PCA-based ANN approach to leak classification in the main pipes of VVER-1000

Abstract

This paper presents a neural network based fault diagnosing approach which allows dynamic crack and leaks fault identification. The method utilizes the Principal Component Analysis (PCA) technique to reduce the problem dimension. Such a dimension reduction approach leads to faster diagnosing and allows a better graphic presentation of the results. To show the effectiveness of the proposed approach, two methodologies are used to train the neural network (NN). At first, a training matrix composed of 14 variables is used to train a Multilayer Perceptron neural network (MLP) with Resilient Backpropagation (RBP) algorithm. Employing the proposed method, a more accurate and simpler network is designed where the input size is reduced from 14 to 6 variables for training the NN. In short, the application of PCA highly reduces the network topology and allows employing more efficient training algorithms. The accuracy, generalization ability, and reliability of the designed networks are verified using 10 simulated events data from a VVER-1000 simulation using DINAMIKA-97 code. Noise is added to the data to evaluate the robustness of the method and the method again shows to be effective and powerful.

Thermal-hydraulic analysis of VVER-1000 residual heat removal system using RELAP5 code, an evaluation at the boundary of reactor repair mode

“…At present the accidents for VVER-1000 reactors are thoroughly investigated where the researchers investigated the basis design improvement for VVER-1000 reactors [5][6][7][8][9][10][11][12][13][14][15]. In reviewing the literature, RELAP5 loop code found a validated tool for thermal hydraulic analysis of nuclear reactors [16,17].…”

SARCS cross section library generator: Part two: Fast and quasi transients evaluations

“…In core and out core fuel management requires the knowledge of fuel composition at any time during its residency in the core. MCNPX used CINDER utility to implement burnup calculations (Hadad and Ayobian, 2006;Hadad et al, 2010Hadad et al, , 2012Hadad et al, , 2013. Although accurate, MCNPX requires substantial time to perform burnup calculations for a typical LWR.…”

PARCS cross-section library generator; part one: Development and verification

“…The VVER-1000 BNPP, as a Russian type LWR, has been extensively studied for different aspects of nuclear safety (Hadad and Ayobian, 2006;Hadad et al, 2010Hadad et al, , 2012Hadad et al, , 2013. For a majority of these studies, Monte Carlo codes (e.g.…”