PREDICAI will be centred on monitoring bearings health state based on information from accelerometers and electrical variables of the motor. This will be accomplished by the combination of supervised and un-supervised machine learning techniques that will include a classification model such as LSTM and the use of time series with the development of a Recurrent Neural Network. Data will be generated by attaching an accelerometer to the bearing. Data will be generated on a test bench where these failures can be caused, both by simulating the errors with a tool with unbalanced weights, and by putting defective bearings on the bench.

This will be the first step to be able to develop algorithms that can predict when a bearing is going to break, since the bearing wear process can be monitored on the test bench. This is essential to develop a predictive maintenance system, in which maintenance can be planned, making the most of the component’s lifetime.

Manufacturing SMEs will benefit from this development since with a very low investment on equipment they will be able to perform predictive maintenance on the bearings of the legacy systems in the factories, which will allow for increased production, reduction in downtime and reduction of maintenance costs.

ALCHEMY (AI SME) and IMH (partner at BDIH) will collaborate to bring this project to life. ALCHEMY´s expertise in the development of AI solutions will be complemented with IMH´s expertise in manufacturing. Furthermore, ALCHEMY will be able to validate the developed solution in a real multi-process machine that integrates milling, turning and drilling operations. By including the AI assets developed in this project in the AI4EU Platform, ALCHEMY aims at initiating their internalisation strategy and to become an AI solution providers references both, locally and at European level.

PREDICAI results will allow ALCHEMY to become a local reference for manufacturing SMEs in relation to AI solutions. At the same time, it will allow IMH to enhance the BDIH services to industry, as well as showcase the benefits of collaborating to innovate.