An automated procedure for selecting response sensor locations for the purpose of experimental modal testing is proposed. Experimental modal test results, which are extracted from the frequency response functions (FRFs) associated with the selected sensor locations, are commonly used for validating analytical models. Therefore, proper response point selection is of paramount importance for obtaining a meaningful correlation between the test results and the analytical model. The proposed research focuses on the application of a genetic algorithm (GA) to the minimization of several candidate error indicator functions. Several new encoding schemes, each of which represents the selected response points in different ways, will also be examined as to their effect on the algorithm's performance. The most promising error indicator and encoding scheme will be implemented upon a multi-processor computer, thereby taking advantage of the GA's ability to process solution families in a parallel fashion. It is anticipated that by using new encoding schemes, effective error indicators, and parallel processor enabled software, a robust, accurate, and efficient response point selection tool can be developed. This tool can be utilized by NASA engineers in the modal testing of all critical structures, such as Space Station modules and launch vehicle components.
Potential Commercial Applications:It is believed that a market niche exists for a software product capable of accurately, and automatically, determining the optimal locations in which to place sensors for modal testing. As the number of industries utilizing experimental modal testing during product development increases (e.g., aircraft, automotive, and bridge construction industries), the market opportunities for such a software tool increases. At this point, our goal is to market both a stand-alone version of the software, and a "component" version of the code. The component version will be tailored to integrate into several of the leading modal analysis software systems, namely, the CADA System from LMS, and the I-DEAS package from SDRC. Both of these leading software vendors have software "hooks" specifically designed to assist in the integration of component software into their systems. The stand-alone version of the product is to be hosted on several computer operating systems, including Windows NT and UNIX.
