Abstract:Nonlinear relationships among different process variables are very common in a pressurized water reactor (PWR). Linear-based monitoring techniques has been proved inefficient and problematic in dealing with nonlinear processes. Moreover, sample imbalance is another issue existing in the diagnosis task of PWRs. When training samples of one fault mode outnumber those of other modes, performance of traditional classification algorithms may deteriorate. A methodology based on kernel principle component analysis (KPCA) and random under-sampling boost (RUSB) for PWRs is proposed. To cope with the nonlinearities in PWRs, KPCA acts as a primary step to distinguish sensor faults and system faults. System faults are further analyzed as a classification problem based on the primary diagnosis results of KPCA. Experiments are conducted using data samples generated from Generic Pressurized Water Reactor Simulator (GPWR) to test and verify the methodology. GPWR can calculate the responses of typical PWR and used to simulate sensor faults and three types of system faults. 36 monitoring variables located in two coolant loops are considered to generate samples. The performance of RUSB is compared with other state-of-art imbalance techniques by four metrics. The results show the superiority of KPCA to detect a univariate fault and the ability of RUSB to cope with different ratios of training samples in a PWR diagnostic task. The proposed methodology releases computation and memory cost due to the ability of KPCA to cope with nonlinearity and univariate detection, and relaxes the variable quality requirement due to the ability of RUSB to cope with imbalance fault samples. The findings in this study is beneficial for diagnostic system development of PWRs.

Research of Dynamic Rod Worth Measurement Technique Based on the Pin-by-pin Response Function

A new dynamic rod worth measurement method based on the transient spatial factor

Research on Control Method for Machining Non-Cylinder Pin Hole of Piston

Real-time Measuring Method to Measure the Micro-Displacement of a Rotating Cutter in Precise Piston Noncylinder Pinhole Boring

An Improved Two-Point Real-Time Measuring Method for Radial Micro-Displacement Measurement on High-Speed Smart Boring Bar

Optimization of Spatial Structure Designs of Control Rod Using Monte Carlo Code RMC

First Application of Large Reactivity Measurement through Rod Drop Based on Three-Dimensional Space-Time Dynamics

A Novel Model for Dynamic Process Study of Reactor Control Rod Drive Mechanism

On-line monitoring the incore power distribution by using excore ion-chambers

Quantification of Control Rod Worth Uncertainties Propagated from Nuclear Data Via a Hybrid High-Order Perturbation and Efficient Sampling Method

Research on Static Characteristics of High Precision Capacitance Rod Position Measurement Sensor

Experimental Study of the Pressure Discharge Process for the Hydraulic Control Rod Drive System Stepped Cylinder

A novel method for diagnosis and de-noising of control rod drive mechanism within floating nuclear reactor

Step Dynamic Process of the Hydraulically-driven Control Rod. (II). Theoretical Model on Step-down Process.

A new algorithm of critical control rod position search in MC code

PWR core calculation based on pin-cell homogenization in three-dimensional pin-by-pin geometry

Dynamic simulation of core power control for SPWR based on dynamic model bank

Shock Analysis on Hydraulic Drive Control Rod During Scram

Studies on the performance of the hydraulic control rod drive for the NHR-200

Kernel Principle Component Analysis and Random Under Sampling Boost Based Fault Diagnosis Method and its Application to a Pressurized Water Reactor

Optimization Analysis of Control Rod Drive Mechanism Electromagnetic Structure