含分布式发电接入的农村电网多目标规划

分布式发电（distributed generation，DG）接入对农村电网损耗、电压及可靠性有很大影响，为了充分发挥DG接入的经济技术效益，提出了一种多目标分布式电源规划方法。该方法以设备投资成本、系统有功损耗、停电损失及购电费用4个指标最小为目标函数，利用判断矩阵获得各目标函数权重，通过加权将多目标优化转化成单目标优化问题，采用改进遗传算法实现了DG位置及容量的优化配置。为了避免因DG可选布点太多、导致算法计算速度慢的问题，根据配电网损耗、电压及可靠性指标改善效果，提出了一种实用的确定DG候选位置的方法。IEEE33节点系统仿真结果表明，本文提出的DG规划模型和求解方法能够有效提高农村电网的投资效益和性能指标。

Multi-objective planning of rural power network incorporating distributed generation

Access of DG to rural power network is closely related to the power energy loss, voltage quality and reliability of rural power network. Therefore, the optimal allocation of DG has important significance for economic and safe operation of rural power grid. In order to make full use of DG，a multi-objective planning method of DG in rural power network is presented in this paper, which takes location and capacity of DG as decision variables, the minimums of the equipment investment cost, system power loss, interruption cost and power purchasing cost as objective functions, power flow balances as equality constraints, maximum line current, upper and lower limit of node voltage, and capacity and quantity limit of DG as inequality constraints. The four objectives are graded according to their importance. The power loss directly reflecting the system operation economy is taken as the first grade, the equipment investment cost reflecting the planning scheme economy as the second grade, the reliability reflecting system operation safety as the third grade, and power purchasing cost as the fourth grade. The judgment matrix can be formed according to the index grade, and then index weights can be calculated by using Judgment Matrix method. The weights of equipment investment cost, system power loss, interruption cost and power purchasing cost are 0.2633, 0.5639, 0.1179 and 0.0549 respectively. The proposed multi-objective optimization is transformed into the single-objective optimization by weighting. The optimal allocation of DG is achieved by the improved genetic algorithm. Each chromosome consists of some gene segments, and each gene segment using 3 binary digits (corresponding to the 8 possible capacity types) represents DG capacity of each candidate location. Adaptive crossover and mutation is adopted, the crossover rate Pc and mutation rate Pm of each individual is adaptively adjusted according to individual fitness value. The individual with high fitness value will have lower Pc and Pm, and is easy to be retained, but the individual with low fitness value will have high Pc and Pm, and is easy to be damaged. In order to avoid the low efficiency of the algorithm caused by too many DG candidate positions, a practical method to determine the DG candidate position is put forward by analyzing the improving effects of the power losses, voltage and reliability. The power loss improvement rate η, means the ratio of the power loss after installing DG and the power loss before installing DG, is used to quantify power loss improvement degree. The voltage improvement rate δ, means the ratio of the voltage quality after installing DG and the voltage quality before installing DG, is used to quantify the voltage improvement degree. The more smaller η is and the more bigger δ is, the more better effect of DG is. The method to determine the DG candidate position is as follows: 1) Sort the power loss improvement rates of all nodes in ascending order，select the front 40% nodes as candidate nodes. 2) Sort the voltage improvement rates of all nodes in descending order, select the front 20% nodes as candidate nodes. 3) Add the load points requiring high reliability. 4) Add candidate nodes for heavy load branch according to the load size. 5) Combine the adjacent nodes selected by above steps. To examine this method’s practical applicability, IEEE 33 node system is used as an example for an empirical research. Two optimal schemes of DG are obtained by using the proposed method, the scheme 1 has larger network loss cost and power purchasing cost, meanwhile the scheme 2 has larger equipment investment cost and interruption cost. Planning engineer can select the optimal scheme based on practical planning objectives and requirements. Simulation results demonstrate that the proposed method is feasible and effective for the optimal allocation of DG in the distribution network, and can effectively improve the investment efficiency and performance index of rural power grid.