Vehicle body-in-white (BIW) crash models are important for crashworthiness analysis. Conventional finite element methods usually deal with a large sized computational model and thus hinder efficient design evaluation. The proposed computation grid based virtual testing lab, with a significant improvement of computation efficiency, is capable of extracting essential safety dynamic characteristics into an artificial neural network (NN). Our analysis shows that the virtual testing lab is capable to provide all the necessary computation resource through computation grid network, and derive the frontal crash characteristics into the NN based knowledge database. The above testing case successfully demonstrates computation grid network is the essential driven technique to accomplish the challenging engineering computation tasks.