor in a pulsing (kinetic)mode. The comparisons give an adequate demonstration of the ability of the code to predict reactor performance and conditions during rapid transient operation. | : 2. "Control of Scientific and Engineering Computer Programs," TRIGA Technologies Engineering Procedure EP-3.7, latest revision 4 21C025, Rev.0 | ||
5 21C025, Rev.0 7.0 ATTACHMENT A -BLOOST (TRIGA VERSION) USERS MANUAL A-1 21G025, Rev.0 THE BLOOST CODE (TRIGA Version).September 16, 1997 A-2; R, o I BLOOST is a combined reactor kinetics-heat transfer code written in the Fortran language for the VAX computer. | |||
The kinetics model employs the space-independent form of the reactor kinetics equations with very general forms of reactivity input and feedback. | 6.0 INDEPENDENT EVALUATION This Validation Report has been reviewed and evaluated and it is concluded that the validation procedures met the requirements of Ref.2. The sample test problems were demonstrated to be available, and the validation results were adequately evaluated and documented. | ||
The first basic assumption of the method is that the shape of the neutron flux does not vary with time regardless of bow large'the change in amplitude may be. While this criteria is rarely satisfied in an exact sense, it is reasonably adequate for a large number of practical cases, including the small cores used in TRIGA reactors.The heat transfer calculation uses simple lumped parameter models to compute the average temperatures in the UZrH fuel material and the water coolant These average temperatures are used in combination with temperature coefficient data appropriate to the whole core to determine the reactivity feedback terms in the kinetics equations. | The evaluator is very familiar with the use of the code for most all past TRIGA design calculations and the comparisons with operating reactor results. | ||
The second basic assumption of the method is thus that a single average fuel element can be used to determine reactivity feedback from temperature changes occurring over the entire. core.The form of the reactor kinetic equations used is given below: | This evaluation consisted of reviewing the benchmark calculations and results, as well as their comparison with operating reactor results. Input was reviewed by sampling random values, checking volumes, and engaging in question and answer sessions with the nuclear designer. | ||
The use of the scram table can be initiated after a specified power level or percentage of full power has been reached in the tranrient. | The validation effort was direct and straightforward; comparing calculated and experimentally determined results for a reactor in a pulsing (kinetic) mode. The comparisons give an adequate demonstration of the ability of the code to predict reactor performance and conditions during rapid transient operation. | ||
A delay time can be specified to simulate electrical or mechanical delays in a control system. The RAMP function may be continued or terminated after the SCRAM starts depending on the value of an input number.The units of S and Ci come about when the kinetics equations, which are usually derived in terrns of the neutron density n, are written in terms of the power level P. A constant proportiotiality between n and P is thus assumed which is not strictly correct since they are coupled by the fission cross section P(Mw) = | 5 21C025, Rev.0 | ||
Mw-sec -3x10 t s (4.)Thus, the source in neutrons/sec-cm3 and the delayed neutron precursor concentrations Ci in | |||
in thrmal, spectrum mea~ctors. | 7.0 ATTACHMENT A - BLOOST (TRIGA VERSION) USERS MANUAL A-1 21G025, Rev.0 | ||
S.tawrp. typc. prmb.crus. | |||
in .which _IO temtpeatmre. | THE BLOOST CODE (TRIGA Version). | ||
feedback occurs can, of course, be run in units of n rather than P by supplying consistent values for the input quantities P and S.The solution of the set of kinetic equations (1) and (2) is done by the method of fourth-order Runge-Kutta integration described by Cohen (1) as modified by Schwartz (2) to include a variable time step. This is done in the subroutines RKCOT, DERIV and COH In the following description, the nomenclature follows the Fortran symbols as far as possible.The set of equations is rewritten using Y for the variable and YP for the time derivative. | September 16, 1997 A-2 ; C*o*" R, o | ||
YP(1) = A(I) Y(1) + R(1)For the power equation Y() =P, A(l) and for the precursor equations Y(I+) =Ci and A(l+l) = -i. The remainder R(I) is always computed as YP(I) -A(I)Y(I). | |||
: 1. Cohen, E.R., "Some Topics in Reactor Kinetics," p/629,2nd Geneva Conference, Vol. II, p.306 (1958).2. Schwartz, A., "Generalized Reactor Kinetics Code -AIREK-il," NAA-SR-Memo 4980, (1960) 3 The sequence of operations is as follows for each equation of the set. A time step H is used, the first value for which is an input number. Subsequent values are chosen by the code. The initial values of Y, YP, A and R are calculated from the input data. The first step of the integrationcomputes Y(t+W2) by assuming R to be constant at its initial value (called Ro). This gives afirst approximation for Y(t+Ht2) from which R(t+H/2) = R, can be computed. | I BLOOST is a combined reactor kinetics-heat transfer code written in the Fortran language for the VAX computer. The kinetics model employs the space-independent form of the reactor kinetics equations with very general forms of reactivity input and feedback. The first basic assumption of the method is that the shape of the neutron flux does not vary with time regardless of bow large'the change in amplitude may be. While this criteria is rarely satisfied in an exact sense, it is reasonably adequate for a large number of practical cases, including the small cores used in TRIGA reactors. | ||
Assuming that the variation of R is linear, a better Y(tt-2) is computed and R(t+H/2) is further improved (called | The heat transfer calculation uses simple lumped parameter models to compute the average temperatures in the UZrH fuel material and the water coolant These average temperatures are used in combination with temperature coefficient data appropriate to the whole core to determine the reactivity feedback terms in the kinetics equations. The second basic assumption of the method is thus that a single average fuel element can be used to determine reactivity feedback from temperature changes occurring over the entire. core. | ||
The equations for this procedure are given below. The symbol A refers to A(l).YO = Y(t), RO = R(YO, t), YPO = YP(t)H H Yt -Yo -Ct(A--)[RO+AYa] | The form of the reactor kinetic equations used is given below: | ||
R 1-=R(Y,,t-+ | dPs | ||
H/ 2)=YP(t+TH/2)-AY, H H | _ p-p+ | ||
+AYO]+2HC 2 (AH)[R 2 -IR,] | .Q+(1.) | ||
and it may be possible to use a fictitious large A value to get around the time step limit Throughout the solution, the subroutine DERIV is called five times to obtain values of the derivatives YP at times (t + H/2) and (t + H). The reactivity components used in Eq. (3) for computing p are read from the ramp, scram, and temperature coefficient tables at these time points. The scram can only be initiated at the end of a kinetics time step (t + H). Once it has been initiated, values are obtained at (t + F/2) and (t + H) in subsequent time steps. The ramp table is read at all time points in the integration. | dt A dt* -zP-AQ (2.) | ||
When the time step H is completed, a check is made on the accuracy of the solution. | p= RAMP (time) - SCRAM (time) - TEMPT (3.) | ||
If satisfactory, the same size step is used again. If the accuracy is very good, the step is doubled, or, if poor, the time step is halved and the step ro.cmpute4. | P= reactor power level -megawatts Ai effective delayed neutron yield for group i. | ||
Doubling is not permitted if A H would then exceed 89. The accuracy criteria is a number Q given by the equation.HCz(AH)f.- | 6 groups are assumed. | ||
They should be about a factor of 10 apart. Values of 10-s, 104 are typical for many TRIGA problems. | i= delayed neutron precursor decay constant for group i - | ||
The user should do some numerical experimentation for the specific problems to be solved.The time variations of the average fuel element temperature, T& and the average coolant temperature, Tl, are expressed in simple lumped parameter equations which have the same form as the 7 kinetic equations and therefore can be solved along with them by the procedures described above.If heat is generated by the total fission power P in all of the fuel and removed by convection heat transfer to the coolant, a simple heat balance gives: d P- (5.) | sec A.= prompt neutron generation time - secs Ci = delayed neutron precursor concentration of group i-megawatts S = constant source term megawatts/sec A-3 aACO3 S7 P~ev 0 | ||
Since the number of fuel elements frequently varies in TRIGA cores, the user supplies the heat capacity per fuel element and the number of elements in the input. The heat capacity may have a linear diependence on the fuel temperature, so that Cý= | |||
The product of the heat transfer coefficient and the area, hA, is represented by a thermal 1 resistance, R, -hA ' which is supplied by another input table as a function of fuel temperature. | 2 RAMP (time) and SCRAM (time) are reactivity changes measured from time zero which are read from user-supplied input tables at each time step using linear interpolation. The use of the scram table can be initiated after a specified power level or percentage of full power has been reached in the tranrient. A delay time can be specified to simulate electrical or mechanical delays in a control system. The RAMP function may be continued or terminated after the SCRAM starts depending on the value of an input number. | ||
It is typically derived from more sophisticated heat transfer calculations for the core at various constant power levels so that T -T.Power per element The coolant flowing in the core channels is assumed to transfer its heat to the water in the reactor tank which is assumed to be a constant temperature sink that remains at the initial coolant temperature. | The units of S and Ci come about when the kinetics equations, which are usually derived in terrns of the neutron density n, are written in terms of the power level P. A constant proportiotiality between n and P is thus assumed which is not strictly correct since they are coupled by the fission cross section t | ||
This is a suitable approximation for the short time scale of transients typically studied with BLOOST.A-?- | P(Mw) = 3xlIfission/ Mw-sec - 3x10 t s (4.) | ||
Thus, the source in neutrons/sec-cm3 and the delayed neutron precursor concentrations Ci in cm3 are multiplied by the factor Ef nvV,. | |||
3X10" which is constant with time to the extent that ,r varies as IN. The validity of this assumption is | |||
..olygrea.r. in thrmal, spectrum mea~ctors. S.tawrp. typc. prmb.crus. in .which _IO temtpeatmre. | |||
feedback occurs can, of course, be run in units of n rather than P by supplying consistent values for the input quantities P and S. | |||
The solution of the set of kinetic equations (1) and (2) is done by the method of fourth-order Runge-Kutta integration described by Cohen (1) as modified by Schwartz (2) to include a variable time step. This is done in the subroutines RKCOT, DERIV and COH In the following description, the nomenclature follows the Fortran symbols as far as possible. | |||
The set of equations is rewritten using Y for the variable and YP for the time derivative. | |||
YP(1) = A(I) Y(1) + R(1) | |||
For the power equation Y() =P, A(l) and for the precursor equations Y(I+) =Ci and A(l+l) = -i. The remainder R(I) is always computed as YP(I) -A(I)Y(I). | |||
: 1. Cohen, E.R., "Some Topics in Reactor Kinetics," p/629,2nd Geneva Conference, Vol. II, p.306 (1958). | |||
: 2. Schwartz, A., "Generalized Reactor Kinetics Code - AIREK-il," NAA-SR-Memo 4980, (1960) | |||
3 The sequence of operations is as follows for each equation of the set. A time step H is used, the first value for which is an input number. Subsequent values are chosen by the code. The initial values of Y, YP, A and R are calculated from the input data. The first step of the integrationcomputes Y(t+W2) by assuming R to be constant at its initial value (called Ro). This gives afirst approximation for Y(t+Ht2) from which R(t+H/2) = R, can be computed. Assuming that the variation of R is linear, a better Y(tt-2) is computed and R(t+H/2) is further improved (called R2). A linear extmplation of R is now used to compute Y(t+f) and R(t+H) = R, for the completetim step. A quadratic fit is now made with Ro, R 2 and R3 to determine an improved | |||
-4 and from this the final Y(t+H) is calculated. | |||
The equations for this procedure are given below. The symbol A refers to A(l). | |||
YO = Y(t), RO = R(YO, t), YPO = YP(t) | |||
H H Yt -Yo - Ct(A--)[RO+AYa] | |||
R 1 -=R(Y,,t-+H/ 2)=YP(t+TH/2)-AY, H H Y2 = Y1 + - C2 (Aj -)[Rl -Re] | |||
R 2 =R(Y2 ,t+H/2) | |||
Y= Yo +HC(AH)[RO +AYO]+2HC 2 (AH)[R 2 -IR,] | |||
R3 = R(Y 3 ,t + H) | |||
Y= Y3 +H{2C 3 (AH) -C,(AH))[R 2 Ra+ R 3] | |||
0 R4 = R(Y4 ,t + H) | |||
Ys=Y 4 +H(2C 3 (AH)-C 2 (AH)}[R 4 R3 ] | |||
In these equations, the functions C(x) are integrating factors defined as follows: | |||
For xi < 1.0 CA(x) = + 1) + 1) + 1) + 1) + 1) + 1) +l)1 | |||
+1 C2(x) - xC,(x) 2 a O | |||
4 c,( - xC2 (X)+ I 2 | |||
For 14> 1.0 C,(x) - e' - 1 c3(x)- 2c,(x)-lt K | |||
The argument x is either A H or A M/2/. The C(x) functions are computed in the function subroutine COH(Nx) where N is the subscript of the function (1,2 or 3). The function computed by COH for N = 3 is actually 2C 3(x) - C2(x), the term occurring in the equations for Y4 and Y. | |||
above. | |||
Since the C 1(x) function involves an exponential for large x, the argument x cannot be allowed to exceed 88. Since A(l)=(p-13)/A, this implies a fundamental time step limit of: | |||
B8A H< | |||
This can be a significant limit for fast reactor problems where A may be of the order of 10. secs. | |||
In some cases when the net reactivity is known to be well below prompt critical throughout the transient, the kinetic behavior is governed by the delayed neutron precursors. and it may be possible to use a fictitious large A value to get around the time step limit Throughout the solution, the subroutine DERIV is called five times to obtain values of the derivatives YP at times (t + H/2) and (t + H). The reactivity components used in Eq. (3) for computing p are read from the ramp, scram, and temperature coefficient tables at these time points. The scram can only be initiated at the end of a kinetics time step (t + H). Once it has been initiated, values are obtained at (t + F/2) and (t + H) in subsequent time steps. The ramp table is read at all time points in the integration. When the time step H is completed, a check is made on the accuracy of the solution. If satisfactory, the same size step is used again. If the accuracy is very good, the step is doubled, or, if poor, the time step is halved and the step ro.cmpute4. Doubling is not permitted if A H would then exceed 89. The accuracy criteria is a number Q given by the equation. | |||
HCz(AH)f.-2o8 +o,,I Q=- I + C,(AH) | |||
,..,IC AL t-o :S"gO | |||
5 The s'sare inverse periods defined by M=YPO 0 I I(Y(t + R) | |||
SyoYO YP(t+H) o -Y(t+H) | |||
The criteria is evaluated only for the first equation of the set so that Y is the reactor power level. | |||
Thus eo is the instantaneous inverse period at the beginning of the step, c), is the instantaneous inverse period at the end of the step, and ota is an average inverse period over the step. The inverse of cn is printed in the problem output as the reactor period. | |||
The value of Q obtained at the end of a time step is compared with two input numbers QI and Q2. | |||
An additional test quantity is called DI and is the first increment of Y calculated in starting the timestop (i.e., Y, -Yo). | |||
Dl= H HtA! | |||
D2 )R + AY-J The use of the Q and DI criteria is as follows: | |||
If Dl <2x0-1 4 YO and Q >:QI keep same H for next step. | |||
If Dl < 2xl 0 "4yo and Q < Q 1 double H for next step. | |||
IfDl >2xl0-xY 0 and Q<QI double H for next step. | |||
If Dl> 2xl0- 4 Y and Q>Qi and/or Q< Q2 keep same H for next step. | |||
If DI>2x10-4 Y0 and Q_>Q2 halve H and recompute the time step, unless the new reduced H is smaller than the input number SMALLH. In that case the problem is stopped. | |||
A-7 atco*r"Ize- 0 | |||
6 Aflow diagram of the kinetics solution is shown in Fig. 1. | |||
Values of QI and Q2 to use are best determined by excperience. They should be about a factor of 10 apart. Values of 10-s, 104 are typical for many TRIGA problems. The user should do some numerical experimentation for the specific problems to be solved. | |||
The time variations of the average fuel element temperature, T&and the average coolant temperature, Tl, are expressed in simple lumped parameter equations which have the same form as the 7 kinetic equations and therefore can be solved along with them by the procedures described above. | |||
If heat is generated by the total fission power P in all of the fuel and removed by convection heat transfer to the coolant, a simple heat balance gives: | |||
d P- -T. (5.) | |||
dt C1 where Cp is the heat capacity of all fuel material. Since the number of fuel elements frequently varies in TRIGA cores, the user supplies the heat capacity per fuel element and the number of elements in the input. The heat capacity may have a linear diependence on the fuel temperature, so that Cý=C0 +yT~ | |||
where the user supplies Co and y. The units of C. are Mw-sec/*C. | |||
The product of the heat transfer coefficient and the area, hA, is represented by a thermal 1 | |||
resistance, R, - hA ' which is supplied by another input table as a function of fuel temperature. | |||
It is typically derived from more sophisticated heat transfer calculations for the core at various constant power levels so that T -T. | |||
Power per element The coolant flowing in the core channels is assumed to transfer its heat to the water in the reactor tank which is assumed to be a constant temperature sink that remains at the initial coolant temperature. This is a suitable approximation for the short time scale of transients typically studied with BLOOST. | |||
A-? -1kc-OAr Qtv 0 | |||
7 The equation used is dT. | |||
dt Cp. | |||
,where C., and R, are, respectively, the heat capacity of the water in the coolant channel (per ful element) and a suitable resistance derived from steady state conditions and assumed constant. | |||
The average fuel and coolant temperatures Tr and Tw are used with input tables of pre-computed reactivity feed-back contributions from the fuel and coolant temperature coefficients. | |||
These contributions are summed to give the term TEMPT in equation 3. The input data for the tables consist of pairs of temperature values (deg c) and reactivity change (A&p). The code automatically normalizes the table supplied by the user so that the reactivity change is zero for the-initial value of the temperature. | These contributions are summed to give the term TEMPT in equation 3. The input data for the tables consist of pairs of temperature values (deg c) and reactivity change (A&p). The code automatically normalizes the table supplied by the user so that the reactivity change is zero for the-initial value of the temperature. | ||
While the average fuel element temperature determines the reactivity feed-back, the quantity of most interest in a transient is usually the maximum fuel temperature, which must not reach a level that would cause damage, The user can supply a pre-computed value for the ratio of maximum to average power density in the hottest fuel element (P/ P). BLOOST uses this in an equation similar to Eq.5 to compute the maximum fuel temperature throughout the transient The power generated during the transient is integrated over time as the calculation progresses and printed at each time step in the output as energy in Mw sec.BLOOST was a fixed address format for the input data. Each card image starts with the address of the first data entry and the number of entries on the card. Data not needed can be omitted.Problems can be stacked. The last card of a problem has the integer 1 in column 1.The algebraic signs of reactivity terms to be supplied in the various input tables will be positive for the most common situations; i.e.,,negative temperature coefficients, positive ramp insertions and negative scrams following the sense of Eq.3.The input card image format is as follows: (II, [5,11, 7E9.4)Column 1 Reserved for integer I on last card of set Columns 2-6 Address of Ist variable on card Column 7 No. of words to read on this card A -9:t | While the average fuel element temperature determines the reactivity feed-back, the quantity of most interest in a transient is usually the maximum fuel temperature, which must not reach a level that would cause damage, The user can supply a pre-computed value for the ratio of maximum to average power density in the hottest fuel element (P/ P). BLOOST uses this in an equation similar to Eq.5 to compute the maximum fuel temperature throughout the transient The power generated during the transient is integrated over time as the calculation progresses and printed at each time step in the output as energy in Mw sec. | ||
BLOOST was a fixed address format for the input data. Each card image starts with the address of the first data entry and the number of entries on the card. Data not needed can be omitted. | |||
-sees | Problems can be stacked. The last card of a problem has the integer 1 in column 1. | ||
The algebraic signs of reactivity terms to be supplied in the various input tables will be positive for the most common situations; i.e.,,negative temperature coefficients, positive ramp insertions and negative scrams following the sense of Eq.3. | |||
power for hottest fiLel element in core (P/1)Number of fuel elements in the core | The input card image format is as follows: (II, [5,11, 7E9.4) | ||
~t 0 9 Address Fortran Name Description of Data 26 RADEL Constant C. in heat capacity of fuel: C,, = | Column 1 Reserved for integer I on last card of set Columns 2-6 Address of Ist variable on card Column 7 No. of words to read on this card A-9 :t VC 0.ý S7 4tv 0 | ||
°C 36 TEMPTM Number of entries in the longer of the 2 temperatures coefficient tables (an entry is a pair of numbers)37-136 TAB A Reactivity contribuxtion associated with fiuel temperature change. Enter in sequence Tt, Ap,, T2, Ap 2 where T is in °C Ap, = lcx(Tr)dT where a (Tr) is the precomputed fuel t dk temperature coefficient I dr k dTr 137-236 TAB D Reactivity contribution associated with water temperature change. Same format as for fuel coefficient 426 RAMP[ Number of entries in ramp or scram table, which ever is longer (an entry is a pair of numbers)A-%I A-DL1CO e. | |||
time, Ap, time, Ap, etc.time in sees. Provide enough entries that linear interpolation is accurate.Last time to look up in scram table -secs Power level (Mw) or % of initial power (depending on value of STARUP, location 4) at which to initiate scram Time elapsed after power POWIRAT before scram is to begin. Used to simulate mechanical or electrical system delays Scram table. Enter in sequence time, Ap, time Ap etc.Number of entries in fuel heat transfer resistance table.Fuel heat transfer resistance table. Enter in sequence values of Tf, Rt, Tf, Rt, etc.Enter integer I in column I of last data card if other problems follow.A-12?a ICO Pee 0 8.0 ATTACHMENT B -SANDIA ACPR CORE PERFORMANCE (PULSE OPERATION) | 8 Columns 8-70 Up to 7 data entries A title card is supplied at the beginning of each case (Format 12A6). | ||
B-1 21C025, Rev.0 7.1.2.5 Core Performance (Pulse Operation). | The input data addresses and descriptions are given in the following pages. | ||
A comparison of the experimental and calculated core performanaee characteristics Is shown in Table 7,1-15 for a $4.40 pulse in the Sandia ACPR. The experimental values Oete obtained through communications with Sandia Corporation reactor operations. | Address Fortran Name Description of Data 1~ | ||
One set of parameters used for the calculated performance was also obtained from Sandia and is used by them because they give a calculated pulse shape which very nearly matches their experimental pulse shape. None of the paremeters (except for the neutron lifetime, 1) is experimentally determined | Title Card (alphanumeric) 1 PB Initial Power Level, P, - Mwt 2 SOURCE Source term if present. S - Mwt/sec 3 TEND Time to stop problem - sees 4- STARUP If STARUP = 1.0, scram level is set at power given in location 630. If STARUP = 0.0, the value in location 630 will be the percentage of initial power PB at which to scram. | ||
5 GENTIM Prompt neutron generation time, A -secs 6-11 ALAM () Delayed neutron precursor constants, | |||
The aef and f/F values used for the Sandia. parameters. | - i=1,6-secs-1 12-17 BETA (I) Delayed neutron yields; P3i i = 1,6 18 H Initial time step to use in Runge-Kutta integration - sees 19 SMALL H Minimum time step to use in Runge-Kutta integration - | ||
*came from the original design. As discussed in Section 7.1.3.2, the more detailed calculational procedures currently used to determine the power.peaking factor give somewhat lower values then the methods used for the orig-inn1 Sandia ACPR analysis. | I se*s 20 Ql Runge-Kutta accuracy criteria (suggest 10") | ||
The parameters used to calculate the performance characteristics in the last column of Table 7.1-15 were all-.generated as a: part of this analysis and are documented in the previous section.The kinetics calculations were done with a simplified version of the BLOOST code. (the general version of the BLOOST code is more fully described-in Section 71..3)whIch is a space-independent combined heat transfer-kinetics code using six delayed neutron groups and temperature feedback.The transfer of heat from the fuel to the coolant channe:L water and from the coolant channel water to the pool water was approximated In this simplified model by thermal resistance values: P where R. Thermal resistance | 21 Q2 Runge-Kutta accuracy criteria (suggest 10 -) | ||
(*C/MW)S5T Temperature difference, between the average fuel element and the average coolant channel water, or between the average coolant channel water and the pool water P Power in the average element (MW)7.1-28 9_2 ;11CO;L-..MO TABLE 7.1-15 SANDIA ACPR PULSING OPERATION PERMORAMNCE CRAMCTIRISTICS | A 22 DELTMX Peak/avg. power for hottest fiLel element in core (P/1) 23 XMTIME Number of fuel elements in the core Aq-1o A-t:0~c~z5S ~t 0 | ||
$4.40 Pulse with transient rods scra ind at 1 see 141 Fuel elements + 6 fueled followers T -25*C.0 | |||
.ParaWSJ:8 EPa4Metesthis aly Peak power, P (14W4)Period, T (msec)Pulse width (at N12) (msec)Energy release (MW-sec)Prompt burst I sec 5 see.10 see Peak adiabatic fuel teamp, TA (C)(Prompt burst)Peak fuel temp (prompt burst), (C)(with heat transfer considered) | 9 Address Fortran Name Description of Data 26 RADEL Constant C. in heat capacity of fuel: C,, = C0 + ' Tf Tindeg C. | ||
Peak temp at -thermocouple., ýTC .-(,C)Average core temperature, T (°C)Prompt burst 1 sec 5sac 7.sec (max)10 sec Parameters Effective delayed neutron fraction, 8 eff.Neutron lifetime, 1, (psec).(total)Prompt negative temperature Coeff, oL, (Ak!*C) | 27 RADCO. Constant y in heat capacity of fuel: units are Mwsec,!degC fuel element 28 RADSP C,, heat capacity of water per fuel element - Mw secPC 30 ELNO Thermal resistance, R, core water to tank water °C/Mw 34 TFUIN Initial average fuel temperature -C | ||
: 35. TELIN Initial average water temperature °C 36 TEMPTM Number of entries in the longer of the 2 temperatures coefficient tables (an entry is a pair of numbers) 37-136 TAB A Reactivity contribuxtion associated with fiuel temperature change. Enter in sequence Tt, Ap,, T2, Ap 2 where T is in °C Ap, = lcx(Tr)dT where a (Tr) is the precomputed fuel t dk temperature coefficient I dr k dTr 137-236 TAB D Reactivity contribution associated with water temperature change. Same format as for fuel coefficient 426 RAMP[ Number of entries in ramp or scram table, which ever is longer (an entry is a pair of numbers) | |||
A-%I A-DL1CO e.S0 | |||
10 I Address 427 Fortran Name TRAMP Description of Data Last time to took up in ramp table - secs 428 END RAM 0.0 Ramp continues after scram starts 1.0 Ramp stops after scram starts 429-628 TABBC Ramp table. Enter in sequence: time, Ap, time, Ap, etc. | |||
time in sees. Provide enough entries that linear interpolation is accurate. | |||
This | 629 TENDSC Last time to look up in scram table - secs 630 POWRAT Power level (Mw) or % of initial power (depending on value of STARUP, location 4) at which to initiate scram 631 TDELAY Time elapsed after power POWIRAT before scram is to begin. Used to simulate mechanical or electrical system delays 632-831 TABB Scram table. Enter in sequence time, Ap, time Ap etc. | ||
$4.76 (0.0348 6k)f 6k.time from beginning fraction of. negative reac-k.of scram scram reactivity tivity inserted ."(ec) inserted from to o 0 0.0.1 "0.025 0.0009 0.2 0.090 0.0631 0.3 0, 15 0.0064 0.4 0.308 0.0107 0.5 0.467 0.0163 0. 6 0.630 0.'0219 0.7 0.771 0.0268 0.8 0.880 0.0306 0.9 0.955 .0.0332 1.0 1.0 -0.0348 The reactivity effect associated with temperature changes in the core water was .the same for these calculations as that derived in Section 7.1.34, The volumetric specific heat for the U- | 909 ENDCOL Number of entries in fuel heat transfer resistance table. | ||
accuracy for the measurements of the pulsing performance. | 910-1109 TABE Fuel heat transfer resistance table. Enter in sequence values of Tf, Rt, Tf, Rt, etc. | ||
Figure 7.1-6 shows the pulse shape and core energy release for a$4.40 pulse in the Sandia ACPR as calculated using the parameters generated in this study. Fig. 7.1-7 shows other important core characteristics | Enter integer I in column I of last data card if other problems follow. | ||
:for the same pulse.For a comparison of flux in the lexperimental cavity, Sandia quotes 1.55 x | A-12? a ICO i'*S Pee 0 | ||
(>10 keV) is 1.55 x 10 7.This compares to the computed value of 1. l. x 10 7./Watt, as shown in Fig. 7.1-5. The flux values given here a"sume an energy release during fission Of 190 MeV/fission. | |||
If the delayed energy from fission is not included, as would be the case during pulsing, the calculated neutron fluence value could increase by as much as 10...7.1-34 8-;r A | 8.0 ATTACHMENT B - SANDIA ACPR CORE PERFORMANCE (PULSE OPERATION) | ||
'ns 1I I I I I io0 | B-1 21C025, Rev.0 | ||
---f1 t O. I I , ,! , .. | |||
(*I) | 7.1.2.5 Core Performance (Pulse Operation). A comparison of the experimental and calculated core performanaee characteristics Is shown in Table 7,1-15 for a $4.40 pulse in the Sandia ACPR. The experimental values Oete obtained through communications with Sandia Corporation reactor operations. One set of parameters used for the calculated performance was also obtained from Sandia and is used by them because they give a calculated pulse shape which very nearly matches their experimental pulse shape. None of the paremeters (except for the neutron lifetime, 1) is experimentally determined - | ||
('C) | as an individual quantity, but their interaction determines the pulse shape I This will. be discusaed in mote detail in Section 7.1.3.3. The negative temp-srature coefficient as actually given In the communication with Sandia.. | ||
",C)300 | was 0.013 to 0.0135 $/6C and when converted using 6 eff 0.0073 this becomes 9.49 to 9.86 x 10-5 Sk/*C. The midpoint value in this range was used as the Sandia parameter. The aef and f/F values used for the Sandia. parameters. | ||
* came from the original design. As discussed in Section 7.1.3.2, the more detailed calculational procedures currently used to determine the power. | |||
peaking factor give somewhat lower values then the methods used for the orig-inn1 Sandia ACPR analysis. The parameters used to calculate the performance characteristics in the last column of Table 7.1-15 were all-.generated as a: | |||
part of this analysis and are documented in the previous section. | |||
The kinetics calculations were done with a simplified version of the BLOOST code. (the general version of the BLOOST code is more fully described | |||
-in Section 71..3)whIch is a space-independent combined heat transfer-kinetics code using six delayed neutron groups and temperature feedback. | |||
The transfer of heat from the fuel to the coolant channe:L water and from the coolant channel water to the pool water was approximated In this simplified model by thermal resistance values: | |||
P where R. Thermal resistance (*C/MW) | |||
S5T Temperature difference, between the average fuel element and the average coolant channel water, or between the average coolant channel water and the pool water P Power in the average element (MW) 7.1-28 9_2 ;11CO;L-..MO | |||
TABLE 7.1-15 SANDIA ACPR PULSING OPERATION PERMORAMNCE CRAMCTIRISTICS | |||
$4.40 Pulse with transient rods scra ind at 1 see 141 Fuel elements + 6 fueled followers T - 25*C | |||
.0 P0 10 watts a a.Calculated Performance Parameter saExperimenta . ParaWSJ:8 EPa4Metesthis aly Peak power, P (14W4) 12,300 14,200 Period, T (msec) 1.31 1.38 1.31 Pulse width (at N12) (msec) 394. 5.0 Energy release (MW-sec) | |||
Prompt burst 76 77 80 I sec 83 88 88 5 see. 92" 10 see 90 90 94 Peak adiabatic fuel teamp, TA (C) 940 (Prompt burst) .. Sa,.: | |||
Peak fuel temp (prompt burst), * (C) 860 (with heat transfer considered) | |||
Peak temp at -thermocouple., ýTC .-(,C) 735 (7 sec) 745 (15 sec) | |||
Average core temperature, T (°C) | |||
Prompt burst 490 .505 1 sec 517 531 5sac 533 547 7.sec (max) 534 548 10 sec 533 547 Parameters Effective delayed neutron fraction, 8 0.0073 0.0073 eff. | |||
Neutron lifetime, 1, (psec) 33 34 32 | |||
.(total) 2.54 2.20 Prompt negative temperature Coeff, oL, (Ak!*C) 9.67 x 10' qee Fig. 7.1-4 (a°3-7oo0'C | |||
- 9. x 10-) | |||
I 4 L 7.1-29 8-3;t1-C SCOAft -j 0 | |||
The thermal resistances used were. | |||
Rt (from fuel to cooling channel) - 2% 105 oC/MW R (from cooling channel to pool) _ 5 x 103 C/MW C | |||
Rt was calculated from measured temperatures given in Section 7.2 and R is a value derived from measurements in the Torrey Pines TRIGA reactor. | |||
The calculated values for peak fuel temperature (with heat transfer) and for peak temperature at the thermocouple were obtained with a two-dimensional heat transfer model more fully described in Section 7.1.3.3. | |||
The reactivity insertion for the kinetics calculations was based on .expel imental rod movement times and rod worths. The fast transient rods are pre-accelerated and, with 120 psi air pressure, the final 15 in. of poison is removed from the core in =0.050 sec. It was assumed that the reactivity. | |||
increase was linear with time over. the 0.050'see rod-removal' time. The Sandia adjustable transient rod is not pre-accelerated,..and with 75 psi air pressure it takes about 0.150 see-for the 15 in. travelof the poison* | |||
rod. It was assumed that the reactivity increase varied as. (time) 2 for the removal of the adjustable transient rod. The three transient rods are timed to finish their withdrawal at the same time; thus, the adjustable rod starts out of the core first. Since the three transient rods are worth' | |||
$4.76 but the maximum pulse is $4.40, about $0,36 of the adjustable tran-sient rod worth isnot used for the maximum pulse. The operating procedure however, is to have all three transient rods fully withdrawn atithe end of a pulse to prevent any flux tilting in the core. To accomplish this, the reactor is pulsed from a sub-critical condition - sub-critical by $0.36 for the maximum pulse. This, In effect, allows a pre-acceleration of the adjustable transient rod, in that about half its removal time is over before the pulse reactivity, above that needed for criticality, begins to be inserted. For the calculated pulses, the reactor was assumed to be at 10 W, and only the time for inserting the reactivity above critical was 7.1-30 4 ;L t C0 ZS1 koq C) | |||
considered. Table 7.1-16 shows the derivation of the time versus reactivity table used for the kinetics calculationsa, TABLE 7.1-16 DERIVATION OF PULSE REACTIVITY INSERTION VERSUS TIME FOR SANDIA ACPR Assumes: Three transient rods, each worth ý1.59 (0.0116 6k) | |||
Pulse size v $4.40 (0,0321 6k) | |||
Two fast rods (0.0232 6k) removed in 0.050 sec where 6kai. | |||
One adjustable rod (0.0089 6k) removed in 0.150 sec where 6kw (6k)t2 -t t to time t (6k) t o (6k) total for 6k 2 to time t for t to 6k above to to time t above critical (see) (sec). 2 total 6k critical time t above critical (see) 0 0 0 0 0 0 0 0.01 0.0001 0.00005 0 0 0 0 0.02 0.0004 0,00021 0 0 0 0 0.03 0.0009 0.00046 0 0 0 0 0.04 0.0016 0.40082 0 0 0. 0 0.05 0.0025 0.00129 0 0- 0 0 0.06' 0.0036 0.00186 0 "0 0 0 | |||
: 0. 07 0.00049 0.00253. 0 0 0 0 0.0-08 0..0064 0.00330 0.0006 0 0.0006 0.01 0.09 0.0081 0.00418 0.0015 0 0.0015 '0.02 0.10 0.0100. 0.00516 0.0025 0 0.0025 0.03 0.11 0.0121 0,00624 0.0035 0.00464 0.001. 0.04 0.12 0.0144 0.00742 0.0047 0.00928 0.0140 0.05 0.13 .0.0169 0.00871 0.0060. 0.01392 0.0199 0.06 0.14 0.0196 0.01010 0.0074 0.01856 0.0260 0.07 0.15 0.0.225 0.01160 0.0089 0.02320 0.0321 0.08 Standard operation for the Sandia ACPR is to scram the transient rods 1 sec after the time of full withdrawal. This practice gives a well-defined pulse dose by clipping the long-delayed tail energy from the pulse | |||
: 7. la-31 Is-s" I-sO~ | |||
7.A3ACo~ ' t-i | |||
and shortens the cooling time for the fuel eleMents allowing further.pul$ing in a shorter time. The safety of the reactor is in -no way. dependent upon the reactor scramming after a pulse. The reactivity scram table used in the kinetics calculations is given in Table 7.1-17. The values for the fractions of inserted reactivity are derived in Section 7.1.3.3. A total drop time of 1 sec is used, consistant with measured values. | |||
TABLE 7.1-17 REACTIVITY VERSUS TIME FOR TRANSIENT ROD SCRAM | |||
:FOLLOWING A PULSE IN THE SANDIA ACPR Assumes: Total worth of rods to be scrammed- $4.76 (0.0348 6k) f 6k. | |||
time from beginning fraction of. negative reac-k. | |||
of scram scram reactivity tivity inserted . | |||
"(ec) inserted from to o 0 0 | |||
.0.1 "0.025 0.0009 0.2 0.090 0.0631 0.3 0, 15 0.0064 0.4 0.308 0.0107 0.5 0.467 0.0163 0.6 0.630 0.'0219 0.7 0.771 0.0268 0.8 0.880 0.0306 0.9 0.955 .0.0332 1.0 1.0 -0.0348 The reactivity effect associated with temperature changes in the core water was .the same for these calculations as that derived in Section 7.1.34, The volumetric specific heat for the U-ZrH1 60 fuel elements is: | |||
C = 749 + 1.52 T W-sec/PC element (from OC) p This value is also derived in Section 7.1.3.3. | |||
: 7. l' 32 S-(,. q I C_01.s' kko 0 | |||
The following relationships are given to show qualitatively how the pulsing performance of the Sandia ACPR is influenced by the parameters already mentioned; T-- reactor period 6k p. | |||
26k 2C~k" | |||
* P-- total energy release in the prompt burst P UO peak power where: ' . prompt neutron lifetime c - prompt negative temperature coefficient C = total heat capacity of the core available to the. prompt burst energy release 5S = change in average core temperature produced bythe prompt burst 6k that portion of the step reactivity insertion which is p | |||
above prompt critical The equations given result from the Fuchs-Nordheim space-independent kinetics model in which heat transfer and delayed neutrons are neglected and the core heat capacity and temperature coefficient are invarient with temperature The applicability of this relatively simple model has been demonstrated by its performance predictions for pulses generated in ,TRICGA reactors (Ref. 7,1-7.1-ý33. 6-7-,O* eO | |||
With reference to the comparisons in Table 7.1-15, the calculated performance, with parameters from this analysist is genetally within about 5% of the experimental values, with the calculated values always being larger and with the calcul&ted peak power about 20% larger than the experi-mental value. The manner 'in which the parameters are interrelated in the Fuchs-Nordheim model would indicate that no single calculated value is greatly different from the real value, They can all be at varying, rela-tively small differences from the real values, but the combination of errors for the quantities as they are related to influence the pulsing petforMance is within a reasonable. accuracy for the measurements of the pulsing performance. | |||
Figure 7.1-6 shows the pulse shape and core energy release for a | |||
$4.40 pulse in the Sandia ACPR as calculated using the parameters generated in this study. Fig. 7.1-7 shows other important core characteristics :for the same pulse. | |||
For a comparison of flux in the lexperimental cavity, Sandia quotes 1.55 x 10 15 nvt (>10 keV) for a $4.40 pulse. Using a 100 MW-sec energy release, to include energy .beyond 10 sec which could be measured by-the detectors, the flux/Watt (>10 keV) is 1.55 x 10 7 . This compares to the computed value of 1.l. x 10 7 ./Watt, as shown in Fig. 7.1-5. The flux values given here a"sume an energy release during fission Of 190 MeV/fission. If the delayed energy from fission is not included, as would be the case during pulsing, the calculated neutron fluence value could increase by as much as 10... | |||
7.1-34 8-;r A&iO;S-C IkvO | |||
*n5 'ns 1I I I I I to! | |||
io0 * " " -0 | |||
,1o 102 ENERGY tO2 energy released versus------ - -- | |||
f1 t O. I I , ,!I , .. . I o 0 0.05 0.10 0.15 . 0,20 0;25 0,30 0.35 BEGINNING OF ROD,0 7IO01O FROII TItlE (SEC) | |||
BL-0619 Fig. 7.1-6, Sand+/-a ACPR standard pugse .p~er ane energy reieaeed versu~s time from first rod mot~ion 7.1-35 8-1 7.1-35. | |||
aco:19 89 eeO | |||
N a 10 - - | |||
5 MAXIMUM tHEAT FIUX (W/CM7) 500 x ".- | |||
' | |||
-0 FUEl. TEMPERATURE (*I) | |||
MAXIMUM 800 TUEL s.D. TEMPERATURE ('C) 700 THERMOCUPLE TEMPAERATURE (C) 5800 600 ac 500 FUELCENTERLINE TEMPERATURE ",C) 300 O 20"CA TEPRTUE(C wfoe TiEFO IG NI1G TRO OO SC I0 o-2 IO-t * , oo iol o | |||
*1 TIME FROM BEGINNING OF ROD It0TlON (SEe) . | |||
0 EL-0620 Fig. 7.1-7. Sandia ACPR t!eperaslture and heat 'flux time verus t4-& f[rom first r"* motion}} |
Revision as of 14:40, 14 November 2019
ML082380268 | |
Person / Time | |
---|---|
Site: | Washington State University |
Issue date: | 08/06/2008 |
From: | Wall D Washington State Univ |
To: | Document Control Desk, Office of Nuclear Reactor Regulation |
References | |
21C025, Rev 0 | |
Download: ML082380268 (34) | |
Text
WASHINGTON STATE
%;UNIVERSITY Nuclear Radiation Center August 6, 2008 United States Nuclear Regulatory Commission Document Control Desk Washington D.C. 20555-0001 Docket Number 50-27 Facility License Number R-76 This purpose of this letter is to transmit a communication from General Atomics declaring that a document submitted by Washington State University, "BLOOST Code Validation Report" on June 13, 2008, may be made available for public release.
A copy of the General Atomics letter authorizing release of the BLOOST Code Validation Report is included with this letter.
I declare under penalty of perjury that the foregoing is true to the best of my knowledge.
Respectfully Submitted Donald Wall, Ph.D.
Director Nuclear Radiation Center Washington State University P.O. Box641 300, Pullman, WA 99164-1 300 509-335-8641 - Fax: 509-335-4433 * *ww.wsu.edu/nrc
Q GENERAL ATOMICS From: Ken Mushinski Refer To: kjm08:018 Project: WSU To: Don Wall, Washington State University Date: 5 Aug. 2008
SUBJECT:
Bloost Code Validation Report, 21C025 Revision 0 Please use this memo as official notification that GeneralAtomics approves the unrestricted release of General Atomics document 21C025 Rev. 0, Bloost Code Validation Report, regardless of the proprietary denotation on the Issue/Release Summary page contained therein.
1 of I Kjm08-018 Bloost Validation Report
21C025 Revision 0 ONGKHARAK NUCLEAR RESEARCH CENTER PROJECT BLOOST Code Validation Report Prepared for the Government of Thailand Ministry of Science, Technology and Environment Office of Atomic Energy for Peace Under Contract No. 56/2540 April, 2000
+ GENERAL ATOMICS, Marubeni HITACHI Qnsto zdfnT +TRIVA TECH!
ONGHARAK NUCLEAR RESEARCH CENTER PROJECT BLOOST Code Validation Report Prepared for the Government of Thailand Ministry of Science, Technology and Environment Office of Atomic Energy for Peace Under Contract No. 56/2540 April, 2000 Prepared By TRIGA Technologies Under Contract to General Atomics
+. TRIGA TECH ISSUE/RELEASE
SUMMARY
T-1495 (REV. 0/99)
SAFE.CLS QA LVL/CLS DOC. TYPE PROJECT DOCUMENT NO. REV.
EJ R&D APPVL LI DV&S LEVEL SR-A IIA R 2641 2641 21 C025 21 C025 00 0 DESIGN Q N/A 2 SYSTEM SEIS. CLS. SEIS. CAT. E. CLASS 21 N/A N/A N/A II TITLE:
BLOOST Code Validation Report CM APPROVAL(S) REVISION APPROVAL/ REV. PREPARED DESCRIPTION DATE BY ENGINEERING QA -. PROJECT IRELEASED-I 1 R. Sherman R. Sherman K. Maxwell AR ca Initial Release 74.?,0VO gA~- ,qjqeo1 Independent Director of Elor: E gi*'rig USE CONTINUATION SHEET NEXT INDENTURED DOCUMENT (S)
- See List of Effective Pages COMPUTER PROGRAM PINs PG000258 APPROVAL BY GENERAL ATOMICS :ý N/A APPROVAL BY CUSTOMER N/A Z GA PROPRIETARY INFORMATION THIS DOCUMENT IS THE PROPERTY OF GENERAL ATOMICS. ANY TRANSMITTAL OF THIS DOCUMENT OUTSIDE OF GA WILL BE IN CONFIDENCE. EXCEPT WITH THE WRITTEN CONSENT OF GA, (I) THIS DOCUMENT MAY NOT BE COPIED IN WHOLE OR IN PART AND WILL BE RETURNED UPON REQUEST OR WHEN NO LONGER NEEDED BY RECIPIENT AND (2) INFORMATION CONTAINED HEREIN MAY NOT BE COMMUNICATED TO OTHERS AND MAY BE USED BY RECIPIENT ONLY FOR THE PURPOSE FOR WHICH IT WAS TRANSMITTED.
L] NO GA PROPRIETARY INFORMATION PAGE iOF*
LIST OF EFFECTIVE PAGES Page Number Page Count Revision Issue/Release Summary 5 0 List of Effective Pages 1 0 Table of Contents 1 0 Pages 1-5 5 0 Attachment A 12 0 Attachment B 10 0 Total Pages 30 ii 21C025, Rev.O
/
TABLE OF CONTENTS 1.0 CODE IDENTIFICATION ............................................................................. 1 2.0 PROGRAM DESCRIPTION .................................. 1 3.0 VALIDATION TESTS ................................................................................... 2
4.0 CONCLUSION
S ............................................................................................ 4
5.0 REFERENCES
............................................................................................. 4 6.0 INDEPENDENT EVALUATION ................................................................. 5 7.0 ATTACHMENT A-BLOOST (TRIGA VERSION) USERS MANUAL ...A-1 8.0 ATTACHMENT B - SANDIA ACPR CORE PERFORMANCE (PULSE OPERATION) ............................................................................................ B-1 iii 21C025, Rev.0
1.0 CODE IDENTIFICATION Program ID No. (PIN) - PG000258 Program Name - The BLOOST Code (TRIGA Version), referred to as TRIGA-BLOOST Computer Platform - Dec Alpha with UNIX operating system Executable Location - /gaprod/decalpha/tbloost/bin/tbloost3 Sample Test Problems - /gaprod/decalpha/tbloost/examples 2.0 PROGRAM DESCRIPTION BLOOST is a combined reactor kinetics-heat transfer computer code developed by General Atomics. A special version, utilizing a space-independent heat transfer model, was developed for pulsing TRIGA reactor design work. A short description of the TRIGA-BLOOST code's purpose and capabilities is given in the user manual:
"The kinetics model employs the space-independent form of the reactor kinetics equations with very general forms of reactivity input and feedback.
The first basic assumption of the method is that the shape of the neutron flux does not vary with time regardless of how large the change in amplitude may be. While this criterion is rarely satisfied in an exact sense, it is reasonably adequate for a large number of practical cases, including the small cores used in TRIGA reactors.
The heat transfer calculation uses simple lumped parameter models to compute the average temperatures in the UZrH fuel material and the water coolant. These average temperatures are used in combination with temperature coefficient data appropriate to the whole core to determine the reactivity feedback terms in the kinetics equations. The second basic assumption of the method is thus that a single average fuel element can be used to determine reactivity feedback from temperature changes occurring over the entire core."
With respect to TRIGA reactor core design, TRIGA-BLOOST is used for calculating the maximum fuel temperature in the core as a function of time reached during a fast transient (similar to a short power pulse). These types of transients are generated by rapid positive reactivity insertions into the core under cold startup or full power operating conditions, followed very quickly by a reactor control rod system scram. Please see the User's Manual, 1 21C025, Rev.0
included in this document as Attachment. A, for more detailed information on the solution algorithms used, as well as the input instructions.
3.0 VALIDATION TESTS Many of the TRIGA reactors built by General Atomics were designed to operate safely in a pulsing mode of operation. With such designs, it was imperative that an accurate and reliable means of calculating core performance during rapid transients (power pulses), be available. The TRIGA-BLOOST code was developed for just that purpose. TRIGA-BLOOST has been used at General Atomics for over thirty years to calculate maximum fuel temperatures and total energy release during rapid reactivity transients. The code has been applied to both pulsing and non-pulsing TRIGA reactors.
A useful benchmark problem for TRIGA-BLOOST has been documented (Ref.1) as part of the safety analysis report for the TRIGA Annular Core Pulse Reactor (ACPR). A comparison is made of the measured versus calculated core performance characteristics of the Sandia Laboratories ACPR for a transient pulse of $4.40 reactivity insertion. Conditions at the onset of the $4.40 transient were as follows:
- To =25C SPo 10watts
- 141 Fuel elements + 6 fueled followers
- $4.40 pulse with transient rods fully scrammed after I second Using this basic data, as well as important calculated input core parameters such as the effective delayed neutron fraction, P3 lf, the prompt neutron lifetime, e, and the prompt negative temperature coefficient, a, TRIGA-BLOOST can be used to model core performance during the transient. The results, taken from Table 7.1-15 of Ref.1, are summarized in Table 3.1.
2 21C025, Rev.0
Table 3.1 Sandia ACPR Pulsing Operation, TRIGA-BLOOST versus Experimental GA Calculated Performance Parameter Sandia Experimental Sandia GA Parameters Parameters Peak Power (MW) 12,100 12,500 14,200 Period, T (msec) 1.31 1.38 1.31 Energy Release (MW-sec)
Prompt Burst 76 77 80 I sec - 83 86 5 sec 88 88 92 10 sec 90 90 94 Peak Fuel Temp (*C) - - 860 Peak Temp at Thermocouple (*C) 735 (7 sec) 745 (15 sec)
Parameters Effective Delayed Neutron 0.0073 0.0073 Fraction, Peff Neutron Lifetime, £ (psec) 33 34 32 Power Peak/Avg. (total) 2.54 2.20 Prompt Negative Temp 9.67x 10-' 9.6x 10-'
Coefficient, ct (8k/'C) 3 21C025, Rev.0
Note the very excellent agreement between the measured performance parameters and those calculated with TRIGA-BLOOST. For TRIGA reactor calculations, the experience at General Atomics has been that the TRIGA-BLOOST code generates results that are somewhat conservative (slightly higher peak powers, energy release, and maximum fuel temperatures) relative to experimental values. The results shown in Table 3.1 confirm this observation. The entire section from Ref.1 concerning the Sandia ACPR transient is included in this document as Attachment B.
This benchmark case shows that TRIGA-BLOOST provides accurate core performance results for TRIGA reactors undergoing rapid reactivity insertion transients. It also shows the adequacy of the GA methods for generating important input to TRIGA-BLOOST, such as i3eff, £, and a.
4.0 CONCLUSION
S Based on its history of successful usage in TRIGA reactor core design and its successful application to an experimental (Sandia ACPR) benchmark (i.e., validation test methods #4 and #3 in Ref.2, respectively), the TRIGA-BLOOST code can be considered validated for use on the ONRC project.
5.0 REFERENCES
- 1. "TRIGA Annular Core Pulse'Reactor Safety Analyses and Preliminary Design for the Nuclear Safety Research Reactor," TRIGA Reactors, Gulf Energy and Environmental Systems, E- 117-174, July 1972.
- 2. "Control of Scientific and Engineering Computer Programs," TRIGA Technologies Engineering Procedure EP-3.7, latest revision 4 21C025, Rev.0
6.0 INDEPENDENT EVALUATION This Validation Report has been reviewed and evaluated and it is concluded that the validation procedures met the requirements of Ref.2. The sample test problems were demonstrated to be available, and the validation results were adequately evaluated and documented.
The evaluator is very familiar with the use of the code for most all past TRIGA design calculations and the comparisons with operating reactor results.
This evaluation consisted of reviewing the benchmark calculations and results, as well as their comparison with operating reactor results. Input was reviewed by sampling random values, checking volumes, and engaging in question and answer sessions with the nuclear designer.
The validation effort was direct and straightforward; comparing calculated and experimentally determined results for a reactor in a pulsing (kinetic) mode. The comparisons give an adequate demonstration of the ability of the code to predict reactor performance and conditions during rapid transient operation.
5 21C025, Rev.0
7.0 ATTACHMENT A - BLOOST (TRIGA VERSION) USERS MANUAL A-1 21G025, Rev.0
THE BLOOST CODE (TRIGA Version).
September 16, 1997 A-2 ; C*o*" R, o
I BLOOST is a combined reactor kinetics-heat transfer code written in the Fortran language for the VAX computer. The kinetics model employs the space-independent form of the reactor kinetics equations with very general forms of reactivity input and feedback. The first basic assumption of the method is that the shape of the neutron flux does not vary with time regardless of bow large'the change in amplitude may be. While this criteria is rarely satisfied in an exact sense, it is reasonably adequate for a large number of practical cases, including the small cores used in TRIGA reactors.
The heat transfer calculation uses simple lumped parameter models to compute the average temperatures in the UZrH fuel material and the water coolant These average temperatures are used in combination with temperature coefficient data appropriate to the whole core to determine the reactivity feedback terms in the kinetics equations. The second basic assumption of the method is thus that a single average fuel element can be used to determine reactivity feedback from temperature changes occurring over the entire. core.
The form of the reactor kinetic equations used is given below:
dPs
_ p-p+
.Q+(1.)
dt A dt* -zP-AQ (2.)
p= RAMP (time) - SCRAM (time) - TEMPT (3.)
P= reactor power level -megawatts Ai effective delayed neutron yield for group i.
6 groups are assumed.
i= delayed neutron precursor decay constant for group i -
sec A.= prompt neutron generation time - secs Ci = delayed neutron precursor concentration of group i-megawatts S = constant source term megawatts/sec A-3 aACO3 S7 P~ev 0
2 RAMP (time) and SCRAM (time) are reactivity changes measured from time zero which are read from user-supplied input tables at each time step using linear interpolation. The use of the scram table can be initiated after a specified power level or percentage of full power has been reached in the tranrient. A delay time can be specified to simulate electrical or mechanical delays in a control system. The RAMP function may be continued or terminated after the SCRAM starts depending on the value of an input number.
The units of S and Ci come about when the kinetics equations, which are usually derived in terrns of the neutron density n, are written in terms of the power level P. A constant proportiotiality between n and P is thus assumed which is not strictly correct since they are coupled by the fission cross section t
P(Mw) = 3xlIfission/ Mw-sec - 3x10 t s (4.)
Thus, the source in neutrons/sec-cm3 and the delayed neutron precursor concentrations Ci in cm3 are multiplied by the factor Ef nvV,.
3X10" which is constant with time to the extent that ,r varies as IN. The validity of this assumption is
..olygrea.r. in thrmal, spectrum mea~ctors. S.tawrp. typc. prmb.crus. in .which _IO temtpeatmre.
feedback occurs can, of course, be run in units of n rather than P by supplying consistent values for the input quantities P and S.
The solution of the set of kinetic equations (1) and (2) is done by the method of fourth-order Runge-Kutta integration described by Cohen (1) as modified by Schwartz (2) to include a variable time step. This is done in the subroutines RKCOT, DERIV and COH In the following description, the nomenclature follows the Fortran symbols as far as possible.
The set of equations is rewritten using Y for the variable and YP for the time derivative.
YP(1) = A(I) Y(1) + R(1)
For the power equation Y() =P, A(l) and for the precursor equations Y(I+) =Ci and A(l+l) = -i. The remainder R(I) is always computed as YP(I) -A(I)Y(I).
- 1. Cohen, E.R., "Some Topics in Reactor Kinetics," p/629,2nd Geneva Conference, Vol. II, p.306 (1958).
- 2. Schwartz, A., "Generalized Reactor Kinetics Code - AIREK-il," NAA-SR-Memo 4980, (1960)
3 The sequence of operations is as follows for each equation of the set. A time step H is used, the first value for which is an input number. Subsequent values are chosen by the code. The initial values of Y, YP, A and R are calculated from the input data. The first step of the integrationcomputes Y(t+W2) by assuming R to be constant at its initial value (called Ro). This gives afirst approximation for Y(t+Ht2) from which R(t+H/2) = R, can be computed. Assuming that the variation of R is linear, a better Y(tt-2) is computed and R(t+H/2) is further improved (called R2). A linear extmplation of R is now used to compute Y(t+f) and R(t+H) = R, for the completetim step. A quadratic fit is now made with Ro, R 2 and R3 to determine an improved
-4 and from this the final Y(t+H) is calculated.
The equations for this procedure are given below. The symbol A refers to A(l).
YO = Y(t), RO = R(YO, t), YPO = YP(t)
H H Yt -Yo - Ct(A--)[RO+AYa]
R 1 -=R(Y,,t-+H/ 2)=YP(t+TH/2)-AY, H H Y2 = Y1 + - C2 (Aj -)[Rl -Re]
R 2 =R(Y2 ,t+H/2)
Y= Yo +HC(AH)[RO +AYO]+2HC 2 (AH)[R 2 -IR,]
R3 = R(Y 3 ,t + H)
Y= Y3 +H{2C 3 (AH) -C,(AH))[R 2 Ra+ R 3]
0 R4 = R(Y4 ,t + H)
Ys=Y 4 +H(2C 3 (AH)-C 2 (AH)}[R 4 R3 ]
In these equations, the functions C(x) are integrating factors defined as follows:
For xi < 1.0 CA(x) = + 1) + 1) + 1) + 1) + 1) + 1) +l)1
+1 C2(x) - xC,(x) 2 a O
4 c,( - xC2 (X)+ I 2
For 14> 1.0 C,(x) - e' - 1 c3(x)- 2c,(x)-lt K
The argument x is either A H or A M/2/. The C(x) functions are computed in the function subroutine COH(Nx) where N is the subscript of the function (1,2 or 3). The function computed by COH for N = 3 is actually 2C 3(x) - C2(x), the term occurring in the equations for Y4 and Y.
above.
Since the C 1(x) function involves an exponential for large x, the argument x cannot be allowed to exceed 88. Since A(l)=(p-13)/A, this implies a fundamental time step limit of:
B8A H<
This can be a significant limit for fast reactor problems where A may be of the order of 10. secs.
In some cases when the net reactivity is known to be well below prompt critical throughout the transient, the kinetic behavior is governed by the delayed neutron precursors. and it may be possible to use a fictitious large A value to get around the time step limit Throughout the solution, the subroutine DERIV is called five times to obtain values of the derivatives YP at times (t + H/2) and (t + H). The reactivity components used in Eq. (3) for computing p are read from the ramp, scram, and temperature coefficient tables at these time points. The scram can only be initiated at the end of a kinetics time step (t + H). Once it has been initiated, values are obtained at (t + F/2) and (t + H) in subsequent time steps. The ramp table is read at all time points in the integration. When the time step H is completed, a check is made on the accuracy of the solution. If satisfactory, the same size step is used again. If the accuracy is very good, the step is doubled, or, if poor, the time step is halved and the step ro.cmpute4. Doubling is not permitted if A H would then exceed 89. The accuracy criteria is a number Q given by the equation.
HCz(AH)f.-2o8 +o,,I Q=- I + C,(AH)
5 The s'sare inverse periods defined by M=YPO 0 I I(Y(t + R)
SyoYO YP(t+H) o -Y(t+H)
The criteria is evaluated only for the first equation of the set so that Y is the reactor power level.
Thus eo is the instantaneous inverse period at the beginning of the step, c), is the instantaneous inverse period at the end of the step, and ota is an average inverse period over the step. The inverse of cn is printed in the problem output as the reactor period.
The value of Q obtained at the end of a time step is compared with two input numbers QI and Q2.
An additional test quantity is called DI and is the first increment of Y calculated in starting the timestop (i.e., Y, -Yo).
Dl= H HtA!
D2 )R + AY-J The use of the Q and DI criteria is as follows:
If Dl <2x0-1 4 YO and Q >:QI keep same H for next step.
If Dl < 2xl 0 "4yo and Q < Q 1 double H for next step.
IfDl >2xl0-xY 0 and Q<QI double H for next step.
If Dl> 2xl0- 4 Y and Q>Qi and/or Q< Q2 keep same H for next step.
If DI>2x10-4 Y0 and Q_>Q2 halve H and recompute the time step, unless the new reduced H is smaller than the input number SMALLH. In that case the problem is stopped.
A-7 atco*r"Ize- 0
6 Aflow diagram of the kinetics solution is shown in Fig. 1.
Values of QI and Q2 to use are best determined by excperience. They should be about a factor of 10 apart. Values of 10-s, 104 are typical for many TRIGA problems. The user should do some numerical experimentation for the specific problems to be solved.
The time variations of the average fuel element temperature, T&and the average coolant temperature, Tl, are expressed in simple lumped parameter equations which have the same form as the 7 kinetic equations and therefore can be solved along with them by the procedures described above.
If heat is generated by the total fission power P in all of the fuel and removed by convection heat transfer to the coolant, a simple heat balance gives:
d P- -T. (5.)
dt C1 where Cp is the heat capacity of all fuel material. Since the number of fuel elements frequently varies in TRIGA cores, the user supplies the heat capacity per fuel element and the number of elements in the input. The heat capacity may have a linear diependence on the fuel temperature, so that Cý=C0 +yT~
where the user supplies Co and y. The units of C. are Mw-sec/*C.
The product of the heat transfer coefficient and the area, hA, is represented by a thermal 1
resistance, R, - hA ' which is supplied by another input table as a function of fuel temperature.
It is typically derived from more sophisticated heat transfer calculations for the core at various constant power levels so that T -T.
Power per element The coolant flowing in the core channels is assumed to transfer its heat to the water in the reactor tank which is assumed to be a constant temperature sink that remains at the initial coolant temperature. This is a suitable approximation for the short time scale of transients typically studied with BLOOST.
A-? -1kc-OAr Qtv 0
7 The equation used is dT.
dt Cp.
,where C., and R, are, respectively, the heat capacity of the water in the coolant channel (per ful element) and a suitable resistance derived from steady state conditions and assumed constant.
The average fuel and coolant temperatures Tr and Tw are used with input tables of pre-computed reactivity feed-back contributions from the fuel and coolant temperature coefficients.
These contributions are summed to give the term TEMPT in equation 3. The input data for the tables consist of pairs of temperature values (deg c) and reactivity change (A&p). The code automatically normalizes the table supplied by the user so that the reactivity change is zero for the-initial value of the temperature.
While the average fuel element temperature determines the reactivity feed-back, the quantity of most interest in a transient is usually the maximum fuel temperature, which must not reach a level that would cause damage, The user can supply a pre-computed value for the ratio of maximum to average power density in the hottest fuel element (P/ P). BLOOST uses this in an equation similar to Eq.5 to compute the maximum fuel temperature throughout the transient The power generated during the transient is integrated over time as the calculation progresses and printed at each time step in the output as energy in Mw sec.
BLOOST was a fixed address format for the input data. Each card image starts with the address of the first data entry and the number of entries on the card. Data not needed can be omitted.
Problems can be stacked. The last card of a problem has the integer 1 in column 1.
The algebraic signs of reactivity terms to be supplied in the various input tables will be positive for the most common situations; i.e.,,negative temperature coefficients, positive ramp insertions and negative scrams following the sense of Eq.3.
The input card image format is as follows: (II, [5,11, 7E9.4)
Column 1 Reserved for integer I on last card of set Columns 2-6 Address of Ist variable on card Column 7 No. of words to read on this card A-9 :t VC 0.ý S7 4tv 0
8 Columns 8-70 Up to 7 data entries A title card is supplied at the beginning of each case (Format 12A6).
The input data addresses and descriptions are given in the following pages.
Address Fortran Name Description of Data 1~
Title Card (alphanumeric) 1 PB Initial Power Level, P, - Mwt 2 SOURCE Source term if present. S - Mwt/sec 3 TEND Time to stop problem - sees 4- STARUP If STARUP = 1.0, scram level is set at power given in location 630. If STARUP = 0.0, the value in location 630 will be the percentage of initial power PB at which to scram.
5 GENTIM Prompt neutron generation time, A -secs 6-11 ALAM () Delayed neutron precursor constants,
- i=1,6-secs-1 12-17 BETA (I) Delayed neutron yields; P3i i = 1,6 18 H Initial time step to use in Runge-Kutta integration - sees 19 SMALL H Minimum time step to use in Runge-Kutta integration -
I se*s 20 Ql Runge-Kutta accuracy criteria (suggest 10")
21 Q2 Runge-Kutta accuracy criteria (suggest 10 -)
A 22 DELTMX Peak/avg. power for hottest fiLel element in core (P/1) 23 XMTIME Number of fuel elements in the core Aq-1o A-t:0~c~z5S ~t 0
9 Address Fortran Name Description of Data 26 RADEL Constant C. in heat capacity of fuel: C,, = C0 + ' Tf Tindeg C.
27 RADCO. Constant y in heat capacity of fuel: units are Mwsec,!degC fuel element 28 RADSP C,, heat capacity of water per fuel element - Mw secPC 30 ELNO Thermal resistance, R, core water to tank water °C/Mw 34 TFUIN Initial average fuel temperature -C
- 35. TELIN Initial average water temperature °C 36 TEMPTM Number of entries in the longer of the 2 temperatures coefficient tables (an entry is a pair of numbers)37-136 TAB A Reactivity contribuxtion associated with fiuel temperature change. Enter in sequence Tt, Ap,, T2, Ap 2 where T is in °C Ap, = lcx(Tr)dT where a (Tr) is the precomputed fuel t dk temperature coefficient I dr k dTr 137-236 TAB D Reactivity contribution associated with water temperature change. Same format as for fuel coefficient 426 RAMP[ Number of entries in ramp or scram table, which ever is longer (an entry is a pair of numbers)
A-%I A-DL1CO e.S0
10 I Address 427 Fortran Name TRAMP Description of Data Last time to took up in ramp table - secs 428 END RAM 0.0 Ramp continues after scram starts 1.0 Ramp stops after scram starts 429-628 TABBC Ramp table. Enter in sequence: time, Ap, time, Ap, etc.
time in sees. Provide enough entries that linear interpolation is accurate.
629 TENDSC Last time to look up in scram table - secs 630 POWRAT Power level (Mw) or % of initial power (depending on value of STARUP, location 4) at which to initiate scram 631 TDELAY Time elapsed after power POWIRAT before scram is to begin. Used to simulate mechanical or electrical system delays 632-831 TABB Scram table. Enter in sequence time, Ap, time Ap etc.
909 ENDCOL Number of entries in fuel heat transfer resistance table.
910-1109 TABE Fuel heat transfer resistance table. Enter in sequence values of Tf, Rt, Tf, Rt, etc.
Enter integer I in column I of last data card if other problems follow.
A-12? a ICO i'*S Pee 0
8.0 ATTACHMENT B - SANDIA ACPR CORE PERFORMANCE (PULSE OPERATION)
B-1 21C025, Rev.0
7.1.2.5 Core Performance (Pulse Operation). A comparison of the experimental and calculated core performanaee characteristics Is shown in Table 7,1-15 for a $4.40 pulse in the Sandia ACPR. The experimental values Oete obtained through communications with Sandia Corporation reactor operations. One set of parameters used for the calculated performance was also obtained from Sandia and is used by them because they give a calculated pulse shape which very nearly matches their experimental pulse shape. None of the paremeters (except for the neutron lifetime, 1) is experimentally determined -
as an individual quantity, but their interaction determines the pulse shape I This will. be discusaed in mote detail in Section 7.1.3.3. The negative temp-srature coefficient as actually given In the communication with Sandia..
was 0.013 to 0.0135 $/6C and when converted using 6 eff 0.0073 this becomes 9.49 to 9.86 x 10-5 Sk/*C. The midpoint value in this range was used as the Sandia parameter. The aef and f/F values used for the Sandia. parameters.
- came from the original design. As discussed in Section 7.1.3.2, the more detailed calculational procedures currently used to determine the power.
peaking factor give somewhat lower values then the methods used for the orig-inn1 Sandia ACPR analysis. The parameters used to calculate the performance characteristics in the last column of Table 7.1-15 were all-.generated as a:
part of this analysis and are documented in the previous section.
The kinetics calculations were done with a simplified version of the BLOOST code. (the general version of the BLOOST code is more fully described
-in Section 71..3)whIch is a space-independent combined heat transfer-kinetics code using six delayed neutron groups and temperature feedback.
The transfer of heat from the fuel to the coolant channe:L water and from the coolant channel water to the pool water was approximated In this simplified model by thermal resistance values:
P where R. Thermal resistance (*C/MW)
S5T Temperature difference, between the average fuel element and the average coolant channel water, or between the average coolant channel water and the pool water P Power in the average element (MW) 7.1-28 9_2 ;11CO;L-..MO
TABLE 7.1-15 SANDIA ACPR PULSING OPERATION PERMORAMNCE CRAMCTIRISTICS
$4.40 Pulse with transient rods scra ind at 1 see 141 Fuel elements + 6 fueled followers T - 25*C
.0 P0 10 watts a a.Calculated Performance Parameter saExperimenta . ParaWSJ:8 EPa4Metesthis aly Peak power, P (14W4) 12,300 14,200 Period, T (msec) 1.31 1.38 1.31 Pulse width (at N12) (msec) 394. 5.0 Energy release (MW-sec)
Prompt burst 76 77 80 I sec 83 88 88 5 see. 92" 10 see 90 90 94 Peak adiabatic fuel teamp, TA (C) 940 (Prompt burst) .. Sa,.:
Peak fuel temp (prompt burst), * (C) 860 (with heat transfer considered)
Peak temp at -thermocouple., ýTC .-(,C) 735 (7 sec) 745 (15 sec)
Average core temperature, T (°C)
Prompt burst 490 .505 1 sec 517 531 5sac 533 547 7.sec (max) 534 548 10 sec 533 547 Parameters Effective delayed neutron fraction, 8 0.0073 0.0073 eff.
Neutron lifetime, 1, (psec) 33 34 32
.(total) 2.54 2.20 Prompt negative temperature Coeff, oL, (Ak!*C) 9.67 x 10' qee Fig. 7.1-4 (a°3-7oo0'C
- 9. x 10-)
I 4 L 7.1-29 8-3;t1-C SCOAft -j 0
The thermal resistances used were.
Rt (from fuel to cooling channel) - 2% 105 oC/MW R (from cooling channel to pool) _ 5 x 103 C/MW C
Rt was calculated from measured temperatures given in Section 7.2 and R is a value derived from measurements in the Torrey Pines TRIGA reactor.
The calculated values for peak fuel temperature (with heat transfer) and for peak temperature at the thermocouple were obtained with a two-dimensional heat transfer model more fully described in Section 7.1.3.3.
The reactivity insertion for the kinetics calculations was based on .expel imental rod movement times and rod worths. The fast transient rods are pre-accelerated and, with 120 psi air pressure, the final 15 in. of poison is removed from the core in =0.050 sec. It was assumed that the reactivity.
increase was linear with time over. the 0.050'see rod-removal' time. The Sandia adjustable transient rod is not pre-accelerated,..and with 75 psi air pressure it takes about 0.150 see-for the 15 in. travelof the poison*
rod. It was assumed that the reactivity increase varied as. (time) 2 for the removal of the adjustable transient rod. The three transient rods are timed to finish their withdrawal at the same time; thus, the adjustable rod starts out of the core first. Since the three transient rods are worth'
$4.76 but the maximum pulse is $4.40, about $0,36 of the adjustable tran-sient rod worth isnot used for the maximum pulse. The operating procedure however, is to have all three transient rods fully withdrawn atithe end of a pulse to prevent any flux tilting in the core. To accomplish this, the reactor is pulsed from a sub-critical condition - sub-critical by $0.36 for the maximum pulse. This, In effect, allows a pre-acceleration of the adjustable transient rod, in that about half its removal time is over before the pulse reactivity, above that needed for criticality, begins to be inserted. For the calculated pulses, the reactor was assumed to be at 10 W, and only the time for inserting the reactivity above critical was 7.1-30 4 ;L t C0 ZS1 koq C)
considered. Table 7.1-16 shows the derivation of the time versus reactivity table used for the kinetics calculationsa, TABLE 7.1-16 DERIVATION OF PULSE REACTIVITY INSERTION VERSUS TIME FOR SANDIA ACPR Assumes: Three transient rods, each worth ý1.59 (0.0116 6k)
Pulse size v $4.40 (0,0321 6k)
Two fast rods (0.0232 6k) removed in 0.050 sec where 6kai.
One adjustable rod (0.0089 6k) removed in 0.150 sec where 6kw (6k)t2 -t t to time t (6k) t o (6k) total for 6k 2 to time t for t to 6k above to to time t above critical (see) (sec). 2 total 6k critical time t above critical (see) 0 0 0 0 0 0 0 0.01 0.0001 0.00005 0 0 0 0 0.02 0.0004 0,00021 0 0 0 0 0.03 0.0009 0.00046 0 0 0 0 0.04 0.0016 0.40082 0 0 0. 0 0.05 0.0025 0.00129 0 0- 0 0 0.06' 0.0036 0.00186 0 "0 0 0
- 0. 07 0.00049 0.00253. 0 0 0 0 0.0-08 0..0064 0.00330 0.0006 0 0.0006 0.01 0.09 0.0081 0.00418 0.0015 0 0.0015 '0.02 0.10 0.0100. 0.00516 0.0025 0 0.0025 0.03 0.11 0.0121 0,00624 0.0035 0.00464 0.001. 0.04 0.12 0.0144 0.00742 0.0047 0.00928 0.0140 0.05 0.13 .0.0169 0.00871 0.0060. 0.01392 0.0199 0.06 0.14 0.0196 0.01010 0.0074 0.01856 0.0260 0.07 0.15 0.0.225 0.01160 0.0089 0.02320 0.0321 0.08 Standard operation for the Sandia ACPR is to scram the transient rods 1 sec after the time of full withdrawal. This practice gives a well-defined pulse dose by clipping the long-delayed tail energy from the pulse
- 7. la-31 Is-s" I-sO~
7.A3ACo~ ' t-i
and shortens the cooling time for the fuel eleMents allowing further.pul$ing in a shorter time. The safety of the reactor is in -no way. dependent upon the reactor scramming after a pulse. The reactivity scram table used in the kinetics calculations is given in Table 7.1-17. The values for the fractions of inserted reactivity are derived in Section 7.1.3.3. A total drop time of 1 sec is used, consistant with measured values.
TABLE 7.1-17 REACTIVITY VERSUS TIME FOR TRANSIENT ROD SCRAM
- FOLLOWING A PULSE IN THE SANDIA ACPR Assumes: Total worth of rods to be scrammed- $4.76 (0.0348 6k) f 6k.
time from beginning fraction of. negative reac-k.
of scram scram reactivity tivity inserted .
"(ec) inserted from to o 0 0
.0.1 "0.025 0.0009 0.2 0.090 0.0631 0.3 0, 15 0.0064 0.4 0.308 0.0107 0.5 0.467 0.0163 0.6 0.630 0.'0219 0.7 0.771 0.0268 0.8 0.880 0.0306 0.9 0.955 .0.0332 1.0 1.0 -0.0348 The reactivity effect associated with temperature changes in the core water was .the same for these calculations as that derived in Section 7.1.34, The volumetric specific heat for the U-ZrH1 60 fuel elements is:
C = 749 + 1.52 T W-sec/PC element (from OC) p This value is also derived in Section 7.1.3.3.
- 7. l' 32 S-(,. q I C_01.s' kko 0
The following relationships are given to show qualitatively how the pulsing performance of the Sandia ACPR is influenced by the parameters already mentioned; T-- reactor period 6k p.
26k 2C~k"
- P-- total energy release in the prompt burst P UO peak power where: ' . prompt neutron lifetime c - prompt negative temperature coefficient C = total heat capacity of the core available to the. prompt burst energy release 5S = change in average core temperature produced bythe prompt burst 6k that portion of the step reactivity insertion which is p
above prompt critical The equations given result from the Fuchs-Nordheim space-independent kinetics model in which heat transfer and delayed neutrons are neglected and the core heat capacity and temperature coefficient are invarient with temperature The applicability of this relatively simple model has been demonstrated by its performance predictions for pulses generated in ,TRICGA reactors (Ref. 7,1-7.1-ý33. 6-7-,O* eO
With reference to the comparisons in Table 7.1-15, the calculated performance, with parameters from this analysist is genetally within about 5% of the experimental values, with the calculated values always being larger and with the calcul&ted peak power about 20% larger than the experi-mental value. The manner 'in which the parameters are interrelated in the Fuchs-Nordheim model would indicate that no single calculated value is greatly different from the real value, They can all be at varying, rela-tively small differences from the real values, but the combination of errors for the quantities as they are related to influence the pulsing petforMance is within a reasonable. accuracy for the measurements of the pulsing performance.
Figure 7.1-6 shows the pulse shape and core energy release for a
$4.40 pulse in the Sandia ACPR as calculated using the parameters generated in this study. Fig. 7.1-7 shows other important core characteristics :for the same pulse.
For a comparison of flux in the lexperimental cavity, Sandia quotes 1.55 x 10 15 nvt (>10 keV) for a $4.40 pulse. Using a 100 MW-sec energy release, to include energy .beyond 10 sec which could be measured by-the detectors, the flux/Watt (>10 keV) is 1.55 x 10 7 . This compares to the computed value of 1.l. x 10 7 ./Watt, as shown in Fig. 7.1-5. The flux values given here a"sume an energy release during fission Of 190 MeV/fission. If the delayed energy from fission is not included, as would be the case during pulsing, the calculated neutron fluence value could increase by as much as 10...
7.1-34 8-;r A&iO;S-C IkvO
- n5 'ns 1I I I I I to!
io0 * " " -0
,1o 102 ENERGY tO2 energy released versus------ - --
f1 t O. I I , ,!I , .. . I o 0 0.05 0.10 0.15 . 0,20 0;25 0,30 0.35 BEGINNING OF ROD,0 7IO01O FROII TItlE (SEC)
BL-0619 Fig. 7.1-6, Sand+/-a ACPR standard pugse .p~er ane energy reieaeed versu~s time from first rod mot~ion 7.1-35 8-1 7.1-35.
aco:19 89 eeO
N a 10 - -
5 MAXIMUM tHEAT FIUX (W/CM7) 500 x ".-
'
-0 FUEl. TEMPERATURE (*I)
MAXIMUM 800 TUEL s.D. TEMPERATURE ('C) 700 THERMOCUPLE TEMPAERATURE (C) 5800 600 ac 500 FUELCENTERLINE TEMPERATURE ",C) 300 O 20"CA TEPRTUE(C wfoe TiEFO IG NI1G TRO OO SC I0 o-2 IO-t * , oo iol o
- 1 TIME FROM BEGINNING OF ROD It0TlON (SEe) .
0 EL-0620 Fig. 7.1-7. Sandia ACPR t!eperaslture and heat 'flux time verus t4-& f[rom first r"* motion