Pressure Vessel Diagnosis by Eliminating Undesired Signal Sources and Incorporating GA-Based Fault Feature Evaluation

This paper proposes a reliable fault detection model for a pressure vessel under low pressure conditions. To improve the diagnostic performance, signals of different vessel health conditions are purified by eliminating noise so that signals of different categories are much more distinguishable. This de-noising technique uses a blind source separation (BSS) technique in which an initial noise-contaminated signal is separated into constituent sources. These individual sources are either device status-characterizing signal sources or interfering sources. Noise is removed and an unimpaired signal is regenerated from the characteristic sources. An abundant pool of heterogenous features are extracted from a refined signal to avoid omitting important fault-related properties. However, some of these fault features may be redundant and irrelevant, and they are likely to cause classification performance degradation. To choose the most discriminative fault-signatures from a plentiful feature set as well as reduce the dimensions of the feature input, this study proposes a new feature selection algorithm that associates a genetic algorithm (GA)-based discriminative feature analysis to a k-nearest neighbors (k-NN) classifier. The efficacy of involved techniques and the overall fault diagnostic model is examined in terms of visual and qualitative evaluations. Experimental results illustrated in this study justify that the proposed fault detection model is promising and outstanding compared to other state-of-the-art counterparts.