High-Speed Camera based Experimental Modal Analysis for Dynamic Testing of an Automotive Coil Spring

Experimental modal analysis (EMA) is a measurement technique to assess the dynamical properties of mechanical components and systems in various phases of their life cycle, e.g. for design, end-of-line testing and health monitoring. The most common EMA uses accelerometers, which provide high frequency acceleration measurements at a few discrete locations. However, attached accelerometers may alter the systems mass and damping properties and multiple tests are required to obtain spatially dense information. To overcome these issues, in this paper we use high-speed cameras and video processing algorithms. In fact, cameras as contact-less sensors do not modify the dynamics of the system under test. Furthermore, cameras provide full-field displacement data, allowing to obtain spatially dense transfer functions with a single excitation, which reduces the experiment duration. On the downside, camera measurements are suitable for relatively low-frequency applications only and require optical contrast on the component surface. While previous camera based research was focused on flat, plate-like components, we demonstrate the methodology on a 3D automotive coil spring. We use a stereo vision setup to measure the 3D displacement field, employing Lucas-Kanade optical flow as feature tracker. Thereby, we make use of local averaging for noise reduction. As cameras are able to capture static information the geometry of the component is obtained in addition to the modal parameters. This allows for intuitive visualization of the EMA results. For the automotive coil spring under investigation we obtain the displacement field up to 140 Hz with an estimated displacement accuracy in the range of a few micrometer. The EMA results are compared to an accelerometer based EMA highlighting the advantages of camera based EMA. Furthermore, we investigate the sensitivity of the camera based EMA with respect to excitation and environmental conditions and discuss two alternative markers to enhance image contrast.

Publication year:2022

Accessibility:Open