Topological classification of small DC motors

In this paper we propose a new methodology based on signal embedding and applied topology for studying real noisy signals. Even if signal embedding is a useful tool, it is not sufficient for studying long range noisy signals. We argue that embedded signal in ℝ^m space can be properly analysed with topology based techniques. We obtained numerical evidences that our procedure properly classifies small DC motors into good/faulty, using the vibration data acquired from the bench. Small DC motors are largely employed in automotive to drive fans in HVAC (Heating, Ventilation and Air Conditioning) systems. Because of their high-speed, they can produce noise and vibration, perceived by the final users as a lack of quality. This problem is becoming relevant with the advent of the hybrid vehicles, in which the electrical motor is completely silent if compared to a traditional internal combustion engine, so all other noise and vibration source become noticeable by the passengers and felt as discomfort or a nuisance.