5. update on inl modeling & simulation dehartresearch.engr.oregonstate.edu/treat-irp/sites/...number...

ww

w.in

l.gov

Update on INL Modeling and SimulationMark D. DeHart, PhDDeputy Director for Reactor Physics Modeling and SimulationNuclear Science and Technology DirectorateIdaho National Laboratory

Approved for public release; distribution is unlimited.

INL/MIS-16-40269

This image cannot currently be displayed.

Outline• Why Are We Doing This?

• Overview of MAMMOTH Project• Status of M8CAL Transient

Simulations• Need For Three-Dimensional Methods• MAMMOTH Methods Development• Startup Testing Timeline



Is Modeling and Simulation Even Needed?• Operations has pointed out (more than once) that they don’t need new

advanced M&S to restart.– True, they can perform operations in exactly the same fashion as was

done historically– Inefficient, time consuming, limited accuracy, but it worked:

• 6,604 reactor startups • 2,885 transient irradiations

– No predictive capabilities on fuel performance– Science was based on push it until it breaks.

• The truth is operations doesn’t need advanced M&S, but does want these capabilities.

– Reduce a week’s worth of pre-transient testing to perhaps a day.– Better quantify the behavior of the full core to reduce conservatisms– Tools for experiment, experiment vessel and instrumentation design– Increased throughput, better science coming out of measurements.

• DOE want the most bang for the buck on future TREAT operations– Increase throughput– Better utilization of reactor– Improved materials performance knowledge– Validation of single- and multi-physics methods



Overview of the MAMMOTH Project• MAMMOTH has just begun its third year of funding under the DOE/NE Advanced

Modeling and Simulation (NEAMS) Program.• It is now one of six focus areas for NEAMS:

– Fuels Product Line– Reactors Product Line– Integration Product Line– MAMMOTH Reactor Physics Product Line – Accident Tolerant Fuels High Impact Problem– Steam Generator FIV High Impact Problem

• The current focus of MAMMOTH wrt TREAT is three-fold:– Validation– Methods development– Deployment


Moving Forward on Advanced Modeling and Simulation

• Development of methods to handle cross section challenges– 3D effects – base cross sections generated using Serpent 2– Strong neutron streaming in hodoscope slot

• Direction-dependent diffusion coefficients using Larsen-Trahan method added to Rattlesnake

– Strong absorption near control rods• Superhomogenization correction • Also corrects for vertical leakage through air channels.


Fission Wire Fluxes


Transient Power Simulation – Transient 2855Linear Power (MW) Log Power (MW)

A B C ARCS A B C ARCS Ave MAMMOTH

Maximum 1290.26 1303.25 1281.37 1208.64 1222.00 1247.00 1227.50 1283.40 1257.93 1363.82

Integral 6.88E+05 6.49E+05 6.34E+05 6.13E+05 5.96E+05 6.03E+05 6.11E+05 6.54E+05 6.31E+05 6.81E+05

This image cannot currently be displayed.This image cannot currently be displayed.


Historical Approach for TREAT Calibration1. Heat balance measurements (calorimetry) were used to determine steady

state power at one or more flux levels. (Detector calibration)2. A sample fuel rod(s) was placed within TREAT, and a steady-state test was

performed for a set amount of time. The test rig was then removed and the number of fissions/sec/gm determined by destructive analytical chemistry techniques or gamma scan. (Power Coupling Factor)

3. Fission wires of uranium alloy were irradiated at steady state and also assayed to obtain burnup data. (Flux and flux shape)

4. TREAT would be operated in transient mode with a second set of fission wires with the planned transient. (Transient Correction Factor)

PCF can be expressed in units of Because the spectrum, and hence the coupling, change with time, the PCF is insufficient to calculate the total energy deposition knowing total core energy.

Hence the need for a TCF:


Limitations of Historical Approach

• In reality,

• The TCF is effectively

• Using calibration measurements,

• Any change in the spectrum between the calibration and the experiment itself could significantly alter the effective TCF

• For example, voiding in the actual experiment that didn’t occur in calibration would have a strong effect on PCF(t), which effects the actual TCF.

• All these issues go away if we can calculate J/gfuel sample directly, factoring in multi-physics effects in both core and experiment.


The Need for Validated Three-Dimensional Methods• Multi-SERTTA – Multi Static Environment Rodlet Transient Test Apparatus• Core energy release is being predicted by point kinetics using historic temp.

feedback tables. • PCF values are predicted using MCNP5.

• Point kinetics can provide reasonable estimates of average P(t) (insofar as the feedback tables are accurate), but lack the ability to predict 3D spectral or spatial shifts.

• Estimates of 3D behaviors are particularly important for the Multi-SERTTA’s intricate design, since none of its geometries can be adequately represented by simplified models

• Monte Carlo methods are used to predict 3D steady state behavior, but lack the ability for full coupling to implicitly predict time-dependent effects.


Three-Dimensional Control Rod Effects

• The Multi-SERTTA flux collar design is currently based on counteracting the steady state solution for the case 3 axial flux profile.

• This assumption is believed to be reasonable because the transient initiates from very low power, the transient rods move very quickly, and the reactor period will be on the order of tens of milliseconds. As a result, the transient rod should be entirely withdrawn before the core reaches appreciable power levels.


Three-Dimensional Control Rod Effects• The transient will be terminated (clipped) by rapid reinsertion of the transient rods. • It is hypothesized that the effective local transient experienced by each vessel unit

within Multi-SERTTA will be successively clipped as the rods proceed downward• No attempt has yet been made (or really is possible to do) to model this effect with

current methods. • This complication, along with complex spatial/spectral interactions between the

water within Multi-SERTTA, delayed neutrons, etc. all strongly suggest that the initial flux collar design (based entirely on a steady state solution for case 3 rod positions), will be found somewhat off-target during the initial Multi-SERTTA-CAL measurements.

• Predictive tools with the ability to inform the designer to the effects of these complicated phenomena can help design Multi-SERTTA flux collars.

• Additionally, these tools can be used to describe conditions for transient experiments post-test based on as-run reactor power, rod positions, etc.

• Along these lines, MAMMOTH-based foundational work has already been initiated to describe the axial dependence of PCF and TCF values in Multi-SERTTA transients, and to ultimately eliminate the needs for these terms.


Multi-SERTTA Simulation• Nic Woolstenhulme doesn’t have a reactor to test his rig in.• Nic had a full head of hair when this process started.• We are here to help Nic keep his remaining hair.• Near end of FY16 we were working to satisfy ourselves that we

were generating good cross-sections for steady state (checking against MCNP and Serpent 2)

• We are now in CR mode and waiting for funding to be able to proceed with transient calculations.

• Based on M8-CAL, we have a high level of confidence.• This work is of highest priority (and highly visible) within NE and

NEAMS (I used to have hair too).


Startup Testing Timeline and MAMMOTH

~Nov 2017 ~Jan 2018 ~Jun 2018 ~Dec 2018

Present ~May 2018 ~Oct 2018 ~Mar 2018 ~Sep 2016 ~Dec 2018 ~Oct 2016

Present

Present


Learning to Walk• TREAT has been idle for 22 years • We’re going to start off crawling first!• The first tests to be performed in TREAT will be to

reproduce the last set of measurements performed in TREAT – M8CAL

• These measurements will be performed with existing equipment

– No additional detectors– No additional flux/fission wires– Existing data acquisition system

(DAS) and Automatic Reactor Control System (ARCS)

– Wires and foils in test position will be gamma scanned and surveyed for FP inventory

• DOE-ID wants INL to demonstrate that we can reproduce historical data before resuming operations

• This data should ultimately be available to improve the M8CAL benchmark specification.


Planning for Validation Measurements• Measurements to support validation from Feb – Nov 2018 (~130 working days)

1. Develop Neutron Flux Map of the TREAT Core • Flux wires in strategic locations (x, y and z)• Fission wires together with flux wires• Strategy for wire insertion, removal, tracking, counting

2. Characterize Neutron Spectrum (steady state)• Core center/experiment location• Performed as a function of temperature• Activation of foils with and without filters• Use procedure for flux unfolding and code comparison using techniques

developed in ATR in 2011-2014• Temperature measurements

3. Develop TREAT Core Temperature Profile/Negative Temperature Coefficient• Non-trivial – only clad temperature is readily measured• Thermocouple or infrared, other?

4. Perform Neutron Lifetime and Beta Measurement• Noise techniques• Oscillation

5. Evaluate new detector technologies (in-core and ex-core)


Planning for Validation Measurements• For wires and foils, post-irradiation counting will be performed (betas and gamma)• TREAT currently has no in-house facilities. • Counting labs are available at the Advanced Test Reactor (20 min. drive) and at

the Materials and Fuels Complex (>5 min. drive)– Unfortunately, both facilities are being heavily utilized– Measurements should be made within 24 hours

• Justification for new counting facilities at TREAT– Budget is available– Decisions need to be made soon!

• February 2018 is just around the corner and much remains to be done by all parties!


Questions?

5. update on inl modeling & simulation dehartresearch.engr.oregonstate.edu/treat-irp/sites/...number...

Documents