Research On Evaluation Of Power System Reliability Based On Maintenance State

As China’s power grid construction continues to accelerate, the grid has become increasingly resilience, the structure more complex. In order to ensure a safe and reliable power grid, dispatch department should arrange a suitable maintenance plan for each component. During component maintenance, as the grid operates under maintenance state, redundancy decreased causing severe challenge to grid reliability. Most of the major chain blackouts in recent years are a result of reduction of redundancy caused by maintenance.Therefore, the reliability evaluation of grid under maintenance state are attracting increasing attention from the dispatchers.This essay looks into the characteristics of grid maintenance, proposes the difference between the reliability evaluation of grid under maintenance and traditional reliability evaluation in the following points:The reliability of grid maintenance state is a form of operational reliability. Using the static failure rate of traditional reliability evaluation to evaluate reliability of grid both before and during component rate, will result in relatively large deviation in the rate of sampling state, thus affecting the credibility of reliability index; On the other hand, the reliability evaluation of maintenance state has to be acquired in a relatively short period, with a relatively high accuracy. It will be difficult to acquire either of these using the traditional reliability evaluation method.According to these two characteristics, this essay presents a fault rate prediction algorithm using support vector machine and a system fault status quick sort algorithm, combined with the load shedding strategy based on sensitivity of power flow, forming the process of reliability evaluation of grid under maintenance state. The prediction algorithm based on support vector machine provide revised data for the system state fast sorting algorithm according to the changes of weather and power flow. The underlying data of reliable indicators provided by load shedding strategy based on sensitivity of power flow. The end of the essay validate the scientific and validation of the algorithm by using data form LanZhou electric power company.