Coal washing equipment image processing intelligent monitoring system based on strategy optimization

Abstract

The present invention discloses an intelligent monitoring system for coal washing equipment image processing based on strategy optimization, which relates to the technical field of coal washing equipment image processing. It constructs a closed-loop intelligent optimization system consisting of modules such as image acquisition, process monitoring, simulation modeling, multi-objective optimization, strategy scoring, control execution and feedback adjustment. By introducing response difference calculation and response space redistribution models, the Pareto solution set is effectively sparsely processed to improve strategy discrimination; multi-factor dynamic scoring is performed based on image features, historical performance and response trajectory to ensure that the selected strategy has the best execution effect under the current working conditions; the system also has the ability of model self-correction and scoring parameter adaptation during operation, realizing high-precision, steady-state and self-learning control of the entire process, and significantly improving the intelligence level and operation reliability of coal washing equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of coal washing equipment image processing, and in particular to a coal washing equipment image processing intelligent monitoring system based on strategy optimization. Background Art

[0002] Intelligent monitoring of coal washing equipment using image processing based on strategy optimization involves using image processing technology to monitor the equipment's operating status in real time during the coal washing process, combined with strategy optimization algorithms to intelligently adjust and control the monitoring system. Its core goal is to improve monitoring efficiency and accuracy through intelligent identification and analysis of equipment image data, promptly identifying equipment anomalies or failures. Furthermore, monitoring parameters can be dynamically adjusted based on preset or self-learning optimization strategies, achieving intelligent, stable, and efficient management of coal washing equipment operations.

[0003] The existing technology has the following shortcomings:

[0004] In the existing technology, the Pareto multi-objective optimization algorithm is used to train and optimize the simulation model to achieve refined control of the coal washing process and intelligent decision support. However, in the multi-objective optimization based on the Pareto frontier algorithm, if there are high-dimensional objective combinations, the Pareto solution distribution may be too dense, which is called Pareto front overcrowding. This problem can cause multiple optimization strategies to be approximately equivalent in mathematics, but in actual operation, due to the system's excessive sensitivity to small adjustments, it causes slow response and even unstable operation. Because this deviation is difficult to detect during the training phase, it often only gradually appears after the system has been deployed and run for a period of time, posing a hidden and serious challenge to the stability and accuracy of the intelligent monitoring system. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent monitoring system for coal washing equipment image processing based on strategy optimization to solve the shortcomings of the background technology.

[0006] In order to achieve the above-mentioned object, the present invention provides the following technical solution: an intelligent monitoring system for coal washing equipment image processing based on strategy optimization, comprising an image acquisition module, a process parameter monitoring module, a simulation modeling module, a multi-objective optimization module, a strategy scoring module, a control execution module and a feedback adjustment module;

[0007] Image acquisition module, used to obtain real-time image information of different parts of the coal washing equipment, and perform recognition and structural processing on it to form image feature data;

[0008] Process parameter monitoring module, used to collect several process parameters during coal washing;

[0009] The simulation modeling module builds a dynamic simulation model based on image feature data and process monitoring data to simulate the response relationship between multiple optimization objectives in the coal washing process;

[0010] The multi-objective optimization module generates multiple control strategies based on the Pareto frontier algorithm, calculates the response differences between the strategies, establishes a response space redistribution model, and performs sparse processing on the Pareto solution set to obtain a sparse optimization solution set;

[0011] The strategy scoring module dynamically scores the sparse optimization solution set based on image features, historical operation data, and response trajectories, and selects the optimal strategy based on the scoring results;

[0012] The control execution module adjusts the operating parameters of the coal washing equipment according to the selected optimal strategy;

[0013] The feedback adjustment module is used to monitor the control execution results and automatically correct the simulation model or adjust the strategy scoring parameters based on the deviation between the device response and the target.

[0014] Preferably, the image acquisition module includes: multiple groups of image sensors installed on the screen surface, sorting slots and conveyor belt discharge ports of the coal washing equipment, and the image sensors include high-resolution industrial cameras, infrared thermal imagers and laser profile scanners.

[0015] Preferably, the image acquisition module is used to collect original image data during the operation of the equipment; and performs grayscale processing, edge enhancement and noise removal on the image, identifies the coal block boundary and extracts image features through the target recognition algorithm, including particle size distribution, stacking thickness, movement direction and coal gangue ratio; outputs the processed image features in the form of structured vectors and passes them into the simulation modeling module as input variables.

[0016] Preferably, the process parameter monitoring module is used to collect process parameters in the coal washing process in real time, including medium concentration, water-coal ratio, feed particle size, vibration frequency, motor current, power consumption, water supply flow and system temperature.

[0017] Preferably, the simulation modeling module is used to build a data-driven model:

[0018] Extracting particle size distribution and material flow rate image features from the image acquisition module;

[0019] Obtain process data on water-coal ratio, medium concentration and power consumption from the process parameter monitoring module;

[0020] Normalize and process outliers on the collected data to form multidimensional modeling samples;

[0021] Build data-driven models using neural networks, support vector machines, or random forest algorithms;

[0022] Output the predicted values ​​of each target parameter under the simulation control strategy for transmission to the multi-objective optimization module.

[0023] Preferably, after generating the Pareto solution set, the multi-objective optimization module performs a sparse processing step:

[0024] Calculate the difference between the target responses of any two control strategies in the solution set, where the difference is measured by Euclidean distance or Mahalanobis distance;

[0025] Establish a target spatial redistribution model based on response differences;

[0026] Use clustering algorithm to cluster the solution set;

[0027] Select the representative solution with the largest response difference in each cluster to form a sparse optimization solution set;

[0028] Output the sparse optimization results to the strategy scoring module.

[0029] Preferably, the strategy scoring module includes:

[0030] Calculate the matching score between the strategy parameters and the current image features: the matching score is calculated by the Euclidean distance between the current image features and the mean of the strategy's historical image features;

[0031] Calculate the average effectiveness score of the strategy under similar working conditions: Calculate the average execution effect of the strategy under similar working conditions and take the average of multiple scores;

[0032] Calculate the deviation control capability score after the strategy is executed: Fit the actual system response to a first-order response model, extract the response time constant, delay time, and fitting error, and construct a scoring function;

[0033] The matching score, average effect score and deviation control ability score are integrated according to the weighted formula to output the comprehensive score result;

[0034] Strategies are sorted according to their scores, and the one with the highest score is selected as the optimal control strategy.

[0035] Preferably, the feedback adjustment module includes:

[0036] Sampling indicators from control execution results;

[0037] Compare with the prediction results of the simulation model and calculate the target deviation vector;

[0038] If the deviation exceeds the preset threshold, the modeling module fine-tuning mechanism is triggered, including incremental model training, parameter update or structure reconstruction.

[0039] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0040] 1. This invention achieves intelligent control of the entire coal washing process, from perception to decision-making, by constructing an intelligent monitoring system consisting of modules including image acquisition, process parameter monitoring, simulation modeling, multi-objective optimization, strategy scoring, control execution, and feedback adjustment. The system combines image information with process data, establishes a multi-objective simulation model through a data-driven approach, and introduces a Pareto frontier sparsification mechanism during control strategy generation. This effectively addresses the sluggish response and control instability caused by the overcrowding of Pareto solutions in existing technologies, thereby improving the system's accuracy, responsiveness, and strategy differentiation capabilities.

[0041] 2. This invention incorporates a multi-factor scoring mechanism for strategy scoring and feedback adjustment, including image matching, historical performance, and response fitting. This is complemented by first-order system response modeling and an adaptive feedback optimization mechanism to ensure the control strategy's feasibility and stable execution under current operating conditions. The overall system boasts a complete structure, strong learning capabilities, high real-time performance, and excellent control precision, significantly improving the operating efficiency, coal quality control, and intelligent operation of coal washing equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0043] Figure 1 This is the system mind map of the present invention. DETAILED DESCRIPTION

[0044] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0045] For examples, see Figure 1 As shown, the coal washing equipment image processing intelligent monitoring system based on strategy optimization described in this embodiment includes an image acquisition module, a process parameter monitoring module, a simulation modeling module, a multi-objective optimization module, a strategy scoring module, a control execution module, and a feedback adjustment module;

[0046] Image acquisition module, used to obtain real-time image information of different parts of the coal washing equipment, and perform recognition and structural processing on it to form image feature data;

[0047] Process parameter monitoring module, used to collect several process parameters during coal washing;

[0048] The simulation modeling module builds a dynamic simulation model based on image feature data and process monitoring data to simulate the response relationship between multiple optimization objectives in the coal washing process;

[0049] The multi-objective optimization module generates multiple control strategies based on the Pareto frontier algorithm, calculates the response differences between the strategies, establishes a response space redistribution model, and performs sparse processing on the Pareto solution set to obtain a sparse optimization solution set;

[0050] The strategy scoring module dynamically scores the sparse optimization solution set based on image features, historical operation data, and response trajectories, and selects the optimal strategy based on the scoring results;

[0051] The control execution module adjusts the operating parameters of the coal washing equipment according to the selected optimal strategy;

[0052] The feedback adjustment module is used to monitor the control execution results and automatically correct the simulation model or adjust the strategy scoring parameters based on the deviation between the device response and the target.

[0053] The image acquisition module mainly consists of the following sub-components:

[0054] Image sensor components: including high-resolution industrial cameras (such as CCD / CMOS cameras); optional infrared thermal imagers and laser profile scanners for capturing non-visible light images (for example, detecting material temperature and humidity changes); with autofocus and dust protection mechanisms, suitable for dusty environments in coal mines.

[0055] Image acquisition controller: connected to the main system PLC or embedded control chip; controls the acquisition frequency (frame rate), exposure time, and shutter parameters; has edge computing capabilities and supports preliminary image preprocessing.

[0056] Image acquisition location and viewing angle design: Installed in key locations of coal washing equipment, such as the screening section (to detect material distribution and blockage on the screen surface); the sorting slot (to detect flow pattern and concentration); and the conveyor belt outlet (to determine particle size and coal-gangue separation efficiency). Multi-view deployment supports stereoscopic visual reconstruction.

[0057] High-resolution visible light camera (CCD / CMOS): Mainly used to collect dynamic images of coal blocks and materials on the screen surface, belt and sorting trough. Combined with image recognition algorithms, it can extract the following information:

[0058] Coal particle size and particle size distribution (used to evaluate screening and sorting effects), material stacking thickness (reflecting feed uniformity), flow velocity and direction (evaluating flow stability), equipment surface operating status (identifying abnormalities such as blockage and deviation).

[0059] Infrared camera or thermal imaging sensor: used to analyze the thermal characteristics of the material surface, which can assist in determining: changes in moisture content (identifying wet coal through differences in thermal radiation), heating conditions of conveyor belts or vibrating screens (used to predict equipment abnormalities or overloads).

[0060] Color image sensor: By analyzing image color and brightness, it can identify: the color difference between coal and gangue, and determine the coal quality composition; material mixing or mudification trend (detected by grayscale and saturation changes).

[0061] The image processing and feature extraction process includes:

[0062] Image preprocessing: grayscale conversion, edge enhancement, image denoising; adapt to uneven lighting and particle dust interference in coal mine environments.

[0063] Image segmentation and recognition: Apply deep learning algorithms (such as YOLOv5, Mask R-CNN) or traditional algorithms (such as threshold segmentation, contour detection) to achieve coal block boundary recognition, particle size statistics, and distribution area extraction.

[0064] Structured feature extraction: Output vectorized features, such as: average particle size, particle size variance; coal block ratio, gangue ratio; image area density distribution; dynamic flow trajectory characteristics (such as velocity field, flow direction); output results are uploaded to the modeling module to participate in simulation modeling or optimization input.

[0065] Interfacing with the simulation modeling module: Image features are used as one of the input variables to participate in multi-objective modeling; physical modeling of non-numerical parameters is supported, such as coal slime adhesion state, flow pattern, etc.

[0066] Linked with the control system: If the image identifies that the particle size is too large, the vibration frequency will be automatically adjusted; if the pile offset is detected, the feeding direction or speed can be adjusted in conjunction; image anomaly detection can trigger the safety shutdown logic.

[0067] The process parameter monitoring module uses a variety of industrial sensors placed at key locations on the equipment to obtain physical quantities, energy consumption indicators, and process parameters that are closely related to the operating status of the coal washing process. These mainly include the following categories:

[0068] Medium concentration monitoring, purpose: to determine the concentration fluctuation of the sorting medium (such as magnetic fluid or heavy medium) and control the sorting accuracy; sensor type: conductivity sensor, optical densitometer, ultrasonic densitometer; collection location: medium circulation system, sorting tank inlet and outlet.

[0069] Feed particle size and flow monitoring, purpose: to control feed uniformity and prevent particle size fluctuations from affecting screening efficiency and coal gangue separation; sensor types: laser particle size analyzer, belt scale, flow meter; collection location: vibrating feeder, belt conveyor system.

[0070] Water-coal ratio and water supply monitoring, purpose: regulating water used in coal washing, affecting coal slime sedimentation and washing effect; sensor type: electromagnetic flowmeter, ultrasonic level meter, online water quality analyzer; collection location: washing water pipeline, water pool, flotation tank.

[0071] Screen surface operating status parameters, purpose: used to monitor whether the vibrating screen is running stably; sensor type: acceleration sensor, vibration sensor; collection location: both sides of the screen body and the bottom vibration device.

[0072] Energy consumption and motor status monitoring, purpose: monitoring system energy consumption efficiency and identifying abnormal equipment load or failure trends in advance; sensor types: current / voltage transformers, smart meters, thermistors; collection locations: main motors, pumps, fans and other key drive equipment.

[0073] System temperature and ambient humidity monitoring, purpose: to ensure safe equipment operation, especially in harsh environments such as wet coal and high dust; sensor type: integrated temperature and humidity sensor, infrared thermometer; collection location: machine body casing, conveyor belt channel, equipment indoor environment.

[0074] Execution status parameters such as valve opening and pump speed. Purpose: Real-time recording of the current status of the actuator, used as feedback to determine whether the adjustment result is effective. Sensor type: Angle encoder, rotational tachometer. Collection location: Automatic valve, water supply pump, electromagnetic flow control mechanism.

[0075] It's important to note that the data collected by the process parameter monitoring module is uploaded to the system's central controller in real time as digital signals and processed synchronously with the image feature data. These collected parameters are not only used for dynamic modeling and calculation of the optimization objective function, but also provide a basis for closed-loop control of the system. After strategy execution, the module can also compare actual results with predicted values, enabling system self-correction and continuous optimization.

[0076] The simulation modeling module is the core decision support unit in the present invention. Its function is to construct a mathematical model or data-driven model that can simulate the coal washing process based on the collected image feature data and process parameter data. It is used to predict the changes and mutual influences between multiple optimization objectives (such as yield, energy consumption, medium consumption, sorting efficiency, etc.) under different control strategies.

[0077] The specific implementation methods of simulation modeling include:

[0078] Model type (select according to application):

[0079] Physical modeling (such as mass conservation and energy conservation equation modeling): suitable for basic coal washing processes such as dense medium separation and flotation;

[0080] Data-driven models (mainstream approach): use historical and real-time data for training, and are suitable for systems with complex coupling and difficult-to-accurate modeling. Typical algorithms include multivariate regression, random forests, neural networks (such as DNN and RNN), and support vector machines (SVM).

[0081] Hybrid modeling: Use physical models for part of the process and machine learning to fill in the remaining parts.

[0082] The modeling steps are as follows:

[0083] Step 1: Extract image features such as particle size distribution, bulk density, and material flow rate from the image acquisition module; obtain real-time data such as medium concentration, water-coal ratio, screening frequency, and power consumption from the process parameter monitoring module; clean, normalize, and standardize the above data to form a modeling sample.

[0084] Step 2: Clarify the output objectives of the simulation, such as: Objective 1: Maximize clean coal recovery rate; Objective 2: Minimize energy consumption; Objective 3: Minimize dielectric loss; Objective 4: Minimize sorting error; Express the objective function in mathematical form, such as linear weighting, sum of squared errors, etc.

[0085] Step 3: Select a suitable algorithm (such as a neural network) to establish a mapping function: Input: X = [image features + process parameters + control strategy variables]; Output: Y = [optimization target values]; perform cross-validation and error evaluation to prevent model overfitting.

[0086] Step 4: Train the model with historical data and update it in real time with the latest data; set the training set and validation set; adjust hyperparameters (such as the number of network layers and learning rate) to improve simulation accuracy.

[0087] Step 5: Perform simulation predictions on different control strategy combinations and output corresponding multiple target response values; analyze the mutual influence and trade-off relationship between the targets.

[0088] The multi-objective results of the simulation output are passed to the multi-objective optimization module (such as the Pareto algorithm) to generate the optimal control strategy.

[0089] The multi-objective optimization module is used in the intelligent monitoring system for coal washing equipment image processing to generate the optimal control strategy combination based on multiple optimization objectives (such as clean coal recovery rate, energy consumption, medium consumption, and sorting efficiency). The system uses the Pareto frontier algorithm to form an optimized solution set. It further uses response differentiation calculation to establish a response space redistribution model, which thins out the dense solution set. The final output is a sparse optimized solution set with good response differentiation, improving system response efficiency and control stability.

[0090] Receive input data from the simulation modeling module, including: image feature parameters (such as particle size distribution, stacking thickness); process operation parameters (such as medium concentration, water-coal ratio, screening frequency, etc.); control strategy variables (such as adjusting water supply flow, screening machine speed, etc.).

[0091] Identify multiple goals that need to be optimized simultaneously: Goal 1: maximize clean coal yield; Goal 2: minimize system energy consumption; Goal 3: minimize medium consumption; Goal 4: minimize sorting error; standardize the objective function into a comparable numerical function form to facilitate unified optimization.

[0092] An initial population containing multiple sets of feasible control strategy parameters is constructed, where each set of control parameters is a solution vector; each solution vector is input into the simulation model, and the predicted value of each target is output.

[0093] Use NSGA-II, MOEA / D or other Pareto frontier algorithms to iterate the population; select a group of solutions with no obvious disadvantages under multiple objectives according to the non-dominated sorting principle to form a preliminary Pareto optimal solution set.

[0094] For any two solutions in the preliminary Pareto solution set, the response difference (such as Euclidean distance, Manhattan distance or Mahalanobis distance) in the output target space is calculated respectively; the difference indicates the distinguishability of the two sets of strategies in actual operation. The larger the value, the more obvious the difference in control behavior.

[0095] Taking the response difference as a measurement indicator, a response distribution map is constructed in the target space; a clustering algorithm (such as K-means and DBSCAN) is used to cluster the solution set; redundant solutions with similar responses in each cluster are merged or representative points are extracted to obtain a more representative sparse solution set.

[0096] The representative solution set obtained after response redistribution is output to form the final sparse optimization solution set; this solution set retains the overall diversity of the solution set while excluding similar, repeated, and strategy combinations with no response differences.

[0097] The strategy scoring module is used to comprehensively evaluate each control strategy in the sparse optimization solution set output by the multi-objective optimization module. By introducing multi-source data such as image features, historical operation data, and system response trajectories, the actual adaptability and control effectiveness of each strategy are dynamically scored, and the optimal strategy is selected based on the scoring results to achieve intelligent decision-making and dynamic adjustment of the system.

[0098] Receive the sparse optimization solution set output by the multi-objective optimization module, where each control strategy contains a set of specific control parameters (such as screening frequency, water volume, medium concentration adjustment value, etc.);

[0099] The image recognition features under the current working conditions are obtained from the image acquisition module, including coal particle size, material distribution, motion trajectory, etc.; the image features are converted into structured vectors for subsequent model matching and scoring.

[0100] Call the historical operation records stored in the system database, including: the use records of various control strategies under similar working conditions; the corresponding actual effect evaluation (such as recovery rate, energy consumption, error rate); environmental and equipment conditions (such as season, humidity, material properties);

[0101] Read the actual response trajectory of the system after executing the strategy from the feedback adjustment module, including: system execution delay; process indicator response time; adjusted fluctuation amplitude; actual target deviation change;

[0102] A scoring function S(i) is established based on three factors: image features, historical data, and response trajectory, which is used to comprehensively evaluate each strategy i.

[0103] For example, the calculation expression of the scoring function can be: ;in: Score the matching degree between the strategy parameters and the current image features, is the average effectiveness score of the strategy under similar working conditions, Score the ability to control deviations after strategy execution; The weights are adjustable and optimized adaptively according to the scenario. The scoring function is calculated for each strategy in the sparse optimization solution set one by one; all scoring results are sorted from high to low to generate a strategy priority list.

[0104] in, The acquisition method is: set the image feature vector collected by the current system to: ; Each represents an image feature index (such as average particle size, bulk density, motion directionality, etc.); the average value of image features that strategy i has applied in the history is: ;but It can be calculated as: Where, is the standard deviation vector modulus of the entire historical image feature data, indicating the overall fluctuation range, and ε is a very small constant to prevent the denominator from being zero.

[0105] The historical performance factor evaluates the actual performance of a control strategy under similar past operating conditions. Its core objective is to assess whether a strategy that has been frequently used and performed well under similar coal quality, equipment conditions, and process conditions will maintain high adaptability and reliability under current operating conditions. This factor prevents the system from selecting a theoretically optimal strategy with a high historical failure rate.

[0106] For example, The method of obtaining is: let the target effect score of strategy i in the past m similar working conditions be: , then the historical performance score is defined as: Where, is the comprehensive effect score of strategy i in the jth similar working condition (normalized, 0~1), and m is the number of matched similar working condition samples.

[0107] When a system executes a control strategy, its key outputs (such as clean coal yield, dielectric loss, and motor power) will change dynamically over a certain period of time. By treating this output as a dynamic response curve and fitting a first-order inertial system model, we can extract parameters reflecting the system's dynamic characteristics, such as response time constant and delay, to quantify response stability.

[0108] For example, The method for obtaining y(t) is as follows: after control strategy i is executed, record the time series data of the target response y(t) within the control period; the data sampling period is once per second, and the duration is recommended to be no less than 3 to 5 expected time constants. The system response is modeled as a standard first-order system response form: Where y(t) is the actual output of the system (such as clean coal yield, power consumption, dielectric consumption, etc.), K is the system gain, which indicates the improvement of the final stable value relative to the initial value, and τ is the time constant, which describes the time required for the system to reach 63% steady-state response. The smaller the value, the faster the response. The response delay is the time delay after the control strategy starts to change the system. is the initial response value, that is, the initial output value of the system before the control strategy.

[0109] The parameters of the first-order inertial model are estimated using least squares fitting (such as the Levenberg-Marquardt algorithm) to obtain the optimal fitting parameters: response time constant , delay time ; At the same time, the fitting error (residual sum of squares) is calculated as the goodness of fit indicator. The fitting result is converted into a score, that is, the calculation is obtained , the expression is: Where, is the sum of squares of the residuals fitted for strategy i, which measures the error between the model and the true response curve, A reference maximum response time constant set for the system (such as a historical average or industry standard), is the maximum allowed response delay time (such as 30 seconds), is the maximum residual value, which is used to normalize the fitting error. α, β, and γ are weight coefficients, reflecting the influence of the three items on the score (the default value is 1 / 3).

[0110] The control module prioritizes the highest-scoring strategy based on the ranking scores. If the system configuration allows for multiple strategies (e.g., dual-control mode), the top N highest-scoring strategies can be retained for subsequent switching. The control parameters (e.g., pump speed, sieving frequency, target medium concentration, etc.) of the highest-scoring strategy are sent to the control execution module, which then performs parameter adjustments.

[0111] The feedback regulation module monitors and analyzes the execution results of the control strategies issued by the control execution module in real time, comparing actual device responses with the simulation model's predictions. If the system detects deviations between actual and expected operational results, it automatically corrects the parameter model in the simulation modeling module or dynamically adjusts the scoring weights and factor structure in the strategy scoring module, thereby enabling adaptive learning and closed-loop optimization of the intelligent monitoring system.

[0112] The equipment response information after control execution is obtained from the process parameter monitoring module, such as: clean coal yield, power consumption, dielectric consumption, sorting error; dynamic response time, steady-state value, fluctuation range; the acquisition time range covers the complete action cycle of the control strategy.

[0113] Call the target output value Ypred predicted by the simulation modeling module, compare it with the actual monitoring result Yreal, and calculate the target deviation vector : ; If the deviation exceeds the preset threshold (such as ±5%), the feedback correction mechanism is triggered.

[0114] According to the error To fine-tune the model used for prediction in the modeling module, you can use the following correction methods:

[0115] Update the regression model weights or neural network parameters; add the current sample as incremental data to the training set for retraining; rebuild the local model or add auxiliary input variables when the continuous error is large; the updated model re-participates in the next round of control strategy simulation and evaluation.

[0116] If the feedback analysis results show that the scoring deviation mainly comes from the strategy scoring module (i.e. the scoring results fail to accurately reflect the execution effect): analyze the degree to which scoring factors such as image adaptability, historical performance, and response stability explain the actual effect; Perform automatic corrections; prioritize enhancing factors that are highly correlated with the current response results and weaken invalid scoring items; the updated scoring model will participate in subsequent strategy screening and scoring to improve the accuracy of strategy recommendations.

[0117] All correction records (error vector, correction method, model effects before and after correction) are written into the system log and knowledge base; the system has the ability to continuously learn and can gradually optimize the model structure and strategy screening accuracy; and realize dynamic adaptation of the system to factors such as equipment operation changes, coal quality fluctuations, and environmental interference.

[0118] For example, after implementing strategy control, the system predicted a clean coal yield of 86%, but the actual measured yield was 78%, exceeding the acceptable threshold of 5%. This triggered the feedback module to retrain the model and adjust the scoring mechanism (for example, by increasing the weight of the response stability factor). This prioritized strategies with greater control stability in the next round of strategy generation, successfully bringing the yield within the target range.

[0119] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. The coal washing equipment image processing intelligent monitoring system based on strategy optimization is characterized by: It includes image acquisition module, process parameter monitoring module, simulation modeling module, multi-objective optimization module, strategy scoring module, control execution module and feedback adjustment module; Image acquisition module, used to obtain real-time image information of different parts of the coal washing equipment, and perform recognition and structural processing on it to form image feature data; Process parameter monitoring module, used to collect several process parameters during coal washing; The simulation modeling module builds a dynamic simulation model based on image feature data and process monitoring data to simulate the response relationship between multiple optimization objectives in the coal washing process; The multi-objective optimization module generates multiple control strategies based on the Pareto frontier algorithm, calculates the response differences between the strategies, establishes a response space redistribution model, and performs sparse processing on the Pareto solution set to obtain a sparse optimization solution set; The strategy scoring module dynamically scores the sparse optimization solution set based on image features, historical operation data, and response trajectories, and selects the optimal strategy based on the scoring results; The control execution module adjusts the operating parameters of the coal washing equipment according to the selected optimal strategy; The feedback adjustment module is used to monitor the control execution results and automatically correct the simulation model or adjust the strategy scoring parameters based on the deviation between the device response and the target.

2. The coal washing equipment image processing intelligent monitoring system based on strategy optimization according to claim 1 is characterized by: The image acquisition module includes: multiple groups of image sensors installed on the screen surface, sorting slots and conveyor belt discharge port of the coal washing equipment, and the image sensors include high-resolution industrial cameras, infrared thermal imagers and laser profile scanners.

3. The coal washing equipment image processing intelligent monitoring system based on strategy optimization according to claim 2 is characterized by: The image acquisition module is used to collect raw image data during the operation of the equipment; it also performs grayscale processing, edge enhancement and noise removal on the image, identifies the coal block boundaries through a target recognition algorithm, and extracts image features, including particle size distribution, accumulation thickness, movement direction and coal gangue ratio; the processed image features are output in the form of structured vectors and passed as input variables to the simulation modeling module.

4. The coal washing equipment image processing intelligent monitoring system based on strategy optimization according to claim 1 is characterized by: The process parameter monitoring module is used to collect process parameters in the coal washing process in real time, including medium concentration, water-coal ratio, feed particle size, vibration frequency, motor current, power consumption, water supply flow and system temperature.

5. The coal washing equipment image processing intelligent monitoring system based on strategy optimization according to claim 4 is characterized in that: The simulation modeling module is used to build a data-driven model: Extracting particle size distribution and material flow rate image features from the image acquisition module; Obtain process data on water-coal ratio, medium concentration and power consumption from the process parameter monitoring module; Normalize and process outliers on the collected data to form multidimensional modeling samples; Build data-driven models using neural networks, support vector machines, or random forest algorithms; Output the predicted values ​​of each target parameter under the simulation control strategy for transmission to the multi-objective optimization module.

6. The coal washing equipment image processing intelligent monitoring system based on strategy optimization according to claim 5 is characterized by: After generating the Pareto solution set, the multi-objective optimization module performs a sparse processing step: Calculate the difference between the target responses of any two control strategies in the solution set, where the difference is measured by Euclidean distance or Mahalanobis distance; Establish a target spatial redistribution model based on response differences; Use clustering algorithm to cluster the solution set; Select the representative solution with the largest response difference in each cluster to form a sparse optimization solution set; Output the sparse optimization results to the strategy scoring module.

7. The coal washing equipment image processing intelligent monitoring system based on strategy optimization according to claim 6 is characterized by: The strategy scoring module includes: Calculate the matching score between the strategy parameters and the current image features: the matching score is calculated by the Euclidean distance between the current image features and the mean of the strategy's historical image features; Calculate the average effectiveness score of the strategy under similar working conditions: Calculate the average execution effect of the strategy under similar working conditions and take the average of multiple scores; Calculate the deviation control capability score after the strategy is executed: Fit the actual system response to a first-order response model, extract the response time constant, delay time, and fitting error, and construct a scoring function; The matching score, average effect score and deviation control ability score are integrated according to the weighted formula to output the comprehensive score result; Strategies are sorted according to their scores, and the one with the highest score is selected as the optimal control strategy.

8. The coal washing equipment image processing intelligent monitoring system based on strategy optimization according to claim 7 is characterized in that: The feedback adjustment module includes: Sampling indicators from control execution results; Compare with the prediction results of the simulation model and calculate the target deviation vector; If the deviation exceeds the preset threshold, the modeling module fine-tuning mechanism is triggered, including incremental model training, parameter update or structure reconstruction.

Images