A random optimization strategy of microgrid dispatching based on stochastic response surface method considering uncertainty of renewable energy supplies and load demands

Abstract

The stochastic response of microgrid regulation under the influence of uncertainty should be considered in the day-ahead optimal dispatching. This paper focuses on the Stochastic Response Surface Method (SRSM) modelling and Second-Order Cone Programming (SOCP) optimal solution for the stochastic optimization strategy of microgrid dispatching considering the random fluctuations of renewable energy supplies and load demands. Based on the SRSM theory, the random distributions of uncertainty are converted into independent standard normal distributions by Nataf transformation, then by means of a small amount of standardized SRSM samples of random fluctuations, the Hermite Chaotic Polynomials can be formulated to describe the random response process of microgrid adjustment. And the Hermite Chaotic Matrix establishes the linear constraint functions for the probability distribution characteristics of stochastic response process of microgrid adjustment considering uncertainty. On this basis, the SRSM based stochastic optimization (SO) model with multi-objective functions are constructed for the optimal economic operation, control of cost fluctuation and lower carbon emissions. In addition, to reduce operation risk, the constraints of extreme power shortage are introduced into the model. To ensure the convexity of the optimization model, the Second-Order Cone Relaxation is applied to all the quadratic terms in the model. Thus, the proposed SO model of microgrid dispatching can be transformed into an SOCP optimization problem. And the Yalmip-Gurobi solver is adopted for the solution of the proposed SO model, which has an efficient operation speed and stability. The effectiveness of the proposed scheme is demonstrated by the case studies using Monte Carlo sampling simulation.