Effect of sample rate and effective number of bits on integrated system health management algorithms

Oftentimes when a scientist or engineer is asked to do data collection to measure a system, the logic that is used is 'bigger and faster must be better.' This logic can end up being counterproductive to achieving the goal of quickly and accurately measuring and analyzing the system. The reasons for this are that a data collection system that is 'bigger and faster' has increased cost, increased power, increased bandwidth, and due to larger data sets created, greatly increases the demands on the data analysis hardware. Clearly, in some cases this type of 'state of the art' data collection is necessary. However, this paper demonstrates that for even relatively complex acoustic and vibration measurements, the data can be gathered using a data acquisition system that meets the minimum needs of the system being measured without severely impacting the results of the data analysis. Clearly this requires the scientist or engineer to closely examine the type of data being collected, the system which is being measured, and the type of analysis the data will undergo. All of these factors help to determine what is needed to collect good data and produce reliable results. This paper will examine these concepts as they are applied to the data collection and analysis used for Integrated System Health Management (ISHM). Specifically, this paper will examine the impact of sampling rate and the effective number of bits (ENOB) of the data acquisition system on the design of the ISHM systems and the output of ISHM diagnostic and prediction algorithms.