Calculation of the circuit breaker’s opening angle based on Deep Forest Regression

Magnetizing inrush current in transformer has rich harmonics and large amplitude due to core saturation when the transformer is energized. Large inrush current can have a negative impact on power equipment, system operation, and electrical protection systems. The key of suppressing the inrush current is to make accurate control of circuit breaker’s closing Angle which is determined by the opening angle. Therefore getting the accurate opening angle plays decisive effect in suppressing the inrush current. Due to the core saturation effect of the transformer and the interference signal, it is hard to determine the actual opening angle. However, in the process of measuring the residual magnetization and the calculating of the optimal closing angle, the breaker’s opening angle is needed. In the laboratory, we can obtain the actual opening angle through the reverse measurement of the residual magnetism after opening. The opening and closing data and the actual opening angle labels are obtained through dynamic simulation experiments. Deep forest is a recent deep learning framework based on tree model ensembles, which does not rely on backpropagation. We consider the advantages of deep forest models are very appropriate for solving problem of getting opening angle. Therefore we design the Deep Forest Regression(DFR) with 2 modules: completely random regression forest and random forest regression. It mainly solves two problems: 1. Obtain the actual opening angle from the recording and broadcasting waveform in the actual operation monitoring; 2. Maintain the Accuracy under the interference signal Higher accuracy.