Johnson’s SU distribution using Gray Wolf Optimizer algorithm for fitting gas turbine reliability data

Controlling failures and degradation of gas turbines is crucial for optimizing efficiency, productivity, and maintaining safe operations in the oil and gas industry. Reliability indices play a vital role in supporting these goals by enabling informed decisions about gas turbine lifespan extension and operational safety. This study proposes a novel approach to estimate reliability indices for a GE MS5002C gas turbine. It leverages the Johnson SU distribution applied to operating data and optimizes the obtained model using the Gray Wolf algorithm to improve prediction accuracy. We compare the proposed method with the three-parameter Weibull distribution to validate its effectiveness. By employing the Johnson SU transformation alongside the Gray Wolf Optimizer, this work offers a more accurate and robust method for determining reliability indicators. This approach, based on survival analysis, unlocks the full operating potential of the turbine while addressing uncertainties and errors in reliability modeling. Consequently, it allows for enhanced control of failure sources throughout the turbine's life cycle, ensuring availability and minimizing environmental impact.