Design And Implementation Of Data-driven Composite Machining Scheduling System

The goal of composite material processing and production scheduling is to ensure the smooth execution of production plans,optimize production efficiency,reduce manufacturing costs,and improve product quality.However,there are several challenges encountered in actual production processes.Firstly,in the hot pressing stage,the internal residual stresses generated in composite material components can cause deformation and prevent the components from meeting quality standards,thereby affecting the successful execution of production plans.Secondly,in the lay-up stage,inaccurate estimation of lay-up time results in unpredictable entry times of components into the hot press,leading to situations where the hot press is either idle or components are delayed outside the hot press.To address the aforementioned issues and assist composite material factories in developing effective production plans,this study focuses on the curing process in the hot press stage of composite material manufacturing.By analyzing historical production data of the hot press equipment and utilizing machine learning techniques,a temperature prediction model for composite material components is established.The prediction model is validated using actual data from the composite material factory,demonstrating an average absolute error of less than7 degrees Celsius,meeting industrial requirements.Furthermore,integrating the prediction model with the scheduling problem in the hot press stage enables the development of tankloading plans that consider both spatial utilization and component quality.Validation with realworld data confirms that the tank-loading plans obtained through the integrated prediction model exhibit minimal changes in spatial utilization while significantly reducing the average temperature difference of components by nearly 10 degrees Celsius.To address the issues in the stacking stage,an Artificial Neural Network(ANN)based on Backpropagation(BP)is introduced to accurately predict the completion time of individual components under different numbers of operators and at different workstations.The stacking completion time prediction model is established.Actual production data is used to validate the prediction model,demonstrating an average absolute error of less than 3 minutes,meeting industrial requirements.Furthermore,by solving the scheduling model for the stacking stage,the workforce and workstations required for each component are determined,forming the stacking stage plan.A data-driven composite material manufacturing scheduling system was designed and implemented to meet the specific requirements of the composite material factory.The system includes features such as basic data management,automated scheduling,and production plan Gantt chart visualization.Through comprehensive testing of the system,we have verified the reliability and stability of its functionalities.The system has successfully met the needs of the composite material factory,providing reliable support for improving the efficiency and accuracy of production planning,optimizing resource utilization,reducing manufacturing costs,and enhancing product quality.