Study On The Methods Of Power Load Modeling Considering Time-variant Characteristic

The power load is an important part of power system. In modern power system studies, digital simulation based on mathematical models of each component has became the major scientific means to planning, scheduling, operation and decision, the model directly affects the simulation accuracy of the results. Now, as an important model in power system, load modeling study has been popular concerned by scholars all over the world. The thesis studies the load modeling methods and theories based on actual electric power network. At first, it introduces the basic theory on load modeling, including identification theory, how to choose model and deal with data,which is the base of the thesis. Then the thesis adopts the ways of Component-Based Modeling Approach and Measure-Based Modeling Approach for load modeling.Based on the fundamental of component-based method in load modeling, the accurate synthesis load static characteristic parameters of substation are obtained by combining analysis of the typical daily load curve. Get the investigation data of the typical consumer's load characteristics on the premise of the rational partition to load consumption industries, synthesis load characteristics of various industries are confirmed by statistical synthesis theory. Next an analysis of the typical daily load curve of the substations to different season, the load composition in each substation can be obtained; Finally the synthesis load static characteristic parameters are obtained take the synthesis load characteristics of industry and the load composition in each substation as foundation. Load modeling practice shows that the method is feasible and effective.Accurate load modeling is a particularly difficult task in dynamic modeling of synthesis load, the most difficult aspect is that the load is time-variant. In view of current research and application problems of load modeling, the classification and synthesis of power load dynamic characteristics is the necessary step of load modeling and an effective way of resolving the random time-variant problem and area decentralization problem. We propose the improved Fuzzy C-Means clustering algorithm as a clustering method of load dynamic characteristics. A classification example illustrate that the algorithm automatically determine the optimal number of clusters by the validity function, avoid the subjectivity; As premise of the classification, parameters weighting-mean synthesis is used in power load dynamic characteristics synthesis. The practical modeling results show that the classification and synthesis means of power load dynamic characteristics are validity.