Fault diagnosis of vacuum cleaner motors

Abstract

In-depth and automatic quality end-tests in modern manufacturing represent an important means for the assurance of top-quality and flawless products. Tough competition on the market of vacuum cleaner motors is increasing the need for fast, reliable and objective quality assessment of every single unit at the end of the assembly cycle. As a step towards meeting these objectives, a prototype version of the diagnostic system for quality tests of vacuum cleaner motors has been designed. The core of the system contains four modules for features extraction that employ, respectively: analysis of commutation, vibration analysis, sound analysis and check of parity relations. The symptoms resulting therefrom are processed by an approximate reasoning module, which utilises the technique referred to as the transferable belief model (TBM). The comprehensive diagnostic procedure is able to clearly distinguish a faulty motor from a non-faulty one and to infer about the tentative fault location. Main contributions of the paper refer to the novel feature extraction procedures, which provide a reliable estimate of the motor's condition. The system performance has been surveyed on a set of about 100 motors subjected to a detailed experimental study. An excerpt is also presented, reflecting the key properties of the diagnostic system performance, such as precision, accuracy, robustness and reliability.