Reliability design of composite generation and transmission system based on Latin Hypercube Sampling with GRNN state adequacy evaluation

The crux of reliability engineering is to analyze the causes of failures, identify the consequences of failures and building of reliable systems by utilizing reliability basic design concepts. Optimal reliability level of the power utility depends on the constraints such as cost factors, available resources, economic benefits and profitability of such exercises. Thus optimization techniques have to be applied while designing reliable systems. This paper presents an approach based on Latin Hypercube Sampling with state adequacy analysis using Generalized Regression Neural Network (GRNN), and Particle Swarm Optimization (PSO) for determining the optimal reliability parameters of the composite generation and transmission system. The cost-benefit based design model has been formulated as an optimization problem of minimizing the system interruption cost and the component investment cost. The design model requires the analysis of several reliability levels which need to evaluate Expected Demand Not Supplied (EDNS) index for those levels. This approach reduces the computational burden for EDNS evaluation by applying GRNN for state adequacy analysis of the sampled states. The optimal reliability design model which has non-linear objective function and constraints is solved using PSO algorithm. Case studies carried out for Modified Stagg & El-Abiad 5-bus system and IEEE 14-bus system.