基于混合聚类算法的高压配电网网架规划

H-V distribution network planning based on hybrid clustering algorithm

以高压配电网为研究对象,建立了含建设运营、故障停电损失及资产回收等成本的全寿命周期成本规划模型,提出一种基于混合聚类算法的高压配电网规划方法。首先,根据规划区域的负荷分布情况,利用DBSCAN算法与K-means算法相结合的混合聚类算法对供电分区及变电站进行划分;然后,结合区域容载比的要求,确定各供电区待建变电站总容量;最后,利用Floyd-Warshall算法求解满足目标函数要求的变电站站址和接线形式,得到整个供电区域内的高压配电网规划方案。通过某实际规划区作为算例,验证了模型和方法的实用性和有效性。

Within the scope of the High-Voltage distribution network, a life cycle cost model, which includes construction cost, operation cost, outage cost, and assets recovery cost, was established. A planning method based on hybrid clustering algorithm was proposed. First, according to the spatial load forecast distribution in the planning area, a hybrid clustering algorithm that combined the DBSCAN algorithm and K-means algorithm were used to solve the division of power supply area and substation. Then, considered the capacity and the load ratio in each supply area, the capacity of all the substations had been determined. Finally, the Floyd-Warshall algorithm was used to solve the model and got the scheme of substation site and wiring in the whole area with the constraint of the objective function. The practicability and effectiveness of the model and method are verified through the actual planning area.