InViDia - Intelligent model-based vibration diagnostics for automotive NVH applications

New methods in vehicle development allow for lower noise and vibration levels for modern cars and especially electric cars. However, this positive development increase the sensitivity of users to deviations from the normal behavior of the vehicle. Vibration and noise problems are therefore becoming an ever-increasing issue in both the development process and the after sales area. The identification of the actual source of noise and vibration problems in vehicles however is often a time-consuming and costly task that requires a lot of expertise and experience. The use of sensors to measure vibration levels at different positions in the car provides useful assistance in this challenging task. The aim of this project is the development of methods for combining the measurement data from sensors with models of vibration propagation paths in a vehicle. These methods enable a precise, reliable and time efficient determination of the sources of noise and vibration problems in an automotive context. The methods of choice have their roots within the field of control engineering and include fault detection and isolation and so called unknown input observers. These methods provide a good tradeoff between necessary model accuracy, necessary computational power and achievable precision. In addition, it is planned to integrate signal-based approaches to combine the advantages of both methods. The increased precision allows for efficient finding of the root causes of vibration problems and therefore quick fixes of the issues (“first time right”).