Knowledge base operator support system for nuclear power plant fault diagnosis

The Sixth International Symposium on Neural Networks (ISNN 2009)

Emergency evacuation time model of nuclear power plant pedestrians

Generalization analysis and improvement of CNN-based nuclear power plant fault diagnosis model under varying power levels

The thermal-hydraulic characteristics of Nuclear Power Plant (NPP) during Steam Generator Tube Rupture (SGTR) accident are of great concern. In this paper, the thermal-hydraulic characteristics of CP300 during SGTR accident with no operator actionsduring the first 30 min are investigated using the best estimate RELAP5/MOD3.4. Modeling and nodalization of CP300 was executed, including the vessel, pumps, pressurizer, steam generations and necessary auxiliary systems. Some main transient parameters were obtained, such as Reactor Coolant Pump (RCP) coolant temperature and pressure, steam generator flow rate and pressure. The calculation results gives the sequences of the NPP during the SGTR as described below. As the tube rupture occurs, the primary pressure drops and secondary pressure increases. When the primary pressure drops down to the set-point of the scram, the control rods drop down and the power of the NPP begins to decrease, causing the primary coolant temperature to decrease. The primary pressure continues to drop. When it drops down to 10.78MPa, the High pressure Safety Injection System is put into operation and the accident is mitigated, even without operator’s actions. In conclusion, the calculated results indicate that the key thermal hydraulic parameters of CP300 during the SGTR accident are in acceptable ranges and the accident is effectively controlled.

Stock price is one of intricate non-linear dynamic system. Typically, Elman neural network is a local recurrent neural network, having one context layer that memorizes the past states, which is quite fit for resolving time series issues. Given this, this paper takes Elman network to predict the opening price of stock market. Considering that Elman network is limited, this paper adopts self-adapting variant PSO algorithm to optimize the weights and thresholds of network. Afterwards, the optimized data, regarded as initial weight and threshold value, is given to Elman network for training, accordingly the prediction model for opening price of stock market based on self-adapting variant PSO-Elman network is formed. Finally, this paper verifies that model by some stock prices, and compares with BP network and Elman network, so as to draw the result that shows the precision and stability of this predication model both are superior to the traditional neural network.

Keywords: Time series estimation, Flow Estimation, Elman Neural Networks Abstract: This paper investigates the application of partially recurrent artificial neural networks (ANN) in the flow estimation for Sao Francisco River that feeds the hydroelectric power plant of Sobradinho. An Elman neural network was used suitably arranged to receive samples of the flow time series data available for Sao Francisco River shifted by one month. For that, the neural network input had a delay loop that included several sets of inputs separated in periods of five years monthly shifted. The considered neural network had three hidden layers. There is a feedback between the output and the input of the first hidden layer that enables the neural network to present temporal capabilities useful in tracking time variations. The data used in the application concern to the measured Sao Francisco river flow time series from 1931 to 1996, in a total of 65 years from what 60 were used for training and 5 for testing. The obtained results indicate that the Elman neural network is suitable to estimate the river flow for 5 year periods monthly. The average estimation error was less than 0.2 %

In this paper, a neuro-genetic approach is proposed for the inverse kinematics problem solution of robotic manipulators. The proposed solution method is based on using neural networks and genetic algorithms in a hybrid system. Neural networks have been used by many researchers in the inverse kinematics solution. Since the neural networks work with an acceptable error, the error at the end of learning has to be minimized for sensitive applications. This study is based on using genetic algorithms to minimize this error. A case study is presented for a 6 degree of freedom robot. In the neural network part, three Elman networks are separately trained and then used in parallel since one Elman network may give better result than the other two ones. These three results are placed in the initial population of the genetic algorithm. The end effector position error is defined as the fitness function and genetic algorithm is implemented. Thus, the error is reduced in micrometer levels. Key words: Elman neural networks, error minimization, six-degree-of-freedom robot, genetic algorithms, robotics.

Nuclear power plant (NPP) fault diagnosis is critical to ensure the safe, stable and economic operation of nuclear facilities. Existing deep learning-based NPP fault diagnosis methods primarily extract features from the temporal and spatial domain. Given recent advances in frequency-domain learning and lightweight models for time-series modeling, this paper proposes a frequency-domain learning-based fault diagnosis method for multi-modular high-temperature gas-cooled reactor (mHTGR) modules. It extracts temporal and spatial frequency-domain features using a lightweight complex-valued neural network (CVNN), which are then applied for fault classification and severity estimation. Under typical fault detection tasks for an mHTGR, comparative experiments verify that the proposed method outperforms existing methods that utilize long short-term memory (LSTM) and graph neural networks (GNNs) for spatiotemporal feature extraction in terms of classification accuracy, mean squared error (MSE) loss, and inference time.

Human factors engineering focuses on human-machine-environment coordination. The purpose is to prevent and reduce human errors in complex human-machine system, and improve the reliability of human-machine system. Based on the latest engineering practice of British generic design assessment for “HPR1000”, this paper introduces a systematic human factor analysis and argument technology by an integration of human factors engineering elements, as well as its engineering application. The engineering practice showed that this systematic method that integrates multiple elements of human factors engineering, can fully combine with the design characteristics of nuclear power plant, comprehensively analyze and consider the impact of human factors on nuclear power plant design, and can transmit and iterate risk information among various professional fields s. It is an effective way to realize the in-depth application of human factors engineering.

The accuracy of the diagnosis obtained from a nuclear power plant fault-diagnostic advisor using neural networks is addressed in this paper in order to ensure the credibility of the diagnosis. A new error estimation scheme called error estimation by series association provides a measure of the accuracy associated with the advisor's diagnoses. This error estimation is performed by a secondary neural network that is fed both the input features for and the outputs of the advisor. Our error estimation by series association outperforms previous error estimation techniques in providing more accurate confidence information with considerably reduced computational requirements. We demonstrate the extensive usability of our method by applying it to a complicated transient recognition problem of 33 transient scenarios. The simulated transient data at different severities consists of 25 distinct transients for the Duane Arnold Energy Center nuclear power station ranging from a main steam line break to anticipated transient without scram (ATWS) conditions. The fault-diagnostic advisor system with the secondary error prediction network is tested on the transients at various severity levels and degraded noise conditions. The results show that our error estimation scheme provides a useful measure of the validity of the advisor's output or diagnosis with considerable reduction in computational requirements over previous error estimation schemes.

A greenhouse control with feed-forward and recurrent neural networks

Model reference control by recurrent neural network built with paraconsistent neurons for trajectory tracking of a rotary inverted pendulum

In order to estimate daily solar radiation, this paper proposes Elman (ENN) and Feed forward backpropagation (FNN) neural networks. The time series data from the location of Kenitra City, Morocco is used to train the created models. Fletcher-Powell Conjugate Gradient (CGF), Scaled Conjugate Gradient (SCG), Resilient Backpropagation (RB), Conjugate gradient with Powell-Beale restarts (CGB), Levenberg-Marquardt (LM), Polak-Ribi´ere Conjugate Gradient (CGP), and One Step Secant (OSS) are also used with both, the ENN and FNN to identify the best and effective training function for each. The models with various training algorithms are tested by using evaluation metrics root mean square error (RMSE), mean square error (MSE) and mean absolute error (MAE). The study showed that the Elman Neural Network had an excellent daily solar radiation prediction for Kenitra city. Good results have been obtained with Levenberg-Marquardt algorithm.

In recent years commercially available minicomputers have been used to study the performance characteristics of nuclear power plants. Both frequency domain and time domain signatures can be used to accomplish this task. In this paper we present the methodologies based on time series analysis to monitor stability of boiling water reactors, response monitoring of system sensors, and surveillance of certain characteristics whose changes could provide system diagnostic information. The stochastic time series modeling algorithms and estimation of performance parameters have been implemented on minicomputers (such as PDP-II) and can be used for on-line monitoring of the dynamic processes. These techniques are applicable for both steady-state response monitoring and for tracking anomalous behavior of the power plant.

Elman neural network was one of the dynamic recurrent neural networks. In this paper, a modified Elman network was introduced first. Then we proposed a PID Elman neural network and its learning algorithms are discussed in detail. Simulation results based on ideal mathematical model and hydro-turbine governing system model show that the PID Elman network is prior to the modified Elman network in identifying nonlinear dynamic system.