Optimal energy management of microgrids‐integrated nonconvex distributed generating units with load dynamics

Abstract

Summary Microgrid (MG) energy management is a complex task for MG operators to integrate and utilize consumer‐based power sources. The MG energy management systems’ problem will become tedious by considering distributed generation (DG) units' nonconvex characteristics. Therefore, a novel attempt is made to solve the day‐ahead dispatch problem of grid‐connected MG with the nonconvex cost function of DG units, including weekend and weekday load dynamics. At first, the utility‐induced flexible load shaping strategy is implemented to enhance the DG units' operation cost and reduce the peak loads. Then, demand‐side management (DSM) programs are plausibly the essential form of energy management to regulate the consumers' energy usage without violating grid price policies. Next, the DSM program is implemented to study the impact of DSM participation levels with convex and nonconvex cost functions. Further, the day‐ahead scheduling time duration with a resolution of 15 minutes is considered to examine the impact of a typical weekend and weekday load dynamics on DG units' nonconvex cost function. Finally, the Quantum Teaching‐Learning‐Based Optimization algorithm (QTLA) is devised to handle the nonconvex cost function of DG units and optimize MG's total operating costs for the first time. The proposed QTLA algorithm is compared with other metaheuristic optimization techniques such as differential evolution (DE), real‐coded genetic algorithm (RCGA), and Teaching‐Learning‐based Optimization (TLBO). The results show that the proposed strategy reduces the MG operating cost by 3.14% compared to the case study, where no DSM participation is considered. Finally, the QTLA algorithm outperforms in terms of efficacy, convergence characteristics, and computational time.

Analysis of a linear programming‐based decision‐making model for microgrid energy management systems with renewable sources

“…CPLEX and CBC solvers are based on branch‐and‐cut algorithm, while IPOPT and BDMLP are based on interior‐point and branch‐and‐bound algorithm respectively 51 . CPLEX, the chosen solver to the problem as in References 5,9–14,20–35,40,50 has outperformed the remaining solving algorithms regarding the computation time with 12.5 seconds as depicted in the convergence diagram Figure 3. The worst scenario was obtained with the BDMLP algorithm which found the optimal solution after 17 minutes and 97 453 iterations.…”

“…The amount of energy in each electrical energy storage i at time t + 1 is the sum of the energy level of each energy storage at time t and the charging/discharging decisions at time t, as stated in Equation (32). 48 However, the energy level of each energy storage is bounded within a minimum and a maximum value [Equation (33)].…”

Hybrid renewable/nonrenewable/storage resources in electrical grid considering active‐reactive losses and depth of discharge

“…The ESSs and DGs have broad applications in the electrical distribution grids. 31 The study conducted in 32 presents a multi-objective nonlinear optimization which optimizes the siting and sizing of storage systems in the distribution systems. The model excludes depth of discharge (DOD) and reactive power of energy storage system (ESS).…”

A swarm intelligence approach for energy management of grid‐connected microgrids with flexible load demand response

“…The above state of the art summarizes various mathematical, 22 heuristic, 19 multi-agent-based, 14,15 and reinforcement learning-based 16 EMS strategies with a wide range of objectives. Furthermore, the quantum computational-based algorithms [28][29][30] are also adopted to resolve the problems of optimal scheduling in MG. In addition, applying various DR strategies unified with the EMS problem has enhanced operational costs.…”