Hot-rolled strip steel mathematical model data layer judgment method

Abstract

The invention relates to a hot-rolled strip steel mathematical model data layer judgment method. The method comprises the following steps: considering main factors, such as steel grade, thickness, width, final rolling temperature and coiling temperature, which influence the rolling stability of strip steel or influence the characteristics of the strip steel, and dividing mathematical model layers; considering different strip steel characteristic factors for different mathematical models when the layers are divided, wherein a mathematical model layer F is a function of a plurality of characteristic factor layer numbers and F=f(Gi, Hi, Wi, FTi, CTi); and finally, performing layer judgement of hot-rolled strip steel data according to mathematical model layer judgment rules. According to the hot-rolled strip steel mathematical model data layer judgment method, the continuity of self-learning coefficients can be retained effectively and self-learning and setting chaos caused by change fluctuations of components of molten steel of the same steel grade in different furnaces and micro-changes of specifications in schedule free rolling, particularly, cross rolling, is avoided through dynamic layer judgment; the setting reliability of a mathematical model is ensured to the greatest extent; and the mathematical model setting precision and the rolling stability are improved.

Description

technical field [0001] The invention relates to the technical field of hot-rolled steel strip production, in particular to a method for judging the data layers of a hot-rolled strip mathematical model. Background technique [0002] In the highly automated hot-rolled strip production line, the mathematical model plays a very important role, and has a great influence on the rolling stability and the dimensional accuracy of the strip. [0003] A conventional hot-rolled strip steel production line produces tens of thousands of steel strip varieties and specifications. In the model calculation, in principle, each variety specification requires a different set of raw data to participate in the model calculation. In this case, in order to support the mathematical model For normal calculations, a huge database is required to store the raw data required for model calculations. In order to save storage space and facilitate model setting and calculation, the original data required for...

AI Technical Summary

Problems solved by technology

If a strip of another layer is suddenly inserted between the strips of the same layer (the probability of strip steel crossing layers caused by a small change in material is relatively high), due to the use of different model parameters (including original data and self-learning parameters), the rolling mill parameter settings calculated by the model will have large differences, and these differences will reduce the calculation accuracy of the model, thereby affecting the product dimensional accuracy and even rolling stability

[0009] If the layer division of the model is coarser (i.e. reducing the classification interval of the layer), the probability of cross-layer steel strips can be reduced to a certain extent, but the occurrence of cross-layer phenomena cannot be avoided, and this will also reduce the model calculation accuracy

Images