International Peer-Reviewed JournalOpen AccessISSN 2456-8880
irejournals@gmail.com+91-7433024337

Home / Current Issue / Paper 1722203

1722203 Vol 10 · Issue 2 Download Paper

Multi-Dimensional Electrical Signature Analysis for Performance Monitoring and Fault Detection in Industrial Chemical Pump Systems

Adeyinka, Olusegun Alfred Alimi, Teslim Adekunle Abdulazeez, Kabiru

Subject area: Science,Engineering and Technology  ·  Area of research: Electrical Engineering

Abstract

Unplanned downtime of chemical dosing and low-lift pump systems remains a major source of production loss and safety risk in industrial and municipal process plants. This paper presents a multi-dimensional electrical signature analysis (ESA) framework for continuous performance monitoring and early fault detection in a three-phase, motor-driven chemical/low-lift pump installation. High-resolution trend data comprising phase voltages, currents, total harmonic distortion (THD), active/apparent/non-active power, frequency, and power factor were logged at one-minute intervals over a representative operating window and processed to derive engineered health indicators, namely current and voltage imbalance, average current and voltage THD, calculated power factor, and reactive power ratio. Statistical process control (SPC) limits were established for phase current imbalance, and unsupervised learning — principal component analysis (PCA) combined with k-means clustering — was applied to the engineered feature space to reveal distinct operational states. The analysis identifies a dominant steady-state operating regime, a high-load/imbalance regime, and a transient start-up/shutdown regime characterized by depressed power factor and elevated harmonic distortion. Correlation analysis further shows that current harmonic distortion rises sharply as active loading falls, consistent with light-load harmonic amplification in the driven induction motor. The results demonstrate that non-invasive electrical signature analysis, combined with simple statistical and unsupervised machine-learning techniques, can provide plant engineers with an interpretable, low-cost early-warning capability for pump health monitoring without additional vibration or acoustic instrumentation.

Keywords

Chemical Pump, Low-Lift Pump, Electrical Signature Analysis, Condition Monitoring, Fault Detection, Power Quality, Total Harmonic Distortion, Phase Imbalance, Statistical Process Control, Principal Component Analysis, K-Means Clustering.

How to cite this paper

Adeyinka, Olusegun Alfred, Alimi, Teslim Adekunle, Abdulazeez, Kabiru "Multi-Dimensional Electrical Signature Analysis for Performance Monitoring and Fault Detection in Industrial Chemical Pump Systems" Iconic Research And Engineering Journals Volume 10 Issue 2 2026 Page 996-1004
Adeyinka, Olusegun Alfred, Alimi, Teslim Adekunle, Abdulazeez, Kabiru "Multi-Dimensional Electrical Signature Analysis for Performance Monitoring and Fault Detection in Industrial Chemical Pump Systems" Iconic Research And Engineering Journals, vol. 10, no. 2, Aug. 2026
Adeyinka, Olusegun Alfred, Alimi, Teslim Adekunle, Abdulazeez, Kabiru (2026). Multi-Dimensional Electrical Signature Analysis for Performance Monitoring and Fault Detection in Industrial Chemical Pump Systems. Iconic Research And Engineering Journals, 10(2).
Adeyinka, Olusegun Alfred, Alimi, Teslim Adekunle, Abdulazeez, Kabiru "Multi-Dimensional Electrical Signature Analysis for Performance Monitoring and Fault Detection in Industrial Chemical Pump Systems" Iconic Research And Engineering Journals, vol. 10, no. 2, Aug. 2026.
@article{1722203,
      author = {Adeyinka, Olusegun Alfred, Alimi, Teslim Adekunle, Abdulazeez, Kabiru},
      title = {Multi-Dimensional Electrical Signature Analysis for Performance Monitoring and Fault Detection in Industrial Chemical Pump Systems},
      journal = {Iconic Research And Engineering Journals},
      year = {2026},
      volume = {10},
      number = {2},
      pages = {996-1004},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1722203.pdf},
      abstract = {Unplanned downtime of chemical dosing and low-lift pump systems remains a major source of production loss and safety risk in industrial and municipal process plants. This paper presents a multi-dimensional electrical signature analysis (ESA) framework for continuous performance monitoring and early fault detection in a three-phase, motor-driven chemical/low-lift pump installation. High-resolution trend data comprising phase voltages, currents, total harmonic distortion (THD), active/apparent/non-active power, frequency, and power factor were logged at one-minute intervals over a representative operating window and processed to derive engineered health indicators, namely current and voltage imbalance, average current and voltage THD, calculated power factor, and reactive power ratio. Statistical process control (SPC) limits were established for phase current imbalance, and unsupervised learning — principal component analysis (PCA) combined with k-means clustering — was applied to the engineered feature space to reveal distinct operational states. The analysis identifies a dominant steady-state operating regime, a high-load/imbalance regime, and a transient start-up/shutdown regime characterized by depressed power factor and elevated harmonic distortion. Correlation analysis further shows that current harmonic distortion rises sharply as active loading falls, consistent with light-load harmonic amplification in the driven induction motor. The results demonstrate that non-invasive electrical signature analysis, combined with simple statistical and unsupervised machine-learning techniques, can provide plant engineers with an interpretable, low-cost early-warning capability for pump health monitoring without additional vibration or acoustic instrumentation.},
      keywords = {Chemical Pump, Low-Lift Pump, Electrical Signature Analysis, Condition Monitoring, Fault Detection, Power Quality, Total Harmonic Distortion, Phase Imbalance, Statistical Process Control, Principal Component Analysis, K-Means Clustering.},
      month = {August},
  }