Regressão aplicada na estimativa do comportamento térmico de um motor de indução trifásico

Abstract

The polyphase induction machines have a huge application in several production field of society, specially the squirrel cage induction motor due to its robustness, its low cost and reliability. However, certain motors often work above the operational condition in which they were specified. In this sense, it is worth mentioning the necessity in knowing the rise internal temperature of induction motors for determined applications, since its overheating causes not only the reduction of the insulation life, it can also cause a reduction of machine’s efficiency. some techniques search for finding the temperature rise on stator winding invasively, such as by installing thermal sensors in huge machines, by the utilization of DC injection, thermal modeling and through machine thermal rise tests. In accordance with showed above, the proposed work offers the utilization of a regression algorithm to estimate, non-invasively, the rise temperature in stator of a polyphasic induction motor by using only easy acquisition magnitudes, such the phase current and the surface carcass temperature. For this purpose, one regression algortihm was considered, being the normal equation. the experimental temperature data, collected from a 7.5 kW polyphasic induction motor which is installed in laboratory, were compared with those estimated by the proposed regression algorithm. the found results show a good accuracy reached to the statistical models, suggesting that the approach can represent an interesting non-invasive alternative, in order to estimate the internal rise temperature of a polyphasic induction motor.