Continuous annealing deviation prediction method and device based on neural network

• Main Authors: LI LIGANG, HE ANRUI, YUAN YUTIAN, YUAN TIEHENG, SUN WENQUAN
• Format: Patent
• Language: Chinese; English
• Published: 10.06.2022
• Subjects: CALCULATING; CHEMISTRY; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; GENERAL DEVICES FOR HEAT TREATMENT OF FERROUS OR NON-FERROUSMETALS OR ALLOYS; MAKING METAL MALLEABLE BY DECARBURISATION, TEMPERING OR OTHERTREATMENTS; METALLURGY; METALLURGY OF IRON; MODIFYING THE PHYSICAL STRUCTURE OF FERROUS METALS; PHYSICS

The invention discloses a neural network-based continuous annealing deviation prediction method and apparatus. The method comprises the steps of obtaining historical production data; the reference deviation probability of the strip steel is calculated based on the historical production data, and a deviation warning value is determined based on the reference deviation probability according to the deviation correction cylinder position mean value and the strip steel position mean value; establishing a neural network prediction model, and training the neural network prediction model based on historical production data; the trained neural network prediction model is used for predicting the position of each deviation correction cylinder and the position of the strip steel of the to-be-produced strip steel, and when the predicted position of each deviation correction cylinder or the position of the strip steel exceeds a deviation warning value, it is judged that the current to-be-produced strip steel deviates; and