用改进的MOEA/D算法求解微电网电力调度多目标优化问题

Multi-Objective Optimization Problem of the Microgrid Power Dispatch Using Improved Multi-Objective Evolutionary Algorithm Based on Decomposition

针对包含可再生能源发电装置,微型燃气轮机(microturbine,MT),燃烧电池(fuel cell,FC)以及镍氢电池(nickel-metal hydride battery,Bat)的典型低压微电网电力调度多目标优化问题,文章提出一种改进的基于分解的多目标进化算法IMOEA/D(improved multi-objective evolutionary algorithm based on decomposition),最小化发电成本和污染物气体排放量.IMOEA/D算法将多目标问题按照切比雪夫权重射线均匀展开分解为若干子问题进行优化;基于Pareto占优概念以及模糊聚类策略更新子问题非劣解,引入基于欧氏距离的解的稀疏度评估法对非劣解保留集进行更新和对其储存空间大小进行控制.最后,将提出的算法与其它进化算法如GA(遗传算法)和AMPSO(自适应改进粒子群优化算法)等分别进行单多目标性能比较,结果表明,IMOEA/D算法在多目标优化性能比较中能较稳定的得到质量更高,范围更广,分布更均匀的非劣解保留集,即该算法在解搜索的深度和广度上都有较好的表现.

Microgrid is an effective way to increase the efficiency of using energy and develop and utilize clean and renewable energy because it can distribute power more independently,efficiently,economically,and safely.The microgrid power dispatch refers to the rational allocation of power generation of different power generating units in the microgrid under the relevant constraints to meet the power load requirements.Therefore,the microgrid power dispatch optimization problem has attracted more and more attention of scholars,and research on the microgrid power dispatch has important significance of economy and environment.In the first part,for the multiobjective optimization problem of power dispatch in a typical low-voltage microgrid which includes renewable energy distributed generation consisting of Photo Voltaic and Wind Turbine,this paper establishes a general multi-objective model about a day24 hours power dispatch in microgrid and proposes an improved decomposition-based multi-objective evolutionary algorithm(IMOEA/D) to minimize Simultaneously generation costs and polluted gas emissions.In the second part,this article constructs the basic framework of the IMOEA/D.In order to improve the performance of the original MOEA/D and solve the multi-objective power dispatch optimization problem of the microgrid,the IMOEA/D algorithm optimizes the multi-objective problem according to the Chebyshev weight ray uniform decomposition into several sub-problems;Based on Pareto dominant concept and fuzzy clustering strategy updating non-dominated solution set of subproblems,introducing a sparsity assessment method based on the Euclidean distance solution to update the non-dominanted retention solution set and control its storage space size,and designing a program into the IMOEA/D to correct new solution to a feasible solution for the model.In the third part,in order to evaluate the performance of the IMOEA/D,this paper introduces a test case and runs the algorithm to obtain the simulation result and compare the result with the simulation results from the other algorithms for single-target and multi-target performance.The statistical results show that the IMOEA/D performs well both in single-objective performance and multi-objective performance.Especially in the performance of multi-objective,the solution of the non-dominanted solution retention set obtained by the IMOEA/D has the advantages of higher quality,more even and broader distribution than others.So the algorithm has a good performance in the depth and breadth of the solution search.Therefore,the IMOEA/D proposed in this paper can effectively solve the multi-objective power dispatch optimization problem of the microgrid,provide various choices for the power system operators,and formulate an appropriate power dispatch plan based on comprehensive environmental and economic considerations.