Exact Simulation And Improving Methods Of Complex Power System

With the structure of modern power system becomes more and more complex, the classical analysis methods and equipment models are unfit to this condition. Change is inevitable. Two contents are discussed in this paper. One is the necessary and feasibility on replacing the mechanic-electrical transient simulation by electric-magnetic transient simulation. The other is on the improvements of the modeling of load and generator.Detailed expressions are shown as follows:1. This paper proposes using ATP-EMTP as the simulation tool for large power system stability calculation on the base of analyzing development condition, trend and advantage and disadvantage of current simulation software. Feasibility is verified by calculating the power flow and transient stability of WSCC-9 buses system and North Jiangsu 220kV and above power system. Comparisons with the results from PSASP, the power flow results are similar and the qualitative conclusions of transient simulation are the same too. Furthermore, equipments model selection is discussed and some useful conclusions were obtained.2. This paper presents a general modeling method of impact loads. Test results verify the validity of this modeling method. After that, the influence of large impact loads on the power system is discussed. Simulation results show that this modeling method can be used to evaluate the impacts of existing or planned nonlinear loads, and select the insertion points of new impact load.3. This paper proposes using immune algorithm (IA) as a parameter identification method of compositive motor model. Some modification is implemented in order to fit the characteristics of load parameter identification problem. Case studies show that IA method searches more quickly and accurately. The improvements about the memory pool strengthen this advantage. Fuzzy immune network is introduced as load classifying method, case studies showed that this method is good at dealing bulk samples and unclear boundary samples.4. This paper presents a generator model considering saturation mutual inductance by adaptive correction neural network. It is easily to inject into power system simulation software and its dynamic compensative ability increases the simulation accurate. Case studies show the validity and feasibility of this method.