基于城市用地性质及开发时间的改进空间负荷预测方法研究

Research on Improved Spatial Load Forecasting Method Based on Land Use and Development Time

传统的空间负荷预测方法涉及数据量大、模型复杂度高、预测速度及精度难以保证。考虑城市用地性质及开发时间,提出了一种基于Logistic回归的改进空间负荷预测方法。该方法将负荷预测问题转换为网格化分区的参数训练和模型的整合预测两个问题,在预测模型中重新构建基于发展速度以及中位年份的参数训练方案,并通过改进Logistic函数的最大似然估计来降低负荷预测的复杂度。仿真结果验证,提出的改进空间负荷预测方法在简化计算复杂度和提升预测精度两方面都优于传统方法,能够有效应用于片区负荷预测。

Traditional spatial load forecasting method involves large data volume,high model complexity,and difficult prediction accuracy.In this paper,considering land use and development time,an improved spatial load forecasting method based on Logistic regression is proposed.Original power load forecasting is transformed into the parameter training of the gridding partition and the integrated forecasting of the model.In the prediction model,a parameter training scheme based on the development speed r and the median year t0 is reconstructed.Based on the maximum likelihood estimation of improved Logistic function,a load forecasting method with low complexity is proposed.Simulation result shows that the proposed improved method is superior to the traditional method in simplifying computational complexity and improving prediction accuracy.The proposed method can be effectively applied to region load forecasting.