Electrical equipment incipient faults simulation using electromagnetic field emission

Abstract

Predicting failures of electrical equipment is an increasingly important topic in wide range of settings including industry as well as household appliances. The protection from failures can be achieved through early preventive actions while faults are in early stages. Identification of such incipient faults needs a thorough study and understanding of their patterns. Several methods exist for collecting necessary incipient fault data for further analysis and extraction of those patterns. One approach is to use the Electromagnetic Fields (EMF) emission from electrical equipment as a signature of incipient fault. However, this parameter requires conducting measurements to capture the EMF emission from equipment experiencing one or more incipient faults, which might require interrupting or damaging the equipment or parts of it, potentially leading to hazardous consequences. Another method is to design an antenna that generates faulty signals like those of equipment experiencing real incipient faults. However, this method requires detailed knowledge about the faults and the EMF emission propagation mechanism and is not in favor due to cost and design complexity. Hence, in this paper, we propose to simulate incipient faults based on a derived model of the EMF emission of a normal operating equipment. To generate the simulated signals, the coefficients of the model transfer function are systematically varied to yield different incipient fault patterns. The results are validated with two measured incipient faults, namely, inverter and winding faults and it was found that the correlation between the simulated and actual faults was significant which indicates that this method is viable as another source of data for incipient fault analysis.