CN-115659576-B - Copper smelting optimization method and device, electronic equipment and storage medium

CN115659576BCN 115659576 BCN115659576 BCN 115659576BCN-115659576-B

Abstract

The invention provides a copper smelting optimization method, a device, electronic equipment and a storage medium, and relates to the technical field of copper smelting; the method comprises the steps of inputting an initial data set into a corresponding copper smelting proxy model, outputting target predicted values for representing copper smelting index predicted results, determining a target solution set based on a plurality of merging predicted groups constructed by the target predicted values corresponding to copper smelting indexes according to the copper smelting indexes based on a depth Gaussian process and Gaussian noise, and determining a merging predicted group optimal solution of copper smelting optimization and a decision parameter optimal solution corresponding to the merging predicted group optimal solution based on expected ultra-volume improvement quantity under the condition that the new merging predicted group is added into the target solution set. The method can realize copper smelting optimization under the multi-competition target and acquire the optimal technological parameter combination in real time.

Inventors

WANG XUELEI
KANG LIWEN
XU BAOWEN
TAN JIE

Assignees

中国科学院自动化研究所

Dates

Publication Date: 20260505
Application Date: 20220826

Claims (9)

1. A copper smelting optimization method, characterized by comprising: Constructing an initial data set, wherein the initial data set is obtained by respectively combining a plurality of copper smelting indexes based on sample decision parameters; Inputting the initial data set into a corresponding copper smelting proxy model, and outputting a target predicted value for representing a copper smelting index predicted result, wherein the copper smelting proxy model is determined for each copper smelting index based on a deep Gaussian process and Gaussian noise; determining a target solution set based on a plurality of merging prediction groups constructed by target prediction values corresponding to copper smelting indexes, and determining an optimal solution of the merging prediction group and an optimal solution of a corresponding decision parameter based on an expected super-volume improvement amount under the condition that the target solution set is added into a new merging prediction group, wherein the target solution set is determined based on a pareto front edge approximation set of a current iteration period, the pareto front edge approximation set of the current iteration period is obtained by adding the optimal solution of the decision parameter which is determined by a previous iteration period and enables the expected super-volume improvement amount to be the largest on the basis of the pareto front edge approximation set of the previous iteration period, and the expected super-volume improvement amount is used for representing the expected improvement amount of the new merging prediction group on the current copper smelting proxy model; under the condition that a new merging prediction group is added into the target solution set, determining an optimal solution of the merging prediction group optimized by copper smelting and a decision parameter optimal solution corresponding to the optimal solution based on the expected super-volume improvement quantity, wherein the method comprises the following steps: Determining a first supersvolume index of the copper smelting proxy model at present based on the target solution set, wherein the first supersvolume index is a supersvolume value formed by a combined prediction group and a reference prediction group in the target solution set; determining a second supersvolume index of a new copper smelting proxy model under the condition that the target solution set is added with a new merging prediction group; Determining an amount of hypervolume improvement based on a difference between the second hypervolume index and the first hypervolume index; determining an expected super-volume improvement based on the super-volume improvement and a corresponding multivariate probability density for the new merged prediction set; And determining a combined prediction group optimal solution for copper smelting optimization and a decision parameter optimal solution corresponding to the combined prediction group optimal solution based on the expected super-volume improvement amount, wherein the combined prediction group optimal solution is used for representing each copper smelting index corresponding to the combined prediction group optimal solution to be better than each copper smelting index corresponding to the combined prediction group except for the combined prediction group optimal solution.
2. The copper smelting optimization method according to claim 1, wherein the determining a combined prediction set optimal solution for copper smelting optimization and a decision parameter optimal solution corresponding thereto based on the expected super-volume improvement amount comprises: Determining a multi-objective optimization evaluation index based on the first supersolume index; And under the condition that the multi-objective optimization evaluation index converges, determining the maximum value of the expected super-volume improvement quantity and a corresponding merging prediction group, determining the merging prediction group as a merging prediction group optimal solution, and determining decision parameters corresponding to the merging prediction group optimal solution as a decision parameter optimal solution.
3. The copper smelting optimization method according to claim 2, wherein the determining a combined prediction set optimal solution for copper smelting optimization and a corresponding decision parameter optimal solution based on the expected super-volume improvement amount further comprises: under the condition that the multi-objective optimization evaluation index is not converged, determining the maximum value of the expected super-volume improvement quantity and a corresponding merging prediction group; Determining decision parameters corresponding to the combined prediction group, and determining copper smelting indexes corresponding to the decision parameters; Adding the decision parameters and the copper smelting indexes to the initial data set to construct an updated data set; And inputting the updated data set into the copper smelting proxy model again for training until the multi-objective optimization evaluation indexes are converged, and determining the combined prediction group optimal solution and the corresponding decision parameter optimal solution.
4. A copper smelting optimization method according to any one of claims 1 to 3, wherein the deep gaussian process comprises an L-layer gaussian process; The copper smelting proxy model is obtained based on the following steps: Constructing an L-layer Gaussian process, wherein the output of the upper-layer Gaussian process is the input of the lower-layer Gaussian process, and the input of the first-layer Gaussian process is a sample decision parameter; and determining a copper smelting proxy model corresponding to the corresponding copper smelting index based on the output result of the L-layer Gaussian process and the sum of Gaussian noise.
5. The copper smelting optimization method of claim 4, wherein the constructing an initial data set includes: Acquiring a sample sampling parameter; Simulating the sample sampling parameters to determine a plurality of copper smelting indexes corresponding to the sample sampling parameters; Combining the sample sampling parameters with the copper smelting indexes respectively to obtain data pairs corresponding to the copper smelting indexes; based on each of the data pairs, an initial data set is constructed.
6. A copper smelting optimization device, comprising: the construction module is used for constructing an initial data set, and the initial data set is obtained by respectively combining a plurality of copper smelting indexes based on sample decision parameters; The output module is used for inputting the initial data set into a corresponding copper smelting proxy model and outputting a target predicted value for representing a copper smelting index predicted result, wherein the copper smelting proxy model is determined for each copper smelting index based on a deep Gaussian process and Gaussian noise; The determining module is used for determining a target solution set based on a plurality of merging prediction groups constructed by target prediction values corresponding to copper smelting indexes, and determining a merging prediction group optimal solution for copper smelting optimization and a decision parameter optimal solution corresponding to the merging prediction group optimal solution based on an expected super-volume improvement amount under the condition that the target solution set is added into a new merging prediction group, wherein the target solution set is determined based on a pareto front edge approximation set of a current iteration period, the pareto front edge approximation set of the current iteration period is obtained by adding the decision parameter optimal solution which is determined by the previous iteration period and enables the expected super-volume improvement amount to be the largest on the basis of the pareto front edge approximation set of the previous iteration period, and the expected super-volume improvement amount is used for representing the expected improvement amount of the new merging prediction group on the current copper smelting proxy model; The determining module is specifically configured to determine a first hypervolume index of the current copper smelting proxy model based on the target solution set, wherein the first hypervolume index is a hypervolume value formed by a merging prediction group and a reference prediction group in the target solution set, determine a second hypervolume index of the new copper smelting proxy model when a new merging prediction group is added to the target solution set, determine an hypervolume improvement amount based on a difference value between the second hypervolume index and the first hypervolume index, determine a desired hypervolume improvement amount based on the hypervolume improvement amount and a multivariate probability density corresponding to the new merging prediction group, and determine a merging prediction group optimal solution for copper smelting optimization and a decision parameter optimal solution corresponding to the merging prediction group based on the desired hypervolume improvement amount, wherein the merging prediction group optimal solution is used for representing each copper smelting index corresponding to the merging prediction group optimal solution to be better than each copper smelting index corresponding to the merging prediction group except the merging prediction group optimal solution.
7. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the copper smelting optimization method according to any one of claims 1 to 5 when executing the program.
8. A non-transitory computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when executed by a processor, implements the copper smelting optimization method according to any one of claims 1 to 5.
9. A computer program product comprising a computer program, characterized in that the computer program, when executed by a processor, implements the copper smelting optimization method according to any one of claims 1 to 5.

Description

Copper smelting optimization method and device, electronic equipment and storage medium Technical Field The invention relates to the technical field of copper smelting, in particular to a copper smelting optimization method, a copper smelting optimization device, electronic equipment and a storage medium. Background Copper smelting is a complex industrial process and has the characteristics of multiple processes, long flow, inter-process associated coupling and the like. In the copper smelting process, the components of raw materials, the production state of equipment, production industrial parameters and the like can influence the operation of the flow, so that decision optimization of the comprehensive production index in the copper smelting process is a key research problem. Copper smelting optimization is to search a group of process parameters such as raw material components, production working conditions, operation parameters and the like, so that a plurality of production indexes are optimized, and the conditions of copper yield and minimum exhaust gas quantity are satisfied simultaneously. In actual production, the decision is mainly made by virtue of experience accumulated for a long time and related process knowledge by technicians, the manual decision has large randomness and lower accuracy, the optimization of production indexes in the copper smelting process cannot be ensured, meanwhile, as the copper smelting process involves complex physicochemical reaction, and the characteristics of nonlinearity, uncertainty, complex mechanism and the like are provided between the production indexes and decision variables, the decision optimization is difficult to establish a mathematical model in an analytic form, and when the evolution algorithm is adopted for copper smelting optimization, the method is mostly used for optimizing a single production index, a large number of samples are required to be generated, the calculation is complex, the time consumption is long, and the real-time requirement cannot be met. Disclosure of Invention The invention provides a copper smelting optimization method, a copper smelting optimization device, electronic equipment and a storage medium, which are used for solving the defects of low fitting degree and low instantaneity of copper smelting problems in the prior art, realizing copper smelting optimization under a multi-competition target and acquiring optimal technological parameter combinations in real time. The invention provides a copper smelting optimization method, which comprises the following steps: Constructing an initial data set, wherein the initial data set is obtained by respectively combining a plurality of copper smelting indexes based on sample decision parameters; Inputting the initial data set into a corresponding copper smelting proxy model, and outputting a target predicted value for representing a copper smelting index predicted result, wherein the copper smelting proxy model is determined for each copper smelting index based on a deep Gaussian process and Gaussian noise; Determining a target solution set based on a plurality of merging prediction groups constructed by target prediction values corresponding to copper smelting indexes, and determining an optimal solution of the merging prediction group and an optimal solution of a corresponding decision parameter based on an expected super-volume improvement amount under the condition that the target solution set is added into a new merging prediction group, wherein the target solution set is determined based on a pareto front edge approximation set of a current iteration period, the pareto front edge approximation set of the current iteration period is obtained by adding the optimal solution of the decision parameter which is determined by a previous iteration period and enables the expected super-volume improvement amount to be the largest on the basis of the pareto front edge approximation set of the previous iteration period, and the expected super-volume improvement amount is used for representing the expected improvement amount of the new merging prediction group on the current copper smelting proxy model. According to the copper smelting optimization method provided by the invention, under the condition that a new merging prediction group is added to the target solution set, based on the expected super-volume improvement amount, the optimal solution of the merging prediction group and the optimal solution of the corresponding decision parameter of the optimal solution of the merging prediction group for copper smelting optimization are determined, and the method comprises the following steps: Determining a first supersvolume index of the copper smelting proxy model at present based on the target solution set, wherein the first supersvolume index is a supersvolume value formed by a combined prediction group and a reference prediction group in the target solution set; determining a second supers