CN-121978936-A - AI collaborative optimization and predictive control system and method for electrolytic manganese full-flow production process

CN121978936ACN 121978936 ACN121978936 ACN 121978936ACN-121978936-A

Abstract

The invention discloses an AI collaborative optimization and prediction control system and method for an electrolytic manganese full-flow production process, which relate to the technical field of nonferrous metal metallurgy and comprise the following steps: and establishing a real-time acquisition link in the continuous operation process of electrolytic manganese production, synchronously recording granularity, temperature, flow and concentration of the ore grinding stage, the leaching stage and the purifying stage, and generating a characteristic data set reflecting the running state of the whole process of electrolytic manganese. According to the invention, by establishing real-time data acquisition and cross-procedure association analysis of the electrolytic manganese full flow, the dynamic response and parameter cooperation based on the instant state is realized, and the granularity fluctuation hysteresis transmission is avoided. The electrolysis quality target is used as a core to carry out reverse decomposition and virtual simulation optimization, and the energy consumption and the quality balance are kept through double time scale correction, so that the whole process is stable, efficient and predictable operation is realized.

Inventors

YANG MAOFENG
CHEN TING
PAN GUOXIAN
LUO FUYUAN
YAN ZHIJUN
WU SONG
HUANG LUNING
NONG YANLI

Assignees

广西新振锰业集团有限公司

Dates

Publication Date: 20260505
Application Date: 20260130

Claims (10)

1. The AI collaborative optimization and prediction control method of the electrolytic manganese full-flow production process is characterized by comprising the following steps: Establishing a real-time acquisition link in the continuous operation process of electrolytic manganese production, synchronously recording granularity, temperature, flow and concentration of an ore grinding stage, a leaching stage and a purifying stage, and generating a characteristic data set reflecting the running state of the whole process of electrolytic manganese; Based on the characteristic data set, executing a grinding optimization analysis program and a leaching purification collaborative analysis program, and carrying out joint calculation on the granularity difference, the reaction speed and the chemical balance in the data to generate an operation suggestion table; According to the operation suggestion list and combining the quality target and the productivity constraint of the electrolysis stage, reversely pushing the key parameter requirements of the purifying liquid, the leaching liquid and the ore pulp step by step from the quality target of the electrolysis product to form an intermediate product control index set; Around the intermediate product control index set, executing overall optimization calculation in a virtual simulation environment, comprehensively analyzing energy consumption, material consumption and yield balance, obtaining coordination parameter combinations among an ore grinding stage, a leaching stage, a purifying stage and an electrolysis stage, and generating an execution rhythm scheme; And executing control operation according to the execution rhythm scheme, and executing double-time-scale correction on parameter offset of the leaching stage, the purifying stage and the electrolysis stage caused by granularity mutation in the ore grinding stage by adopting a mode of combining a time-delay response instruction, an alternate disturbance rhythm and periodic feeding and backspacing.
2. The AI collaborative optimization and prediction control method for an electrolytic manganese full-flow production process according to claim 1, characterized in that the characteristic data set generation steps are as follows: In the ore grinding stage, a continuous monitoring device is arranged at a feed inlet, a discharge outlet and a classifying device of the mill to synchronously collect the particle size distribution of ore pulp, the ore feeding rate, the internal temperature of the mill and the circulating water flow; In the leaching stage, temperature probes, pH sensing elements and flow meters are arranged at each process node of a leaching tank, the temperature distribution, pH value change and pulp flow state of the leaching solution are synchronously recorded, and the concentration of main metal ions in the solution is detected in real time; In the purification stage, measuring points are arranged at the inlet, the reaction zone and the liquid outlet of the purification tank, the flow rate, the temperature and the concentration of impurity elements in the solution are synchronously collected, and the data are kept consistent with the time index of the leaching stage; Synchronously collecting data of grinding, leaching and purifying stages, converging and integrating the data according to time marks and process sequences, and establishing a characteristic data set.
3. The AI collaborative optimization and prediction control method for the electrolytic manganese full-flow production process according to claim 2 is characterized in that when a characteristic data set is established, the granularity, the temperature and the flow parameters output by an ore grinding stage are corresponding to the input conditions of a leaching stage, the pH value and the concentration parameters output by the leaching stage are corresponding to the liquid inlet state of a purifying stage, and all data are sequentially ordered through a unified time reference, so that the granularity, the temperature, the flow and the concentration parameters form a continuous corresponding relation in the full-flow range.
4. The AI collaborative optimization and predictive control method for an electrolytic manganese full-process production process according to claim 2, wherein the steps of executing a grinding optimization analysis program and a leaching purification collaborative analysis program based on a characteristic data set are as follows: carrying out data normalization and condition screening on the characteristic data set according to the sequence of the working procedure, establishing a corresponding relation between the granularity information of the grinding stage and the reaction speed information of the leaching stage, and enabling the grinding output and the leaching input to form continuous mapping on a time axis; after finishing grinding and leaching data correspondence, carrying out sectional analysis on the reaction rate according to the temperature, flow, pH and concentration information of the leaching stage, and reversely pushing out the residual concentration interval of main ions in the leaching solution by combining the primary concentration data of the purifying stage; After analysis of the leaching reaction process is completed, matching the chemical equilibrium state between the leaching solution and the purified liquid, and determining the acidity adjustment range of the leaching stage by comparing the concentration change amplitude of the purified inlet and the purified outlet; After the combined results of the particle size difference, the reaction speed and the chemical equilibrium are obtained, the associated results of the stages are summarized to form an operation suggestion table.
5. The AI collaborative optimization and predictive control method for a full-flow electrolytic manganese production process according to claim 4, wherein the granularity control parameters in the operation suggestion table are determined based on the corresponding relation between granularity and leaching rate in the grinding stage, the acidity adjustment parameters are determined according to chemical equilibrium analysis of leaching solution and purifying solution, the reactant delivery parameters are determined according to concentration change characteristics in the purifying stage, and all parameters are kept consistent with the time index of the characteristic data set.
6. The AI collaborative optimization and prediction control method for an electrolytic manganese full-flow production process according to claim 4, wherein the step of step-by-step back-pushing key parameter requirements according to an operation suggestion table and combining quality targets and productivity constraints of an electrolytic stage is as follows: Analyzing the component characteristics of the electrolyte according to the product quality target in the electrolysis stage, and determining target parameter intervals of manganese ion concentration, impurity element content, conductivity and acidity level to form a quality target reference set in the electrolysis stage; after the quality target reference set of the electrolysis stage is obtained, according to the influence relation of the purifying liquid on the input condition of the electrolyte, reversely deducing the key parameter requirement of the purifying liquid, and determining the upper limit of the concentration of impurity ions, the pH range and the temperature interval in the purifying liquid; after the control requirement of the purifying liquid is determined, reversely pushing the composition condition and the reaction characteristic of the leaching liquid by taking the allowable concentration of impurities and the ion purity standard as targets, and determining the acidity range, the liquid-solid ratio and the temperature control interval of the leaching liquid; after the leaching liquid control index is determined, the particle size distribution characteristics, the temperature and the fluidity parameters of the ore pulp are reversely pushed from the leaching condition to the ore grinding stage, and a cross-process control intermediate product control index set is formed.
7. The AI collaborative optimization and prediction control method for the electrolytic manganese full-flow production process according to claim 6 is characterized in that in the process of reversing the key parameter requirements of the purifying liquid, the impurity concentration of the electrolyte and the impurity removal efficiency of the purifying reaction are mapped correspondingly, the allowable concentration upper limit of impurity ions is determined by comparing the concentration changes of the purifying inlet and the purifying outlet, and the pH range and the liquid temperature range of the purifying liquid are limited on the basis, so that the purifying output quality and the electrolyte input condition are continuously matched.
8. The AI collaborative optimization and predictive control method for an electrolytic manganese full-flow production process according to claim 6, wherein the overall optimization calculation steps are performed in a virtual simulation environment around an intermediate product control index set as follows: constructing a process mapping structure in a virtual simulation environment, and logically mapping an intermediate product control index set according to a process sequence to enable an ore grinding stage output parameter, a leaching stage reaction condition, a purification stage output characteristic and an electrolysis stage liquid inlet condition to form a continuous traceable process chain; After the process mapping relation is formed, comprehensively analyzing the energy consumption characteristics and the material consumption characteristics in the virtual simulation environment, and realizing the dynamic coupling of input and output by establishing a unified framework of the cross-process energy flow and the material flow; After obtaining the relation between the energy consumption and the material flow, comprehensively analyzing the yield balance of the ore grinding stage, the leaching stage, the purifying stage and the electrolysis stage, determining the influencing factors of the yield change and obtaining feasible parameter combinations in the balance state; After the balance analysis of energy consumption, material consumption and yield is completed, generating a cross-process coordination parameter combination and formulating an execution rhythm scheme, so that the material flow, the energy input and the reaction rate form a periodic coordination relationship in the whole process range.
9. The AI collaborative optimization and prediction control method for the electrolytic manganese full-flow production process according to claim 8, wherein the control operation steps are executed according to an execution rhythm scheme as follows: Establishing an initial time reference for executing a rhythm scheme, performing time positioning and response window division on a change interval of granularity fluctuation in an ore grinding stage, and establishing a corresponding relation between a granularity state and a liquid-solid ratio parameter in a leaching stage to form a double-layer time structure; After a time reference is established, delay reaction instruction control is implemented for the influence of the particle size mutation of the ore grinding stage on the leaching stage, so that the leaching reaction condition is gradually transited according to an execution rhythm scheme after a stable state is maintained in a delay window; After the delay control of the leaching stage is executed, implementing a control mode of combining alternate disturbance rhythm and periodic feeding back, and realizing disturbance balance of the purification and electrolysis stage through reverse micro-amplitude adjustment and periodic callback; Under the comprehensive actions of time delay reaction, alternate disturbance and periodic feeding and backspacing, double time scale correction is carried out on the whole flow operation parameters, so that synchronous and stable operation of grinding, leaching, purifying and electrolysis stages is kept in a short period and long period range.
10. An AI collaborative optimization and prediction control system for an electrolytic manganese full-process production process, which is used for realizing the AI collaborative optimization and prediction control method for the electrolytic manganese full-process production process according to any one of claims 1-9, is characterized by comprising a data acquisition module, a collaborative analysis module, a target thrust module, a simulation optimization module and an execution control module: The data acquisition module establishes a real-time acquisition link in the continuous operation process of electrolytic manganese production, synchronously records granularity, temperature, flow and concentration of an ore grinding stage, a leaching stage and a purifying stage, and generates a characteristic data set reflecting the running state of the whole flow of electrolytic manganese; the collaborative analysis module executes a grinding optimization analysis program and a leaching purification collaborative analysis program based on the characteristic data set, and performs joint calculation on the granularity difference, the reaction speed and the chemical balance in the data to generate an operation suggestion table; the target back-pushing module is used for gradually back-pushing key parameter requirements of the purified liquid, the leaching liquid and the ore pulp from the electrolytic product quality target according to the operation suggestion table and combining the quality target and the productivity constraint of the electrolytic stage to form an intermediate product control index set; The simulation optimization module is used for executing overall optimization calculation in a virtual simulation environment around the intermediate product control index set, comprehensively analyzing energy consumption, material consumption and yield balance, obtaining coordination parameter combinations among an ore grinding stage, a leaching stage, a purifying stage and an electrolysis stage, and generating an execution rhythm scheme; the execution control module executes control operation according to an execution rhythm scheme, and adopts a mode of combining a delay response instruction, an alternate disturbance rhythm and periodic feeding back to implement double time scale correction on parameter offset of a leaching stage, a purifying stage and an electrolysis stage caused by granularity mutation of a grinding stage.

Description

AI collaborative optimization and predictive control system and method for electrolytic manganese full-flow production process Technical Field The invention relates to the technical field of nonferrous metal metallurgy, in particular to an AI collaborative optimization and prediction control system and method for an electrolytic manganese full-flow production process. Background The artificial intelligent collaborative optimization and predictive control of the electrolytic manganese full-flow production process is characterized in that under the unified scheduling and data bearing frame of a factory control system, the continuous production process from ore treatment, leaching and purification to electrolytic finished products of electrolytic manganese is surrounded, data such as granularity, flow, temperature, components, electricity consumption and the like collected in each process are concentrated, concentrated and associated analysis is carried out, the quality index and comprehensive energy consumption target of the final electrolytic product are taken as starting points, key parameter intervals which are required to be met by each upstream process are reversely deduced, the parameters are collaborative set and adjusted in real time in the production operation process by depending on the factory control system, on the basis, trend prediction is carried out on raw material fluctuation, impurity change or working condition disturbance which possibly occur in the production process, and correction instructions are issued to related processes in advance by the factory control system, prior intervention is completed before influence is not transmitted to the downstream, each process is no longer isolated, linkage cooperation is formed under unified target constraint and unified control rhythm, and finally the integral operation state with stable product quality, controlled consumption and continuous controllable production rhythm is realized. The prior art has the following defects: Under the prior art condition, each procedure in the electrolytic manganese production process is dependent on periodically updated artificial intelligent analysis results to carry out parameter adjustment. When the ore property changes in the upstream ore grinding process, and the ore grinding granularity rapidly fluctuates in a short time, if the change speed exceeds the update and correction frequency of the artificial intelligent model, the system still carries out reverse index decomposition based on the granularity state of the last period, and time response lag is easy to generate. The deduced target state of the leaching liquid is not matched with the actual pulp condition, so that the step-out phenomenon occurs in the liquid viscosity adjustment and acidity control in the leaching process. The deviation can be amplified step by step in the process of transferring among the working procedures, so that the stability of the purifying liquid is further influenced, the quality prediction and regulation judgment of the subsequent electrolysis working procedures are interfered, the misalignment of the whole-flow collaborative optimization result is finally easily caused, and the stable overall running state is difficult to maintain. The above information disclosed in the background section is only for enhancement of understanding of the background of the disclosure and therefore it may include information that does not form the prior art that is already known to a person of ordinary skill in the art. Disclosure of Invention The invention aims to provide an AI collaborative optimization and prediction control system and method for an electrolytic manganese full-flow production process, so as to solve the problems in the background technology. In order to achieve the purpose, the invention provides the following technical scheme that the AI collaborative optimization and prediction control method of the electrolytic manganese full-flow production process comprises the following steps: Establishing a real-time acquisition link in the continuous operation process of electrolytic manganese production, synchronously recording granularity, temperature, flow and concentration of an ore grinding stage, a leaching stage and a purifying stage, and generating a characteristic data set reflecting the running state of the whole process of electrolytic manganese for providing a uniform data source; Based on the characteristic data set, executing an ore grinding optimization analysis program and a leaching purification cooperative analysis program, and carrying out joint calculation on the granularity difference, the reaction speed and the chemical balance in the data to generate an operation suggestion table containing granularity control parameters, acidity adjustment parameters and reactant feeding parameters so as to support cross-procedure adjustment; According to the operation suggestion list and combining the quality target and the p