Predicting flatness of strip tandem cold rolling using a general regression neural network optimized by differential evolution algorithm

Flatness prediction is a critical technical concern in flatness feedforward control during strip cold rolling. This work realized a high-precision prediction of flatness for strip cold rolling by the data-driven and industrial Internet of Things (IIoT) technology and provided an effective mode for industrial data utilization. A flatness prediction model based on general regression neural network (GRNN) optimized by the differential evolutionary (DE) algorithm was proposed; an intact dataset was established by collecting data from the hot rolling and cold rolling production lines by developing a cross-process IIoT platform, and the proposed model and other common data-driven models are trained and tested based on that. The experiment results obtained based on a dataset with 50,000 samples show that the proposed model is feasible and can achieve accurate prediction of flatness during the strip cold rolling, and compared with the BP and SVM model, it has a better performance.