A study on the robustness of neural network models for predicting the break size in LOCA

• Published in: Progress in nuclear energy (New series) 2018-11, Vol.109, p.12-28
• Main Authors: Tian, Xiange, Becerra, Victor, Bausch, Nils, Santhosh, T.V., Vinod, Gopika
• Format: Article
• Language: English
• Subjects: Accident prediction; Accidents; Brain; Computer simulation; Condition monitoring; Cooling systems; Datasets; Diagnosis; Diagnostic models; Diagnostic systems; Diagnostics; Group method of data handling; Interpolation; Loss of coolant accident; Loss of coolant accidents; Mathematical models; Model robustness; Multilayer perceptrons; Neural networks; Nuclear power plant; Nuclear power plants; Nuclear reactors; Robustness; Training
• ISSN: 0149-1970; 1878-4224
• DOI: 10.1016/j.pnucene.2018.07.004

The diagnosis of loss of coolant accidents in nuclear reactors has attracted a great deal of attention in condition monitoring of nuclear power plants given that the health of the cooling system is crucial to the nuclear reactor's stable operation. Many different types of neural networks have commonly been applied to loss of coolant accident diagnosis. It is important to select a suitable architecture for the neural network that delivers robust results, in that the predicted break size is deemed to be accurate even for break sizes that are not included in the training data sets. The robustness metric proposed in our previous work is applied to compare the robustness of different diagnostic models. The data used for training these models consists of a number of time-series data sets, each for a different break size, with the transient behavior of different measurable variables in the coolant system of a nuclear reactor, following a simulated loss of coolant accident in a high-fidelity simulator. Given the simulation data for different break sizes, four different neural network architectures are investigated and their properties are compared and discussed. These models include a fully-connected multilayer perceptron with one hidden layer, a multilayer perceptron with one hidden layer that is pruned using the optimal brain surgeon algorithm, a fully-connected multilayer perceptron with two hidden layers, and a group method of data handling neural network. In this paper, an interpolation pre-processing method is investigated and shown to be effective to further improve the capability of neural networks for robustly predicting the break size of a loss of coolant accident. Both linear interpolation and cubic spline interpolation are studied as alternatives for the pre-processing approach. The performance of models developed with and without interpolation pre-processing are compared with the previously proposed robustness metric. Moreover, three blind cases are introduced to evaluate and compare the performance of the diagnostic models. Finally, a combined diagnostic model is proposed based on three different architectures to obtain high prediction accuracy and good robustness. •The robustness of neural networks for predicting the break size in LOCA is studied.•An interpolation pre-processing approach improves the robustness of the predictions.•The 2-hidden layer MLPs trained with pre-processing was the best individual model.•GMDH models with pre-processing perfo