Research on Optimization of Cigarette Stick Weight Parameters of Cigarette Machine Based on Neural Network

In this paper, a new tobacco weight control system based on the tobacco rolling mechanism frame was designed. By combining the neural network architecture, a time series parameter prediction model based on LSTM network was constructed, real-time dynamic optimization was implemented for the three parameters of PID control (proportion, integration and differential), the recurrent neural network was used to process the time series related features in the production process, and the network weight was dynamically adjusted through the backpropagation algorithm to effectively capture the nonlinear time-varying characteristics of equipment operation. At the same time, the BP neural network was used to construct a twochannel regression model with the quality deviation of online cigarette sticks as the response variable and the equipment parameter cluster as the explanatory variable, so as to realize the quantitative prediction of the impact of parameter adjustment on quality. Finally, through a case analysis, the performance of the traditional PID control and the neural network optimized control system in the weight control of cigarette sticks of the winding machine is compared, and the effectiveness of the neural network algorithm in parameter dynamic optimization and system stability control is verified.