The invention discloses a compensation method for a high-strength cold-rolled
strip steel deformation resistance model, and belongs to the field of cold-rolled
strip steel production. According to the method, regression fitting is carried out on different deformation rates and corresponding yield strengths to obtain a basic deformation resistance model, an
artificial neural network is trained in combination with data of
strip steel not lower than 450 coils, an
artificial neural network model is established to obtain a deformation resistance curve, and
back calculation is carried out on multiple deformation rates corresponding to a platform period and a rapid rising period of the deformation resistance curve to obtain a deformation resistance curve. The deformation resistance Ka of the strip steel is calculated,
reduction rate correction factors Ff corresponding to different true strains are calculated, a deformation resistance correction factor C is obtained, and therefore a corrected deformation resistance model K is obtained. According to the method, through a'
physical model + data driving 'compensation or optimization method combining
field data with laboratory data, a deformation resistance model with clear physical significance and high precision is obtained. Therefore, the purposes of improving the precision of the deformation resistance model, further improving the setting precision of the rolling force and the roll
bending force and improving the thickness fluctuation and the plate shape defect of the strip steel are achieved.