Semiconductor Refrigeration Self-learning Intelligent Control System On ARM

In military, medical and scientific research, precise environment temperature control is often required in a limited space to meet specific applications. Traditional mechanical compression refrigeration is not able to meet the mentioned requirement above for its large volume, high power consumption, noise and difficulty to control. Semiconductor refrigeration technology based on the Peltier effect features fast cooling, non-polluting and high precision. Therefore, It has grown up to be a preferred solution applied in this field. However, the technologies, which are used to how to perform the best cooling efficiency of semiconductor, control the refrigeration precisely and extend its service life has not been solved up to now. Based on the small-sized semiconductor refrigeration, this paper does research on semiconductor cooling performance and self-learning control algorithm in terms of improving the speed of semiconductor refrigeration, cooling accuracy and stability, the optimal cooling efficiency, and prolonging its working life.Firstly, this paper introduces the research progress of the semiconductor refrigeration, analyzes the basic principles and thermal calculation of semiconductor refrigeration under different working conditions. The optimal design of its working conditions is achieved.Secondly, through numerical simulation analysis of semiconductor radiator temperature and time constant under different power consumption, the cooling experiment of semiconductor refrigeration is designed.This paper establishes the self-learning control algorithm by analyzing numerical simulation and using experimental data. Then combining the self-learning control algorithm with a PID control algorithm, a self-learning PID control algorithm is designed. This self-learning PID control algorithm enables the semiconductor refrigerator to achieve constant temperature control of high stability, high accuracy in the fastest cooling conditions..To a certain extent, the design of semiconductor refrigerator self-learning control algorithm in this paper would optimize the working performance of the semiconductor refrigeration and prolong its service life. It is expected to be applied in the future engineering designs.