Research And Application Of Temperature Control System For High Stability Laser Source

High-stability laser sources are widely used a large number of important applications in optical fiber communications,wireless laser communications,and fiber loss measurement.Due to the obvious temperature properties obtained by the semiconductor laser,its working state is highly sensitive to the changes in ambient temperature.Hence the laser source with high stability temperature control system in this paper develops from the theoretical model of the semiconductor laser,based on the traditional temperature control techniques,using the basic PID(Proportion-Integral-Differential)control and neural network tuning PID parameter technology.It is combined with negative feedback technology to improve the stability of the operating temperature of the semiconductor laser and the stability of Output Light Power by raising the accuracy of temperature control in the system,Traditional PID control possesses a number of advantages,such as simplicity,intuition and good robustness.However,the setting of key parameters of PID is overly dependent on the experience guidance,meanwhile many PID parameter tuning methods proposed are limited by the usage scenario.Because the controlled object is non-linear and time-varying,the traditional PID control is unable to reach the optimal state.The neural network possesses the standout abilities of self-learning and nonlinear processing and has been widely used in the domain of control.The neural network and PID control are combined together to form a neural network tuning PID parameter control system,which is applied to the semiconductor laser temperature control system to improve the accuracy of operating temperature control of semiconductor lasers and the stability of Output Light Power of semiconductor lasers.This paper mainly optimizes the traditional PID control algorithm,and proposes a neural network tuning PID parameter control algorithm to improve the accuracy of temperature control,the stability of Output Light Power and the dynamic performance of the system at the same time to restrain the outside interference,to accelerate the response rate of the system and to maintain the stable performance.This paper is composed of various parts,including the operating principles,system theory modeling,neural network structure design,control algorithm optimization,algorithm application,and experiment analysis,details are as follows:1.Establish a model of semiconductor laser temperature control system,optimize the theoretical analysis of semiconductor laser refrigerator,and lay the foundation for the design of basic the PID algorithm and neural network structure.2.According to the principle of negative feedback control and the working principle of the semiconductor laser,the traditional PID control algorithm is analyzed and optimized.The three-layer neural network structure designed in this paper correspond to the three key parameters of the PID controller via the nodes of the output layer.The simulation results illustrate that the improved neural network PID controller has benign step signal tracking performance.3.Complete the experimental system construction,optimize the key modules and make a design of the sliding average filter to reduce the effects of electrical and random noise on the temperature control system,so that the overall performance of the system will improve in the future.4.Apply the optimization algorithm to the system and measure the long-term and short-term working temperature output light power of the semiconductor laser,the results of comparative analysis lead to the conclusion that the accuracy of the temperature control of the neural network PID temperature control algorithm designed in this paper is better than 0.001 ?,the long-term(8 hours)fluctuation range of the output power of the semiconductor laser is less than 0.01 dB,and the short-term(15 minutes)fluctuation range is less than 0.005 dB,which illustrates the superiority of the design algorithm.