Kalman-based force identification applied to vibration transfer path analysis

One of the main goals and critical steps of a transfer path analysis campaign is the identification of operational forces. Installing force sensors for direct measurements in the transfer paths is either impractical or may significantly change the system’s dynamics, which leads to a range of estimation methods with varying degrees of accuracy and complexity. Recently, a Kalman filtering technique with dummy measurements was proposed as an inverse estimation method based only on acceleration data. To avoid drifts in the estimated forces, the dummy sensors’ experimental covariance is included in the uncertainty model. Such an approach seeks to reduce the design and validation phases of a product’s life cycle through non-intrusive force measurement techniques. This work analyses the effectiveness of this method to identify multiple inputs in an academic powertrain test rig, which allows for direct force measurements. Results show a good correlation between actual forces and those estimated via the proposed method, validating the Kalman filter with dummy measures as a reliable alternative to the classic TPA method.