石墨烯/碳纳米管复合电热膜制备过程工艺优化及预测模型

目的 本文利用响应面法和神经网络遗传算法对石墨烯/碳纳米管复合电热膜的固化工艺进行优化，并对2种方法的优化结果进行比较，为复合电热膜制备提供了最佳的工艺参数。方法 通过单因素实验探讨浆料定量、固化温度和固化时间对复合电热膜体积电阻率的影响，在此基础上进行BB试验设计，在BB试验结果上进行响应面法（RSM）和BP神经网络分析及优化。结果 单因素实验结果表示随电热膜定量增加，体积电阻率先下降后上升，随着固化温度的升高或固化时间增加，体积电阻率逐渐下降直至趋于稳定。对BB响应面法和GA-BP遗传神经网络法优化获得的最佳工艺进行实验验证，GA-BP遗传神经网络模型优化的结果相对误差较小为1.06%，因此得出最佳固化工艺参数：定量为0.056 g/cm2、固化温度为85.71℃、固化时间为11.13 h。该研究结果对石墨烯碳纳米管复合电热膜的制备工艺具有参考价值。结论 通过响应面方差分析表明，定量、固化温度和固化时间三因素对体积电阻率既有显著的线性影响，也有极其显著的平方影响。BP神经网络预测模型的准确性很好，可用于石墨烯/碳纳米管复合电热膜体积电阻率的预测。

Process Optimization and Prediction Model for the Preparation of Graphene/Carbon Nanotube Composite Electric Heating Film

The work aims to optimize the curing process of graphene/carbon nanotube composite electric heating film by response surface method and neural network genetic algorithm and compare the optimization results of the two methods, so as to provide the optimal process parameters for preparing the composite electric heating film. The effects of slurry weight, curing temperature and curing time on the volume resistivity of composite electric heating film were discussed through single factor experiments. On this basis, the BB test design was carried out, and the response surface method (RSM) and BP neural network were analyzed and optimized based on the BB test results. The single factor experiment results showed that with the increase of the weight of the electric heating film, the volume resistance firstly decreased and then increased. With the increase of the curing temperature or the curing time, the volume resistance gradually decreased until it became stable. Experimental verification was conducted on the optimal curing process obtained by optimized BB response surface method and GA-BP genetic neural network method. The relative error of the optimized GA-BP genetic neural network model was relatively small at 1.06%, so the optimal curing process parameters were weight of 0.056 g/cm2, curing temperature of 85.71 ℃ and curing time of 11.13 h. The research results had a reference value for the preparation process of graphene/carbon nanotube composite electric heating film. The response surface analysis of variance shows that the three factors of weight, curing temperature, and curing time have significant linear and square effects on the volume resistivity. BP neural network prediction model has good accuracy and can be used to predict the volume resistivity of graphene/carbon nanotube composite electric heating film.