Multi-objective Planning for Siting and Sizing of Distributed Generators Considering Correlations

A multi-objective planning model for siting and sizing of Distributed Generators (DG) in distribution network which considers the correlations among uncertain parameters as well as their volatility is proposed. These uncertain parameters including wind and solar irradiation can be viewed as random variables. The model has two planning objective functions. One objective function is to find the minimal value of network loss, which taking into account uncertain parameters. The other is to find the minimal cost of DG investment and operating. Considering that the correlations among actual random variables are uncertain and the probability distribution functions of random variables are difficult to be described accurately, this article references to the method of Multivariate Polynomial Normal Transformation and Latin Hypercube Sampling (MPNT-LHS) to generate sample matrix with correlations and randomness. Incorporate MPNT-LHS with Monte Carlo Simulation Method (MCSM) and Non-domination Sorting Genetic Algorithms-II (NSGA-II), to solve the proposed multi-objective planning problem and get the Pareto optimal solution sets. This article validates the model and method on the IEEE 33-bus distribution network example. The results of the simulation have demonstrated the model’s correctness and method’s effectiveness.