# MOC Efficiency Improvements Using a Jacobi Inscatter Approximation

## Abstract

In recent weeks, attention has been given to resolving the convergence issues encountered with TCP _{0} by trying a Jacobi (J) inscatter approach when group sweeping, where the inscatter source is constructed using the previous iteration flux. This is in contrast to a Gauss-Seidel (GS) approach, which has been the default to-date, where the scattering source uses the most up-to-date flux values. The former is consistent with CASMO, which has no issues with TCP _{0} convergence. Testing this out on a variety of problems has demonstrated that the Jacobi approach does indeed provide substantially more stability, though can take more outer iterations to converge. While this is not surprising, there are improvements that can be made to the MOC sweeper to capitalize on the Jacobi approximation and provide substantial speedup. For example, the loop over groups, which has traditionally been the outermost loop in MPACT, can be moved to the interior, avoiding duplicate modular ray trace and coarse ray trace setup (mapping coarse mesh surface indexes), which needs to be performed repeatedly when group is outermost.

- Authors:

- Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
- Univ. of Michigan, Ann Arbor, MI (United States)

- Publication Date:

- Research Org.:
- Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States). Consortium for Advanced Simulation of LWRs (CASL)

- Sponsoring Org.:
- USDOE Office of Nuclear Energy (NE)

- Contributing Org.:
- Univ. of Michigan, Ann Arbor, MI (United States)

- OSTI Identifier:
- 1325480

- Report Number(s):
- ORNL/TM-2016/406; CASL-U-2016-1056-002

NT0304000; NEAF343; CASL-U-2016-1056-002; TRN: US1700079

- DOE Contract Number:
- AC05-00OR22725

- Resource Type:
- Technical Report

- Country of Publication:
- United States

- Language:
- English

- Subject:
- 97 MATHEMATICS AND COMPUTING; 21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS; APPROXIMATIONS; CONVERGENCE; EFFICIENCY; SCATTERING; STABILITY; TESTING; ITERATIVE METHODS; M CODES

### Citation Formats

```
Stimpson, Shane, Collins, Benjamin, and Kochunas, Brendan.
```*MOC Efficiency Improvements Using a Jacobi Inscatter Approximation*. United States: N. p., 2016.
Web. doi:10.2172/1325480.

```
Stimpson, Shane, Collins, Benjamin, & Kochunas, Brendan.
```*MOC Efficiency Improvements Using a Jacobi Inscatter Approximation*. United States. doi:10.2172/1325480.

```
Stimpson, Shane, Collins, Benjamin, and Kochunas, Brendan. Wed .
"MOC Efficiency Improvements Using a Jacobi Inscatter Approximation". United States.
doi:10.2172/1325480. https://www.osti.gov/servlets/purl/1325480.
```

```
@article{osti_1325480,
```

title = {MOC Efficiency Improvements Using a Jacobi Inscatter Approximation},

author = {Stimpson, Shane and Collins, Benjamin and Kochunas, Brendan},

abstractNote = {In recent weeks, attention has been given to resolving the convergence issues encountered with TCP0 by trying a Jacobi (J) inscatter approach when group sweeping, where the inscatter source is constructed using the previous iteration flux. This is in contrast to a Gauss-Seidel (GS) approach, which has been the default to-date, where the scattering source uses the most up-to-date flux values. The former is consistent with CASMO, which has no issues with TCP0 convergence. Testing this out on a variety of problems has demonstrated that the Jacobi approach does indeed provide substantially more stability, though can take more outer iterations to converge. While this is not surprising, there are improvements that can be made to the MOC sweeper to capitalize on the Jacobi approximation and provide substantial speedup. For example, the loop over groups, which has traditionally been the outermost loop in MPACT, can be moved to the interior, avoiding duplicate modular ray trace and coarse ray trace setup (mapping coarse mesh surface indexes), which needs to be performed repeatedly when group is outermost.},

doi = {10.2172/1325480},

journal = {},

number = ,

volume = ,

place = {United States},

year = {Wed Aug 31 00:00:00 EDT 2016},

month = {Wed Aug 31 00:00:00 EDT 2016}

}