Control System Development For Carbon Fiber Precursor Production Line

Carbon fiber is a kind of important and high-performance fiber material, while PAN precursor with high quality is the key to its production. However, the production process of PAN precursor is time consuming and complicated, and numbers of control parameters need to be controlled highly precious. According to the present situation and development needs of domestic carbon fiber industry, a new type of production line control system for carbon fiber precursor was developed in this paper.The overall structure of the control system was designed according to the production process of PAN precursor. An ordinary PC was used as the superior computer, and the automation product of Beckhoff was used in the PLC control system. The parameters such as temperature, pressure, flow and speed were controlled by the soft PLC control technology, and a series of anti-interference methods were used to ensure those machines and the control system running well and stably.The host computer software used the dynamic link library provided by Beckhoff to realize the communication between the host computer and PLC, which is based on the platform of Microsoft.net with the programming language of Visual C#. It has many functions, such as saving, querying, modifying and printing。In order to upgrade and maintain the hardware easily, the modular structure was used by the control system. The system owns the functions of temperature monitoring and control, movement driving control, press monitoring and control, flow monitoring and control, etc. The PID control strategy was used to control temperature precisely, which got a result of fast temperature rising speed and accurate control level, and the PID parameters can be obtained by manual input or the system self-tuning. Servo motor system was used to adjust movement speed instead of traditional stepping motors and frequency motor, which is more flexible and precise with low noise.According to the Artificial Neural Networks technology, the BP(Back Propagation) neural network for the performance prediction of precursor was established with the software of Matlab, whose input were the processing parameters and output was the properties of the PAN precursor. The network was tested with the experimental data and the deviations between the measured data and the forecast were less than 3%, this means that the neural network had a good fitting.The system has been put into practice in the cooperative company in 2009.The field application results show that the system exhabited many advantages, such as rational design, accurate control, steady running, friendly software interface and perfect functions, high performance of the precursor manufactured by this production line, all of these satisfied the design requirement.