Research On Pulling Machine's Intelligent Control System Based On DSP

Aiming at improving the pulling technology of self-focus lens during their manufacturing, a pulling machine intelligent control system based on DSP has been developed. While the DSP is integrated as hardware controller, fuzzy neutral network theorie is applied in the control system, and close loop control on the pulling machine is realized by feedback component of laser calliper. Through the intelligentizing on the pulling machine, the automatic control on fibre diameter and conicity is implemented, which would change the current situation of manual work on the pulling technology and improve productivity greatly.This thesis mainly introduces the system's intelligent control algorithm, hardware platform, software system, experimental methods and result data analysis. During the control algorithm study, the fuzzy neutral network model for the control on the pulling machine motor is designed using MATLAB neutral network toolbox.The key part of hardware design is the DSP model TMS320LF2407A from TI, based on which the level converter circuit, reset circuit, clock circuit, SCI serial interface circuit, JTAG interface circuit and AD modulation circuit are designed. This provides a physical foundation for software design. The software system includes the DSP control program design under CCS environment and the upper-end computer software interface design by C++ Builder 6. The DSP programs comprise: fuzzy neutral network algorithm design, AD conversion program design, serial communication design and motor control program design. The upper-end computer interface consists of the main control interface and other sub-interfaces for parameter setting. The design methods, its realization and the detection techniques are introduced in details in the thesis. All the function modules have been working properly after the laboratory experiment and then the system has been applied on the pulling machine in the field. The analysis results meet the requirement of fibre pulling standards.The pulling machine intelligent control system developed in this thesis need not precise mathematical model. For application, only the offline results from experienced operator are needed as input of the fuzzy neutral network model. After the parameters are determined and the online feedback value is known, the control value will be outputted by the network calculation. The system efficiently decreases the influence of unknown factors on the pulling procedure, and has a good practicability and commercial importance.