| High-power laser diode(LD)-pumped all-solid-state lasers play an important role in scientific research and industrial production because of their advantages,such as high output power and good beam quality.With the rapid development of laser technology,there is an urgent demand for optical source of high-power solid-state laser in various fields,which must first increase the power of the pump source of solid-state lasers.In this case,the modules such as the pumped LD and laser crystals will generate a large amount of heat,which will affect laser output and even damage the laser system if not dissipated properly.Therefore,it is necessary to design a set of high-performance temperature control device.In addition,the vibrant development of advanced industries such as aerospace has laid the foundation for further use of quantum light sources for information detection,which requires us to achieve remote and intelligent control of laser control systems.For this reason,in this paper,a laser control system was developed by introducing the machine learning algorithm based on Particle Swarm Optimization(PSO)Back Propagation Neural Network(BPNN)to achieve fast and precise temperature control.Meanwhile,the remote-control method was used to optimize the system.Thus,an intelligent,high-performance,and more complete temperature control system was developed,which improved the performance of an all-solid-state laser system.The main work of this paper is:1.Based on the Proportion Integral Differential(PID)algorithm,the temperature control requirements of different modules in the laser temperature control system were analyzed.Then a PSO-BPNN-PID temperature controller was designed to achieve the temperature control of high-power LD.In addition,a high-power constant-current-source fast temperature control unit was developed because of the characteristics of LD with high heat dissipation.An adaptive dynamic adjustment strategy was proposed to enable this control unit to work closely with the PID accurate temperature control unit,so that the system could intelligently allocate control quantities between the two parts,thereby improving the speed and stability of temperature control.2.An intelligent,all-solid-state laser multi-terminal control system was developed based on remote control technology.The system was designed based on the DSP processor and the We Studio platform,by combining with the USR Io T terminal and the Alibaba cloud server,and a computer software designed with C# language was used for auxiliary control.With the above methods,a control system was designed to realize remote data processing,machine learning algorithm calculation,and instruction assignment of laser modules.The system realized the intellectualization of the laser temperature control system and improved the reliability and practicability of the equipment.Using the above principles and technologies,this paper proposed a high-power allsolid-state laser remote temperature control system based on machine learning and manufactured a prototype for experimental testing.The experimental results showed that the control speed of the LD temperature controller optimized by the machine learning algorithm was improved by 73.1%,the temperature stability was improved by ±0.02 ℃,the step response time was improved by 78.1%,and the overshoot was reduced by 65.3%compared with the traditional PID temperature controller.The results demonstrated that the LD with thermal power ≤ 135 W could be controlled after an adaptive dynamic adjustment strategy was introduced.In addition,the remote-control system was employed to monitor and collect data from other modules in real-time.The results also showed that the temperature stability of normal-temperature module was better than ±0.0035 ℃,and that of the high-temperature module was better than ±0.0045 ℃.The system could greatly improve the speed and accuracy of the temperature controller and would further improve the performance of an all-solid-state laser system. |
PDF Full Text Request