Optimization Of GFRP Pultrusion Process And Research On Saving Energy

One of the important tasks in pultrusion of FRP (Fiber Reinforce Plastic) is the confirmation of appropriate process parameters, the die temperature is one of the most important parameters. If the die temperature is too low, it will affect the quantity of productions. If the die temperature is too high, it will cost too much energy. On the basis of numerical simulation, this paper optimized the Glass Fiber Reinforce Plastic (GFRP) pultrusion process by the method of BP neural network and genetic algorithm, and thus obtained the most energy-efficient die temperature settings.The pultrusion process was researched, according to the theories of heat transfer and chemical kinetics, the non-steady state heat conduction equation and the curing kinetics equation were established. By the means of the differential scanning calorimetry method and least square method, the epoxy's kinetics parameters were calculated. Through finite element technique the equation of thermal conduction and curing kinetics were dispersed to ordinary differential equation in both time domain and spatial domain. The coupled equation between temperature and curing degree was also solved. Utilization of ANSYS soft to simulate GFRP pultrusion process, simulate and analysis many different processing parameters that affect temperature and degree of cure, such as the die temperature and pull speed.The paper also measured the non-steady temperature field of GFRP by the Fiber Bragg Grating experimental, and obtained the degree of cure of GFRP by the extraction method. The ANSYS simulation and the experimental results were compared. The agreement was excellent, so it proof ANSYS is true and reliable.The paper trained and tested the dates which were obtained from simulation, then established the BP neural network models between degrees of cure and die temperature using the Ann toolbox of MATLAB 6.5. On the basis of the model, we obtained an optimization result through optimizing a single objective function using genetic algorithm with floating point encoded. The result indicated that using of syllogism die can save more energy than single die, when the pultrusion speed is 200mm/min, using the syllogism die can reduce the consumption of power 384.91W than using the single die.According to the study of this paper, when the degree of cure of GFRP meets the requirements, we can make the consumption of power at the least, thereby reduce the energy consumption. Therefore, it is of great significance for enterprises to reduce the production cost and respond to the state's calling for saving energy.