ACCELERATED LIFE ESTIMATION FOR LCD MODULES USING ARTIFICIAL NEURAL NETWORK BASED ON TAGUCHI METHOD

摘要

It is a quite important issue for manufacturer to choose an appropriate reliability evaluation method for getting a good balance between the needs of consumer and the cost, quality and time in competing environments. Usually, the reliability of products is evaluated by the accelerated life test to analyze and extrapolate it to normal operating conditions. This study introduces a back-propagation neural network methodology to estimate the accelerated life, and Taguchi methods to optimize the learning parameters for the neural networks. The results show that the method combined with Taguchi method and neural network is a prospective method for reliability estimation. Numerical example of accelerated life testing data of LCD modules is given to illustrate the validation and application of the presented method.