基于并行技术的在线监测系统燃耗计算优化开发

Optimization Development of Online Monitoring System Depletion Calculation Based on Parallel Technology

1.
Science and Technology on Reactor System Design Technology Laboratory, Nuclear Power Institute of China, Chengdu, 610213, China
2.
Hainan Nuclear Power Co. Ltd., Haikou, 572733, China

摘要: 为了提高在线监测系统程序DESCAR模块在燃耗区数目较大时燃耗计算求解的计算效率，本文使用开放式多处理（OpenMP）和信息传递接口（MPI）并行技术，对DESCAR模块进行优化重构，以提高在线监测系统程序的计算效率。验证结果表明，优化重构后的DESCAR模块在保证计算结果正确性的同时，可以有效地提高在线监测系统程序的计算速度，加速效果显著，满足在线监测系统实时性的要求。
- 并行技术 /
- 燃耗计算 /
- 开放式多处理(OpenMP) /
- 信息传递接口(MPI)
Abstract: In order to improve the calculation efficiency of the DESCAR module of the online monitoring system program when the number of depletion zones is large, the DESCAR module is optimized and reconstructed by using the parallel technology of Open Multi-processing (OpenMP) and Message Passing Interface (MPI) to improve the calculation efficiency of the online monitoring system program. The verification results show that the optimized and reconstructed DESCAR module can effectively improve the calculation speed of the online monitoring system program while ensuring the correctness of the calculation results, and the acceleration effect is remarkable, which meets the requirements of the real-time online monitoring system.
- Parallel technology /
- Depletion calculation /
- Open Multi-Processing (OpenMP) /
- Message Passing Interface (MPI)

图 1 DESCAR模块计算热点

Figure 1. DESCAR Module Calculation Hotspots

下载: 全尺寸图片幻灯片

图 2 DESCAR模块性能变化

Figure 2. DESCAR Module Performance Changes

下载: 全尺寸图片幻灯片

图 3 数据流并行计算方案示意图

Figure 3. Schematic Diagram of Data Flow Parallel Computing Scheme

下载: 全尺寸图片幻灯片

图 4 各进程计算范围划分方案实例

Figure 4. Example of Calculation Range Division Scheme For Each Process

下载: 全尺寸图片幻灯片

表 1 MPI并行技术k_eff计算结果

Table 1. MPI Parallel Results of k_eff

核数/个燃耗步数

第0燃耗步第1燃耗步

单核 1.0753903847 1.0572208056

16核 1.0753903847 1.0572208056

45核 1.0753903847 1.0572208056

下载: 导出CSV

表 2 MPI并行技术计算耗时结果

Table 2. MPI Parallel Results of Time-consuming

核数/个计算时间/s 加速比并行效率/%

单核 139.9 1 100

16核 11.0 12.7 79.6

45核 5.7 24.8 55.0

下载: 导出CSV

[1]	CODY W J, MEINARDUS G, VARGA R S. Chebyshev rational approximations to e^{- x} in[0, +∞) and applications to heat-conduction problems[J]. Journal of Approximation Theory, 1969, 2(1): 50-65. doi: 10.1016/0021-9045(69)90030-6
[2]	陈国华,俞陆林,蒋校丰,等. 反应率线性外推含钆燃料燃耗计算方法的改进[J]. 核科学与工程,2011, 31(1): 9-14.
[3]	郭凤晨,柴晓明,芦韡,等. 微观燃耗方程求解方法研究与数值验证[J]. 核动力工程,2018, 39(S1): 20-23. doi: 10.13832/j.jnpe.2018.S1.0020
[4]	涂晓兰,刘东,王鑫,等. 基于并行框架的KYLIN-II程序输运模块并行开发[J]. 核动力工程,2020, 41(S1): 18-21.

图(4) / 表(2)