| Summary: | This thesis is a study of freeplay - gaps between notionally connected components in mechanical systems. For some systems any amount of freeplay can complicate closed loop control design and introduce energetic impacts between bearing components. The work was motivated by freeplay present in the control systems-of rotorcraft and explores the viability of developing sensing devices to detect freeplay during operation. In this work a generic freeplay model is introduced from first principles which predicts potential measures of a system that could indicate the level freeplay. To validate these measures a mechanical experiment is constructed which represents a minimum working example of a freeplay system. It is designed to exclude as many sources of variability as possible and offers unparalleled completeness and acuity of generated data. A novel mathematical model is derived (and found to be in good agreement with the gathered experimental data) which validates the theorised damage sensitive features. Freeplay sensitive features are used as detection metrics in two real systems. First, on data gathered from the experiment; A freeplay detection system was created which determined the level of freeplay to a high accuracy using data gathered from a single accelerometer. Second, to strain measurements taken from a rotorcraft in flight; It was found that some metrics derived from earlier work were not good predictors of freeplay on rotorcraft data, but, using a selection of data processing techniques, a bespoke freeplay detection system was successful for many flight conditions. The work in this thesis is a strong indicator that detection of freeplay is possible for certain classes of systems using autonomous devices. Positive identification of freeplay in a simple and a complex freeplay system have been shown and a critical review is offered of the conclusions, limitations and some opportunities for future study.
|
|---|
