CN-116303499-B - Sample library updating method and device of self-learning coefficient calculation model

CN116303499BCN 116303499 BCN116303499 BCN 116303499BCN-116303499-B

Abstract

The embodiment of the application provides a sample library updating method and device of a self-learning coefficient calculation model, and relates to the technical field of rolling force calculation model control; the method comprises the steps of creating a plurality of sample libraries, judging whether the sample libraries meet requirements based on the minimum effective capacity, calling the sample libraries meeting the requirements, and updating other sample libraries. According to the application, the freshness of the sample number is ensured by periodically updating the sample data of the sample library, the taking precision of the self-learning coefficient can be improved when the method is applied to variable-specification production, the condition that the original accumulated self-learning coefficient influences the current production due to equipment parameter change is avoided, and further, the steel rolling production is more stable and smooth.

Inventors

CAO JING
LIU YUJIN
LI YUNLAI
LIU XIAOQIANG
XIE YU
HU ZHIYUAN
WANG DIANLONG
LI WENPENG
HE XULING

Assignees

首钢智新迁安电磁材料有限公司

Dates

Publication Date: 20260508
Application Date: 20230303

Claims (6)

1. The method for updating the sample library of the self-learning coefficient calculation model is characterized by comprising the following steps of: determining a minimum effective capacity of the sample library; Creating a plurality of sample libraries, wherein each sample library is provided with a corresponding set value, and the set value is a sample library capacity threshold; Comparing the plurality of sample libraries with the minimum effective capacity one by one according to the capacity sequence of the sample libraries; when the number of samples in the sample libraries for comparison is not lower than the minimum effective capacity and the number of samples in other sample libraries exceeds a set value, judging that the sample libraries for comparison meet the requirements; The method comprises the steps of calling a sample library meeting requirements, simultaneously emptying other sample libraries and re-accumulating data, taking out original data for comparison from the sample library in the process of re-accumulating data, comparing and checking the original data with new data to be added into the sample library, and putting the new data into the sample library after checking the new data meets the requirements; And (3) performing cycle reciprocation on the process of comparing the plurality of sample libraries with the minimum effective capacity, and the process of emptying other sample libraries and re-accumulating data.
2. The method of claim 1, wherein the minimum effective capacity is half the capacity of the smallest-capacity sample library of the plurality of sample libraries.
3. The method of claim 1, wherein the comparing the original data with new data to be added to the sample library comprises: acquiring the credibility of the new data through a formula (1), and acquiring the self-learning coefficient of the new data through a formula (2); ; (1) ; (2) Wherein, the For the degree of trustworthiness of the original data, For the degree of trustworthiness of the new data, Is the self-learning coefficient of the new data, Is the self-learning coefficient of the original data, The number of the strip steel rolls in the current batch is the number of the strip steel rolls in the current batch, The self-adaptive related data are obtained from actual production data in the rolling process; When (when) When the new data is judged to meet the verification requirement, wherein Is a constant value.
4. A sample library updating apparatus of a self-learning coefficient calculation model, wherein the sample library updating apparatus is configured to implement the sample library updating method according to claim 1, the sample library updating apparatus comprising: The system comprises a creating unit, a storage unit and a storage unit, wherein the creating unit is used for creating a plurality of sample libraries, and each sample library is provided with a corresponding set value which is a sample library capacity threshold; The judging unit is used for comparing the plurality of sample libraries with the minimum effective capacity one by one according to the capacity sequence of the sample libraries, and judging that the sample libraries subjected to comparison meet the requirement when the number of samples in the sample libraries subjected to comparison is not lower than the minimum effective capacity and the number of samples in other sample libraries exceeds a set value; the calling unit is used for calling the sample library meeting the requirements; the updating unit is used for emptying other sample libraries and re-accumulating data, taking out the original data for comparison from the sample libraries in the process of re-accumulating the data, comparing and checking the original data with new data to be added into the sample libraries, and placing the new data into the sample libraries after the check meets the requirements.
5. A computer readable storage medium having stored thereon computer instructions which, when run on a computer, cause the computer to perform the method of any of claims 1-3.
6. An electronic device comprising a memory and a processor; The memory is used for storing instructions; The processor for invoking instructions in the memory to cause the electronic device to perform the method of any of claims 1-3.

Description

Sample library updating method and device of self-learning coefficient calculation model Technical Field The present application relates to the field of rolling force calculation model control technology, and in particular, to a method for updating a sample library of a self-learning coefficient calculation model, a sample library updating device of a self-learning coefficient calculation model, a computer-readable storage medium, and an electronic device. Background In the process of strip steel production, a traditional calculation model calculates a rolling force to obtain a self-learning coefficient according to an actual rolling force and the model, when strip steel is produced in a variable specification, the model refers to the self-learning coefficient accumulated in the previous production of the steel coil, and participates in the calculation of the preset rolling force of the next steel coil, wherein the closer the preset rolling force is to the actual rolling force of the steel coil, the smoother the starting is. However, with continuous production of steel coils, some original technical parameters of steel rolling equipment need to be modified, after the technical parameters are changed, the magnitude of the self-learning coefficients is affected, but a traditional calculation model cannot recognize the change, and the method can calculate all the self-learning coefficients before and after the parameters are modified in an accumulated manner. As it cannot react in time. After the parameters are modified, the self-learning coefficient calculated in the model cannot be corrected in time, and the calculated preset rolling force is either higher or lower, so that the phenomenon of difficult starting occurs, and the production rhythm is influenced. Therefore, an updating method is urgently needed, and the parameters are modified, and meanwhile, the sample library of the self-learning coefficient calculation model is modified, so that the self-learning coefficient is corrected and calculated. Disclosure of Invention Embodiments of the present application provide a sample library updating method of a self-learning coefficient calculation model, a sample library updating apparatus of a self-learning coefficient calculation model, a computer-readable storage medium, and an electronic device, according to the application, in the continuous production and adjustment process of the steel coil, the sample library of the self-learning coefficient calculation model is continuously updated, and then the accumulated self-learning coefficient is quickly corrected, so that the rolling production is smoother. Other features and advantages of the application will be apparent from the following detailed description, or may be learned by the practice of the application. According to a first aspect of an embodiment of the present application, there is provided a sample library updating method of a self-learning coefficient calculation model, including: determining a minimum effective capacity; Creating a plurality of sample libraries; judging whether the sample library meets the requirement or not based on the minimum effective capacity; Sample libraries meeting the requirements are called, while other sample libraries are updated. In some embodiments of the application, based on the foregoing, the minimum effective capacity is half the capacity of the smallest of the plurality of sample libraries. In some embodiments of the present application, based on the foregoing solution, the determining whether the sample library meets the requirement based on the minimum effective capacity includes: Comparing the plurality of sample libraries with the minimum effective capacity one by one according to the capacity sequence of the sample libraries; and when the number of samples in the sample libraries for comparison is not lower than the minimum effective capacity and the number of samples in other sample libraries exceeds a set value, judging that the sample libraries for comparison meet the requirements. In some embodiments of the application, based on the foregoing, the updating of the other sample libraries includes emptying the other sample libraries and re-accumulating the data. In some embodiments of the present application, based on the foregoing scheme, in the process of re-accumulating the data, the data verification is further included. In some embodiments of the present application, based on the foregoing scheme, the verifying the data includes: taking out the original data for comparison from the sample library; Comparing and checking the original data with new data to be added into a sample library; and after the verification meets the requirements, placing the new data into a sample library. In some embodiments of the present application, based on the foregoing scheme, the comparing and checking the original data with the new data to be added to the sample library includes: acquiring the credibility of the new