Short-term operation of a distribution company: A pseudo-dynamic tabu search-based optimisation

Abstract

This work presents a probabilistic sequential framework for short-term operation of distribution companies (DisCos) participating in the day-ahead (DA) and real-time (RT) markets. In the proposed framework, the DisCo's operating decisions are sequentially optimised; first, in a DA operation stage, and then in RT. The DA decisions are driven by the DisCo's profit maximisation, while the DisCo aims to minimise the actions required to accommodate deviations from forecasted quantities (i.e. the DA decisions) in the RT operation stage. This sequential approach considers realistic voltage-sensitive loads and full ac power flow equations to represent the realistic network's active and reactive power injections. In addition, the operation of stationary batteries and the demand elasticity under time-varying retail prices are explicitly modelled. The two resulting models are large-scale highly non-linear non-convex mathematical problems with continuous and discrete variables. A pseudo-dynamic tabu-search-based solution algorithm is used as an alternative to conventional optimisation solvers in order to tackle the problem in an effective manner, without linearisations. Numerical results from 69- and 135-bus distribution systems illustrate the performance and the scalability of the proposed approach.