Vacuum Circuit Breaker State Evaluation System Research

As an important control equipment of power system, vacuum circuit breaker has a very important significance for the safe operation of the power system. And as a form of status maintenance, vacuum circuit breaker on-line monitoring can predict the working state of the vacuum circuit breaker according to the indicators of on-line monitoring of power equipment and have a purpose maintenance, which can save resources, manpower and material resources.At first, the paper is based on predecessors’ research of the status indicators of power equipment on-line monitoring, combination with the characteristics of vacuum circuit breaker, so four states indicators including vacuum degree, points/closing coil current electricity, contact temperature and remaining life of four states are taken as indicators, and indicators feasibility, theoretical and specific methods were introduced.In the premise of the characteristics of the evaluation model determining, this paper designed sampling circuit of vacuum circuit breaker online monitoring and condition assessment which can implement acquisition, processing and analysis for vacuum circuit breaker online monitoring indicators. And using FPGA as the main control chip of acquisition circuit of vacuum circuit breaker realizes acquisition, processing and analysis of four online monitoring indexes for laboratory prototype. The experimental results show that the hardware circuit meets the experimental conditions, the index data are obtained.Four characteristics online collection of hardware acquisition circuit is completed, this paper realizes the classification of status indicators based on least squares support vector machines with hybrid kernel function and fuzzy theory, and parameter optimization of least squares support vector machines with hybrid kernel function uses PSO and GA.Through the simulation analysis, using PSO optimization based on hybrid kernel function algorithm of least squares support vector machine (SVM) classification accuracy rate was 98.5% into 200 groups of training set and testing set, but the GA- LSSVM based on hybrid kernel function classification accuracy is only 90%.so the PSO based on hybrid kernel- LSSVM function has higher reliability. The PSO- LSSVM algorithm model is applicable to solve problems with large sample and high dimension, and the model can be expanded with the constant improvement of characteristics.At last, the paper uses the fuzzy algorithm based on the optimal combination weights selecting gauss function as membership function to evaluate the vacuum circuit breaker condition, the simulation results also meet the requirements and achieve reliable classification. Two kinds of evaluation method can evaluate the state of vacuum circuit breaker:The fuzzy algorithm based on the optimal combination weights model is suitable for the need to introduce expert opinion to guide the evaluation modeling, modeling method using two models improves the faults of single modeling of original research and the adaptability of evaluation model, and provides a reference for intelligent vacuum circuit breaker condition assessment and maintenance. In the process of actual operation and we can select different algorithms according to specific circumstances.