A Multi-Domain Diagnostics Approach for Solenoid Pumps Based on Discriminative Features

Abstract

Accurate condition monitoring of industrial cyber-physical systems/components demands the use of reliable fault detection and isolation (FD&I) methodologies. Meta-heuristic algorithms for feature selection have good exploration capability for optimal discriminative feature selection for fault isolation/classification of which the Binary Particle swarm optimization (BPSO) is superior to its counterparts. This study presents a robust approach for vibration-based failure diagnostics of electromagnetic/solenoid pumps which employ a multi-domain feature extraction procedure (statistical time-domain and frequencydomain features, Mel frequency cepstral coefficients, and continuous wavelet coefficients) for capturing linear and nonlinear properties from the signals. Compared with other filter and wrapper methods for supervised feature selection, a hybrid filter-wrapper (Pearson's correlation-BPSO (ρ-BPSO)) feature selection procedure is proposed for global search of optimal discriminative (uncorrelated) features for fault diagnosis with an RBF-kernel support vector machine (SVM*). Subsequently, a practical case study involving five VSC63A5 solenoid pumps at various operating/fault conditions is presented for validating the performance of the proposed approach. Results show the superior performance of the proposed hybrid filterwrapper approach against filter-based and wrapper-based techniques for discriminative feature selection. Also, the proposed ρ-BPSO-SVM* diagnostics model performance was compared with other standard fault isolation/classification methods.

A Cost-Efficient MFCC-Based Fault Detection and Isolation Technology for Electromagnetic Pumps

“…Recently, vibration monitoring has become a highly reliable condition monitoring paradigm for fluid pumps (and most systems) [11]; consequently, pump manufacturers are recently integrating on-board vibration (and temperature) sensors into their products; however, interpreting and drawing accurate conclusions from the complex big data has been one of the setbacks facing optimized fluid pump FDI and prognostics [6]. As a result that raw vibrational signals are non-stationary with background noise, extracting salient features from these signals remains an on-going challenge for accurately understanding the underlying dynamics of targeted systems.…”

An Automated Sensor Fusion Approach for the RUL Prediction of Electromagnetic Pumps

“…Electromagnetic pumps, like other types of pumps are usually operated for long periods which expose them to diverse failure modes from sources ranging from unfavorable environmental conditions, fluid contamination, electrical and/or mechanical stress, fatigue, uncertainty, filter clogging, overuse/underuse, etc. [26], [34]. It is on these grounds that the need for real-time condition monitoring (and assessment) becomes crucial for accurate predictive maintenance of these pumps.…”

“…Interestingly, solenoid pump and valves are identical in their mode of operation-solenoid magnetization by electrical current through the coil which causes the electromagnetic core to move against a spring to slide a diaphragm into the discharge position. Although solenoid valves have witnessed a reasonable number of research studies on FDI and prognostics, unfortunately, only a few exist for solenoid pumps out of which, a majority cover only the design and fault diagnosis [26], [5], [34].…”

“…Meta-heuristic algorithms like the GA, particle swarm optimization, Ant Colony Optimization, fruit fly optimization, and a host of hybrid methods have been reported (and compared) in many studies for several optimization problems including but not limited to feature selection [26], fault detection and identification [15], structural optimization of predictive algorithms [27], [28], and functional process/system control [29]. Particularly for multi-objective probelms, the MOGA has proven its superiority over the multi-objective PSO (MOPSO) as verified in [30] and has also shown reliable efficiencies for global parameterization of time-series predictors [27], [28].…”

“…This work, being the first attempt, aims at exploring vibrational and pressure signals for the RUL prediction. Following earlier studies on failure causes in pumps [34], [39], and failure diagnostics of electromagnetic pumps in particular [26], filter clogging has shown to be one of the most critical failure modes of solenoid pumps; hence the motivation behind this work-to simulate a natural operating condition of the pump under fluid contamination till full filter clogging occurs. Fig.…”