Generation of detection indices and data management for a miniature health and usage monitoring system

The aim of this research is to search for generic algorithms that are diverse, which will allow the development of a small and low cost ‘Health Usage and Monitoring Systems’ (HUMS) solution called SmartHUMS. The achievement of this research will expand the capabilities of modern HUMS, in a way that allows HUMS benefits to reach areas which are traditionally physically and cost-wise impossible to be applied. In this research, the small to median size ‘Unmanned Aerial Vehicles’ (UAVs) are the primary targets for the application of this low cost HUMS solution. <br><br>A prototype SmartHUMS unit was used as the data acquisition system to evaluate the use of Detection Indices to generate a generic HUMS capability which, unlike existing HUMS technologies, does not rely on identifying all possible failure modes and their fault signatures. The process followed involved the selection of candidate algorithms, the testing of those algorithms on two different test rigs and the final quick check that such a result is possible by flying the prototype unit in a manned helicopter.<br><br>The generic algorithms selected are ‘Autocorrelation’ and ‘Cross-Correlation’ methods. Both algorithms are being used to analyse vibration signals to detect signal anomalies that correspond to system behavioural changes, which is why this research refers to the generic algorithms as ‘Detection Indices’ (DI). The fundamental characteristic behaviour of the DI analysis results are examined, which help in the differentiation process of separating detected system anomalies into control induced change and non-control induced change. The detected non-control induced anomalies are further analysed with Cyclostatic analysis methods to identify the possible cause of system behaviour change.<br>