The inlet air system is critical to the performance and reliability of gas turbines. Particle accumulation, environmental variation, and filter degradation can increase pressure drop, reduce performance, and raise maintenance costs. However, conventional filter management relies on inspections, expert judgment, and indicators such as pressure drop, without integrating filter health, system performance, environmental conditions, and economic considerations. This study reviews applications of artificial intelligence (AI) and machine learning (ML) in condition monitoring, performance assessment, and filtration management of gas turbine inlet air systems. The review identifies a gap in integrating filter health, system performance, sustainability, and economic considerations into a unified decision process. Accordingly, a conceptual framework for intelligent filtration management is proposed, incorporating four indices: the Filter Health Index (FHI), Air Filtration Performance Index (AFPI), Smart Inlet Air Health Index (SAIHI), and Filter Economic Index (FEI). These indices integrate performance, environmental, and economic factors. AI/ML methods are considered tools for data analysis, condition assessment, and decision support, rather than a trained and validated model in this study. The indices are preliminary and have not been field-validated. Their coefficients, thresholds, and performance require real data and field studies. The framework provides a basis for data-driven decisions regarding operation, maintenance, replacement, or filter-class upgrading. |
