We know that conventional thermal stress estimation schemes for the components on a system board will usually arrive at over-predicted results inefficiently. In this presentation, we will propose one fast and accurate estimation scheme to predict the thermal stress of the components on a TFT-LCD system board. We use the adaptive filtering methodology for the thermal stress estimation of the components on system board. The simulation results show that the proposed scheme can achieve prediction results accurately while the error between the predicted value and the measured value will be under 3°C with 95% confidence. We use a visual image to extract the thermal information of any component on the system board, so we can predict the thermal stress more efficiently. We eventually used Deming’s Plan, Do, Check, Act (PDCA) cycle methodology to address the issues of design for reliability.

Key Words: PDCA, CST, Adaptive Filter, LMS, SD, MSE