Mill collaborative optimization control method and system based on digital twinning

Abstract

The invention discloses a mill collaborative optimization control method and system based on digital twinning, belongs to the technical field of intelligent manufacturing, and is used for solving the technical problems that an existing mill control system has obvious defects in the aspects of perception of complex working conditions in a mill, self-adaptive control, collaborative optimization of upstream and downstream equipment and the like. The method comprises the following steps: constructing a digital twinborn model of a target mill; collecting multi-source operation data of the target mill in real time through a sensor group arranged on the target mill; performing data fusion on the multi-source operation data to obtain multi-source fusion data; transmitting the multi-source fusion data to a digital twin model, and simulating the running state of the mill; generating a control strategy of the target mill based on mill load parameters of the digital twin model in the mill running state simulation process; and the control strategy is sent to a control module of the target mill, so that the running state of the target mill is adjusted in real time.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing technology, and in particular to a mill collaborative optimization control method and system based on digital twins. Background Technology

[0002] In modern industrial production, grinding mills, as a key grinding equipment, are widely used in many industries such as cement, mining, and power. The operating status of the grinding mill directly affects product quality, output, and energy consumption. However, traditional grinding mill control systems have many problems.

[0003] On the one hand, the internal working process of a mill is extremely complex, involving multiple stages such as material crushing, grinding, and conveying, and encompassing multiple disciplines including mechanics, fluid dynamics, and thermodynamics. Current monitoring methods struggle to comprehensively and accurately obtain real-time, precise operating parameters within the mill, such as material filling rate, grinding media motion state, and material-to-ball ratio. This prevents operators from clearly understanding the true operating conditions inside the mill, making it difficult to make accurate control decisions.

[0004] On the other hand, most existing mill control systems rely on manual adjustments based on operator experience or employ simple automatic control strategies based on fixed parameters. This control method cannot dynamically and adaptively adjust to the constantly changing operating conditions during mill operation. For example, when the properties of the raw materials (such as hardness and particle size distribution) fluctuate, the optimal operating parameters of the mill will also change accordingly. However, traditional control systems struggle to quickly and accurately track these changes, leading to decreased mill efficiency, unstable product quality, and increased energy consumption.

[0005] Furthermore, mills are typically a component of a production line, working in conjunction with upstream and downstream equipment. However, currently, there is a lack of effective collaborative optimization mechanisms between these devices, making it impossible to achieve globally optimal control of the entire production process. For example, the coordination between the mill and feeding and classifying equipment is often imprecise, easily leading to problems such as overfeeding or underfeeding, and low classification efficiency, further impacting the performance of the entire production system.

[0006] In summary, existing mill control systems have significant shortcomings in sensing complex internal working conditions, adaptive control, and collaborative optimization with upstream and downstream equipment, urgently requiring an innovative technology to solve these problems. Summary of the Invention

[0007] This invention provides a mill collaborative optimization control method and system based on digital twins to solve the following technical problem: existing mill control systems have significant shortcomings in sensing complex working conditions inside the mill, adaptive control, and collaborative optimization with upstream and downstream equipment.

[0008] The embodiments of the present invention adopt the following technical solutions:

[0009] On the one hand, embodiments of the present invention provide a mill collaborative optimization control method based on digital twins, the method comprising: constructing a digital twin model of the target mill;

[0010] The multi-source operating data of the target mill are collected in real time through a sensor array deployed on the target mill.

[0011] The multi-source operational data is fused to obtain multi-source fused data;

[0012] The multi-source fusion data is transmitted to the digital twin model to simulate the mill's operating status;

[0013] Based on the mill load parameters obtained during the simulation of mill operation using a digital twin model, a control strategy for the target mill is generated.

[0014] The control strategy is sent to the control module of the target mill to adjust the operating status of the target mill in real time.

[0015] In one feasible implementation, constructing a digital twin model of the target mill specifically includes:

[0016] Based on 3D modeling technology, a 3D model of the mechanical structure of the target mill is created, and the attribute parameters of each mechanical structure are defined to obtain a mechanical structure model; wherein, the mechanical structure includes at least a cylinder, a liner, grinding media, and a transmission device; the attribute parameters include at least geometric dimensions, material properties, and connection relationships;

[0017] Based on the internal operating principle of the mill, a mathematical model of the material grinding process inside the target mill is established to obtain a physical process model;

[0018] The mechanical structure model and the physical process model are fused together to obtain a digital twin model of the target mill.

[0019] The digital twin model is verified and calibrated using actual production data to ensure that it accurately reflects the actual operating status of the mill.

[0020] In one feasible implementation, multi-source operating data of the target mill is collected in real time through a sensor array deployed on the target mill, specifically including:

[0021] A sensor array is installed at key locations of the target mill, including:

[0022] A microphone array is installed on the surface of the target mill cylinder to collect the collision sound signals between the grinding media and the material;

[0023] Temperature sensors, triaxial vibration sensors, and material level sensors are deployed at key locations of the target mill to collect data on temperature, vibration acceleration, and material level during equipment operation, respectively.

[0024] The sensor array is used to collect real-time operating data of the target mill.

[0025] Simultaneously, the design drawings, process parameters, and historical operating data of the target mill are acquired, which, together with the real-time operating data, constitute the multi-source operating data.

[0026] In one feasible implementation, data fusion is performed on the multi-source operational data to obtain multi-source fused data, specifically including:

[0027] The multi-source operational data is preprocessed; wherein the preprocessing operation includes at least: data cleaning, data denoising, and data normalization.

[0028] The preprocessed multi-source operational data is fused using a data fusion algorithm to form a unified data representation from different types and sources, thus obtaining preliminary fused data.

[0029] Data mining and feature extraction are performed on the preliminary fused data to obtain the mill's implicit information; wherein, the mill's implicit information includes at least the mill's operating rules, mill fault characteristics, and mill performance indicators;

[0030] The preliminary fusion data and the mill's implicit information are combined to construct the multi-source fusion data.

[0031] In one feasible implementation, the multi-source fusion data is transmitted to the digital twin model for mill operation state simulation, specifically including:

[0032] The multi-source fused data is transmitted to the digital twin model to drive the digital twin model to simulate the operating state of the target mill.

[0033] Record the predicted load parameters corresponding to the target mill within a preset time period in the future; wherein, the predicted load parameters include at least the predicted wear degree of the grinding media, the predicted grinding efficiency of the material, the predicted grinding fineness of the material, the predicted thermal load of the mill, and the predicted power consumption of the mill.

[0034] In one feasible implementation, a control strategy for the target mill is generated based on the mill load parameters during the mill operation simulation using a digital twin model, specifically including:

[0035] Based on the predicted load parameters, the operating cost function of the target mill is constructed;

[0036] With the optimization objectives of minimizing the operating cost function and maximizing the mill production efficiency, a multi-objective optimization mathematical model is constructed, and the constraints are determined based on the operating requirements of the target mill.

[0037] Under the premise of satisfying the above constraints, the optimal solution of the multi-objective optimization mathematical model is obtained by solving the particle swarm optimization algorithm to obtain the optimal operating parameters of the mill.

[0038] Based on the optimal operating parameters of the mill and the pre-built control strategy library, a mill control strategy for the future preset time period is generated.

[0039] In one feasible implementation, a mill control strategy for the future preset time period is generated based on the mill's optimal operating parameters and a pre-built control strategy library, specifically including:

[0040] Based on the actual load values ​​of the target mill in daily operation and expert experience, a control strategy library is pre-built; the control strategy library includes a rule library and a strategy library.

[0041] The optimal operating parameters are input into the control strategy library for matching, and the corresponding mill control strategy is output.

[0042] In one feasible implementation, the control strategy is sent to the control module of the target mill to adjust the operating status of the target mill in real time, specifically including:

[0043] The control strategy is converted into control commands and output to the control module of the target mill, so that the control module can adjust the operating parameters of the target mill in a future preset time period to achieve the optimization goal.

[0044] In one feasible implementation, after sending the control strategy to the control module of the target mill to adjust the operating state of the target mill in real time, the method further includes:

[0045] Based on Petri net theory, an equipment collaboration model is created for the production system consisting of the target mill and upstream and downstream equipment;

[0046] Based on the predicted load parameters output by the equipment collaboration model and the digital twin model, an equipment collaboration optimization strategy is formulated with the goal of optimizing the overall performance of the production system.

[0047] The equipment collaborative optimization strategy is applied to the production system.

[0048] On the other hand, embodiments of the present invention also provide a mill collaborative optimization control system based on digital twins, the device comprising:

[0049] The simulation module is used to construct a digital twin model of the target mill; multi-source operating data of the target mill is collected in real time through a sensor group deployed on the target mill; the multi-source operating data is fused to obtain multi-source fused data; the multi-source fused data is transmitted to the digital twin model to simulate the mill's operating status.

[0050] The intelligent control module is used to generate a control strategy for the target mill based on the mill load parameters during the mill operation state simulation process using a digital twin model; and to send the control strategy to the control module of the target mill to adjust the operating state of the target mill in real time.

[0051] This invention provides a mill collaborative optimization control method and system based on digital twins, which not only improves the operating efficiency and stability of the mill system but also significantly reduces maintenance costs, promoting the intelligent transformation of industrial production. The following are the specific beneficial effects of this system in practical applications:

[0052] (1) Based on the predictive capabilities of digital twin models (such as mill load and power consumption prediction), the intelligent control module dynamically adjusts parameters such as feed rate and rotation speed through adaptive model predictive control algorithms to maintain a stable grinding fineness qualification rate and reduce over-grinding or under-grinding of products. Furthermore, the particle swarm optimization algorithm is adopted to achieve the overall optimal goal among targets such as unit energy consumption, throughput, and equipment wear, thereby improving the mill's production efficiency and product quality.

[0053] (2) By fusing data from multiple sensors such as vibration, temperature, and acoustic emission, and combining it with digital twins to construct a health status model of the equipment, early warnings of faults such as bearing overheating and abnormal gear meshing are provided, reducing downtime due to faults. Extreme working conditions are simulated in the virtual environment of the digital twin model, generating safety thresholds and feeding them back to the control module, avoiding equipment damage caused by human error and enhancing the reliability of equipment operation and fault early warning capabilities.

[0054] (3) Through the cloud-based digital twin platform, managers can view the mill's 3D visualization model and key parameters in real time, supporting remote adjustment of control strategies, reducing response time from hours to minutes. This enables managers to monitor the mill's operating status anytime, anywhere, and to conduct remote operation and decision-making. It not only improves management efficiency but also reduces the labor intensity and safety risks for on-site personnel. Simultaneously, the system can automatically generate operation reports and data analysis, providing strong support for the company's refined management and realizing full-process intelligent and remote control. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0056] Figure 1 A flowchart of a mill collaborative optimization control method based on digital twins is provided for an embodiment of the present invention;

[0057] Figure 2 This is a schematic diagram of a mill collaborative optimization control system based on digital twins, provided as an embodiment of the present invention. Detailed Implementation

[0058] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this invention.

[0059] This invention provides a mill collaborative optimization control method based on digital twins, such as... Figure 1 As shown, the mill collaborative optimization control method based on digital twins specifically includes steps S101-S105:

[0060] S101. Construct a digital twin model of the target mill.

[0061] Specifically, based on 3D modeling technology, the mechanical structure of the target mill is modeled in 3D, and the attribute parameters of each mechanical structure are defined to obtain the mechanical structure model; wherein, the mechanical structure includes at least the cylinder, liner, grinding media and transmission device; the attribute parameters include at least geometric dimensions, material properties and connection relationships.

[0062] Furthermore, based on the internal operating principle of the mill, a mathematical model of the material grinding process inside the target mill is established, thus obtaining a physical process model.

[0063] Finally, the mechanical structure model and the physical process model are fused to obtain a digital twin model of the target mill. Then, the digital twin model is verified and calibrated using actual production data to ensure that it accurately reflects the actual operating status of the mill.

[0064] As a feasible implementation method, 3D modeling technology is used to accurately model the mechanical structure of the mill, including the cylinder, liners, grinding media, and transmission device, defining in detail the geometric dimensions, material properties, and connection relationships of each component. Then, based on principles of mechanics, kinematics, and thermodynamics, a mathematical model of the material grinding process inside the mill is established. A digital twin model of the mill is constructed to achieve a virtual mapping of the physical mill entity, supporting the simulation of the material's trajectory and particle size changes under the impact and grinding of the grinding media. Finally, model fusion and verification are performed: the mechanical structure model and the physical process model are merged to form a complete digital twin model of the mill. The model is verified and calibrated using actual production data to ensure that it accurately reflects the actual operating state of the mill.

[0065] S102. Real-time collection of multi-source operating data of the target mill through the sensor group deployed on the target mill.

[0066] Specifically, sensor assemblies are installed in key parts of the target mill, including:

[0067] A microphone array is installed on the surface of the target mill cylinder to collect the sound signals of the collision between the grinding media and the material.

[0068] Temperature sensors, triaxial vibration sensors, and material level sensors are deployed at key parts of the target mill to collect data on temperature, vibration acceleration, and material level during equipment operation.

[0069] Furthermore, real-time operating data of the target mill is collected through a sensor array. Simultaneously, design drawings, process parameters, and historical operating data of the target mill are acquired, which, together with the real-time operating data, constitute multi-source operating data.

[0070] As one feasible implementation method, sensor group deployment includes the following:

[0071] Grinding sound detection module: A microphone array is installed on the surface of the mill cylinder to collect the collision sound signals between the grinding media and the materials.

[0072] Temperature and vibration sensor module: Temperature sensors (accuracy ±0.5℃) and triaxial vibration sensors are deployed in key parts such as bearings and transmission devices to collect temperature and vibration acceleration during equipment operation.

[0073] Furthermore, various sensors are installed on the mill, including speed sensors, mill noise sensors, temperature sensors, vibration sensors, and material level sensors, to collect the mill's operating parameters in real time. Simultaneously, static data such as the mill's design drawings and process parameters, as well as historical operating data, are acquired.

[0074] S103. Perform data fusion on the multi-source operational data to obtain multi-source fused data.

[0075] Specifically, preprocessing operations are performed on multi-source operational data; the preprocessing operations include at least: data cleaning, data denoising, and data normalization.

[0076] Furthermore, the preprocessed multi-source running data is fused using a data fusion algorithm to form a unified data representation of data of different types and sources, thus obtaining preliminary fused data.

[0077] Furthermore, data mining and feature extraction are performed on the preliminary fused data to obtain the mill's implicit information; the mill's implicit information includes at least the mill's operating rules, mill fault characteristics, and mill performance indicators.

[0078] Finally, the preliminary fusion data and the mill's implicit information are combined to construct multi-source fusion data.

[0079] As a feasible implementation method, the data collected by the sensor array undergoes preprocessing operations such as cleaning, noise reduction, and normalization to remove outliers and noise interference, ensuring data accuracy and consistency. Then, data fusion algorithms (such as Kalman filtering, Bayesian inference, machine learning methods, deep learning models, feature-level / decision-level fusion, etc.) are used to fuse data of different types and sources, forming a unified data representation to provide comprehensive data support for subsequent analysis and decision-making. Furthermore, data mining and machine learning techniques are used to conduct in-depth analysis of the fused data, uncovering information such as mill operating patterns, fault characteristics, and performance indicators. Acoustic signals from inside the mill are collected using a microphone array, and spectral features are extracted and correlated with the impact intensity of the grinding media. The mill's operating status is monitored in real time using grinding sound detection and temperature / vibration sensors.

[0080] S104. Transmit the multi-source fusion data to the digital twin model to simulate the mill's operating state. Based on the mill load parameters obtained from the digital twin model during the mill's operating state simulation, generate a control strategy for the target mill.

[0081] Specifically, multi-source fused data is transmitted to the digital twin model to drive the digital twin model to simulate the operating status of the target mill.

[0082] Furthermore, the predicted load parameters corresponding to the target mill within a preset time period are recorded; wherein, the predicted load parameters include at least the predicted wear degree of the grinding media, the predicted grinding efficiency of the material, the predicted grinding fineness of the material, the predicted thermal load of the mill, and the predicted power consumption of the mill.

[0083] As a feasible implementation method, the digital twin model calculates the mill's load based on real-time collected data, including the load of the grinding media and the mill's thermal load. Based on these load conditions, the digital twin model uses an algorithm to calculate the optimal mill operating parameters, such as the mill's rotational speed and the amount of grinding media added.

[0084] Furthermore, based on the predicted load parameters, the operating cost function of the target mill is constructed.

[0085] As a feasible implementation method, the operating cost function of the target mill is: Among them, P pred x is the predicted power consumption value for the mill. d To predict the grinding fineness of materials, X ref α represents the standard grinding fineness of the material, and β represents the weighting coefficients.

[0086] With the optimization objectives of minimizing the aforementioned operating cost function and maximizing the mill's production efficiency, a multi-objective optimization mathematical model is constructed, and constraints are determined based on the target mill's operating requirements. Under the premise of satisfying the constraints, the optimal solution of the multi-objective optimization mathematical model is obtained using the particle swarm optimization algorithm, yielding the optimal operating parameters of the mill.

[0087] As a feasible implementation method, a multi-objective optimization mathematical model is established with the minimization of the mill's cost function (which can be decomposed into energy consumption minimization and product quality optimization) and the maximization of production efficiency as the main optimization objectives. Let the mill's operating parameter vector be x = [x1, x2, ..., x...]. n ], where x i These represent different operating parameters such as mill speed, feed rate, and grinding media addition. Each objective function can be expressed as:

[0088] The production efficiency objective function f1(x) can be defined, for example, as the output of qualified products per unit time.

[0089] The product quality objective function f2(x) can be measured by indicators such as the uniformity of product particle size distribution and the proportion of products with specific particle sizes.

[0090] The energy consumption objective function f3(x) is expressed as the total power consumption or unit product energy consumption during the operation of the mill.

[0091] So the objective optimization model can be expressed as: min{f1(x),f2(x),f3(x)}.

[0092] Meanwhile, considering the physical constraints and process requirements of mill operation, such as mill speed range limitations and feed rate upper limit constraints, constraint condition g is introduced. j(x)≤0, (j=1,2,…,m), ensuring that the optimization result is within the feasible range.

[0093] Furthermore, based on the optimal operating parameters of the mill and a pre-built control strategy library, a mill control strategy for a future preset time period is generated, specifically including:

[0094] Based on the actual load values ​​of the target mill during daily operation and expert experience, a control strategy library is pre-built; the control strategy library includes a rule library and a strategy library. Optimal operating parameters are input into the control strategy library for matching, and the corresponding mill control strategy is output.

[0095] As a feasible implementation method, the control strategy library contains over 200 built-in process rules (such as "reduce the feed rate by 10% when the vibration amplitude is >5mm / s"). The fuzzy controller inputs are current deviation and deviation change rate, and the output is the feed rate adjustment. The membership function adopts a Gaussian type.

[0096] S105. Send the control strategy to the control module of the target mill to adjust the operating status of the target mill in real time.

[0097] Specifically, the control strategy is converted into control commands and output to the control module of the target mill, so that the operating parameters of the target mill can be adjusted by the control module in the future within a preset time period to achieve the optimization goal.

[0098] As a feasible implementation method, the optimal operating parameters calculated by the optimization algorithm are sent to the mill's control device through the output of the control system, enabling dynamic adjustment and optimization of the mill. The collaborative control module adjusts the mill's operating parameters based on the output of the digital twin model, using a programmable logic controller or other types of industrial controller.

[0099] Furthermore, based on Petri net theory, an equipment collaboration model is created for the production system consisting of the target mill and upstream and downstream equipment.

[0100] Based on the predicted load parameters output by the equipment collaboration model and the digital twin model, an equipment collaboration optimization strategy is formulated with the goal of optimizing the overall performance of the production system. This strategy is then applied to the production system.

[0101] As a feasible implementation method, a production system consisting of a mill and upstream and downstream equipment (such as crushers, classifiers, and conveying equipment) is considered, and a Petri net theory is used to establish a collaborative model of the equipment. Petri nets intuitively describe the operational logic and collaborative relationships of each piece of equipment in the production system through places, transitions, and directed arcs. For example, the discharge rate of the crusher affects the feed rate of the mill; the product particle size of the mill is related to the classification effect of the classifier; and the conveying capacity of the conveying equipment restricts the material flow of the entire production process. By modeling these relationships, the collaborative operation rules of each piece of equipment under different operating conditions are clarified.

[0102] Then, based on the equipment collaboration model, a collaborative optimization strategy is formulated with the goal of optimizing the overall performance of the production system. By optimizing the operating parameters and working sequence of each piece of equipment, efficient flow and rational allocation of materials within the production system are achieved. For example, based on the grinding capacity of the mill and the product particle size requirements, the crushing ratio and feeding speed of the crusher are dynamically adjusted to make the particle size distribution of the material entering the mill more suitable for grinding. At the same time, based on the mill's output particle size and output, the classification parameters of the classifier are adjusted in real time to ensure that the classification effect matches the mill's production capacity, reducing over-grinding and under-grinding. Through this collaborative optimization among equipment, the operating efficiency and product quality of the entire production system are improved, while energy consumption and production costs are reduced.

[0103] As a feasible implementation method, a multi-source data fusion and analysis module is used to monitor in real time the fluctuations in material properties (such as changes in material hardness and moisture content) and changes in market demand (such as adjustments to product particle size requirements and changes in output demand) during the production process. This real-time operating condition data is rapidly fed back to the collaborative optimization module via a sensor network and communication system. When a change in operating conditions is detected, the collaborative optimization module quickly initiates a dynamic optimization adjustment mechanism. First, based on the real-time operating condition data, the weights and constraints of the objective function in the multi-objective optimization model are re-evaluated. For example, if the market demand for product particle size increases, the weight of the product quality objective function is appropriately increased; if the material hardness increases, the constraint range of the mill operating parameters is adjusted. Then, based on the updated multi-objective optimization model, the NSGA-II algorithm is used to re-solve the problem, obtaining the optimal combination of mill operating parameters and equipment coordination scheme adapted to the new operating conditions. Simultaneously, the optimization results are transmitted to the intelligent control module to adjust the operating status of the mill and upstream and downstream equipment in real time, ensuring that the mill maintains efficient and stable operation under dynamic conditions, achieving continuous optimization of the production system.

[0104] In addition, embodiments of the present invention also provide a mill collaborative optimization control system based on digital twins, such as... Figure 2 As shown, the mill collaborative optimization control system 200 based on digital twins specifically includes:

[0105] The simulation module 210 is used to construct a digital twin model of the target mill; multi-source operating data of the target mill are collected in real time through a sensor group deployed on the target mill; the multi-source operating data is fused to obtain multi-source fused data; and the multi-source fused data is transmitted to the digital twin model to simulate the mill's operating status.

[0106] The intelligent control module 220 is used to generate a control strategy for the target mill based on the mill load parameters during the mill operation state simulation process using a digital twin model; and to send the control strategy to the control module of the target mill to adjust the operating state of the target mill in real time.

[0107] The various embodiments in this invention are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0108] The foregoing has described specific embodiments of the present invention. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0109] The above description is merely an embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, various modifications and variations can be made to the embodiments of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the embodiments of the present invention should be included within the protection scope of the present invention.

Claims

1. A mill collaborative optimization control method based on digital twins, characterized in that, The method includes: Construct a digital twin model of the target mill; The multi-source operating data of the target mill are collected in real time through a sensor array deployed on the target mill. The multi-source operational data is fused to obtain multi-source fused data; The multi-source fusion data is transmitted to the digital twin model to simulate the mill's operating status; Based on the mill load parameters obtained during the simulation of mill operation using a digital twin model, a control strategy for the target mill is generated. The control strategy is sent to the control module of the target mill to adjust the operating status of the target mill in real time.

2. The mill collaborative optimization control method based on digital twins according to claim 1, characterized in that, Constructing a digital twin model of the target mill specifically includes: Based on 3D modeling technology, a 3D model of the mechanical structure of the target mill is created, and the attribute parameters of each mechanical structure are defined to obtain a mechanical structure model; wherein, the mechanical structure includes at least a cylinder, a liner, grinding media, and a transmission device; the attribute parameters include at least geometric dimensions, material properties, and connection relationships; Based on the internal operating principle of the mill, a mathematical model of the material grinding process inside the target mill is established to obtain a physical process model; The mechanical structure model and the physical process model are fused together to obtain a digital twin model of the target mill. The digital twin model is verified and calibrated using actual production data to ensure that it accurately reflects the actual operating status of the mill.

3. The mill collaborative optimization control method based on digital twins according to claim 1, characterized in that, The target mill uses a sensor array deployed on its surface to collect multi-source operational data in real time, including: A sensor array is installed at key locations of the target mill, including: A microphone array is installed on the surface of the target mill cylinder to collect the collision sound signals between the grinding media and the material; Temperature sensors, triaxial vibration sensors, and material level sensors are deployed at key locations of the target mill to collect data on temperature, vibration acceleration, and material level during equipment operation, respectively. The sensor array is used to collect real-time operating data of the target mill. Simultaneously, the design drawings, process parameters, and historical operating data of the target mill are acquired, which, together with the real-time operating data, constitute the multi-source operating data.

4. The mill collaborative optimization control method based on digital twins according to claim 1, characterized in that, The multi-source operational data is fused to obtain multi-source fused data, specifically including: The multi-source operational data is preprocessed; wherein the preprocessing operation includes at least: data cleaning, data denoising, and data normalization. The preprocessed multi-source operational data is fused using a data fusion algorithm to form a unified data representation from different types and sources, thus obtaining preliminary fused data. Data mining and feature extraction are performed on the preliminary fused data to obtain the mill's implicit information; wherein, the mill's implicit information includes at least the mill's operating rules, mill fault characteristics, and mill performance indicators; The preliminary fusion data and the mill's implicit information are combined to construct the multi-source fusion data.

5. The mill collaborative optimization control method based on digital twins according to claim 1, characterized in that, The multi-source fusion data is transmitted to the digital twin model to simulate the mill's operating state, specifically including: The multi-source fused data is transmitted to the digital twin model to drive the digital twin model to simulate the operating state of the target mill. Record the predicted load parameters corresponding to the target mill within a preset time period in the future; wherein, the predicted load parameters include at least the predicted wear degree of the grinding media, the predicted grinding efficiency of the material, the predicted grinding fineness of the material, the predicted thermal load of the mill, and the predicted power consumption of the mill.

6. The mill collaborative optimization control method based on digital twins according to claim 5, characterized in that, Based on the mill load parameters in the mill operation simulation process using a digital twin model, a control strategy for the target mill is generated, specifically including: Based on the predicted load parameters, the operating cost function of the target mill is constructed; With the optimization objectives of minimizing the operating cost function and maximizing the mill production efficiency, a multi-objective optimization mathematical model is constructed, and the constraints are determined based on the operating requirements of the target mill. Under the premise of satisfying the above constraints, the optimal solution of the multi-objective optimization mathematical model is obtained by solving the particle swarm optimization algorithm to obtain the optimal operating parameters of the mill. Based on the optimal operating parameters of the mill and the pre-built control strategy library, a mill control strategy for the future preset time period is generated.

7. The mill collaborative optimization control method based on digital twins according to claim 6, characterized in that, Based on the optimal operating parameters of the mill and the pre-built control strategy library, a mill control strategy for the future preset time period is generated, specifically including: Based on the actual load values ​​of the target mill in daily operation and expert experience, a control strategy library is pre-built; the control strategy library includes a rule library and a strategy library. The optimal operating parameters are input into the control strategy library for matching, and the corresponding mill control strategy is output.

8. The mill collaborative optimization control method based on digital twins according to claim 7, characterized in that, The control strategy is sent to the control module of the target mill to adjust the operating status of the target mill in real time, specifically including: The control strategy is converted into control commands and output to the control module of the target mill, so that the control module can adjust the operating parameters of the target mill in a future preset time period to achieve the optimization goal.

9. The mill collaborative optimization control method based on digital twins according to claim 1, characterized in that, After sending the control strategy to the control module of the target mill to adjust the operating status of the target mill in real time, the method further includes: Based on Petri net theory, an equipment collaboration model is created for the production system consisting of the target mill and upstream and downstream equipment; Based on the predicted load parameters output by the equipment collaboration model and the digital twin model, an equipment collaboration optimization strategy is formulated with the goal of optimizing the overall performance of the production system. The equipment collaborative optimization strategy is applied to the production system.

10. A mill collaborative optimization control system based on digital twins, characterized in that, The device includes: The simulation module is used to construct a digital twin model of the target mill; multi-source operating data of the target mill is collected in real time through a sensor group deployed on the target mill; the multi-source operating data is fused to obtain multi-source fused data; the multi-source fused data is transmitted to the digital twin model to simulate the mill's operating status. The intelligent control module is used to generate a control strategy for the target mill based on the mill load parameters during the mill operation state simulation process using a digital twin model; and to send the control strategy to the control module of the target mill to adjust the operating state of the target mill in real time.

Images