基于改进粒子群算法的微电网多目标优化运行研究

发布时间：2019-05-23 13:10

【摘要】：随着全球能源危机加剧、环境的恶化以及大规模电力系统弊端的日益凸显，分布式发电受到越来越多的关注，微电网作为多种分布式发电的有效集成，充分发挥了分布式发电的优势，能够灵活地运行在并网状态或孤岛状态，是解决分布式发电接入大电网的有效技术手段，提高了供电的可靠性和安全性。然而，分布式发电的多样性以及组合的灵活性导致微电网在优化运行、能量管理和运行控制与保护等方面存在较多亟须解决的问题，其中，微电网的优化运行能有效提高能源利用率，减少发电成本和污染物的排放，对微电网系统的经济、环保、可靠运行有重要的意义，因此，研究微电网的优化运行具有重要的实用价值和理论意义。微电网的优化运行问题是一个复杂的多目标，多约束，多变量的非线性优化问题，智能优化算法以其优越的性能在微电网的优化运行领域获得了广泛应用，如遗传算法、免疫算法和粒子群算法等。其中，粒子群算法具有简单、鲁棒性强、精度高、收敛快等优点。本文针对多目标粒子群算法存在的问题进行改进，提出基于全局最优位置自适应选取和变尺度混沌局部搜索的改进多目标粒子群算法(Improved Multiobjective Particle Swarm Optimization Algorithm Based on GlobalBest Adaptive Selection and Mutative Scale Chaotic Local Search，IMOPSO-GL)，并采用ZDT系列标准测试函数和IEEE30节点电力系统无功优化问题测试IMOPSO-GL算法的性能。结果表明，IMOPSO-GL具有良好的收敛性、且搜索到的Pareto最优前端分布宽广且均匀。IMOPSO-GL算法主要改进的策略有： ①全局最优位置对多目标粒子群算法的收敛性和多样性有很大的影响，本文提出了全局最优位置自适应选取策略，在借鉴Sigma法原理的基础上，，采用动态零点技术修正Sigma值的计算，同时引入拥挤距离机制，限制档案粒子被选为全局最优位置的最大次数，最终使种群均匀且快速地向当前Pareto最优前沿飞行； ②外部档案用来保留算法在搜索过程中找到的非劣解，对算法的性能有很大影响，本文通过对外部档案设置最小规模，防止外部档案规模过小时，粒子过于聚集在某些区域；当档案规模超过最大规模时，采用循环拥挤排序策略删除外部档案多余成员，确保外部档案在整个进化过程中的多样性； ③由于多目标粒子群算法容易早熟收敛，本文提出变尺度混沌局部搜索策略，在种群寻优能力减弱时，从外部档案中选择部分特殊粒子进行局部搜索，以提高算法的收敛性。本文以包含光伏电池、风机、燃料电池、微型燃气轮机、柴油发电机以及储能系统的微电网为例，其中储能系统采用由蓄电池和超级电容组成的蓄电池超级电容混合储能系统。针对并网运行的微电网系统，建立了考虑经济收益与环境成本的微电网多目标优化运行数学模型，针对孤岛运行的微电网系统，建立了考虑发电成本与环保费用的微电网多目标优化运行数学模型，并考虑了正负旋转备用约束。在满足功率平衡、分布式电源出力约束、蓄电池和超级电容荷电状态和输出功率约束以及与配电网传输功率约束的前提下，利用模糊专家系统和IMOPSO-GL进行模型求解，通过算例仿真，验证了所建立模型的有效性。
[Abstract]:With the aggravation of the global energy crisis, the deterioration of the environment and the abuse of the large-scale power system, the distributed power generation has attracted more and more attention, and the micro-power grid is the effective integration of a variety of distributed power generation, and the advantages of the distributed power generation are fully realized, Can operate flexibly in a grid state or an isolated island state, and is an effective technical means for solving the distributed power generation access large power grid, and the reliability and the safety of the power supply are improved. However, the diversity of the distributed power generation and the flexibility of the combination lead to the problems that the micro-power grid has to solve in the aspects of optimal operation, energy management and operation control and protection, and the optimized operation of the micro-power grid can effectively improve the energy utilization rate, It is of great practical value and theory significance to study the optimization and operation of the micro-power grid, which is of great significance to the economic, environmental and reliable operation of the micro-grid system. The optimization and operation of the micro-power grid is a complicated multi-objective, multi-constraint and multi-variable nonlinear optimization problem. The intelligent optimization algorithm has been widely used in the optimization and operation of the micro-power grid with its superior performance, such as genetic algorithm, immune algorithm and particle swarm optimization. And the like, wherein the particle swarm algorithm has the advantages of simplicity, strong robustness, high precision, fast convergence, and the like. An improved multi-objective particle swarm optimization algorithm based on global optimal position adaptive selection and variable-scale hybrid local search is proposed in this paper. The improved Multiobjective Particle Swarm Optimization Algorithm Based on GlobalBest Adaptive Selection and Mutative Scale Chaotic Local Search is proposed. and the performance of the IMOPPSO-GL algorithm is tested by adopting a ZDT series standard test function and an IEEE30 node power system reactive power optimization problem The results show that the IMAOPSO-GL has good convergence and the Pareto optimal front-end distribution is wide and uniform. An improved strategy of the IMAOPSO-GL algorithm It is shown that the global optimal position has great influence on the convergence and diversity of the multi-objective particle swarm optimization algorithm. In this paper, a global optimal position adaptive selection strategy is proposed, which is based on the principle of the Sigma method. on the basis of the calculation of the sigma value by adopting the dynamic zero-point technique, the congestion distance mechanism is introduced, the maximum number of the file particles is restricted to the global optimal position is limited, the population is finally made to be uniform and fast, In flight, an external file is used to keep the non-inferior solution found in the search process, and has a great impact on the performance of the algorithm. in that case of the most large scale, the redundant members of the external file are removed by a circular congestion sort policy to ensure that the external files are throughout the evolution process In this paper, the local search strategy of variable-scale hybrid is presented in this paper, because the multi-objective particle swarm optimization algorithm is easy to converge. when the optimization capability is reduced, part of the special particles are selected from the external files for partial search, The convergence of the high algorithm is as an example of the micro-power grid which contains the photovoltaic cell, the fan, the fuel cell, the micro-gas turbine, the diesel generator and the energy storage system. Aiming at the micro-grid system running in grid-connected operation, a multi-objective optimal operation mathematical model of a micro-power grid considering the economic benefit and the environmental cost is established, and the micro-grid multi-objective optimization operation mathematical model which takes into account the economic benefit and the environmental cost is established, and the micro-grid multi-objective optimization operation model which takes into account the power generation cost and the environmental protection cost is established for the micro-grid system running on the island. The Mathematical Model of the Optimization of the Standard Operation and the Consideration Under the premise of satisfying the power balance, the distributed power output constraint, the charge state and the output power of the storage battery and the super capacitor and the transmission power of the power distribution network, the fuzzy expert system and the IMO-GL are used for the model. The solution is simulated and verified by an example.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM732

【参考文献】