Predictive Maintenance of High-Voltage GIS Assets for Saudi Smart Grids: Using Partial Discharge Monitoring, Equipment Diagnostics, and Failure-Risk Assessment under Vision 2030
Abstract
- High-voltage gas-insulated switchgear (GIS) is central to compact, reliable transmission substations, yet its apparent reliability can conceal insulation defects, mechanical degradation, gas-system anomalies, thermal problems, and ageing mechanisms that become costly when detected only after failure. This review develops a predictive-maintenance framework for GIS assets in Saudi smart grids by integrating partial-discharge (PD) monitoring, multimodal equipment diagnostics, health assessment, and failure-risk prioritisation. Evidence from 2020 to 2025 was synthesised through a structured integrative review of PD sensing and localisation, machine-learning diagnosis, health indices, switchgear condition monitoring, and Saudi smart-grid transformation. The synthesis shows that PD monitoring provides high-value early evidence but should not be treated as a standalone maintenance trigger because interference, sensor position, operating context, defect type, and model uncertainty can materially change interpretation. Stronger decisions emerge when PD evidence is fused with gas, thermal, mechanical, switching-duty, inspection and historical failure information. Health indices create an interpretable bridge between heterogeneous measurements and asset condition, while probabilistic failure-risk assessment adds consequence and criticality to maintenance ranking. Data-driven models can improve defect recognition, especially under complex signal conditions, but field deployment requires explicit treatment of scarce labelled data, domain shift, calibration, explainability, and confidence. The review proposes a closed-loop architecture that connects sensing to diagnosis, risk-ranked intervention, and post-maintenance learning. For Saudi Arabia, the framework supports Vision 2030 priorities by improving grid reliability, reducing forced outages, extending justified asset life, and strengthening local diagnostic and analytics capability.