DIAGFLU

The aim of this project is to investigate and develop new technologies that enable an early and reliable diagnosis of different types of failures in electric motors. These will be based on the analysis of the stray flux in the vicinity of the machine. This diagnosis technique has shown a high potential and is drawing an increasing attention among the academic and industrial communities due to its simplicity, low cost and to the continuous evolution in the technology of the flux sensors which give it a high versatility for its future implementation in the industry. The project is intended to go beyond the current state of the art regarding the application of this technology, which has been strictly circumscribed to stationary regimes. In this context, the project pursues the application of the technique under any operation regime of the motor, including transient regimes, which are rather frequent in many industrial applications. For the development and subsequent validation of the techniques, it is primordial to have a significant amount of flux data available. To this end, the project will rely on three main data sources: on the one hand, multiple simulations will be carried out with the aid of finite element based programs that will enable to obtain flux data for the motor operating under diverse operation regimes and faults. On the other hand, the multiple laboratory tests that will be developed will allow for obtaining data in small-medium sized motors. Finally, the available industrial collaborators will enable to obtain data of real large motors which will be of indubitable interest to validate the generality of the developed techniques. It is crucial that the obtained flux data are referred to different sensor positions, since the information provided by the technique is different depending on the location of the sensor. Once the flux signals will be available, the different diagnosis techniques will be developed. These will be based on the application of advanced time-frequency signal processing tools to the aforementioned flux signals. One of the challenges of the project relies on determining the most suitable tool for the analysis, among the wide variety of available alternatives. In this regard, both continuous and discrete transforms will be applied and assessed. Moreover, it is intended to introduce new fault severity indicators based on the applied tools. This will be a novel contribution of the project since there are very few indicators in the literature relying on flux data analysis. The developed techniques will be designed to detect a wide range of possible faults, including rotor damages, eccentricities, insulation faults, bearing failures and core damages.

Duration: January 2019- December 2021 Principal Investigator: Prof. JOSE A ANTONINO-DAVIU, Prof. LARISA DUNAI