Modeling and Optimizing the Composite Prepreg Tape Winding Process Based on Grey Relational Analysis Coupled with BP Neural Network and Bat Algorithm.

As a significant way to manufacture revolving body composite, the composite prepreg tape winding technology is widely applied to the domain of aerospace motor manufacture. Processing parameters, including heating temperature, tape tension, roller pressure, and winding velocity, have considerable effects on the void content and tensile strength of winding products. This paper was devoted to studying the influence of process parameters on the performances of winding products including both void content and tensile strength and trying to provide the optimal parameters combination for the objectives of lower void content and higher tensile strength. In the experiments, tensile strength and void content were selected as the mechanical property and physical performance of winding products to be tested, respectively. An integrated approach by uniting the Grey relational analysis, backpropagation neural network, and bat algorithm was presented to search the optimal technology parameters for composite tape winding process. Then, the composite tape winding process model was provided by backpropagation neural network utilizing the results of Grey relational analysis. According to the bat algorithm, the optimal parameter combination was heating temperature with 73.8 °C, tape tension with 291.2 N, roller pressure with 1804.1 N, and winding velocity with 9.1 rpm. The value of tensile strength increased from 1215.31 to 1329.62 MPa. Meanwhile, the value of void content decreased from 0.15 to 0.137%. At last, the developed method was verified to be useful for optimizing the composite tape winding process.