电网的快速发展给电力负荷建模带来了巨大的挑战，为了克服其全面性、时效性、准确性和自适应性四个方面的问题，本项目拟以交互多模型算法为基本理论工具，建立一个新的广义电力负荷在线建模理论体系。在建立子模型的环节，先考虑新能源、电动汽车等新成分的影响，提出一个新型的广义电力负荷子模型结构，再依据离线信息，采用人工蜂群优化算法进行参数辨识，从而全面描述电力负荷的基本特性；在建立在线模型的环节，先利用在线信息对各个子模型进行输入交互、条件滤波、概率更新和融合估计，再依概率将这些子模型折算为标准的广义电力负荷模型以逼近真实负荷特性，在减少参数辨识环节以提高建模的时效性的同时，能通过不同模型的概率变化描述电力负荷的强非线性离散变化特性以提高其准确性；在子模型的调整环节，先根据实时信息对各个子模型与真实负荷模型的匹配程度进行评价，再进一步实现子模型的在线动态增减，从而提高模型在不同应用条件下的自适应性。

Electric load modeling is facing a huge challenge due to the rapid development of power grid. In order to overcome the problems of electric load modeling methods in comprehensiveness, timeliness, accuracy and adaptability, this project intends to set up a novel generalized electric load online modeling theory based on interactive multiple model algorithm. In the link of sub-model construction,a new generalized electric load model structure is proposed taking into account the impact of renewable energy, electric vehicle and other new ingredients.Then the parameters are identified using the offline information and artificial bee colony approach. In this way, the basic characteristics of generalized electric load are descrided comprehensively. In the link of online model construction,the input interaction, conditional filter, probability updating and fusion estimation of sub-models are executed employing the online information.Then the sub-models are combined according to their respective probabilities. This method not only neglecte the online parameter identification and improve the modeling speed, but also describe the nonlinear discrete changes of electric load and enhance the modeling accuracy. In the link of sub-model adjustment, the matching degree of each sub-model and the real electric load is assessed according to the real-time information of application environment. Then the sub-models are dynamic increased or decreased so as to improve the adaptability of generalized electric load model.