Performance optimization of generator in steam turbine power plants using computational intelligence techniques

A generator is the crucial subsystem of steam turbine power plants. Its configuration is very complex, as it is assembled using seven different subsystems. The key objective of the present investigation is to develop an efficient stochastic model for generators under the concepts of cold standby redundancy and exponentially distributed failure and repair laws. The subsystem, namely the cooling and exhaust units, has the provision of cold standby redundancy. For this purpose, a novel stochastic model is proposed using the Markovian methodology, and Chapman–Kolmogorov differential–difference equations are derived. The switch devices are considered perfect, and units after repair work are considered new. To predict the optimal availability and profit of the proposed model, computational intelligence techniques, namely grey wolf optimization, whale optimization algorithm, moth-flame optimizer, dragonfly algorithm, grasshopper optimization algorithm, sine cosine algorithm, black hole algorithm, and ant lion algorithm are used. The impact of various numbers of iterations and population sizes is investigated on the availability, profit, and decision variables of the generator unit. It is revealed that the whale optimization algorithm predicts optimal availability of 0.9999905 after 10 iterations, while in a particular case, the optimal profit is 7199.924. The derived expressions of failure and repair rates, availability, and profit function are useful for system designers and maintenance engineers to design and plan maintenance strategies for generators and steam turbine power plants.