自适应模糊系统理论在负荷预测中的应用研究

Application Research of Adaptive Fuzzy System Theory in Load Forecasting

应用自适应模糊系统理论的最新成果实现实用化的预测系统，并以电力电负荷预测为具体应用背景完成了实验研究。系统的建立和运行分别依赖于对历史数据和实时数据的离线和在线学习，具有明显的自适应性和鲁棒性。通过合理的设计实现简洁的系统结构，通过在线训练确定优化的系统设置，短期负荷预测的日均相对误差小于２％，可以满足现场的实用化要求。

This paper discussed application problems of adaptive fuzzy system theory in power load forecasting. The establishment of the system was based on off-line study form historical data and its running was depended on adaptive study from real time data. Thus, the system is provided with extinguished adaptive feature and self-learning capability. Simple system structure was achieved by reasonable design and optimal parameters were obtained by on-line training. Experiment result with daily short-term load forecasting relative error less than 2% demonstrates that the intelligent system can meet applying engineering demand.