Research on Simulation and State Prediction of Nuclear Power System Based on LSTM Neural Network

Abstract

The nuclear power plant systems are coupled with each other, and their operation conditions are changeable and complex. In the case of an operation fault in these systems, there will be a large number of alarm parameters, which can cause humans to be hurt in the accidents under great pressure. Therefore, it is necessary to predict the values of the key parameters of a device system. The prediction of the key parameters’ values can help operators determine the changing trends of system parameters in advance, which can effectively improve system safety. In this paper, a deep learning long short-term memory (LSTM) neural network model is developed to predict the key parameters of a nuclear power plant. The proposed network is verified by simulations and compared with the traditional grey theory. The simulation and comparison results show that the proposed LSTM neural network is effective and accurate in predicting the key parameters of the nuclear power plant.

Deep learning in nuclear industry: A survey

“…The prediction of key parameter values can help operators judge the changing trend of system parameters in advance and then effectively improve system security. Chen et al [120] proposed an LSTMbased model for predicting critical parameters of NPP. Chen et al [121] proposed a fault diagnosis method of NPP based on Deep Belief Network (DBN), which is trained on numerous original time-domain signal data of process parameters in NPPs.…”

Application of ARIMA and LSTM Model to the Forecast of CSI 300 Close Price

“…For listed companies, changes in stock prices are very crucial because the rise in stock prices can bring more financial support to them and promote their development. For a country, the healthy development of the stock market is conducive to guiding the rational allocation of market funds and resources [1]. Therefore, in order to better predict stock price, it is meaningful and necessary to explore reasonable and effective prediction methods.…”

Interval early warning method for state of engineering structures based on structural health monitoring data

“…The data‐driven methods can avoid above difficulties, which only rely on structural monitoring data of the mechanical equipment to build predictive models, and no prior assumptions on physical model of the structure have to be made beforehand 5 . In recent years, many researchers have developed many data‐driven approaches for health state prediction, structural damage detection, and early warning, including machine learning (ML) methods and statistical data‐driven methods, which currently have become the most popular prediction and early warning methods 6–8 . Nagarajaiah and Yang 9 proposed a new paradigm of explicitly modeling the sparse and low‐rank data structure, which can be used for data‐driven health monitoring and prediction.…”

“…developed many data-driven approaches for health state prediction, structural damage detection, and early warning, including machine learning (ML) methods and statistical data-driven methods, which currently have become the most popular prediction and early warning methods. [6][7][8] Nagarajaiah and Yang 9 proposed a new paradigm of explicitly modeling the sparse and low-rank data structure, which can be used for data-driven health monitoring and prediction. Avci et al 1 presented a literature review of vibration-based structural damage detection methods in civil structures and the transition from the traditional methods to ML and deep learning (DL) methodologies; various studies in the review were summarized and compared.…”