基于负荷预测和非支配排序遗传算法的人工相序优化方法

在对0.4 kV配电台区进行节能降损时发现，调整负荷的接入相序能够有效降低台区的线路损耗和三相负荷不平衡度。文中提出基于负荷预测和非支配排序遗传算法(NSGA2)的人工相序优化方法。首先，利用配电台区出口电流曲线替代法建立用户负荷模型。其次，基于历史数据使用Elman神经网络对调相日的台区内各用户日电量和出口三相电流进行预测。然后，基于预测数据综合考虑以线损最低和调相次数最少为目标函数，建立配电台区多目标相序优化数学模型，使用NSGA2对该模型进行求解，得到优化后各负荷接入相序。最后，通过对比安徽电网某配电台区调相前后的理论线损，验证本文所提方法的有效性。

Optimization Method for Artificial Phase Sequence Based on Load Forecasting and Non-dominated Sorting Genetic Algorithm

It is found that adjusting the access phase sequence of the load can effectively reduce the line loss and three-phase load unbalance during the process of saving energy and reducing loss in 0.4 kV distribution network. This paper proposes an artificial phase sequence optimization method based on load forecasting and non-dominated sorting genetic algorithm (NSGA2). Firstly, the user load model is established by using the outlet current curve substitution method for distribution network. Secondly, based on the historical data, Elman neural network is used to predict the daily electricity consumption of each user and three-phase outlet current in distribution network on the day of phase modulation. Then, a multi-objective phase sequence optimization mathematical model of distribution network is established based on prediction data, taking the minimum line loss and the least number of phase modulation as the objective function. The NSGA2 is applied to solve the model to obtain the optimized phase sequence of each load. Finally, the effectiveness of the proposed method is verified by comparing the theoretical line loss before and after phase sequence adjustment in one distribution network of Anhui power grid of China.