Life prediction method of nuclear power heat exchanger based on improved red kite optimization algorithm and LSTM

doi:10.11996/JG.j.2095-302X.2025051085

Journal of Graphics ›› 2025, Vol. 46 ›› Issue (5): 1085-1093.DOI: 10.11996/JG.j.2095-302X.2025051085

• Digital Design and Manufacture • Previous Articles Next Articles

Life prediction method of nuclear power heat exchanger based on improved red kite optimization algorithm and LSTM

XIAN Siyu¹^,²(), ZHAO Zetian³, WU Xuanyu¹(), FENG Yixiong¹^,²^,³, XUE Yang⁴, ZHANG Zhifeng³

¹ School of Mechanical Engineering, Guizhou University, Guiyang Guizhou 550025, China
² State Key Laboratory of Public Big Data, Co-built by the Province and Ministry, Guizhou University, Guiyang Guizhou 550025, China
³ National Key Laboratory of Fluid Power Fundamental Components and Mechatronic Systems, Zhejiang University, Hangzhou Zhejiang 310027, China
⁴ China General Nuclear Power Engineering Co., Ltd., Shenzhen Guangdong 518116, China

Received:2025-05-07 Accepted:2025-07-10 Online:2025-10-30 Published:2025-09-10
Contact: WU Xuanyu
About author:First author contact:
XIAN Siyu (2000-), master student. His main research interests cover predictive maintenance of nuclear power equipment, etc. E-mail：2985712029@qq.com
Supported by:
Zhejiang Provincial Key Research and Development Program(2024C01029);Zhejiang Provincial Key Research and Development Program(2025C01023);Guizhou University Open Fund(PBD2024-0515)

Abstract

Abstract:

With the increase in the quantity and diversification of nuclear power equipment, predictive maintenance strategies based on condition detection have gradually become a focus of attention for nuclear power plants, especially for predicting the remaining service life of key equipment, such as cooling water heat exchangers. A heat exchanger life prediction method combining an improved red kite optimization algorithm (IROA) with a long short-term memory network (LSTM) was proposed to address the limitations of traditional methods in hyperparameter optimization and improve prediction accuracy. In response to the problem of premature convergence caused by insufficient initial population diversity in existing methods, crossover and mutation operations from genetic algorithms were introduced to improve the ROA algorithm, in order to enhance population diversity and the ability to escape from local optima. With an LSTM model trained using historical degradation data, we conducted a detailed analysis of the impact of different hyperparameter combinations on model performance, demonstrating that the optimized hyperparameter combinations could significantly improve prediction performance. To verify the effectiveness of the proposed method, we conducted comparative experiments between IROA-LSTM and other common prediction methods such as SVM, CNN, RNN, and standard LSTM, as well as comparative experiments with several other optimization algorithms, along with noise interference tests. The results indicated that IROA-LSTM not only performed well in various performance indicators, but also demonstrated strong robustness and stability, maintaining high prediction accuracy under different conditions. This provided reliable data support for developing scientific and reasonable maintenance strategies, thereby contributing to improved safety and economic efficiency of nuclear power plant equipment operation.

Key words: life prediction, predictive maintenance, heat exchanger, improve red kite optimization, long short-term memory

CLC Number:

TH183
TH39

XIAN Siyu, ZHAO Zetian, WU Xuanyu, FENG Yixiong, XUE Yang, ZHANG Zhifeng. Life prediction method of nuclear power heat exchanger based on improved red kite optimization algorithm and LSTM[J]. Journal of Graphics, 2025, 46(5): 1085-1093.

Figures/Tables 9

Fig. 1 Red kite optimization algorithm flow

Fig. 2 LSTM network architecture diagram

Fig. 3 Flowchart of the IROA combined with LSTM methodology

Table 1 Hyperparameter combination

超参数类型	Num_layers	Batch size	Lstm_size
取用值	[1、2、 3、4]	[32、64、 128、256]	[32、64、 128、256]

Fig. 4 Test effect diagram under different hyperparameters ((a) Experimental results of the model with one layer; (b) Experimental results of the model with two layer; (c) Experimental results of the model with three layer;(d) Experimental results of the model with four layer)

Fig. 5 Comparison and validation of IROA-LSTM with different methods

Fig. 6 Comparative effectiveness diagram of IROA with various optimization methods

Fig. 7 The experimental results under Gaussian white noise interference ((a) Effect of 5% Gaussian white noise interference; (b) Effect of 10% Gaussian white noise interference)

Fig. 8 SHAP value analysis for each sensor

References 22

[1]	KAUR S, SAHOTA R S, YING S Y, et al. Emerging trends in industry 4.0 and predictive maintenance[J]. Abhigyan, 2025, 43(1): 54-67.
[2]	MAZZETTO S. Hybrid predictive maintenance for building systems: integrating rule-based and machine learning models for fault detection using a high-resolution Danish dataset[J]. Buildings, 2025, 15(4): 630.
[3]	GUO J, ZHANG X W, CUI Y H. Estimation of fatigue parameters and life prediction for orthotropic steel deck based on reverse Markov theory[J]. Journal of Structural Engineering, 2025, 151(5): 04025037.
[4]	陈卓, 闫明, 金映丽. 基于灰色马尔科夫模型的随机多次冲击损伤累积研究[J]. 振动与冲击, 2024, 43(24): 295-300.
	CHEN Z, YAN M, JIN Y L. A study on damage accumulation ofrandom multiple impact based on a grey Markov model[J]. Journalof Vibration and Shock, 2024, 43(24): 295-300 (in Chinese).
[5]	EZZOUHRI A, CHAROUH Z, GHOGHO M, et al. A data-driven-based framework for battery remaining useful life prediction[J]. IEEE Access, 2023, 11: 76142-76155.
[6]	WANG F K, MAMO T. Hybrid approach for remaining useful life prediction of ball bearings[J]. Quality and Reliability Engineering International, 2019, 35(7): 2494-2505.
[7]	LI N P, LEI Y G, GEBRAEEL N, et al. Multi-sensor data-driven remaining useful life prediction of semi-observable systems[J]. IEEE Transactions on Industrial Electronics, 2021, 68(11): 11482-11491.
[8]	DE MARCO L M, TRIERWEILER J O, FARENZENA M. Determination of remaining useful life in cyclic processes[J]. Industrial & Engineering Chemistry Research, 2019, 58(48): 22048-22063.
[9]	ZHANG W H, WANG Z H, RAÏSSI T, et al. An ellipsoid‐based framework for fault estimation and remaining useful life prognosis[J]. International Journal of Robust and Nonlinear Control, 2023, 33(12): 7260-7281.
[10]	王炳波, 刘赓传. 基于无迹卡尔曼滤波的离心泵剩余寿命研究[J]. 石油机械, 2021, 49(11): 31-38.
	WANG B B, LIU G C. Research on residual life of centrifugal pump based on unscented Kalman filter[J]. China Petroleum Machinery, 2021, 49(11): 31-38 (in Chinese).
[11]	李乃鹏, 蔡潇, 雷亚国, 等. 一种融合多传感器数据的数模联动机械剩余寿命预测方法[J]. 机械工程学报, 2021, 57(20): 29-37, 46. DOI
	LI N P, CAI X, LEI Y G, et al. A model-data-fusion remaining useful life prediction method with multi-sensor fusion for machinery[J]. Journal of Mechanical Engineering, 2021, 57(20): 29-37, 46 (in Chinese). DOI
[12]	WANG S L, FAN Y C, JIN S Y, et al. Improved anti-noise adaptive long short-term memory neural network modeling for the robust remaining useful life prediction of lithium-ion batteries[J]. Reliability Engineering & System Safety, 2023, 230: 108920.
[13]	ZHANG H X, XI X P, PAN R. A two-stage data-driven approach to remaining useful life prediction via long short-term memory networks[J]. Reliability Engineering & System Safety, 2023, 237: 109332.
[14]	TIAN G D, WANG W J, ZHANG H H, et al. Multi-objective optimization of energy-efficient remanufacturing system scheduling problem with lot-streaming production mode[J]. Expert Systems with Applications, 2024, 237: 121309.
[15]	CHEN X F, YE C M, ZHANG Y. Strengthened grey wolf optimization algorithms for numerical optimization tasks and AutoML[J]. Swarm and Evolutionary Computation, 2025, 94: 101891.
[16]	RAEISI-GAHRUEI J, BEHESHTI Z. The electricity consumption prediction using hybrid red kite optimization algorithm with multi-layer perceptron neural network[J]. Journal of Intelligent Procedures in Electrical Technology, 2025, 15(60): 1-22.
[17]	ALSHAREEF S M, FATHY A. Efficient red kite optimization algorithm for integrating the renewable sources and electric vehicle fast charging stations in radial distribution networks[J]. Mathematics, 2023, 11(15): 3305.
[18]	RUIZ R I. A GPU accelerated genetic algorithm for the construction of Hadamard matrices[D]. Rio Grande: The University of Texas Rio Grande Valley, 2022.
[19]	熊思颖, 董黎君. 标记平面立体线图的自适应遗传算法[J]. 图学学报, 2018, 39(6): 1105-1111. DOI
	XIONG S Y, DONG L J. A self-adaptive genetic algorithm for labeling line drawings of planar objects[J]. Journal of Graphics, 2018, 39(6): 1105-1111 (in Chinese).
[20]	GRAVES A. Long short-term memory[M]// GRAVESA. Supervised Sequence Labelling with Recurrent Neural Networks. Berlin, Heidelberg: Springer, 2012: 37-45.
[21]	杨世强, 杨江涛, 李卓, 等. 基于LSTM神经网络的人体动作识别[J]. 图学学报, 2021, 42(2): 174-181.
	YANG S Q, YANG J T, LI Z, et al. Human action recognition based on LSTM neural network[J]. Journal of Graphics, 2021, 42(2): 174-181 (in Chinese). DOI
[22]	SAXENA A, GOEBEL K, SIMON D, et al. Damage propagation modeling for aircraft engine run-to-failure simulation[C]// 2008 International Conference on Prognostics and Health Management. New York: IEEE Press, 2008: 1-9.

Life prediction method of nuclear power heat exchanger based on improved red kite optimization algorithm and LSTM

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 9

References 22

Related Articles 2

Recommended Articles

Metrics

Comments

[1]	WEI Min, YAO Xin. Two-stage storm entity prediction based on multiscale and attention [J]. Journal of Graphics, 2024, 45(4): 696-704.
[2]	LV Ling, LI Hua, WANG Wu. Multi-directional text detection based on the fusion of enhanced feature extraction network and semantic feature [J]. Journal of Graphics, 2024, 45(1): 56-64.