A validation method of complex product supportability simulation results based on image similarity

Supportability simulation is widely used in the supportability analysis of complex products. The accuracy of the simulation results directly affects the quality of the support concept, which determines the performance of the product. In this paper, image similarity is introduced as a new criteria to the validation of complex product supportability simulation results. In the developed method, the simulation results, in the form of multivariate time series, are first preprocessed into gray image. A two-dimensional discrete wavelet transform is then used to filter out random errors. Finally, the image similarity is calculated to identify the quality of the simulation. The feasibility and applicability of the method are verified by the consistency test of the complex product supportability parameters.