Commissioning Reliability for Critical Facilities: A Root-Cause and Preventive-Maintenance Model for Automation, Power Distribution, and Electronic Infrastructure

@article{1719879,
      author = {Godsave Archford Sajanga, Gladman Nhamoinesu Machekera, Munashe Naphtali Mupa},
      title = {Commissioning Reliability for Critical Facilities: A Root-Cause and Preventive-Maintenance Model for Automation, Power Distribution, and Electronic Infrastructure},
      journal = {Iconic Research And Engineering Journals},
      year = {2026},
      volume = {10},
      number = {1},
      pages = {2670-2687},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1719879.pdf},
      abstract = {Critical facilities depend on tightly coupled automation, power distribution, backup power, communication, sensor, electronic-control, and human-interface systems. Where such systems are commissioned without durable performance validation, issue-backlog discipline, root-cause traceability, spare-parts readiness, and preventive-maintenance feedback loops, the facility may become operational on paper while remaining fragile in practice. This paper develops a Commissioning Reliability and Preventive-Maintenance Model (CR-PMM) for automation, power-distribution, and electronic infrastructure in critical facilities. The model links commissioning readiness checks, functional validation, issue prioritisation, fault isolation, condition monitoring, failure-mode classification, spares planning, corrective-action verification, and stakeholder risk communication into one auditable lifecycle. The empirical section uses the public AI4I 2020 Predictive Maintenance Dataset, distributed through UCI and Kaggle, comprising 10,000 machine-operation records, five principal failure flags, and operating variables such as temperature, speed, torque, tool wear, and machine-failure status (Matzka, 2020; UCI Machine Learning Repository, 2020; Kaggle, 2020). Although the dataset describes a synthetic industrial process rather than a specific facility, it is appropriate for testing transferable maintenance-analytics logic because critical facilities face similar problems of class imbalance, early warning, threshold exceedance, component degradation, and the need to convert sensor signals into actionable work orders. The analysis finds that only 339 of 10,000 records (3.39%) record machine failure, confirming a severe class-imbalance condition typical of high-reliability infrastructure. Heat dissipation failure, power failure, and overstrain failure are the dominant failure flags, while failed records show materially higher torque, tool-wear, power-proxy, and wear-torque values than non-failed records. A random-forest benchmark achieves balanced accuracy of 0.897, recall of 0.800, F1-score of 0.805, and ROC-AUC of 0.984 on hold-out data, showing that predictive-maintenance signals can be translated into commissioning and O&M decision support when combined with disciplined engineering governance. The paper contributes a practical lifecycle framework, a data schema, a root-cause taxonomy, heat-map visualisations, model-performance evidence, and a preventive-maintenance control matrix suitable for U.S. critical-facility contexts.},
      keywords = {Commissioning Reliability, Critical Facilities, Preventive Maintenance, Root-Cause Analysis, Power Distribution, Automation Systems, Electronic Infrastructure, Predictive Maintenance, FMEA, AI4I 2020 Dataset.},
      month = {July},
  }