Agile diagnostic tool based on electrical signature analysis

Abstract

AbstractToday, time based maintenance is the most common strategy, to assure a certain level of productivity and reliability in the industrial plants, with a fixed interval length between two maintenance actions. However, given the stochastic nature of the faults, plant equipments and machines usually fail unexpectedly, before the scheduled maintenance action. Therefore, condition based maintenance (CBM) has been introduced, without any fixed interval length between the maintenance actions, as a more efficient maintenance policy. A technique developed for condition based maintenance of electro-mechanical systems is electrical signature analysis. The basic idea behind electrical signature analysis is that, time-dependent load and speed variations in an electro-mechanical system generally induce correlated small variations in the system's electrical response. Electrical signature analysis analyzes these small variations providing diagnostic information. A tool based on this technique is agile in its adoption since no specific sensors must be installed on the machine, but only power signal must be measured and thus the implementation of such tool can be considered simple and flexible. This paper presents a tool that has such kind of characteristics. The tool processes power signals acquired by transducers installed on the electrical lines of electro-mechanical systems, for machine fault diagnosis. The data acquired are processed through some steps, using some signal indicators developed to understand the machine behaviour.