Optimization Frame Body Durability Testing with Backpropagation Neural Network and Genetic Algorithm

There is a frame body durability test that aims to replicate the stress loads received on the road. so that this part can be verified properly. to obtain stress bench conditions many possible combinations of input parameters that influence the stress response in the frame body area being observed, the iteration process on existing input parameters does not produce linear response data. this research will use Backpropagation neural network (BPNN) as a model for testing frame body durability, and genetic algorithms (GA) are very suitable for use in problems of non-linearity, and multi-objective response. with the factors that influence the existing response is, angle of pneumatic (A), angle of pneumatic (B), pressure load (A) and pressure load (B). Each of this factor has 5 level. Design experiment used is orthogonal array Taguchi design L25. The experimental method using strain measurement for getting stress value from lab condition. From the experimental data, BPNN data training was carried out which produced an MSE value of 17.4862, these results were used for the optimization process with GA and produced optimal factor parameters. combination of significant factors obtained with fitness value are: angle of pneumatic (580), angle of pneumatic (00), pressure load (67Kgf) and pressure load (83Kgf).