工况的剧烈变化可能导致一回路系统中法兰连接部位、泵的密封面等处发生泄漏。针对准确的泄漏物理模型很难建立的实际情况,在对不可测的温度应力参数进行正态随机游走的基础上,以高斯过程回归模型为替代模型对一回路泄露率进行预测,并对替代模型的不确定性进行定量分析。结果表明,高斯过程回归模型能够方便地实现对替代模型的不确定性评估,并且在小样本容量的情况下,能够实现对一回路泄漏率较准确的预测。
The leakage may occur at the flange connection part and sealing surface of pump in primary circuit system due to the drastic change of working condition. In view of the difficult to establish an accurate and complete leakage mechanism model, the Gaussian process regression model is used as an alternative model and the uncertainty of the alternative model is calculated quantitatively. Based on the normal random walk of the unmeasurable parameters, the leakage rate of the primary circuit is predicted. The results show that the Gaussian process regression model can calculate the uncertainty of the alternative model, and can accurately predict the leakage rate of the primary circuit in the case of small samples.
2024,46(13): 102-106 收稿日期:2023-09-05
DOI:10.3404/j.issn.1672-7649.2024.13.018
分类号:TL364
作者简介:魏淋东(1995-),男,硕士,工程师,研究方向为核能与核技术工程
参考文献:
[1] 陆晓峰, 沈轶. 高温法兰密封接头的可靠性分析[J]. 压力容器, 2007, 24(9): 20-24.
[2] 喻健良, 张忠华, 闫兴清, 等. 高温下螺栓-法兰-垫片系统密封性能研究[J]. 压力容器, 2012, 29(5): 5-9.
[3] 孔慈宇, 陈春辉, 张斌, 等. 管法兰用柔性石墨金属齿形垫泄漏率及其预测模型[J]. 压力容器, 2019, 36(12): 12–15.
[4] 赵梦恩. 改进高斯过程回归算法及其应用研究[D]. 杭州: 浙江理工大学, 2019.
[5] 郭后. 基于高斯过程的备件消耗预测研究[D]. 武汉: 华中科技大学, 2016.
[6] 刘健. 基于高斯过程回归的锂离子电池剩余寿命预测研究[D]. 上海: 上海交通大学, 2019.
[7] 曾聿赟. 基于状态监测数据的核电厂设备寿命预测算法研究[D]. 北京: 清华大学, 2017.
[8] BARALDI P, COMPARE M, SAUCO S, et al. Ensemble neural network-based particle filtering for prognostics[J]. Mechanical Systems and Signal Processing, 2013, 41(1): 288-300.
[9] 王洪桥. 高斯过程回归在不确定性量化中的应用[D]. 上海: 上海交通大学, 2019.
[10] HU Y, BARALDI P, DI M F, et al. A particle filtering and kernel smoothing-based approach for new design component prognostics[J]. Reliability Engineering and System Safety, 2015, 13(4): 19-31.
[11] 刘守相, 于雷, 鄢炳火. 一体化压水堆强迫循环转自然循环过渡过程特性分析[J]. 原子能科学技术, 2012, 46(增刊): 220-224.