Search

In allusion to the fact that the heavy and inaccuracy workload as well as the problems which are lack of accuracy and practical applicability of model based on conventional component-based approach load modeling, this paper applies the theory of fuzzy C means clustering and pattern recognition, with full use of the power information from consumer side collected by the load control system in power marketing department, to conduct the primary assortment for industry consumer and the choiceness for typical industry consumer. After that, the investigation and statistics of industry parameters of typical consumers should be performed. Associated the substation power information provided by the energy management system (i.e. EMS), from which the industry composition proportion of substation is obtained, with the classic characteristics of electrical equipment, a high efficient component-based approach modeling platform is developed by the proposed method. And now the platform is applied on Hunan power grid.

The accurate degree of power system dynamic simulation is directly related to the stable operation of power system, while the accuracy of load models have a direct influence on simulation results and final decisions. Based on PSASP/UPI, this paper develops an interface method linking the model to the transient stability program. By the transient stability program, load model can be linked to power system for dynamic simulation. Then the design idea of user program and realization steps of dynamic simulation are described. The interface principles between user program and transient stability program are showed. This paper connects induction motor synthetic load model with distribution network lumping to EPRI-36 bus system in PSASP for dynamic simulation, and the calculation results prove the correctness of the user program interface method. Therefore, the way to link load models to power system for dynamic simulation based on PSASP/UPI can be concluded.

The optimal configuration model based on 2-layer programming was built using chance constrained programming method. The upper optimization model determines the optimal location and capacity of DG-C while the lower optimization model determines the optimal switching strategy of C. An improved genetic algorithm which can be easy to handle discrete variables was chosen for the optimal solution. The simulation results show that the 2-layer optimized model has improved expected value of node voltage and the probability within normal range as well as obtained comprehensive optimization of economic and environmental benefit.