Artificial Intelligence Clustering and Discrimination of Power Equipment Fault Types Based on Image Recognition and Big Data Analysis Method

Paper tested error of piezoelectric accelerometer, and the error did not exceed 1 %, meeting the objectives of the countermeasure table. Selecting the set of typical defects with poor contact of the HV equipment isolation switches to test the software and hardware of whole machine. The charge amount reached its maximum value at 5000s, and then showed decreasing trend. Under action of an external electric field, uneven materials exhibit polarization, known as the Maxwell Wagner polarization (MWS polarization). The maximum field strength of the 4-layer pressure ball coating on the outlet device should also be as equal as possible to ensure the uniformity of E-field distribution. Pareto solution set contains 175 feasible solutions, randomly selected solution set includes 3000 elements. The figure shows that Pareto optimal solution distribution of the first front end is uniform, all of which are regions of selected points. Distribution conforms to idea of NSGA-II multi-objective optimization algorithm.