Multi-source monitoring and dynamic regulation and control method and system for mill, terminal and medium

Abstract

The invention relates to a multi-source monitoring and dynamic regulation and control method and system for a mill, a terminal and a medium, and belongs to the field of mill equipment control. According to the method, multi-source data such as spindle vibration, grinding cavity acoustic emission, gearbox temperature and barrel rotating speed are collected, and multi-dimensional features are extracted by combining first-order difference, energy density spectrum and thermal imaging technologies; and constructing a regression model fusing a linear term and a Gaussian kernel function, and synchronously predicting the operation efficiency, the energy consumption and the grinding particle size. And parameters such as the rotating speed and the feeding amount are dynamically adjusted based on the prediction result, and a three-level early warning mechanism is established. The system comprises a template management module, a data acquisition module, a feature extraction module, a modeling and prediction module, a dynamic control module and an early warning management module, and supports scenarized parameter configuration and adaptive regulation and control. Through multi-source data fusion and an intelligent prediction model, the problems of parameter stiffness and response lag of a traditional mill are solved, the working condition adaptability is remarkably improved, and efficient and safe operation of equipment is guaranteed.

Description

Technical Field

[0001] The present invention belongs to the technical field of mill equipment control, and particularly relates to a multi-source monitoring and dynamic regulation method, system, terminal and medium for a mill. Background Art

[0002] In industries such as mining, building materials, and metallurgy, as an important grinding equipment, the operating efficiency, energy consumption level, and product quality of the mill directly affect the overall production efficiency. Most traditional mill equipment adopts static parameter configuration, usually setting fixed rotational speed, water supply, coal supply, and additive dosage based on experience or single working conditions. However, in the actual production process, the hardness, particle size, moisture content, and load of the ore change frequently. The traditional fixed parameter configuration is difficult to adapt to the changing working conditions in real time, easily leading to increased energy consumption, reduced grinding effect, and even equipment failures.

[0003] Some existing technologies attempt to introduce data acquisition means, and realize digital acquisition of the operating state by deploying vibration sensors, temperature probes, current transformers and other devices on the mill body and auxiliary equipment. For example, vibration monitoring technology can capture the vibration frequency spectrum characteristics of the mill cylinder to indirectly reflect the load state, the temperature detection system can track the thermal state changes of the bearing and lubrication system, and current analysis is used to evaluate the motor load characteristics.

[0004] However, the regulation in the existing technology is limited to a single signal source, and only simple alarm prompts are carried out. A multi-source feature correlation analysis and real-time dynamic optimization adjustment mechanism has not been established. In addition, there is a lack of an intelligent regulation system guided by an operating state prediction model in the existing technology. Especially in abnormal working conditions, there is a lack of hierarchical response and adaptive adjustment strategies, resulting in insufficient overall intelligence level and production flexibility of the mill, and it is difficult to meet the high-efficiency and safe operation requirements in the modern complex manufacturing environment. Summary of the Invention

[0005] Aiming at the problems in the existing technology, the present invention provides a multi-source monitoring and dynamic regulation method, system, terminal and medium for a mill, and solves the problems that in the existing technology, there is a lack of hierarchical response and adaptive adjustment strategies in abnormal working conditions, resulting in insufficient overall intelligence level and production flexibility of the mill, and it is difficult to meet the high-efficiency and safe operation requirements in the modern complex manufacturing environment.

[0006] The technical solution adopted by the present invention is as follows: In the first aspect, the present application provides a multi-source monitoring and dynamic regulation method for a mill, including the following steps: Step S1, preset a scenario template, and select the corresponding scenario template according to the corresponding mill usage scenario; Step S2, collect the operating state data of the mill, including the main shaft vibration acceleration signal, the acoustic emission signal of the grinding chamber, the gearbox temperature distribution data, and the cylinder rotational speed signal; Step S3: Extract the change characteristics of the vibration acceleration signal through first-order difference and second-order difference, extract the energy density spectrum characteristics of the acoustic emission signal, extract the local abnormal temperature rise characteristics and rotational speed fluctuation characteristics of the gearbox temperature distribution, and generate a feature data set; Step S4: Based on the feature data set, construct a regression prediction model with the mill operating efficiency, unit energy consumption, and grinding particle size as the target outputs. The model form includes a linear term and a non-linear Gaussian kernel term; Step S5: According to the target values output by the regression prediction model, dynamically adjust the mill rotational speed, water supply, coal supply, and additive dosage in combination with the controller, where the controller output is dynamically updated according to the real-time deviation change; Step S6: Set up a three-level early warning response mechanism based on the degree of abnormality. The first-level early warning triggers fine-tuning of the operating parameters, the second-level early warning triggers joint optimization regulation, the third-level early warning triggers shutdown protection, and synchronize the feature data set with the early warning records.

[0007] Preferably, in step S2, the acquisition of the main shaft vibration acceleration signal is completed by MEMS acceleration sensors symmetrically installed on both sides of the main shaft bearing housing; The acquisition of the acoustic emission signal in the grinding chamber is carried out by deploying wide-band acoustic emission sensors in the feed port area; By performing energy density spectrum analysis on the acoustic signal, extract the high energy density peaks corresponding to abnormal impact events and the changes in the occurrence frequency; The acquisition of the gearbox temperature distribution data uses a non-contact infrared thermal imager.

[0008] Preferably, in step S3, the acquired main shaft vibration acceleration signal is processed by first-order difference to obtain the velocity change trend, and processed by second-order difference to obtain the acceleration mutation amount. The maximum vibration amplitude and vibration trend slope are output in real time as the input parameters for dynamic adjustment; By performing energy density spectrum analysis on the acoustic signal, extract the high energy density peaks corresponding to abnormal impact events and the changes in the occurrence frequency; The gearbox temperature distribution data is acquired in the form of a two-dimensional thermal distribution matrix. By extracting the temperature gradient change rate, local extreme points, and hot spot diffusion speed as the local abnormal temperature rise characteristics of the gearbox temperature distribution; The main shaft rotational speed signal is extracted by FFT transformation to obtain the main frequency and high-order harmonic characteristics, and the harmonic amplitude change amount is used as the rotational speed fluctuation characteristic of the mill.

[0009] Preferably, in step S4, the regression prediction model is:

[0010] Where: is the predicted output result, which includes the unit energy consumption E, the grinding particle size d, and the production efficiency η; is the bias term; is the linear regression coefficient of the i-th feature; is the i-th input feature variable; is the k-th input feature variable; is the j-th input feature variable; is the feature cross-term coefficient, indicating the contribution of the joint action of features and to the predicted value; is the weight coefficient of the Gaussian kernel function, controlling the influence intensity of the Gaussian kernel function of the k-th feature; is the mean of the k-th feature; is the standard deviation of the k-th feature.

[0011] Preferably, in step S5, when performing dynamic adjustment based on the target value output by the regression prediction model, according to the deviation between the predicted value of the unit energy consumption E and the preset value of the preset scenario template, the mill speed is adjusted in real time, and the speed adjustment strategy adopts a piecewise linear function. When the deviation is lower than the preset fine-tuning threshold, the adjustment amplitude is set to ±3% of the base speed. When the deviation exceeds the fine-tuning threshold, the adjustment amplitude increases to ±10%.

[0012] Preferably, in step S6, in the three-level early warning mechanism set based on the degree of abnormality, the trigger condition for the first-level early warning is that the change amplitude of any feature parameter exceeds the corresponding first-level deviation threshold but does not reach the corresponding second-level early warning trigger condition; The trigger condition for the second-level early warning is that the feature parameter changes and simultaneously satisfies at least two first-level early warning conditions or there is a single feature parameter exceeding the corresponding second-level deviation threshold; The trigger condition for the third-level early warning is that the feature parameter changes and simultaneously satisfies at least two second-level early warning conditions or there is a single feature parameter exceeding the corresponding third-level deviation threshold.

[0013] Preferably, after the first-level early warning is triggered, the system performs a first-level amplitude fine-tuning operation on the corresponding single operating parameter, and the adjustment amplitude is automatically controlled in multiple levels according to the deviation; After the second-level early warning is triggered, the system performs a comprehensive optimization adjustment of the linkage parameters, giving priority to adjusting the mill speed and the coal feeding amount, and then adjusting the water supply amount and the additive dosage. Each adjustment amount is dynamically optimized according to the output result of the real-time prediction model and the deviation trend; After the third-level early warning is triggered, an emergency shutdown command is immediately executed, and at the same time, the vibration signal, acoustic emission signal, temperature distribution data, and speed change within the last 60 seconds are recorded, and an exception report is pushed simultaneously.

[0014] In a second aspect, the present application provides a multi-source monitoring and dynamic regulation system for a mill, including: A template management module, configured to preset scenario templates and select a corresponding scenario template according to the usage scenario of the mill; A data acquisition module, configured to acquire the operating state data of the mill, including the main shaft vibration acceleration signal, the grinding chamber acoustic emission signal, the gearbox temperature distribution data, and the cylinder rotation speed signal; A feature extraction module, configured to extract the variation features of the vibration acceleration signal through first-order difference and second-order difference processing, extract the energy density spectrum features of the acoustic emission signal, extract the local abnormal temperature rise features of the gearbox temperature distribution, and extract the rotation speed fluctuation features to generate a feature data set; A modeling and prediction module, configured to construct a regression prediction model based on the feature data set, the model taking the mill operating efficiency, unit energy consumption, and grinding particle size as the target outputs, and the regression model form including a linear term and a non-linear Gaussian kernel term; A dynamic control module, configured to dynamically adjust the rotation speed, water supply, coal supply, and additive dosage of the mill according to the target values output by the regression prediction model in combination with a controller, and the output of the controller being dynamically updated according to the real-time deviation change; An early warning management module, configured to set a three-level early warning response mechanism based on the degree of abnormality, the first-level early warning triggering fine-tuning of the operating parameters, the second-level early warning triggering joint optimization regulation, the third-level early warning triggering shutdown protection, and synchronizing the feature data set and the early warning records to the background management system.

[0015] In a third aspect, the present application provides a terminal, including: A memory, configured to store the multi-source monitoring and dynamic regulation program for the mill; A processor, configured to implement the steps of the multi-source monitoring and dynamic regulation method for the mill as described in the first aspect when executing the multi-source monitoring and dynamic regulation program for the mill.

[0016] In a fourth aspect, the present application provides a computer-readable storage medium, the computer-readable storage medium including a stored program, wherein when the program runs, it controls the device where the computer-readable storage medium is located to execute the multi-source monitoring and dynamic regulation method for the mill as described in the first aspect.

[0017] From the above technical solutions, it can be seen that the present invention has the following advantages: 1. By integrating multi-dimensional data such as vibration, acoustic emission, temperature, and rotation speed, a comprehensive perception of the operating state of the mill is achieved, effectively overcoming the limitations of single-signal monitoring, and significantly improving the accuracy of working condition identification. The prediction method combining linear and non-linear models can accurately capture the complex relationships between different parameters, providing a reliable basis for dynamic regulation. The hierarchical early warning mechanism adopts differential response strategies according to the degree of abnormality, maximizing the maintenance of production continuity while ensuring equipment safety and reducing the losses caused by unplanned shutdowns.

[0018] 2. The sensors with a symmetric layout effectively eliminate the measurement deviation caused by the spindle offset load, ensuring that the vibration data truly reflects the operating state of the equipment. The broadband acoustic emission sensors cover the key frequency bands during the grinding process, accurately identifying material breakage and abnormal collision events. The non-contact temperature detection technology avoids the problem of traditional contact temperature measurement being easily interfered by oil stains, achieving precise positioning and early warning of the thermal anomalies in the gearbox.

[0019] 3. The multi-order difference processing of vibration signals enhances the feature extraction ability of impact faults, significantly advancing the discovery time of potential hazards such as bearing damage. The two-dimensional temperature field analysis technology can accurately track the heat source diffusion path inside the gearbox, providing a quantitative basis for evaluating the tooth surface wear degree. The harmonic feature analysis of the rotational speed signal can sensitively capture the minute fluctuations in the transmission system, timely warning of mechanical looseness or abnormal load problems.

[0020] 4. The regression model integrating linear and non-linear features fully considers the coupling effect between parameters, significantly improving the prediction stability under complex working conditions. By quantifying the influence of multi-variable interactions on targets such as energy consumption and particle size, the collaborative optimization control of process parameters is realized. The introduction of the Gaussian kernel function enables the model to adaptively match different feature distribution characteristics, effectively reducing the prediction deviation caused by non-linear relationships.

[0021] 5. The segmented regulation strategy based on energy consumption deviation takes into account both the response speed and the operating stability, avoiding secondary faults caused by sudden changes in rotational speed. The fine-tuning threshold setting ensures that the system only makes small corrections within the normal fluctuation range, reducing unnecessary parameter perturbations. The gradient design of the dynamic adjustment amplitude enables the system to quickly recover to the optimal state when dealing with significant deviations, significantly reducing the fluctuation amplitude of unit energy consumption.

[0022] 6. The warning trigger logic with multi-condition combination significantly reduces the false alarm rate, avoiding over-response caused by single parameter anomalies. The hierarchical threshold setting fully considers the physical relevance of different characteristic parameters, ensuring that the warning level is strictly matched with the actual risk. Through the cross-validation of multi-dimensional abnormal states, the system can accurately identify potential fault modes and initiate corresponding disposal processes.

[0023] 7. The hierarchical response mechanism realizes a progressive disposal from parameter fine-tuning to emergency shutdown, maximizing the equipment availability. The linkage regulation strategy quickly suppresses the development of abnormal states through multi-parameter collaborative optimization, reducing the risk of fault escalation. The abnormal data recording function provides complete information support for post-event analysis, helping to optimize the maintenance cycle and regulation strategy.

[0024] 8. The modular system architecture realizes the closed-loop management of data collection, analysis, and control, improving the overall response efficiency. The independent functional module design facilitates system expansion and maintenance, and can quickly adapt to the regulation requirements of different models of grinding mills. The background data synchronization function realizes the full life cycle traceability of the production process, providing a data basis for process optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of the present invention, the drawings required for description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0026] Figure 1 It is a flowchart of the method in the embodiment of the specific implementation manner of the present invention. SPECIFIC IMPLEMENTATION MANNER

[0027] In the following detailed description, various embodiments of the present disclosure will be described more fully. The present disclosure may have various embodiments, and adjustments and changes can be made therein. However, it should be understood that there is no intention to limit the various embodiments of the present disclosure to the specific embodiments disclosed herein, but the present disclosure should be understood to cover all adjustments, equivalents, and / or alternative solutions falling within the spirit and scope of the various embodiments of the present disclosure.

[0028] Embodiment 1: The present invention provides a multi-source monitoring and dynamic regulation method for grinding mills in view of the problems in the prior art, as Figure 1 shown, including the following steps: Step S1, preset a scenario template, and select the corresponding scenario template according to the corresponding grinding mill usage scenario; For different industry application scenarios, standardized process parameter templates are pre-established in the system. For example, in the rough ore grinding scenario, a higher rotation speed range and a larger particle size target value are set; in the fine grinding scenario, a lower rotation speed and more stringent particle size control parameters are adopted. Operators can quickly call the matching template through the visual interface according to the actual material characteristics, production capacity requirements, and equipment specifications, and the system automatically loads the preset key parameter combinations such as the rotation speed reference value, the feed rate range, and the energy consumption threshold.

[0029] Step S2, collect the operating state data of the grinding mill, including the spindle vibration acceleration signal, the acoustic emission signal of the grinding chamber, the temperature distribution data of the gearbox, and the cylinder rotation speed signal; Deploy high-precision vibration sensors at symmetric positions of the main shaft bearing housing to capture axial and radial vibration signals in real time. Install a broadband acoustic monitoring device inside the grinding chamber to cover the characteristic frequency range of material crushing and mechanical collisions. Arrange non-contact temperature detection devices on the surface of the gearbox to obtain the temperature field distribution in a two-dimensional thermal imaging manner. Install a rotational speed encoder on the barrel transmission shaft to record the details of rotational speed fluctuations through high-frequency sampling. After anti-interference processing, the multi-source data is synchronously transmitted to the central processing unit.

[0030] Step S3: Extract the change characteristics of the vibration acceleration signal through first-order and second-order differences, extract the energy density spectrum characteristics of the acoustic emission signal, extract the local abnormal temperature rise characteristics and rotational speed fluctuation characteristics of the gearbox temperature distribution, and generate a characteristic data set. Perform multi-order differential processing on the vibration signal to extract the characteristic waveform parameters representing mechanical impacts. The acoustic signal uses spectrum analysis technology to identify the high-energy frequency bands of abnormal collision events and their occurrence frequencies. The temperature data uses an image processing algorithm to calculate the diffusion rate of the local hot area and the change trend of the temperature gradient. After frequency domain transformation of the rotational speed signal, quantify the harmonic distortion degree of the main drive system. After normalization processing of all characteristic parameters, form a multi-dimensional feature vector set.

[0031] Step S4: Based on the characteristic data set, construct a regression prediction model with the operating efficiency, unit energy consumption, and grinding particle size of the mill as the target outputs. The model form includes linear terms and non-linear Gaussian kernel terms. Establish a composite model integrating linear regression and non-linear kernel functions. The input layer receives the multi-dimensional feature vector, the hidden layer quantifies the synergistic effect between parameters through feature cross terms, and the output layer synchronously predicts the energy consumption, particle size, and efficiency indicators.

[0032] Step S5: According to the target values output by the regression prediction model, dynamically adjust the rotational speed, water supply, coal supply, and additive dosage of the mill in combination with the controller, where the output of the controller is dynamically updated according to the real-time deviation change. Adopt a progressive adjustment strategy according to the deviation between the real-time prediction result and the target value. When the predicted energy consumption value deviates from the benchmark, give priority to adjusting the main shaft rotational speed to balance the load; if the particle size index is abnormal, jointly adjust the feeding rate and the grinding medium addition rate. The control algorithm incorporates anti-oscillation logic. When the continuous adjustment fails to achieve the expected effect, automatically switch to the multi-parameter collaborative optimization mode and find the optimal operating point through the gradient descent method.

[0033] Step S6: Set up a three-level early warning response mechanism based on the degree of abnormality. The first-level early warning triggers fine-tuning of the operating parameters, the second-level early warning triggers joint optimization adjustment, and the third-level early warning triggers shutdown protection, and synchronize the characteristic data set and the early warning records.

[0034] The system continuously evaluates the deviation degree of characteristic parameters. When a single parameter slightly exceeds the limit, a first-level warning is triggered, and relevant variables are automatically fine-tuned; if multiple parameters are abnormally linked or a single parameter seriously deviates, a second-level warning is initiated, and a multi-variable joint optimization algorithm is executed; when an extreme working condition that may cause equipment damage is detected, a third-level warning is immediately triggered and protective shutdown is executed. All warning events are associated with timestamps and working condition snapshots to form a traceable abnormal event database.

[0035] In this embodiment, in step S2, the acquisition of the spindle vibration acceleration signal is completed by MEMS acceleration sensors symmetrically installed on both sides of the spindle bearing housing; The acquisition of the acoustic emission signal in the grinding chamber is carried out using broadband acoustic emission sensors deployed in the feeding port area; By performing energy density spectrum analysis on the acoustic signal, the high energy density peaks corresponding to abnormal impact events and the changes in occurrence frequencies are extracted; The acquisition of the temperature distribution data of the gearbox is carried out using a non-contact infrared thermal imager.

[0036] In this embodiment, in step S3, the acquired spindle vibration acceleration signal is processed by first-order difference to obtain the velocity change trend, and by second-order difference to obtain the acceleration mutation amount. The maximum vibration amplitude and the vibration trend slope are output in real time as input parameters for dynamic adjustment; By performing energy density spectrum analysis on the acoustic signal, the high energy density peaks corresponding to abnormal impact events and the changes in occurrence frequencies are extracted; The temperature distribution data of the gearbox is acquired in the form of a two-dimensional thermal distribution matrix, and the local abnormal temperature rise characteristics of the gearbox temperature distribution are extracted by extracting the temperature gradient change rate, local extreme points, and heat spot diffusion speed; The cylinder rotation speed signal is processed by FFT transformation to extract the main frequency and high-order harmonic characteristics, and the change amount of harmonic amplitude is used as the rotation speed fluctuation characteristic of the mill.

[0037] In this embodiment, in step S4, the regression prediction model is:

[0038] Where: is the predicted output result, which includes unit energy consumption E, grinding particle size d, and production efficiency η; is the bias term; is the linear regression coefficient of the i-th feature; is the i-th input feature variable; is the k-th input feature variable; is the j-th input feature variable; is the feature cross-term coefficient, indicating the contribution of the joint action of features and to the predicted value; is the weight coefficient of the Gaussian kernel function, which controls the influence intensity of the k-th characteristic Gaussian kernel function; is the mean value of the k-th feature; is the standard deviation of the k-th feature.

[0039] In this embodiment, in step S5, when performing dynamic adjustment based on the target value output by the regression prediction model, according to the deviation between the predicted value of the unit energy consumption E and the preset value of the preset scenario template, the mill speed is adjusted in real time. The speed adjustment strategy adopts a piecewise linear function. When the deviation is lower than the preset fine-tuning threshold, the adjustment range is set to ±3% of the base speed. When the deviation exceeds the fine-tuning threshold, the adjustment range increases to ±10%.

[0040] In this embodiment, in step S6, in the three-level early warning mechanism set based on the degree of abnormality, the triggering condition for the first-level early warning is that the change amplitude of any characteristic parameter exceeds the corresponding first-level deviation threshold but does not reach the corresponding second-level early warning triggering condition; The triggering condition for the second-level early warning is that the change of the characteristic parameter simultaneously satisfies at least two first-level early warning conditions or there is a single characteristic parameter exceeding the corresponding second-level deviation threshold; The triggering condition for the third-level early warning is that the change of the characteristic parameter simultaneously satisfies at least two second-level early warning conditions or there is a single characteristic parameter exceeding the corresponding third-level deviation threshold.

[0041] In this embodiment, after the first-level early warning is triggered, the system performs a first-level amplitude fine-tuning operation on the corresponding single operating parameter, and the adjustment range is automatically controlled in grades according to the deviation; After the second-level early warning is triggered, the system performs a comprehensive optimization adjustment of the linkage parameters, preferentially adjusting the mill speed and the coal feeding amount, and then adjusting the water supply amount and the additive dosage. Each adjustment amount is dynamically optimized according to the output result of the real-time prediction model and the deviation trend; After the third-level early warning is triggered, an emergency shutdown command is immediately executed, and at the same time, the vibration signal, acoustic emission signal, temperature distribution data, and speed change within the last 60 seconds are recorded, and an exception report is pushed at the same time.

[0042] Embodiment 2: The present application provides a multi-source monitoring and dynamic regulation system for a mill, including: The configuration design module uses visual interaction technology to build a parameter configuration interface, supporting users to quickly build a control panel through drag-and-drop operations. Operators can freely select interactive components such as sliders, knobs, and switches on the browser or mobile terminal and bind them to physical parameters such as mill speed, water supply, and coal supply. The components support multi-level nested management, and users can combine multiple associated controls into a composite function unit. For example, "energy-saving mode" is defined as a parameter combination package with a 5% reduction in speed and an 8% increase in coal supply, and saved as a reusable template. During the configuration process, a real-time simulation operation function is provided. When adjusting parameters, the interface synchronously displays the predicted energy consumption curve, the change trend of grinding efficiency, and the particle size distribution heat map to ensure the rationality of parameter settings.

[0043] The template management module has an industry-standard process library and a user-defined template library built-in, supporting multi-dimensional classification retrieval according to material type, equipment model, production target, etc. Each template contains a complete set of parameter configurations. For example, the iron ore rough grinding template presets 128 parameters such as speed range, steel ball grading ratio, and lubrication period. The system supports the version management and difference comparison functions of templates. When users modify existing templates, modification records are automatically generated and the impact assessment of key parameter changes is prompted.

[0044] The data acquisition module integrates a multi-source sensing network, captures the three-dimensional vibration signals of the main shaft in real time through a high-precision vibration sensor array, monitors the impact events inside the grinding chamber with a broadband acoustic emission sensor group, generates a surface temperature field distribution map of the gearbox with a non-contact infrared thermal imager, and accurately measures the rotational speed fluctuation of the cylinder body with a magnetoelectric encoder. The data acquisition unit is equipped with a signal conditioning circuit and an anti-interference module to ensure stable transmission of raw data in a strong electromagnetic environment.

[0045] The feature extraction module performs joint time-frequency domain analysis on vibration signals to extract 12-dimensional features such as acceleration mutation amount and energy entropy value; acoustic emission data uses wavelet packet decomposition technology to quantify the energy proportion and impact event frequency in different frequency bands; thermal imaging data uses the region growing algorithm to identify abnormal temperature rise regions and calculate the hot spot area expansion rate; the rotational speed signal extracts the main frequency harmonic distortion rate through spectrum analysis. All feature parameters are normalized to form a standardized feature vector.

[0046] The modeling and prediction module deploys a hybrid regression model architecture. The linear term captures the explicit associations between parameters, and the Gaussian kernel function models the non-linear coupling effect. The input layer of the model receives the feature vector, the hidden layer quantifies the synergistic effects of multiple parameters through feature cross terms, and the output layer synchronously predicts the unit energy consumption, grinding particle size, and production efficiency. The online learning mechanism continuously receives new data, dynamically adjusts the model weights to adapt to the working condition drift, and starts the model self-check process at a fixed time every day. When the prediction error exceeds the threshold, a retraining alarm is triggered.

[0047] The dynamic control module is deeply integrated with the PLC control system to convert the output of the prediction model into actual control instructions. When it is detected that the energy consumption deviates from the target value, a segmented adjustment strategy is adopted: when there is a slight deviation, the speed is fine-tuned proportionally; when there is a significant deviation, the coal feed and additive dosage are adjusted in conjunction. The control algorithm has built-in anti-oscillation logic. When three consecutive adjustments do not achieve the expected effect, it automatically switches to the global optimization mode and searches for the optimal parameter combination through the particle swarm algorithm.

[0048] The early warning management module establishes a three-level early warning response system. When the first-level early warning is triggered, the related parameters are automatically fine-tuned and abnormal indicators are highlighted on the interface; the second-level early warning starts the multi-parameter collaborative optimization algorithm and simultaneously pushes SMS notifications to the responsible engineer; the third-level early warning directly issues a shutdown command and fully records the multi-source data snapshot 60 seconds before the shutdown. The early warning event is correlated with historical data, and a root cause analysis report is automatically generated to put forward maintenance suggestions.

[0049] Embodiment 3: The present application provides a terminal, including: A memory for storing a multi-source monitoring and dynamic control program for a grinding mill; The processor is used to implement the steps of the multi-source monitoring and dynamic control method for a mill as described in Example 1 when executing the multi-source monitoring and dynamic control program for the mill.

[0050] Embodiment 4: The present application provides a computer-readable storage medium, which includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the multi-source monitoring and dynamic control method for a mill described in Example 1.

[0051] It is understandable that the systems, devices, modules or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer, and the specific form of the computer can be a personal computer, a laptop computer, a personal digital assistant, a tablet computer, a wearable device, or a combination of any of these devices.

[0052] In a typical configuration, a computer includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0053] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0054] A computer-readable medium includes permanent and non-permanent, removable and non-removable media that can implement information storage by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory, or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD), or other optical storage, magnetic cassette tapes, disk storage, quantum memory, graphene-based storage media, or other magnetic storage devices, or any other non-transitory media that can be used to store information accessible by a computing device. As defined herein, a computer-readable medium does not include transitory computer-readable media, such as modulated data signals and carrier waves.

[0055] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data that have been authorized by the user or fully authorized by all parties. And the collection, use, and processing of relevant data need to comply with the relevant laws, regulations, and standards of relevant countries and regions, and corresponding operation entrances are provided for users to choose to authorize or refuse.

[0056] It should also be noted that the term "comprising", "including", or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, commodity, or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or also includes elements inherent to such process, method, commodity, or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the existence of additional identical elements in the process, method, commodity, or device comprising the said element.

[0057] The specific embodiments of this specification have been described above. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve the desired result. Additionally, the processes depicted in the figures do not necessarily require the particular order or sequential order shown to achieve the desired result. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0058] The terms used in one or more embodiments of this specification are for the purpose of describing particular embodiments only and are not intended to limit one or more embodiments of this specification. The singular forms "a", "the", and "said" used in one or more embodiments of this specification and the appended claims are also intended to include the plural forms unless the context clearly dictates otherwise. It should also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0059] It should be understood that although the terms first, second, third, etc. may be used in one or more embodiments of this specification to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of one or more embodiments of this specification, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "when" or "while" or "in response to determining".

[0060] The above are only the preferred embodiments of one or more embodiments of this specification and are not intended to limit one or more embodiments of this specification. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of one or more embodiments of this specification shall be included within the scope of protection of one or more embodiments of this specification.

Claims

1. A multi-source monitoring and dynamic control method for a grinding mill, characterized in that: The following steps are involved: Step S1, preset a scene template, and select a corresponding scene template according to the corresponding grinding mill usage scenario; Step S2, collecting mill operation status data, including spindle vibration acceleration signal, grinding chamber acoustic emission signal, gear box temperature distribution data and barrel speed signal; Step S3, extracting the change characteristics of the vibration acceleration signal through first-order difference and second-order difference, extracting the energy density spectrum characteristics of the acoustic emission signal, extracting the local abnormal temperature rise characteristics and speed fluctuation characteristics of the gearbox temperature distribution, and generating a feature data set; Step S4, based on the characteristic data set, constructing a regression prediction model with mill operation efficiency, unit energy consumption, and grinding particle size as target outputs, wherein the model form includes a linear term and a nonlinear Gaussian kernel term; Step S5, according to the target value output by the regression prediction model, the mill speed, water supply, coal supply and additive dosage are dynamically adjusted in combination with the controller, wherein the controller output is dynamically updated according to the real-time deviation change; Step S6: Set a three-level warning response mechanism based on the degree of abnormality. The first-level warning triggers fine-tuning of operating parameters, the second-level warning triggers joint optimization adjustment, and the third-level warning triggers shutdown protection, and synchronizes the feature data group with the warning record.

2. The multi-source monitoring and dynamic control method for a grinding mill according to claim 1, characterized in that: In step S2, the spindle vibration acceleration signal is collected by MEMS acceleration sensors symmetrically installed on both sides of the spindle bearing seat; The acoustic emission signal of the grinding chamber is collected using a wide-band acoustic emission sensor deployed in the feed inlet area; By analyzing the energy density spectrum of the acoustic signal, the high energy density peak and the frequency change corresponding to the abnormal impact event are extracted; The temperature distribution data of the gearbox is obtained using a non-contact infrared thermal imager.

3. The multi-source monitoring and dynamic control method for a mill according to claim 2, characterized in that: In step S3, the collected spindle vibration acceleration signal is processed by first-order difference to obtain the speed change trend, and the acceleration mutation is processed by second-order difference to obtain the maximum vibration amplitude and vibration trend slope in real time as input parameters for dynamic adjustment; By analyzing the energy density spectrum of the acoustic signal, the high energy density peak and the frequency change corresponding to the abnormal impact event are extracted; The gearbox temperature distribution data is collected in the form of a two-dimensional heat distribution matrix, and the temperature gradient change rate, local extreme point and hot spot diffusion speed are extracted as the local abnormal temperature rise characteristics of the gearbox temperature distribution; The main frequency and high-order harmonic characteristics of the drum speed signal are extracted through FFT transformation, and the change in harmonic amplitude is used as the speed fluctuation characteristic of the mill.

4. The multi-source monitoring and dynamic control method for a mill according to claim 1, characterized in that: In step S4, the regression prediction model is: in: To predict the output results, which include specific energy consumption E, grinding particle size d, and production efficiency η; is the bias term; is the linear regression coefficient of the i-th feature; is the i-th input feature variable; is the kth input feature variable; is the jth input feature variable; is the characteristic cross-term coefficient, indicating the characteristic and The contribution of the combined effect to the predicted value; is the Gaussian kernel function weight coefficient, which controls the influence intensity of the k-th feature Gaussian kernel function; is the mean of the kth feature; is the standard deviation of the kth feature.

5. The multi-source monitoring and dynamic control method for a grinding mill according to claim 4, characterized in that: In step S5, when dynamic adjustment is performed based on the target value output by the regression prediction model, the mill speed is adjusted in real time according to the deviation between the predicted value of the unit energy consumption E and the preset value of the preset scenario template. The speed adjustment strategy adopts a piecewise linear function. When the deviation is lower than the preset fine-tuning threshold, the adjustment range is set to ±3% of the basic speed. When the deviation exceeds the fine-tuning threshold, the adjustment range is increased to ±10%.

6. The multi-source monitoring and dynamic control method for a grinding mill according to claim 1, characterized in that: In step S6, in the three-level warning mechanism set based on the degree of abnormality, the trigger condition of the first-level warning is that the change amplitude of any characteristic parameter exceeds the corresponding first-level deviation threshold but does not reach the corresponding second-level warning trigger condition; The triggering condition for the second-level warning is that the characteristic parameter changes simultaneously meet at least two first-level warning conditions or a single characteristic parameter exceeds the corresponding second-level deviation threshold; The triggering condition for the third-level warning is that the characteristic parameter changes simultaneously meet at least two second-level warning conditions or a single characteristic parameter exceeds the corresponding third-level deviation threshold.

7. The multi-source monitoring and dynamic control method for a grinding mill according to claim 6, characterized in that: After the first-level warning is triggered, the system performs the first-level amplitude fine-tuning operation corresponding to a single operating parameter, and the adjustment amplitude is automatically controlled in stages according to the deviation amount; After the secondary warning is triggered, the system performs comprehensive optimization and adjustment of linkage parameters, giving priority to adjusting the mill speed and coal feed, followed by adjusting the water feed and additive dosage. Each adjustment is dynamically optimized based on the real-time prediction model output results and deviation trends. When the third-level warning is triggered, the emergency shutdown command is immediately executed, and the vibration signal, acoustic emission signal, temperature distribution data, and speed change within the last 60 seconds are recorded synchronously, and an abnormality report is pushed at the same time.

8. A multi-source monitoring and dynamic control system for a grinding mill, characterized in that: include: The template management module is used to preset scene templates and select the corresponding scene template according to the mill usage scenario; Data acquisition module, used to collect mill operation status data, including spindle vibration acceleration signal, grinding chamber acoustic emission signal, gear box temperature distribution data and barrel speed signal; A feature extraction module is used to extract the vibration acceleration signal change characteristics, extract the acoustic emission signal energy density spectrum characteristics, extract the local abnormal temperature rise characteristics of the gearbox temperature distribution, and extract the speed fluctuation characteristics through first-order difference and second-order difference processing to generate a feature data group; A modeling and prediction module is used to construct a regression prediction model based on the feature data set, wherein the model takes mill operation efficiency, unit energy consumption, and grinding particle size as target outputs, and the regression model form includes linear terms and nonlinear Gaussian kernel terms; Dynamic control module, used to dynamically adjust the mill speed, water supply, coal supply and additive dosage according to the target value output by the regression prediction model in combination with the controller, and the controller output is dynamically updated according to the real-time deviation change; The early warning management module is used to set up a three-level early warning response mechanism based on the degree of abnormality. The first-level early warning triggers fine-tuning of operating parameters, the second-level early warning triggers joint optimization and adjustment, and the third-level early warning triggers shutdown protection, and synchronizes the feature data group and early warning records to the background management system.

9. A terminal, characterized in that: include: A memory for storing a multi-source monitoring and dynamic control program for a grinding mill; A processor is used to implement the steps of the multi-source monitoring and dynamic control method for a mill as described in any one of claims 1 to 7 when executing the multi-source monitoring and dynamic control program for the mill.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the multi-source monitoring and dynamic control method for a mill as described in any one of claims 1-7.

Images