CN118536677A - Ore dressing efficiency prediction and technological parameter intelligent decision method and system of ore dressing system - Google Patents
Ore dressing efficiency prediction and technological parameter intelligent decision method and system of ore dressing system Download PDFInfo
- Publication number
- CN118536677A CN118536677A CN202410962734.1A CN202410962734A CN118536677A CN 118536677 A CN118536677 A CN 118536677A CN 202410962734 A CN202410962734 A CN 202410962734A CN 118536677 A CN118536677 A CN 118536677A
- Authority
- CN
- China
- Prior art keywords
- bees
- food source
- prediction
- technological parameters
- mineral separation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/27—Regression, e.g. linear or logistic regression
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Economics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Human Resources & Organizations (AREA)
- Artificial Intelligence (AREA)
- Strategic Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Primary Health Care (AREA)
- Computing Systems (AREA)
- Marine Sciences & Fisheries (AREA)
- Mining & Mineral Resources (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Animal Husbandry (AREA)
- Agronomy & Crop Science (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a mineral separation efficiency prediction and process parameter intelligent decision method and system of a mineral separation system, wherein a fusion prediction model is constructed by selecting a plurality of prediction models of different types, taking mineral attributes and process parameters of the mineral separation system as inputs, taking mineral separation efficiency indexes as outputs, the accuracy of prediction can be greatly improved through multi-model fusion prediction, and then the input of the fusion prediction model is taken as an optimization variable, and the output of the fusion prediction model is taken as an optimal target, so that an optimization model is constructed; solving the optimal solution of the optimization model, and selecting the minerals with the corresponding attributes according to the technological parameters corresponding to the optimal solution, so that the mineral separation efficiency of the mineral separation system can be greatly improved.
Description
Technical Field
The invention relates to the field of intelligent metallurgy, in particular to a mineral separation efficiency prediction and process parameter intelligent decision method and system of a mineral separation system.
Background
In the existing beneficiation system, most of the technological parameters are set manually according to experience, and the mode of the technological parameters set in the mode is low in beneficiation efficiency, and is prone to unstable in recovery efficiency, and overflow fineness, crushing ratio index and crushing granularity are not in accordance with the standard.
Disclosure of Invention
The invention provides a mineral separation efficiency prediction and technological parameter intelligent decision method and system of a mineral separation system, which are used for solving the technical problem that the mineral separation efficiency of the mineral separation system is low due to a technological parameter setting method.
In order to solve the technical problems, the technical scheme provided by the invention is as follows:
A beneficiation efficiency prediction method of a beneficiation system, comprising:
selecting a plurality of different kinds of prediction models, taking mineral attributes and technological parameters of a mineral separation system as inputs, taking a mineral separation efficiency index as output, constructing a fusion prediction model, and predicting the mineral separation efficiency of the mineral separation system based on the fusion prediction model.
Preferably, the mineral attributes include any combination of several of primary useful minerals, primary gangue minerals, coarse particle distribution, mineral distribution characteristics, sulfidation oxidation conditions, raw ore grade;
And/or
The technological parameters of the mineral separation system comprise one or a combination of any several of the technological parameters of the crushing/preselection system, the technological parameters of the grinding system and the technological parameters of the screening system;
And/or
The mineral separation efficiency index comprises one or a combination of any several of recovery rate, overflow fineness, crushing ratio index and crushing granularity.
Preferably, the fusion prediction model comprises a plurality of different regression prediction modules and a fusion module; the regression prediction modules take mineral attributes and technological parameters of the mineral separation system as input and mineral separation efficiency indexes as output; the outputs of the regression prediction modules are connected with the input of the fusion module, and the fusion module is used for fusing the prediction results output by the regression prediction modules of different types to obtain and output a final prediction result.
Preferably, the regression prediction module comprises any combination of a neural network regression module, a vector support vector regression module and a multiple linear regression module;
And/or
The fusion module is a weighted fusion module, and the weights of the regression prediction modules are equal.
An intelligent decision-making method for technological parameters of a mineral separation system comprises the following steps:
Setting constraints, and constructing an optimization model based on the fusion prediction model; the optimization model
Taking the input quantity of the fusion prediction model as an optimization variable and taking the output quantity optimization of the fusion prediction model as a target;
and solving an optimal solution of the optimization model, and selecting minerals with corresponding attributes according to the technological parameters corresponding to the optimal solution.
Preferably, the technological parameters of the mineral separation system comprise the technological parameters of the crushing/preselection system, the technological parameters of the mineral grinding system and the technological parameters of the screening system;
The process parameters of the crushing/preselection system include: breaking grain size, breaking ratio index and breaking grain size;
the technological parameters of the ore grinding system comprise overflow concentration, ore grinding work index and overflow fineness; the technological parameters of the sorting system comprise: concentrate grade, tailing grade and beneficiation recovery rate;
And/or
The optimization targets of the optimization model are mineral separation efficiency indexes, wherein the mineral separation efficiency indexes comprise recovery rate, overflow fineness, crushing ratio indexes and crushing granularity;
the optimization target is that the recovery rate is maximum;
The constraint is that overflow fineness, crushing ratio index and crushing granularity are all in a preset range.
Preferably, the regression prediction module comprises a neural network regression module, a vector support vector regression module and a multiple linear regression module; the objective function of the optimization model is as follows:
The constraints of the optimization model are as follows:
Wherein, Is the output value of the neural network regression module,For the output value of the vector support vector regression module,The output value of the multiple linear regression module;,, the overflow fineness predicted by the neural network regression module, the vector support vector regression module and the multiple linear regression module is respectively obtained; ,, the crushing ratio indexes are predicted by the neural network regression module, the vector support vector regression module and the multiple linear regression module respectively; ,, The crushing granularity predicted by the neural network regression module, the vector support vector regression module and the multiple linear regression module is respectively obtained; ,,,,, Respectively the minimum and maximum constraints of overflow fineness, crushing ratio index and crushing granularity, Is an input parameter of the three regression prediction modules.
Preferably, the optimal solution of the optimization model is solved through an artificial swarm optimization algorithm, wherein the artificial swarm optimization algorithm takes the objective function as an fitness function thereof, and takes mineral attributes and technological parameters of a mineral separation system as food source coordinates of the swarm.
Preferably, solving an optimal solution of the optimization model by using an artificial bee colony optimization algorithm specifically comprises:
s1, determining the dimension of a bee colony according to a mineral separation efficiency index, and setting the scale and the maximum iteration number of the bee colony, wherein the bee colony comprises employment bees, following bees and reconnaissance bees;
s2, initializing the swarm coordinates, taking the initialized coordinates as a current food source, and calculating the adaptability of the current food source of the swarm;
S3, hiring bees to find candidate solutions in the neighborhood of the current food source, calculating the adaptability of the candidate solutions, judging whether the adaptability of the candidate solutions is larger than the adaptability of the current food source, if not, maintaining the coordinates of the current food source unchanged, and if so, replacing the coordinates of the candidate solutions with new food source coordinates;
S4, searching a candidate solution in the neighborhood of the current food source by following the bees according to the searching probability, calculating the adaptability of the candidate solution, judging whether the adaptability of the candidate solution is greater than the adaptability of the current food source coordinates, if not, maintaining the current food source coordinates unchanged, and if so, iterating the coordinates of the candidate solution into new food source coordinates;
S5, the scout bees are used for comparing the duration period of the currently used food source coordinates with the preset period, randomly searching candidate solutions from the global solving domain when the currently used food source coordinates are not updated in the preset period, if the fitness of the currently used food source coordinates is not greater than that of the currently used food source coordinates, maintaining the currently used food source coordinates unchanged, and if the fitness of the currently used food source coordinates is not greater than that of the currently used food source coordinates, iterating the coordinates of the candidate solutions into new food source coordinates;
s6, updating the pheromone, adjusting the search directions and the distances of the employment bees and the following bees according to the updated pheromone, and adjusting the search probability of the following bees;
s7, evaluating the overall performance of the current bee colony, judging whether the overall performance of the current bee colony is lower than the target performance, increasing the number of bees in the bee colony when the overall performance of the current bee colony is lower than the target performance, and reducing the number of bees in the bee colony when the overall performance of the current bee colony is higher than the target performance;
S8, judging whether the current bee colony reaches the maximum iteration number, if not, returning to S3, otherwise, outputting the current food source coordinates as an optimal solution.
Preferably, the step S6 includes:
for each food source Initializing a pheromone intensityRepresenting the food sourceIs the mass of (3);
after each iteration, according to the fitness of the food source Updating the intensity of the pheromone:
Wherein, Is the evaporation coefficient of the pheromone,Is based on the amount of pheromone increased by the mass of the food source, andProportional to the ratio;
calculating the search probability of the following bees according to the intensity of the pheromone:
Wherein the method comprises the steps of AndIs a parameter controlling the influence of pheromone intensity and fitness, SN is the total number of food sources,Representing pheromone intensityIs adjusted by a weight of the model; whileIs a power exponent for adjusting the influence degree of the intensity of the pheromone; Is the first The fitness value of the individual food sources,Representing a solutionIs a fitness function value of (a);
adjusting search direction and distance according to the pheromone intensity when hiring and following bees to conduct neighborhood searches near food sources;
And/or
The step S7 includes:
Setting an evaluation function Evaluating the overall performance of the bee colony after each iteration;
Wherein the function is evaluated The method comprises the following steps:
Wherein, AndRespectively two different weight coefficients; Representing the iteration number; indicating the optimal food position and the position of the food, Indicating the degree of dispersion of the solution;
setting adaptive parameters ，The number parameters representing employment bees, following bees and scout bees, respectively;
after each iteration, according to the number parameter and the number parameter of the current period of the bee colony The number parameter of the next cycle of employment bees, follow bees and scout bees:
Wherein, The number of iterations is indicated and,The method is a function for adjusting the quantity according to the evaluation result, and is used for increasing the quantity parameter when the current overall performance is lower than the target performance and reducing the quantity parameter when the current overall performance is higher than the target performance;
According to new quantity parameters Updating the number of bees:
Wherein, Is the total number of bees;，， representing the number of employment bees, following bees and spying bees.
A computer system comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the method described above when the computer program is executed.
The invention has the following beneficial effects:
1. according to the invention, a plurality of different prediction models are selected, the mineral attribute and the technological parameters of a mineral separation system are taken as inputs, the mineral separation efficiency index is taken as output to construct a fusion prediction model, the accuracy of prediction can be greatly improved through multi-model fusion prediction, the input of the fusion prediction model is taken as an optimization variable, and the output of the fusion prediction model is taken as an optimal target to construct an optimization model; solving the optimal solution of the optimization model, and selecting the minerals with the corresponding attributes according to the technological parameters corresponding to the optimal solution, so that the mineral separation efficiency of the mineral separation system can be greatly improved.
2. In the preferred scheme, three different reasoning prediction models are constructed by utilizing methods such as Neural Network Regression (NNR), vector Support Vector Regression (VSVR), multiple Linear Regression (MLR) and the like, and the ore dressing efficiency is predicted by integrating the advantages of the models, so that the prediction accuracy is higher.
3. In a preferred scheme, the invention constructs an optimization model which takes the average value of NNR, VSVR, MLR model outputs as a basis, takes the maximum recovery rate as a target, meets constraint conditions of overflow fineness, crushing ratio index and crushing granularity, and adopts an improved bee colony optimization Algorithm (ABC) to solve the optimal solution of the optimization model. The bee colony optimization algorithm combines an information sharing mechanism and an adaptive bee proportion improvement strategy, can enhance the global searching capability and convergence speed of the algorithm, simultaneously keeps the diversity of solutions, and can improve the solving quality and the solving speed of complex problems.
In addition to the objects, features and advantages described above, the present invention has other objects, features and advantages. The invention will be described in further detail with reference to the accompanying drawings.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application. In the drawings:
Fig. 1 is a flow chart of a process parameter intelligent decision method of a beneficiation system provided by the embodiment of the invention.
Detailed Description
Embodiments of the invention are described in detail below with reference to the attached drawings, but the invention can be implemented in a number of different ways, which are defined and covered by the claims.
In this embodiment, a beneficiation efficiency prediction method of a beneficiation system is provided, including:
1. Data collection and preprocessing
1) Data collection
The present invention relates to a step of collecting historical beneficiation data from a plurality of metal mines, the data comprising raw mineral attributes (primary useful minerals, primary gangue minerals, coarse particle embedment, mineral distribution characteristics, sulfidation oxidation conditions, raw ore grade), crushing/preselection systems (entry particle size, crushing ratio index, crushing particle size), grinding systems (overflow concentration, grinding work index, overflow fineness), beneficiation systems (concentrate grade, tailings grade, beneficiation recovery), and the like. These data may come from different mines and different mining periods, and thus there is a difference in consistency and availability of the data.
2) Data integration
To ensure consistency and availability of data, the present invention formats and normalizes data from different mines for subsequent analysis and application.
3) Data preprocessing
And cleaning the integrated data, including removing duplicate records, processing missing values, correcting errors and abnormal values, and the like, so as to improve the data quality.
4) Data normalization
After data preprocessing, the data is converted into a uniform format to ensure that the data of different data sources can be compared and analyzed under the same standard.
5) Data verification
Consistency and integrity checks are performed on the cleaned and standardized data to ensure the quality of the data cleaning step.
The data preprocessing solves the problems of consistency and usability differences in collecting historical beneficiation data from different metal mines. These data cover the original mineral properties, parameters of the crushing/preselection system, the grinding system and the sorting system.
2. Constructing a fusion prediction model:
And an advanced regression analysis technology is utilized to optimize the key technological parameter reasoning result in the beneficiation process, so as to provide a guiding basis for the follow-up parameter intelligent decision.
1) Determination of input parameters
The method comprises the steps of firstly determining input parameters of a prediction model, wherein the parameters comprise mineral properties (mainly useful minerals, main gangue minerals, coarse grain embedding, mineral distribution characteristics, sulfuration oxidation conditions, raw ore grades), a crushing/preselection system (particle size entering), a grinding system (overflow concentration and grinding work index), a sorting system (concentrate grade and tailing grade) and the like.
2) Step two: definition of output metrics
Output indexes of the prediction model, including recovery rate, overflow fineness, crushing ratio index, crushing granularity and the like, are determined, and are key to evaluating the beneficiation efficiency.
3) Step three: construction of predictive model
Selecting a plurality of different kinds of prediction models, taking mineral attributes and technological parameters of a mineral separation system as inputs, taking a mineral separation efficiency index as output, constructing a fusion prediction model, and predicting the mineral separation efficiency of the mineral separation system based on the fusion prediction model.
The fusion prediction model comprises a plurality of regression prediction modules of different types and a fusion module; the regression prediction modules take mineral attributes and technological parameters of the mineral separation system as input and mineral separation efficiency indexes as output; the outputs of the regression prediction modules are connected with the input of the fusion module, and the fusion module is used for fusing the prediction results output by the regression prediction modules of different types to obtain and output a final prediction result. The fusion module is a weighted fusion module, and the weights of the regression prediction modules are equal.
In a preferred embodiment, three different regression prediction modules are constructed using neural network regression (NNR, neuralNetworkRegression), vector support vector regression (VSVR, vectorSupportVectorRegression), multiple linear regression (MLR, multivariateLinearRegression), and the like.
The reasoning result is optimized by integrating the advantages of three different regression prediction modules. Specifically, NNR is capable of handling nonlinear complex relationships, VSVR provides excellent generalization capability, while MLR is suitable for handling linear relationships.
In addition, as shown in fig. 1, in this embodiment, an intelligent decision method for technological parameters of a beneficiation system is further provided, including:
1) Constructing an optimization model
Setting constraints, and constructing an optimization model based on the fusion prediction model; and the optimization model takes the input quantity of the fusion prediction model as an optimization variable and takes the output quantity of the fusion prediction model as an optimal target.
Specifically, the objective function of the optimization model is based on the average value of recovery rate, overflow fineness, crushing ratio index and crushing granularity output by the three models (NNR, VSVR, MLR), aims at maximizing recovery rate, and needs to meet constraint conditions of overflow fineness, crushing ratio index and crushing granularity.
The following is a mathematical model of the objective function:
Is provided with ,,Recovery rates predicted by NNR, VSVR, MLR models respectively;
,, predicted overflow fineness;
,, Respectively predicting crushing ratio indexes;
,, respectively the predicted crushing size.
The objective function is:
The constraint conditions are as follows:
Wherein the method comprises the steps of ,,,,,Respectively the minimum and maximum constraints of overflow fineness, crushing ratio index and crushing granularity,Is an input parameter of the three regression prediction modules.
2) Solving for optimal solution
And solving an optimal solution of the optimization model, and selecting minerals with corresponding attributes according to the technological parameters corresponding to the optimal solution.
In this embodiment, an artificial bee colony optimization method is used to solve the optimal solution.
The artificial bee colony optimization method (ArtificialBeeColony, ABC) is a high-efficiency bionic intelligent optimization technology inspired by the foraging behavior of the bees in the nature. The method is widely applied to solving various complex problems, and the optimization principle is based on the natural foraging mechanism of bees. In the ABC algorithm, bees are divided into three main roles:
employment of bees: these are bees responsible for exploring and finding food sources. They not only evaluate the value of the food sources, but also pass information to the peers with a certain probability, leading the whole colony to gather towards a good quality food source.
Following bees: these bees rely on information provided by the employment bees to select food sources. They search near the selected food sources, evaluate and compare the value of the food to determine if it is worth preserving.
Reconnaissance bees: when a food source is abandoned by employment bees, the scout bees intervene and they randomly explore in the search space to find new potential food sources.
Through the three types of cooperative work, the ABC algorithm simulates the intelligent behaviors of the bee colony, and effectively navigates in the search space to find the optimal solution. The advantages of this approach are its simplicity, efficiency and adaptability to many types of optimization problems. It provides a novel and productive approach to solving the complex optimization challenges in the real world.
In this embodiment, the method for solving the optimal solution by using the artificial bee colony optimization method includes:
S1, setting parameters
Determining the dimension of the bee colony according to a crushing/preselection system (particle size of the inlet), an ore grinding system (overflow concentration and ore grinding work index) and a sorting system (concentrate grade and tailing grade)Setting the scale of the populationMaximum number of iterations.
S2, initializing the position of the population
Each solution can be expressed as oneAnd (5) a dimension vector. Each individual in the populationIs the position of (2)Initialization may be by the following formula:
Wherein:
Is the first The food source is at the firstThe position of the dimension.
AndRespectively the firstUpper and lower bounds of the dimension.
Is a random number in the range of 0, 1.
S3, calculating an objective function value
Location for each beeCalculating an objective function value。
S4, honey source searching is carried out on the bee colony
A) Employment of bees:
Objective function WhereinIs a point in the solution space and,Is the dimension of the optimization objective.
The task of employing bees is to use the food source at the present timeSearching for a new candidate solution in the neighborhood of (a). This new solution is generated by:
Wherein:
Is a new solution In the first placeValues of dimensions.
Is the current solutionIn the first placeValues of dimensions.
AndIs that two points randomly selected from the solution space are at the firstValues of dimensions.
Is an in-intervalAnd (3) a number randomly generated in the search window and used for controlling the step length of the neighborhood search.
Employment of bees according to new solutionsIs adapted to (a)To determine whether to replace the original food source。
B) Following bees:
The mathematical model of the following bees is responsible for selecting and utilizing food source information found by the employing bees to search for new solutions. The following is a description of a mathematical model following a bee:
probability of following bees to select food sources Adaptation to the food sourceProportional, can be calculated by the following formula:
Wherein:
Is the first Probability of individual food sources being selected.
Is the firstFitness value of individual food sources.
Is the total number of food sources.
Once a following bee selects a food source, it will generate a new candidate solution in the neighborhood of the food sourceThis process is similar to employing bees, using the following formula:
Is a new solution In the first placeValues of dimensions.
Is the current food sourceIn the first placeValues of dimensions.
AndIs randomly selected from food sources at the first two pointsValues of dimensions.
Is an in-intervalAnd (3) a number randomly generated in the search window and used for controlling the step length of the neighborhood search.
Following the bee will be according to the new solutionTo determine whether to replace the original food source. If the adaptability of the new solution is higher, the new solution can replace the original solution and become a new food source.
C) Reconnaissance bees:
The role of the scout bee is to explore new potential solutions when the food source is abandoned. The mathematical models of the scout bees are relatively simple because their main task is to search for new food sources randomly in the solution space. The following is a description of the mathematical model of the scout bee:
When a food source Not updated for a certain number of periods (i.e. no other bees select it as a better solution), it would be considered a discarded food source. At this point, the spy bees will intervene and randomly generate a new food sourceInstead of it. This new food source is randomly selected throughout the solution space and can be expressed as:
Wherein:
Is the newly generated food source location.
AndWhich are the lower and upper bounds of the solution space, respectively.
Is a random number in the range of 0, 1.
This process ensures that the algorithm can continue to explore new regions, avoiding excessive aggregation near the locally optimal solution, thereby increasing the likelihood of finding a globally optimal solution.
S5, bee colony information sharing
Aiming at the limitations of the traditional bee colony optimization algorithm in processing multi-peak, high-dimensionality and dynamic optimization, the invention provides an improved strategy combining an information sharing mechanism and a self-adaptive bee proportion. The strategy aims at enhancing the global searching capability and convergence speed of the algorithm, and meanwhile, the diversity of the solution is maintained so as to improve the solving quality.
A pheromone model is introduced to improve the information sharing mechanism between employment bees and following bees, which not only serves as an indicator of food source quality, but also directs bees to more effectively search for and utilize good food sources, the sharing mechanism being as follows:
a) Initializing a pheromone:
-for each food source Initializing a pheromone intensityIndicating the quality of the food source.
B) Pheromone update rules:
after each iteration, according to the fitness of the food source Updating the intensity of the pheromone:
Wherein, Is the evaporation coefficient of the pheromone,Is the amount of pheromone increased according to the quality of food source, and is generally equal toProportional to the ratio.
C) Pheromone dependent search following bees:
The probability of following a bee to select a food source depends not only on the fitness of the food source, but also on the pheromone strength:
Wherein the method comprises the steps of AndIs a parameter controlling the influence of pheromone intensity and fitness, and SN is the total number of food sources.Representing pheromone intensityIs adjusted by a weight of the model; here the number of the elements is the number,Is a basic value representing the concentration or quality index of pheromone of a specific element (such as food source in the swarm optimization algorithm)Is a power exponent for adjusting the influence degree of the intensity of the pheromone; Is the first The fitness value of the individual food sources,Representing a solutionAnd measures the fitness of the solution.
D) Information-guided neighborhood search:
when employing and following bees for neighborhood searching in the vicinity of a food source, the search direction and distance may be adjusted to take into account pheromone strength.
S6, adaptively adjusting the number of bees
S61, self-adaption ratio of bees
The algorithm can adaptively adjust the number of bees according to dynamic changes in the searching process so as to keep the flexibility and efficiency of the algorithm. This helps the algorithm to balance between exploration and development, improves global searching ability and convergence speed, and the mathematical model for adaptively adjusting the number of bees can be realized by the following ways:
a) Evaluation function:
defining an evaluation function For at each iterationThe overall performance of the colony was then assessed.
Adaptive evaluation functionThe aim is to improve the global search capability and convergence of the swarm optimization Algorithm (ABC) when dealing with complex optimization problems. The evaluation function comprehensively considers the adaptability trend and the dispersion degree of the solution in the bee colony so as to realize the aim of enhancing the global searching capability in the early stage of the algorithm and enhancing the convergence rate in the later stage.
It is composed of two parts: optimal solution fitness trendDegree of dispersion of the solution. Wherein,Is the iteration timeFitness of optimal solution in time-swarmIs the standard deviation of the solutions of the swarms at the same time point. These two metrics measure the convergence of the algorithm and the global search capability, respectively.
Construction of an evaluation function:
Evaluation function The specific form of (2) is as follows:
Wherein, AndIs a weight coefficient used to balance the importance of convergence and global search capability. Along withIf the fitness of the optimal solution improves, thenThe increase reflects an increase in the convergence of the algorithm. Conversely, if the degree of dispersion of the solutionAn increase indicates an increase in global search capability. By adjustingAndWe can adjust the behaviour of the algorithm according to different optimized scenarios.
B) Bee number regulation rule:
the number of employment bees, follow bees, and spying bees is dynamically adjusted based on the results of the evaluation function. For example, if The method shows that the diversity of the bee colony is insufficient, and the number of the reconnaissance bees can be increased.
C) Adaptive parameters:
setting adaptive parameters Representing the number of employment bees, follow bees and scout bees, respectively.
D) Updating the formula:
after each iteration, according to the number parameter and the number parameter of the current period of the bee colony The number parameter of the next cycle of employment bees, follow bees and scout bees:
Wherein, Is a function of the number of adjustments based on the evaluation result.
In the artificial colony optimization Algorithm (ABC), the Adjust function is used to dynamically Adjust the proportion of bees according to the evaluation result. This function may increase or decrease the number of employment bees, follower bees, and spy bees depending on the current performance of the algorithm and the intended goal.
Assume thatIs at the time of iterationAnd integrates the convergence and diversity of the bee colony. The Adjust function will be according toTo adjust the bee proportion：
Wherein:
Is at the time of Bee ratio of (a).
Is the learning rate, a small positive number, used to control the magnitude of the adjustment.
Is the target evaluation value, the algorithm attempts to reach this evaluation level.
At t=1, a is the ratio of the initial bees.
If it isBelow is lower thanThis means that the performance of the colony is not as expected, and the Adjust function will increase the proportion of bees to enhance the search capability. Conversely, ifHigher thanThe bee fraction is reduced to promote convergence of the algorithm.
E) Updating the number of bees:
According to new quantity parameters Updating the number of bees:
Wherein, Is the total number of bees.
S7, judging whether the maximum iteration times are reached, otherwise, continuing to iterate to S4, and reaching the output optimal position.
S8, generating optimal process key parameters according to the optimal position obtained by ABC optimization: crushing/preselecting system (particle size of the crushed ore), grinding system (overflow concentration, grinding work index), sorting system (concentrate grade, tailing grade) and the like.
In summary, the invention constructs a key process result prediction model. The model first determines input parameters including mineral properties and key parameters for each beneficiation system. Output indexes such as recovery rate, overflow fineness, crushing ratio index, crushing granularity and the like are defined, and are key to evaluating the beneficiation efficiency. Next, three different inference prediction models are constructed using Neural Network Regression (NNR), vector Support Vector Regression (VSVR), multiple Linear Regression (MLR), etc., and the inference results are optimized by integrating the advantages of these models.
Further, the invention constructs an intelligent optimized mathematical model. The model is based on an average value of NNR, VSVR, MLR model outputs, aims at maximizing recovery rate, and meets constraint conditions of overflow fineness, crushing ratio index and crushing granularity.
Finally, the present invention employs an improved swarm optimization Algorithm (ABC). The algorithm simulates an intelligent group search strategy of the foraging behavior of the bees in the nature, and the optimal solution is quickly found through information sharing among the bees. The research provides an improved strategy combining an information sharing mechanism and a self-adaptive bee proportion, and aims to enhance the global searching capability and convergence speed of an algorithm, and meanwhile, maintain the diversity of solutions so as to improve the solving quality of complex problems.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (11)
1. A beneficiation efficiency prediction method of a beneficiation system, comprising:
selecting a plurality of different kinds of prediction models, taking mineral attributes and technological parameters of a mineral separation system as inputs, taking a mineral separation efficiency index as output, constructing a fusion prediction model, and predicting the mineral separation efficiency of the mineral separation system based on the fusion prediction model.
2. The method of claim 1, wherein the mineral attributes include any combination of a plurality of primary useful minerals, primary gangue minerals, coarse particle distribution, mineral distribution characteristics, sulfidation oxidation conditions, raw ore grade;
And/or
The technological parameters of the mineral separation system comprise one or a combination of any several of the technological parameters of the crushing/preselection system, the technological parameters of the grinding system and the technological parameters of the screening system;
And/or
The mineral separation efficiency index comprises one or a combination of any several of recovery rate, overflow fineness, crushing ratio index and crushing granularity.
3. The beneficiation efficiency prediction method of beneficiation system according to claim 1 or 2, wherein the fusion prediction model comprises a plurality of different kinds of regression prediction modules and a fusion module; the regression prediction modules take mineral attributes and technological parameters of the mineral separation system as input and mineral separation efficiency indexes as output; the outputs of the regression prediction modules are connected with the input of the fusion module, and the fusion module is used for fusing the prediction results output by the regression prediction modules of different types to obtain and output a final prediction result.
4. A beneficiation efficiency prediction method of a beneficiation system according to claim 3, wherein the regression prediction module comprises a combination of any of a neural network regression module, a vector support vector regression module and a multiple linear regression module;
And/or
The fusion module is a weighted fusion module, and the weights of the regression prediction modules are equal.
5. An intelligent decision-making method for technological parameters of a mineral separation system is characterized by comprising the following steps:
Setting constraints, and constructing an optimization model based on the beneficiation efficiency prediction method of the beneficiation system in any one of claims 1-4; the optimization model takes the input quantity of the fusion prediction model as an optimization variable and takes the output quantity of the fusion prediction model as an optimal target;
and solving an optimal solution of the optimization model, and selecting minerals with corresponding attributes according to the technological parameters corresponding to the optimal solution.
6. The intelligent decision-making method for the technological parameters of the mineral separation system according to claim 5, wherein the technological parameters of the mineral separation system comprise the technological parameters of a crushing/preselection system, the technological parameters of a grinding system and the technological parameters of a screening system;
The process parameters of the crushing/preselection system include: breaking grain size, breaking ratio index and breaking grain size;
the technological parameters of the ore grinding system comprise overflow concentration, ore grinding work index and overflow fineness; the technological parameters of the sorting system comprise: concentrate grade, tailing grade and beneficiation recovery rate;
And/or
The optimization targets of the optimization model are mineral separation efficiency indexes, wherein the mineral separation efficiency indexes comprise recovery rate, overflow fineness, crushing ratio indexes and crushing granularity;
the optimization target is that the recovery rate is maximum;
The constraint is that overflow fineness, crushing ratio index and crushing granularity are all in a preset range.
7. The intelligent decision-making method for technological parameters of a beneficiation system according to claim 6, wherein the fusion prediction model comprises a plurality of different kinds of regression prediction modules and a fusion module; the regression prediction module comprises a neural network regression module, a vector support vector regression module and a multiple linear regression module; the objective function of the optimization model is as follows:
The constraints of the optimization model are as follows:
Wherein, Is the output value of the neural network regression module,For the output value of the vector support vector regression module,The output value of the multiple linear regression module;,, the overflow fineness predicted by the neural network regression module, the vector support vector regression module and the multiple linear regression module is respectively obtained; ,, the crushing ratio indexes are predicted by the neural network regression module, the vector support vector regression module and the multiple linear regression module respectively; ,, The crushing granularity predicted by the neural network regression module, the vector support vector regression module and the multiple linear regression module is respectively obtained; ,,,,, Respectively the minimum and maximum constraints of overflow fineness, crushing ratio index and crushing granularity, Is an input parameter of the three regression prediction modules.
8. The intelligent decision-making method for technological parameters of a mineral separation system according to claim 7, wherein an optimal solution of the optimization model is solved by an artificial swarm optimization algorithm, the artificial swarm optimization algorithm uses the objective function as an fitness function thereof, and uses mineral properties and technological parameters of the mineral separation system as food source coordinates of a bee colony.
9. The intelligent decision-making method for technological parameters of a beneficiation system according to claim 8, wherein the optimal solution of the optimization model is solved by an artificial bee colony optimization algorithm, and the method specifically comprises the following steps:
s1, determining the dimension of a bee colony according to a mineral separation efficiency index, and setting the scale and the maximum iteration number of the bee colony, wherein the bee colony comprises employment bees, following bees and reconnaissance bees;
s2, initializing the swarm coordinates, taking the initialized coordinates as a current food source, and calculating the adaptability of the current food source of the swarm;
S3, hiring bees to find candidate solutions in the neighborhood of the current food source, calculating the adaptability of the candidate solutions, judging whether the adaptability of the candidate solutions is larger than the adaptability of the current food source, if not, maintaining the coordinates of the current food source unchanged, and if so, replacing the coordinates of the candidate solutions with new food source coordinates;
S4, searching a candidate solution in the neighborhood of the current food source by following the bees according to the searching probability, calculating the adaptability of the candidate solution, judging whether the adaptability of the candidate solution is greater than the adaptability of the current food source coordinates, if not, maintaining the current food source coordinates unchanged, and if so, iterating the coordinates of the candidate solution into new food source coordinates;
S5, the scout bees are used for comparing the duration period of the currently used food source coordinates with the preset period, randomly searching candidate solutions from the global solving domain when the currently used food source coordinates are not updated in the preset period, if the fitness of the currently used food source coordinates is not greater than that of the currently used food source coordinates, maintaining the currently used food source coordinates unchanged, and if the fitness of the currently used food source coordinates is not greater than that of the currently used food source coordinates, iterating the coordinates of the candidate solutions into new food source coordinates;
s6, updating the pheromone, adjusting the search directions and the distances of the employment bees and the following bees according to the updated pheromone, and adjusting the search probability of the following bees;
s7, evaluating the overall performance of the current bee colony, judging whether the overall performance of the current bee colony is lower than the target performance, increasing the number of bees in the bee colony when the overall performance of the current bee colony is lower than the target performance, and reducing the number of bees in the bee colony when the overall performance of the current bee colony is higher than the target performance;
S8, judging whether the current bee colony reaches the maximum iteration number, if not, returning to S3, otherwise, outputting the current food source coordinates as an optimal solution.
10. The intelligent decision-making method for technological parameters of a beneficiation system according to claim 9, wherein the step S6 comprises:
for each food source Initializing a pheromone intensityRepresenting the food sourceIs the mass of (3);
after each iteration, according to the fitness of the food source Updating the intensity of the pheromone:
Wherein, Is the evaporation coefficient of the pheromone,Is based on the amount of pheromone increased by the mass of the food source, andProportional to the ratio;
calculating the search probability of the following bees according to the intensity of the pheromone:
Wherein the method comprises the steps of AndIs a parameter controlling the influence of pheromone intensity and fitness, SN is the total number of food sources,Representing pheromone intensityIs adjusted by a weight of the model; whileIs a power exponent for adjusting the influence degree of the intensity of the pheromone; Is the first The fitness value of the individual food sources,Representing a solutionIs a fitness function value of (a);
adjusting search direction and distance according to the pheromone intensity when hiring and following bees to conduct neighborhood searches near food sources;
And/or
The step S7 includes:
Setting an evaluation function Evaluating the overall performance of the bee colony after each iteration;
Wherein the function is evaluated The method comprises the following steps:
Wherein, AndRespectively two different weight coefficients; Representing the iteration number; indicating the optimal food position and the position of the food, Indicating the degree of dispersion of the solution;
setting adaptive parameters ，The number parameters representing employment bees, following bees and scout bees, respectively;
after each iteration, according to the number parameter and the number parameter of the current period of the bee colony The number parameter of the next cycle of employment bees, follow bees and scout bees:
Wherein, The number of iterations is indicated and,The method is a function for adjusting the quantity according to the evaluation result, and is used for increasing the quantity parameter when the current overall performance is lower than the target performance and reducing the quantity parameter when the current overall performance is higher than the target performance;
According to new quantity parameters Updating the number of bees:
Wherein, Is the total number of bees;，， representing the number of employment bees, following bees and spying bees.
11. A computer system 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 steps of the method of any of the preceding claims 1 to 10 when the computer program is executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410962734.1A CN118536677B (en) | 2024-07-18 | 2024-07-18 | Intelligent decision-making method and system for technological parameters of mineral separation system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410962734.1A CN118536677B (en) | 2024-07-18 | 2024-07-18 | Intelligent decision-making method and system for technological parameters of mineral separation system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118536677A true CN118536677A (en) | 2024-08-23 |
| CN118536677B CN118536677B (en) | 2024-10-01 |
Family
ID=92388135
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410962734.1A Active CN118536677B (en) | 2024-07-18 | 2024-07-18 | Intelligent decision-making method and system for technological parameters of mineral separation system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118536677B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119916683A (en) * | 2024-12-31 | 2025-05-02 | 长沙矿冶研究院有限责任公司 | Intelligent decision-making method and system for dry magnetic separator |
| CN120068671A (en) * | 2025-04-28 | 2025-05-30 | 山西暖洋洋科技有限公司 | Intelligent mineral separation system and method based on digital brain |
| CN120286168A (en) * | 2025-04-14 | 2025-07-11 | 广东启明机电设备有限公司 | Automatic processing system and method for by-products of producing and processing generator set parts |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104267695A (en) * | 2014-09-23 | 2015-01-07 | 东北大学 | Mineral separation process technological index optimization decision making system and method |
| US20180240018A1 (en) * | 2016-05-19 | 2018-08-23 | Jiangnan University | Improved extreme learning machine method based on artificial bee colony optimization |
| WO2021019551A1 (en) * | 2019-07-26 | 2021-02-04 | Tata Consultancy Services Limited | System and method for real-time self-optimization of manufacturing operations |
| RU2019127282A3 (en) * | 2019-08-29 | 2021-03-01 | ||
| CN115455804A (en) * | 2022-08-02 | 2022-12-09 | 辽宁科技大学 | Blasting block size prediction method based on ore rock crushing mechanism |
| CN116070750A (en) * | 2023-01-04 | 2023-05-05 | 东北大学 | A Multi-objective Robust Optimal Decision-Making Method for Mineral Processing Daily Comprehensive Production Index |
| US20230316069A1 (en) * | 2021-05-24 | 2023-10-05 | Soochow University | Wind speed prediction method based on forward complex-valued neural network |
-
2024
- 2024-07-18 CN CN202410962734.1A patent/CN118536677B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104267695A (en) * | 2014-09-23 | 2015-01-07 | 东北大学 | Mineral separation process technological index optimization decision making system and method |
| US20180240018A1 (en) * | 2016-05-19 | 2018-08-23 | Jiangnan University | Improved extreme learning machine method based on artificial bee colony optimization |
| WO2021019551A1 (en) * | 2019-07-26 | 2021-02-04 | Tata Consultancy Services Limited | System and method for real-time self-optimization of manufacturing operations |
| RU2019127282A3 (en) * | 2019-08-29 | 2021-03-01 | ||
| US20230316069A1 (en) * | 2021-05-24 | 2023-10-05 | Soochow University | Wind speed prediction method based on forward complex-valued neural network |
| CN115455804A (en) * | 2022-08-02 | 2022-12-09 | 辽宁科技大学 | Blasting block size prediction method based on ore rock crushing mechanism |
| CN116070750A (en) * | 2023-01-04 | 2023-05-05 | 东北大学 | A Multi-objective Robust Optimal Decision-Making Method for Mineral Processing Daily Comprehensive Production Index |
Non-Patent Citations (2)
| Title |
|---|
| 景广军, 梁雪梅, 李松仁: "遗传神经网络预测模型在选矿中的应用", 有色金属, vol. 52, no. 04, 15 November 2000 (2000-11-15), pages 27 - 32 * |
| 李春生等: "针对小规模数据集的多模型融合算法研究", 计算机技术与发展, vol. 30, no. 2, 29 February 2020 (2020-02-29), pages 63 - 66 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119916683A (en) * | 2024-12-31 | 2025-05-02 | 长沙矿冶研究院有限责任公司 | Intelligent decision-making method and system for dry magnetic separator |
| CN120286168A (en) * | 2025-04-14 | 2025-07-11 | 广东启明机电设备有限公司 | Automatic processing system and method for by-products of producing and processing generator set parts |
| CN120286168B (en) * | 2025-04-14 | 2025-10-21 | 广东启明机电设备有限公司 | Automatic processing system and method for by-products of producing and processing generator set parts |
| CN120068671A (en) * | 2025-04-28 | 2025-05-30 | 山西暖洋洋科技有限公司 | Intelligent mineral separation system and method based on digital brain |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118536677B (en) | 2024-10-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN118536677B (en) | Intelligent decision-making method and system for technological parameters of mineral separation system | |
| Praveena et al. | A literature review on supervised machine learning algorithms and boosting process | |
| Chen et al. | A hybrid fuzzy time series model based on granular computing for stock price forecasting | |
| Tian et al. | Steel plates fault diagnosis on the basis of support vector machines | |
| CN118096085B (en) | Flour production line equipment operation and maintenance management method based on Internet of things | |
| CN112488283A (en) | Improved multi-target grey wolf optimization algorithm | |
| Hegazy et al. | Comparative study between FPA, BA, MCS, ABC, and PSO algorithms in training and optimizing of LS-SVM for stock market prediction | |
| CN116934220A (en) | Intelligent warehouse layout method based on intelligent data analysis and algorithm optimization | |
| Gong et al. | Interactive evolutionary algorithms with decision-maker׳ s preferences for solving interval multi-objective optimization problems | |
| Li et al. | Optuna-DFNN: An Optuna framework driven deep fuzzy neural network for predicting sintering performance in big data | |
| Feng et al. | Variable coded hierarchical fuzzy classification model using DNA coding and evolutionary programming | |
| Qi et al. | Reservoir flood control operation using multi-objective evolutionary algorithm with decomposition and preferences | |
| CN115271237A (en) | An Industrial Data Quality Prediction Method Based on Improved PSO-GA and SVM | |
| CN110807543A (en) | An investment portfolio optimization method and device based on group decision intelligent search | |
| Liu et al. | IWOA-RNN: An improved whale optimization algorithm with recurrent neural networks for traffic flow prediction | |
| He et al. | Reference point reconstruction-based firefly algorithm for irregular multi-objective optimization | |
| CN110119846A (en) | A kind of multiple target deepness belief network for mid-term electric load forecasting | |
| Ramasubramanian et al. | Strategies to assure optimal trade-offs among competing objectives for the genetic improvement of soybean | |
| CN116304518A (en) | Method and system for constructing heterogeneous graph convolutional neural network model for information recommendation | |
| Basak et al. | RSO: A novel reinforced swarm optimization algorithm for feature selection | |
| CN120029071B (en) | A precise control method and system for intelligent control cabinet based on fusion fuzzy algorithm | |
| Qin et al. | Decomposition-based multiobjective evolutionary algorithm with density estimation-based dynamical neighborhood strategy | |
| Khadem et al. | Improving the performance of adaptive neural fuzzy inference system (ANFIS) using a new meta-heuristic algorithm | |
| Ding et al. | GA-based principal component selection for production performance estimation in mineral processing | |
| Ghasemi et al. | Optimising SAG mill throughput and circulating load using machine learning models: A multi-objective approach for identifying optimal process parameters |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |