Method and system for device control and regulation optimization based on working condition adaptive perception

Abstract

The application provides a device control and adjustment optimization method and system based on working condition adaptive perception, comprising: adopting a dynamic working condition segmentation and perception feature mapping method, automatically dividing the running working condition through unsupervised clustering, and constructing a weight vector to quantify the similarity between the current state and each typical working condition. An improved grey wolf optimization algorithm with deep fusion of working condition perception is adopted, which embeds the working condition characteristics, indicators and adjustment factors into the whole process of population initialization, leader wolf reorganization, position update and diversity disturbance of the algorithm. A compound fitness function with dynamic weighting and exploration reward fusion is adopted, the sub-target weight is dynamically adjusted according to the working condition deviation degree, and an exploration reward item based on population distribution is introduced. A two-stage safe and robust strategy output mechanism is adopted, which contains a posterior decision based on multi-criteria dynamic weight and an instruction generation integrating rate limitation, coupling constraint correction and working condition adaptive smoothing filtering.

Description

Technical Field

[0001] This invention relates to the field of control and regulation optimization technology, and in particular to a method and system for equipment control and regulation optimization based on adaptive sensing of operating conditions. Background Technology

[0002] As modern mining develops towards deeper and more intelligent operations, core mining equipment such as coal mining machines and tunneling machines operate for extended periods in complex, harsh, and dynamically changing geological and production environments. The operational efficiency, energy consumption, and safety stability of this equipment directly impact the overall production efficiency and safety level of the mine. In actual operations, equipment frequently needs to cope with external disturbances such as changes in coal seam thickness, gangue distribution, and geological structures, exhibiting significant multimodal, nonlinear, and highly time-varying characteristics in its operating conditions. Traditional equipment control strategies often rely on preset fixed parameters or empirical rules based on single-condition models, making it difficult to adaptively follow the dynamic changes in operating conditions. This frequently leads to low operating efficiency, increased energy consumption, or overload and wear of critical components under non-design conditions, and can even cause safety accidents. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a method and system for equipment control and regulation optimization based on adaptive perception of operating conditions, so as to improve the control system's ability to perceive and adapt to complex time-varying operating conditions of mining equipment, enhance the search efficiency and accuracy of optimization algorithms, realize intelligent dynamic balance of multi-objective performance of equipment operation, and ensure high reliability, smoothness and safety of control commands.

[0004] In a first aspect, embodiments of the present invention provide a method for equipment control and adjustment optimization based on adaptive working condition perception. The method includes: continuously collecting operational data of the target mining equipment under various working conditions from multiple sensor arrays deployed on the target mining equipment; constructing a working condition perception feature mapping function using a dynamic working condition segmentation strategy; enhancing the representation of the operational data based on the working condition perception feature mapping function; calculating a working condition perception index based on the working condition perception feature mapping function; quantifying the deviation between the current operating state and various working conditions through the working condition perception index; dynamically generating parameter adjustment factors for the Grey Wolf optimization algorithm based on the deviation; and using guided intelligent population initialization based on the Grey Wolf optimization algorithm. The strategy utilizes the characteristic information of historical operating data to generate the initial population position in a region closer to the high-quality solution, while introducing randomness to ensure that the initial solution meets the constraints, thus obtaining the updated population position. A composite fitness function is constructed that integrates multi-dimensional performance indicators, operating condition penalties, and the search state feedback of the Grey Wolf optimization algorithm. By performing posterior decisions based on the operating condition matching degree on the frontier solution set obtained by the algorithm optimization, the parameter set most suitable for the current and near-future predicted operating conditions is selected. Based on the parameter set, and combined with equipment physical constraints, inter-subsystem coupling relationships, and the control commands from the previous moment, a control command sequence is generated. The operating state of the target mining equipment is controlled based on the control command sequence.

[0005] Secondly, embodiments of the present invention also provide an equipment control and adjustment optimization system based on adaptive working condition perception. The system includes: a data layer for continuously collecting operational data of the target mining equipment under various working conditions from multiple sensor arrays deployed on the target mining equipment; a feature extraction layer for constructing a working condition perception feature mapping function using a dynamic working condition segmentation strategy, and enhancing the representation of the operational data based on the working condition perception feature mapping function; a working condition perception layer for calculating working condition perception indicators based on the working condition perception feature mapping function, quantifying the deviation between the current operating state and various working conditions through the working condition perception indicators, and dynamically generating parameter adjustment factors for the Grey Wolf optimization algorithm based on the deviation; and an optimization algorithm layer for employing a guided intelligent swarm optimization algorithm. The initialization strategy utilizes the characteristic information of historical operating data to generate the initial population position in a region closer to the high-quality solution, while introducing randomness to ensure that the initial solution meets the constraints, thus obtaining the updated population position. The optimization algorithm layer is also used to construct a composite fitness function that integrates multi-dimensional performance indicators, operating condition penalties, and the search state feedback of the Grey Wolf optimization algorithm. The control output layer is used to make posterior decisions based on the operating condition matching degree on the frontier solution set obtained by the algorithm optimization, and to select the parameter set that best suits the current and near-future predicted operating conditions. Based on the parameter set, and combined with the physical constraints of the equipment, the coupling relationship between subsystems, and the control commands of the previous moment, a control command sequence is generated. The control output layer is also used to control the operating state of the target mining equipment based on the control command sequence.

[0006] The embodiments of the present invention bring the following beneficial effects: This invention provides a method and system for optimizing equipment control and regulation based on adaptive perception of operating conditions. It can improve the control system's ability to perceive and adapt to complex time-varying operating conditions of mining equipment, enhance the search efficiency and accuracy of optimization algorithms, achieve intelligent dynamic balance of multi-objective performance of equipment operation, and ensure high reliability, smoothness and safety of control commands.

[0007] Other features and advantages of this disclosure will be set forth in the following description, or some features and advantages may be inferred from the description or determined without doubt, or may be learned by practicing the techniques described above.

[0008] To make the above-mentioned objects, features and advantages of this disclosure more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0009] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0010] Figure 1 A flowchart of a device control and adjustment optimization method based on adaptive working condition sensing provided in an embodiment of the present invention; Figure 2 A schematic diagram illustrating normalization and feature extraction enhancement provided in an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating how to calculate operating condition perception indicators and dynamically generate parameter adjustment factors for the Grey Wolf optimization algorithm, as provided in an embodiment of the present invention. Figure 4 A schematic diagram of a scatter plot provided in an embodiment of the present invention; Figure 5 A schematic diagram of a kernel density map provided in an embodiment of the present invention; Figure 6 A schematic diagram of an improved gray wolf optimization algorithm provided in an embodiment of the present invention; Figure 7 A schematic diagram of a device control and regulation optimization system based on adaptive working condition sensing provided in an embodiment of the present invention; Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0012] Currently, existing mining equipment control systems mainly suffer from the following deficiencies: 1. In terms of data preprocessing and feature engineering, simple global normalization is often used, which fails to fully consider the differences in data distribution under multiple operating conditions, making it difficult to effectively extract features that can characterize the essence of different operating modes, thus limiting the learning effect of subsequent models.

[0013] 2. In terms of optimization algorithm application, standard metaheuristic algorithms with fixed parameters are often used. Their exploration and development strategies are rigid and cannot be adaptively adjusted according to the current specific operating status of the equipment. When dealing with dynamically changing working conditions, they are prone to slow convergence or getting trapped in local optima.

[0014] 3. In dealing with multi-objective optimization problems, a linear weighting method with fixed weights is often used to transform multiple objectives into a single objective. This ignores the objective requirement that the priority of each optimization objective should change dynamically under different operating conditions, resulting in the control strategy failing to achieve a contextualized optimal trade-off.

[0015] 4. In terms of optimization result output and application, the theoretical optimal solution is often directly output. There is a lack of comprehensive consideration and post-processing regarding the safety, smoothness and consistency of the solution with the current execution state of the equipment. This may result in infeasible or drastically fluctuating control commands, which poses application risks.

[0016] Based on this, the present invention provides a method and system for equipment control and adjustment optimization based on adaptive perception of operating conditions, which can improve the control system's ability to perceive and adapt to complex time-varying operating conditions of mining equipment, enhance the search efficiency and accuracy of optimization algorithms, realize intelligent dynamic balance of multi-objective performance of equipment operation, and ensure high reliability, smoothness and safety of control commands.

[0017] To facilitate understanding of this embodiment, a detailed description of the equipment control and regulation optimization method based on adaptive working condition sensing disclosed in this embodiment of the invention will be provided first.

[0018] Example 1: This invention provides a method for optimizing equipment control and regulation based on adaptive working condition sensing. (See also...) Figure 1The flowchart shown illustrates a method for equipment control and regulation optimization based on adaptive operating condition sensing. This method includes the following steps: Step S102: Continuously collect operating data of the target mining equipment under various working conditions from multiple sensor arrays deployed on the target mining equipment.

[0019] Regarding the issue of data acquisition and control optimization for mining equipment operation, this embodiment can systematically collect data that comprehensively reflects the equipment's operating status and clearly define the control optimization problem to be solved.

[0020] In some embodiments, the sensor array is installed in key parts of the target mining equipment, including: motor, bearing, hydraulic system, and body; the operating data covers multiple physical dimensions, including: electrical, mechanical, and thermal, and includes: motor current and voltage, temperature of key components, vibration acceleration of body, and working traction force.

[0021] Specifically, historical and real-time operational data under various working conditions are continuously collected from multiple sensor networks deployed on the target mining equipment (such as coal mining machines, tunneling machines, etc.). The process of collecting training data is as follows: By installing sensor arrays in key parts of the equipment (such as motors, bearings, hydraulic systems, and machine bodies), the operating parameters of the equipment under various typical and atypical working conditions are recorded continuously over a long period of time at a fixed sampling frequency, forming a high-dimensional, timestamped historical time-series dataset. The collected data attributes need to cover multiple physical dimensions such as electrical, mechanical, and thermal, including motor current and voltage, temperature of key components, vibration acceleration of the machine body, and working traction force.

[0022] To ensure that the data can effectively support subsequent working condition classification and model learning, the data collection process needs to cover the entire life cycle of the equipment from startup, stable operation, load changes to shutdown, and should include as many operating records as possible under different geological conditions and production tasks to ensure the diversity and representativeness of the data.

[0023] Based on the comprehensive data collected, the core of the control optimization problem defined in this step is: within the feasible domain that satisfies the physical safety constraints and process requirements of the equipment, dynamically find a set of optimal equipment operation adjustment parameters so that the equipment can optimize a comprehensive evaluation index under a specific real-time operating state.

[0024] This optimization problem is a typical complex problem with multiple objectives, multiple constraints, nonlinearity, and dynamic time-varying characteristics. Its challenge lies in the fact that the relative importance of the optimization objectives changes dynamically with the change of equipment operating conditions, and the structure of the feasible solution space also changes accordingly.

[0025] Step S104: Construct a working condition perception feature mapping function using a dynamic working condition segmentation strategy, and enhance the representation of the operating data based on the working condition perception feature mapping function.

[0026] Mining equipment operation data has multi-sensor, multi-condition, nonlinear and time-varying characteristics. In order to effectively cope with the impact of data scale differences and dynamic changes under different conditions, it is necessary to standardize the data and extract features. Conventional techniques directly normalize the raw data, ignoring the inherent condition segmentation information and local fluctuation characteristics of the data. It is difficult to effectively capture multi-scale time series changes, and the performance is poor under the uneven distribution of multi-condition data and noise interference.

[0027] This embodiment can perform normalization and feature extraction enhancement: a dynamic operating condition segmentation strategy is used to construct an operating condition-aware feature mapping function to enhance the representation of the original operating data, thereby solving the problems of uneven distribution of multi-operating condition data and noise interference. (See also...) Figure 2 The diagram illustrates a normalization and feature extraction enhancement method.

[0028] In some embodiments, multi-sensor historical time-series data of the target mining equipment during operation can be collected, and each feature of the historical time-series data can be standardized. An unsupervised clustering algorithm is used to cluster the historical time-series data, and the operating status of the target mining equipment is dynamically divided into multiple working condition segments. A working condition perception feature mapping function is constructed based on the results of the working condition segmentation. The operating data is mapped to a weight vector with similarity to each working condition based on the working condition perception feature mapping function, and the weight vector is used as the enhanced feature representation.

[0029] This embodiment can perform dynamic operating condition segmentation: collect multi-sensor historical time series data of the equipment during normal operation, perform Z-score standardization on each feature to eliminate the influence of the dimension, and then use an unsupervised clustering algorithm to cluster the data, automatically dividing data points with similar operating characteristics into the same cluster, thereby dynamically dividing the continuous equipment operating state into multiple discrete and representative operating condition segments. definition This represents the total number of operating condition segments obtained through dynamic partitioning, characterizing the diversity of equipment operating states; Indicates the first Cluster center vectors for each working condition segment This represents the typical characteristic pattern of equipment operation data within this segment; Indicates a segmented index based on operating conditions; Indicates the first The feature width parameter of each working condition segment is used to adjust the response range of the segment feature to the data points. Its value is usually set according to the average of the squared Euclidean distances of all sample feature vectors in the segment to their cluster centers.

[0030] In practical implementation, the clustering algorithm can be selected according to the characteristics of data distribution. For example, clustering based on the K-means algorithm can provide a basis for accurate division of working conditions.

[0031] This embodiment can construct a working condition-aware feature mapping function: based on the results of dynamic working condition segmentation, a mapping function is constructed to map the original operating data of any sample into a weight vector with similarity to each typical working condition. This weight vector serves as the enhanced feature representation, expressed as: ; In the formula, This represents the feature mapping function for operating condition perception, and its output is: A weight vector of dimension; Indicates the first Each sample's device operation data vector is the device operation data after Z-score normalization of the original collected data, with dimensions denoted as . ; Indicates the sample index; Indicates the first For the nth sample The adaptive weight coefficients for each work condition segment are used to quantify the similarity between the current sample and that work condition. A higher value indicates a higher similarity. The calculation method is expressed as follows: And satisfy ; In the formula, Indicates the transpose operation; This represents the weight decay coefficient, used to adjust the sensitivity of distance to weight calculation. The larger the value, the more concentrated the weight is on the nearest working condition center; the example value is 1.5. Represents the natural exponential function; This represents the L2 norm.

[0032] In one embodiment, taking a mining coal mining machine as an example, the attribute set of the equipment operation data is as follows: ,in, This indicates the motor current, measured in amperes. This indicates the motor voltage, measured in volts. This indicates the bearing temperature, in degrees Celsius. This indicates the hydraulic oil temperature, in degrees Celsius. This indicates the vibration acceleration of the fuselage, measured in meters per second squared. It represents traction force, measured in kilonewtons. In practical applications, Z-score standardization is used to eliminate dimensional differences.

[0033] It should be noted that, As an enhancement feature, it can quantify the degree of matching between the current operating status and various known typical operating conditions, effectively integrating information from multiple operating conditions.

[0034] Step S106: Calculate the working condition perception index based on the working condition perception feature mapping function, quantify the deviation between the current operating state and each working condition through the working condition perception index, and dynamically generate the parameter adjustment factor of the Grey Wolf optimization algorithm based on the deviation.

[0035] Faced with the complex time-varying characteristics of mining equipment operation, the conventional gray wolf optimization algorithm uses fixed control parameters, which makes it difficult to adaptively balance the algorithm's global exploration and local development capabilities under different equipment operating conditions. This can easily lead to slow convergence speed or getting stuck in local optima, making it difficult to achieve fast and accurate optimization of equipment operating parameters.

[0036] This embodiment constructs an adaptive operating condition sensing module based on the operating condition sensing feature mapping function. It quantifies the deviation between the current operating state and typical operating conditions by calculating operating condition sensing indicators, and dynamically generates parameter adjustment factors for the Grey Wolf optimization algorithm accordingly. (See also...) Figure 3 The diagram shows a method for calculating operating condition perception indicators and dynamically generating parameter adjustment factors for the Grey Wolf optimization algorithm.

[0037] In some embodiments, the working condition perception index of the current sample can be calculated based on the cluster centers of each working condition segment and the weight vector obtained by the working condition perception feature mapping function; the working condition perception index is used to quantify the deviation between the current operating state and each working condition by weighted average distance; and the working condition perception index is used to dynamically calculate the parameter adjustment factor of the Grey Wolf optimization algorithm by a smooth mapping function.

[0038] This embodiment can calculate the operating condition perception index: based on the cluster centers of each operating condition segment and the weights obtained by the mapping function, the operating condition perception index of the current sample is calculated. This index quantifies the overall deviation of the current operating state from the known typical operating condition mode by weighted average distance, and is expressed as: ; In the formula, Indicates the first The operating condition perception index for each sample represents the weighted average deviation of the equipment from the center of all typical operating conditions under its current operating state. The larger the value, the more the current state deviates from all typical operating conditions, and it may be in a transitional or abnormal state. The smaller the value, the closer the current state is to a typical working condition.

[0039] This embodiment can calculate the adaptive parameter adjustment factor: using the operating condition perception index, the adjustment factor of key parameters in the Grey Wolf optimization algorithm is dynamically calculated through a smooth mapping function, enabling the algorithm to adaptively adjust the search strategy according to the degree of deviation of the operating state, expressed as: , ; In the formula, Indicates corresponding to the first The convergence factor adjustment factor for each sample is used to dynamically enhance the global exploration capability of the Grey Wolf optimization algorithm; Indicates corresponding to the first The development factor adjustment factor for each sample is used to dynamically adjust the local development capability of the Grey Wolf optimization algorithm; This represents the baseline value of the convergence factor adjustment factor, which is a preset parameter of the algorithm under standard operating conditions. An example value is 1.0. This represents the baseline value of the development factor adjustment factor, which is the preset parameter of the algorithm under standard operating conditions. An example value is 0.5. This represents the adjustment intensity coefficient, used to control the influence of the working condition perception index on parameter adjustment. An example value is 0.2. This represents the hyperbolic tangent activation function, which maps the input to... interval; This represents the mean of all historical sample operating condition perception indicators, used to normalize the operating condition perception indicators of the current sample, so that the value input to the hyperbolic tangent activation function is within a reasonable range.

[0040] In some embodiments, the effect of adaptive sensing of operating conditions can also be analyzed by combining scatter plots and kernel density plots; wherein, the horizontal and vertical axes of the scatter plot represent two dimensions of operating characteristics after standardization, and the kernel density plot is used to show the distribution of operating condition sensing indicators.

[0041] This embodiment can also analyze the effect of the adaptive sensing of the working conditions in this embodiment by combining scatter plots and kernel density plots, and visualize the system's ability to identify and quantify different operating states of the equipment.

[0042] See also Figure 4The diagram shows a scatter plot. In the scatter plot, the horizontal and vertical axes represent two key operational feature dimensions after standardization: "fuselage vibration acceleration" and "traction force." Each point in the plot represents a sample of equipment operating status, and the different shapes and colors of the points represent the four preset operating condition types to which they belong. Four clusters of clusters are clearly visible in the plot, and the black "X" marks represent the centers of each operating condition segment automatically learned through unsupervised clustering in this embodiment. The dashed lines drawn around the centers indicate the identification boundaries of each operating condition. Experimental results show that this embodiment can effectively learn and distinguish different operating modes from high-dimensional operational data.

[0043] See also Figure 5 The diagram illustrates a kernel density plot, which further demonstrates the distribution of the condition-aware index defined in this embodiment. This index quantifies the deviation of the current state from a known typical operating condition. The horizontal axis represents the operating condition type, categorized into "stable operating condition" and "abnormal operating condition." The vertical axis represents the condition-aware index, a continuous non-negative value. By combining the violin plot with the internal box plot, it can be clearly seen that the index values ​​under stable operating conditions are concentrated in a lower range, with a narrow distribution and a low median. Conversely, the index values ​​under abnormal operating conditions have a wider distribution range and tend to shift towards larger values. This distribution difference confirms the effectiveness of the condition-aware index: when the equipment operates smoothly and approaches a typical mode, the index value is small; when the operating state is abnormal or in a transitional phase, the index value increases.

[0044] Step S108: Based on the Grey Wolf Optimization Algorithm, a guided intelligent population initialization strategy is adopted. The feature information of historical running data is used to generate the initial population position in a region that is closer to the high-quality solution. While introducing randomness, the initial solution is guaranteed to meet the constraints, and the updated population position is obtained.

[0045] In this embodiment, an equipment control and regulation optimization system based on the improved gray wolf optimization algorithm can be constructed.

[0046] This embodiment can improve the population initialization of the Grey Wolf Optimization Algorithm. The conventional Grey Wolf Optimization Algorithm uses a completely random method to initialize the population position in the solution space. In the problem of optimizing the operating parameters of mining equipment, due to the high dimensionality of the parameters to be optimized and the complex boundary of the feasible region, completely random initialization may lead to low quality and unreasonable distribution of the generated initial solutions, or even a large number of violations of physical or safety constraints, which seriously affects the convergence speed and final optimization effect of the algorithm.

[0047] This embodiment, based on the output of the condition-aware feature mapping function and the adaptive parameter adjustment factor, employs a guided intelligent population initialization strategy. It utilizes the feature information of historical operating data to generate the initial population position in a region closer to the optimal solution. While introducing an appropriate amount of randomness, it strictly ensures that all initial solutions satisfy the constraints. (See also...) Figure 6 The diagram shows an improved gray wolf optimization algorithm.

[0048] In some embodiments, a device operation adjustment parameter vector can be defined for each operating condition segment. Based on the similarity weight between the sample and each operating condition, multiple device operation adjustment parameter vectors are weighted and fused, and random perturbations modulated by the operating condition perception state are introduced to generate an initial population position. Boundary constraint processing is performed on the initial population position. Through element-by-element comparison and pruning, the parameter values ​​of multiple dimensions of the initial population position are restricted to a preset feasible domain to obtain the updated population position.

[0049] This embodiment can generate an initial population position based on the characteristics of the operating conditions: a typical equipment operation adjustment parameter vector is defined for each operating condition segment, and the samples are weighted and fused based on the similarity weight between the samples and each operating condition. At the same time, random perturbations modulated by the operating condition perception state are introduced to generate the initial population position, represented as: ; In the formula, In the Grey Wolf optimization algorithm, the first... The initial position vector of each individual. Each element in the vector corresponds to a device operation adjustment parameter to be optimized; This represents the index of an individual in the Grey Wolf Optimization Algorithm, with each individual associated with a sample data point during initialization; This represents the population size of the Grey Wolf Optimization Algorithm, with an example value of 30. Indicates the first The typical equipment operation adjustment parameter vector corresponding to each working condition segment is extracted from the historical best operating record under that working condition. Indicates the first For the nth sample The adaptive weight coefficients for the first working condition segment represent the weights used to initialize the first working condition segment. The individual-related sample data for the first The weight of each working condition segment; Indicates corresponding to the first The convergence factor adjustment factor for the sample, characterizing the . Convergence factor adjustment factor for individual related samples; This represents the random disturbance coefficient, which is a preset constant used to control the overall magnitude of the random disturbance term. An example value is 0.05. Represent a A dimensional random vector, each element of which is in It follows a uniform distribution over the interval; This represents the lower bound vector of the equipment's operating and adjustment parameters. This represents the upper bound vector of the equipment's operating and adjustment parameters. This represents element-wise multiplication of vectors, also known as the Hadamard product.

[0050] It should be noted that, The system performs weighted fusion based on the similarity between the current data and various operating conditions to obtain a physically meaningful guiding position. The term is adaptive random disturbance; when the operating condition perception index is large, Larger sizes and stronger disturbances facilitate exploration over a wider range of conditions when deviating from typical operating conditions.

[0051] This embodiment can perform initial population constraint processing: The generated initial position vector is subjected to boundary constraint processing. Through element-by-element comparison and pruning, the parameter values ​​of all dimensions are forced to be limited to a preset feasible region, ensuring that every individual in the population is a feasible solution, expressed as: ; In the formula, This indicates an assignment / update operation; This indicates the operation of retrieving the maximum value; This indicates the operation of taking the minimum value.

[0052] In one embodiment, taking the speed regulation and cooling system control of a coal mining machine as an example, the parameters to be optimized include: traction speed setpoint (m / min), cutting motor power limit (kW), hydraulic system pressure setpoint (MPa), cooling water valve opening (percentage), and spray dust suppression system flow rate (L / min). In this embodiment, the upper and lower bound vectors are exemplified as follows: , , which correspond to the upper and lower limits of the above parameters, respectively.

[0053] The conventional gray wolf optimization algorithm adopts a fixed position update mechanism, and its search behavior depends on the guidance of a pre-set leader wolf. Under the background of multi-condition operation of mining equipment, the fixed update strategy cannot adapt to the dynamic characteristics of the optimization target and the differences in solution space structure under different conditions. This may lead to the algorithm converging too quickly and getting trapped in local optima under some conditions, or underdeveloping and having limited accuracy under other conditions. At the same time, the lack of population diversity in the later stages of iteration weakens the global optimization ability.

[0054] In some embodiments, operating condition sensing information can be integrated, and the exploration and development behavior of the gray wolf optimization algorithm can be dynamically adjusted through adaptive parameter adjustment factors, operating condition sensing feature weights, and real-time load characteristics. This allows the gray wolf optimization algorithm to intelligently balance the following of the leader wolf, random exploration, and guidance toward historical high-quality solution regions based on the stability and typicality of the current equipment operating status.

[0055] This embodiment can also implement a gray wolf optimization algorithm update mechanism based on adaptive working condition perception: the core update mechanism of the gray wolf optimization algorithm is improved, the working condition perception information is deeply integrated, and the algorithm's exploration and development behavior is dynamically adjusted through adaptive parameter adjustment factors, working condition perception feature weights and real-time load characteristics. This enables the algorithm to intelligently balance the following of the leader wolf, random exploration and guidance towards the historical high-quality solution area based on the stability and typicality of the current equipment operating status.

[0056] This embodiment can adaptively reorganize the leader wolf hierarchy based on work condition weights: Before each iteration, the comprehensive evaluation value of an individual is calculated based on its fitness value and the work condition weight vector of associated samples. Then, the leader wolf hierarchy is redefined using this comprehensive evaluation value. This ensures that under a specific work condition, individuals with a higher degree of matching to that work condition, even with slightly lower absolute fitness values, can obtain higher guidance priority, making the search direction more consistent with the optimal solution distribution characteristics under the current work condition. This can be expressed as: ; In the formula, Indicates the first The overall evaluation value of an individual is used to rank all individuals in the population according to their quality; the smaller the value, the better the overall evaluation. This represents the influence coefficient of the working condition weight, used to balance the relative importance of the normalized fitness value and the maximum working condition weight in the leader wolf selection, and its value range is [value range missing]. Example of a value: ; Indicates the first The normalized fitness value of each individual is used to map the original fitness value to a linear transformation. The interval is calculated as follows: ; Indicates the first Sample data associated with each individual for all The maximum value in the adaptive weight vector of each working condition is used to reflect the highest matching confidence of the solution represented by that individual with a certain typical working condition. Indicates the first The final fitness value of each individual; This represents the minimum initial fitness value in the current population. This represents the maximum initial fitness value in the current population; This represents a very small constant used to prevent the denominator from being zero. Examples of its values ​​are shown below. .

[0057] Furthermore, based on the calculated comprehensive evaluation value, all individuals are sorted in ascending order, and the individuals ranked first, second, and third are selected as the current iteration's... Leader of wolves Leader of wolves and The leader is the wolf.

[0058] This embodiment can integrate location updates based on operating condition awareness and real-time load characteristics: In the location update method of the standard Grey Wolf optimization algorithm, a convergence factor and development factor dynamically adjusted by operating condition information are introduced to regulate the encirclement step size. At the same time, an operating condition guidance term is added, and the search direction is softly pulled towards the typical parameter region of historical high-quality operating conditions using operating condition weights. Furthermore, real-time load change rate characteristics are combined to enhance the response capability to sudden operating conditions, expressed as: ; In the formula, Indicates the first The first gray wolf individual in the... The position vector at the next iteration corresponds to a set of equipment operation adjustment parameters to be optimized; Indicates the first The first gray wolf individual in the... The updated position vector at the next iteration; Indicates the iteration count index; Indicates the first In the next iteration The position vector of the leader wolf. Indicates the first In the next iteration The position vector of the leader wolf. Indicates the first In the next iteration The position vector of the leader wolf; express The distance coefficient vector corresponding to the leader wolf. express The distance coefficient vector corresponding to the leader wolf. express The distance coefficient vector corresponding to the leader wolf; express The oscillation factor vector corresponding to the leader wolf. express The oscillation factor vector corresponding to the leader wolf. express The oscillation factor vector corresponding to the leader wolf; The total number of iterations is preset by the user. Indicates the first The working condition guidance term for each individual gray wolf is used to bias and guide the search direction towards the historical high-quality solution region. The calculation method is as follows: ; In the formula, Indicates corresponding to the first The development factor adjustment factor for the sample is related to the first sample. The development factor adjustment factor for each individual associated sample is used to control the strength of the working condition guidance item. The larger the value, the stronger the guidance to historically high-quality areas. Indicates the relationship with the first The real-time load change rate characteristics of individual related samples are used to quantify the instantaneous fluctuation intensity of equipment operating status. The calculation method is as follows: , which is the L2 norm of the difference vector between the current standardized running data vector and the data vector of the previous time window; Indicates the first The device operation data vector of the sample is related to the first sample. Device operation data vectors associated with individual samples, Indicates the first The difference vector of the device operation data vector of each sample relative to the data vector of the previous time window; This represents the moving average of the real-time load change rate characteristics of all historical samples.

[0059] In practical implementation, the distance coefficient vector and sway factor vector during the position update process are calculated based on the dynamic convergence factor. Specifically, firstly, the first step is to calculate the... The dynamic convergence factor of each individual is expressed as: Then, for each leader wolf, generate a random vector, including, for Leader Wolf: , ,for Leader Wolf: , ,for Leader Wolf: , , then, for Leader Wolf: , ,for Leader Wolf: , ,for Leader Wolf: , ,in, Indicates each element in Uniformly distributed in intervals dimensional vector, This indicates the dimension of the parameter to be optimized.

[0060] In practical implementation, taking a coal mining machine as an example, a set of equipment operation adjustment parameters includes the traction speed setting value, the cutting motor power limit value, the hydraulic system pressure setting value, the cooling water valve opening degree, and the spray dust suppression system flow rate. The quality of these parameters is measured by a fitness function, and the smaller the value, the better the overall performance.

[0061] This embodiment can maintain perturbation based on the diversity of working condition perception indicators: In the later stages of each iteration, a random perturbation is applied to non-leader wolf individuals with a certain probability. The perturbation intensity is dynamically controlled by the working condition perception indicators of the individual's associated samples. When the equipment operating state deviates from typical working conditions, a stronger perturbation is applied to improve population diversity and enhance the ability to escape local optima, as expressed as: ; In the formula, This represents the baseline disturbance coefficient, a preset constant used to control the overall amplitude of the random disturbance term. An example value is shown below. ; This represents the average value of the working condition perception index of all associated samples of all individuals in the current population; Indicates the first Working condition perception indicators of individual related samples; This represents a random vector whose elements in each dimension independently follow a standard normal distribution.

[0062] In practice, this perturbation operation is performed after each iteration with a preset probability. An example of the preset probability value is: It primarily functions on non-leader wolf individuals to maintain population diversity.

[0063] Step S110: Construct a composite fitness function that integrates multi-dimensional performance indicators, working condition penalties, and search state feedback from the Grey Wolf optimization algorithm.

[0064] Control optimization of mining equipment is a typical complex problem with multiple objectives and constraints. Conventional methods usually simply linearly weight multiple objectives into a single fitness value, which fails to fully consider the dynamic changes in the priority of each optimization objective under different operating conditions. As a result, the evaluation results cannot accurately reflect the true comprehensive performance of the equipment under specific operating conditions.

[0065] In this embodiment, fitness function calculation and optimization result evaluation can be performed: This embodiment constructs a composite fitness function that integrates multi-dimensional performance indicators, operating condition penalties, and algorithm search state feedback. It not only evaluates the direct effect of the control parameter set, but also promotes the optimization to converge to a more stable and typical operating region by introducing an operating condition deviation penalty term. Combined with an "exploration reward" term based on population distribution characteristics, it guides the algorithm to balance "exploration" and "mining" from the evaluation level.

[0066] In some embodiments, multiple sub-objective components can be defined based on the equipment physical model, operating data, and optimization parameters; wherein, the sub-objective components include: productivity component, energy consumption component, safety risk component, and operating condition deviation penalty component; multiple sub-objective components are aggregated into a scalar fitness value, and the weights are dynamically adjusted according to the operating condition perception index of the current sample; an exploration reward term based on the current population distribution and individual position is introduced to correct the fitness, resulting in a composite fitness function.

[0067] This embodiment allows for comprehensive optimization of target components: defining multiple core optimization target components, each calculated based on the equipment physical model, operational data, and optimization parameters. These target components collectively cover key aspects such as production efficiency, energy consumption, equipment safety, and operational stability. Specifically, let's assume that for the first... Each individual gray wolf corresponds to a set of equipment operation and adjustment parameters, and its associated real-time operation data samples are: Calculate the following four sub-target components: 1) Productivity component: The productivity component is positively correlated with the equipment's operating efficiency. In one embodiment, for a coal mining machine, it can be modeled as a function of traction speed and effective cutting power, and normalized to the [0,1] interval. A larger value indicates higher productivity. To align with the principle of minimizing fitness, it is converted into a negative index, expressed as: ; In the formula, Indicates the first The productivity component of an individual is a negative indicator; the smaller the value, the higher the estimated productivity under that set of parameters. This represents the weighting coefficient of traction speed in productivity assessment, used to adjust the relative importance of traction speed and cutting power to productivity contributions, with a value range of [value range missing]. Example of a value: ; Indicates the first The traction speed setting value corresponding to each individual is the position vector of that individual. One of the dimensions, with units of meters per minute; This indicates the physical lower limit that the traction speed parameter is allowed. This represents the physical upper limit allowed for the traction speed parameter, derived from parameter constraints. and The corresponding value in; Indicates the first The average effective cutoff power estimated under individual parameters is obtained through historical data mapping relationships, and the unit is kilowatt; This represents the empirical lower limit of average effective cutting power. This represents the empirical upper limit of average effective cutting power, which is statistically derived from historical operating data.

[0068] 2) Energy Consumption Component: The energy consumption component assesses the system's energy consumption and is positively correlated with the total input power. It is also normalized and negatively quantified, and expressed as: ; In the formula, Indicates the first The energy consumption component of an individual is a negative indicator; the smaller the value, the lower the energy consumption, and the larger the value, the higher the energy consumption. Indicates the first The total power estimate corresponding to each individual parameter, in kilowatts, is calculated in real time using parameters such as motor current, voltage, and system efficiency. The reference power base value is a constant used for normalization, specifically the average power value of the equipment under rated or typical high-efficiency operating conditions.

[0069] 3) Safety risk components: The safety risk component quantifies the equipment risk that operating parameters may cause; the higher the value, the higher the risk. It is constructed by combining real-time load characteristics and key status parameters, and is expressed as follows: ; In the formula, Indicates the first The individual safety risk component is a negative indicator; the smaller the value, the lower the safety risk. When all critical state parameters are far from the threshold and the load is stable, this component tends towards... This indicates that the risk is extremely low. For example, a set of critical security states This includes the security dimensions that need to be monitored; Indicates a critical security state index, used for traversing the collection. The various security dimensions within; Indicates the first The penalty weighting coefficient for each safety state is used to adjust the importance of different safety states in the total risk. It is preset by the user and satisfies the following conditions: ; Indicates the first The first individual parameter A predicted or estimated value for a safety state, such as a predicted bearing temperature, can be obtained using a lightweight forward model or empirical formula, and needs to be compared with... Maintain the same dimensions; Indicates the first A preset security threshold for a security state, when The risk increases when this threshold is exceeded; This represents the positive part operation, such that only when... A positive penalty term will only be generated when the state exceeds a threshold. Its absolute value increases as the over-limit ratio increases; The scaling factor, representing the effect of the load change rate, is a positive constant used to adjust... Sensitivity to the impact of security risks, with examples of possible values. .

[0070] 4) Operating condition deviation penalty component: The operating condition deviation penalty component uses the operating condition-aware weight vector and operating condition-aware index to penalize operating states that deviate from typical operating conditions, guiding the optimization results towards a stable and familiar operating condition region, as expressed as: ; In the formula, Indicates the first The deviation of an individual's working condition from the penalty amount is a positive indicator; the larger the value, the heavier the penalty. This represents the penalty intensity coefficient, a preset positive constant used to control the overall weight of the deviation penalty term in the total fitness. An example value is shown below. .

[0071] This embodiment can perform dynamic weighted multi-objective aggregation based on fusion of operating condition perception: the above sub-objective components are aggregated into scalar fitness values, and the weights are no longer fixed, but are dynamically adjusted according to the operating condition perception indicators of the current sample to reflect the optimization focus under different operating conditions. When it is close to the typical operating condition, the focus is on efficiency and energy consumption, and when it deviates from the typical operating condition, the focus is on safety and regression stability. Specifically, define a dynamic weight vector. And satisfy ; in, Indicates the first The individual The dynamic weights of each sub-objective component are used to dynamically adjust the relative importance of different optimization objectives during aggregation. Their values ​​are dynamically calculated based on the current running state, and the calculation method is expressed as follows: ; In the formula, Indicates the sub-target component index. These correspond to four components: productivity, energy consumption, safety risk, and operating condition deviation penalty, respectively. Indicates the first The deviation of a sample from the normalized operating condition is represented by a value that indicates a greater deviation from the typical operating condition. The calculation method is expressed as follows: ; Indicates the first The sensitivity coefficient of each sub-target component to the deviation of the operating condition is a preset normal value. When the sensitivity coefficient increases, The larger the sub-objective, the higher its dynamic weight. The faster it decays.

[0072] Further, the initial aggregation fitness is calculated, denoted as... ; in, Indicates the first The initial aggregated fitness value of each individual is combined with the four sub-objective components and dynamically weighted according to the degree of deviation from the current working condition, following the principle that the smaller the value, the better.

[0073] This embodiment can perform adaptive fitness correction of the search state of the fusion algorithm: In order to more explicitly balance exploration and development at the algorithm level, an "exploration reward" item based on the current population distribution and individual position is introduced to correct the fitness. When an individual is far away from the current leader wolf and the algorithm is in the early stage or in a state that requires exploration, a certain fitness reward is given. Specifically, the calculation method for the exploration degree metric is defined as follows: ; in, Indicates the first The individual in the first The exploration measure during generation iteration represents the average Euclidean distance between the individual's position and the positions of the three current leader wolves. The larger the value, the further the individual is from the core leadership area of ​​the population, and the stronger the exploration behavior. Indicates the first The global exploration demand factor during iteration quantifies the overall demand intensity of the algorithm for exploration behavior in the current iteration. It is jointly determined by the iteration progress and the population average convergence factor adjustment factor, and is calculated as follows: ; Indicates the current population number The convergence factor and adjustment factor of all individuals are averaged in the iterative process, inheriting the working condition perception information of the group. When the group generally deviates from the typical working condition, Increase The increase indicates a stronger need for global exploration.

[0074] It should be noted that in calculating the first... The individual in the first Exploration degree measurement during iteration Before that, the components of the position vector are normalized to the same scale to eliminate dimensions.

[0075] Furthermore, the calculation method for the revised final fitness value is expressed as follows: ; in, Indicates the first The final fitness value of an individual is a scalar evaluation index used to guide algorithm iteration and individual selection. The smaller the value, the better the overall performance. This represents the exploratory measure of all individuals in the current population. The minimum value; This represents the exploratory measure of all individuals in the current population. The maximum value; This represents the exploration reward intensity coefficient, a preset positive constant used to control the maximum impact of exploration rewards on the final fitness value. An example value is shown below. .

[0076] It should be noted that, The normalized value of the individual exploration degree is calculated by the term representation. Mapped to The interval is used to fairly measure an individual's relative exploration level in the current population. Furthermore, individuals who move away from the leader wolf when exploration is needed receive negative rewards, resulting in a smaller final fitness value. This increases their probability of being retained in the next generation, thereby enhancing population diversity.

[0077] Step S112: By performing a posteriori decision based on the working condition matching degree on the frontier solution set obtained by the algorithm optimization, the parameter set that best suits the current and near-future predicted working conditions is selected; based on the parameter set, and combined with the physical constraints of the equipment, the coupling relationship between subsystems and the control command of the previous moment, a control command sequence is generated.

[0078] After evaluation and multiple rounds of iterative optimization, a set of high-performance equipment operation adjustment parameters are output. However, directly applying the parameters corresponding to the optimal individual to the equipment may be risky because the optimization process is based on models and historical data, and the operating conditions may continue to change. Conventional methods directly output the optimal solution, lacking consideration for the smoothness, safety, and consistency with preceding control commands of the output results.

[0079] This embodiment can output equipment control and adjustment strategies: it adopts a two-stage strategy output mechanism, and performs posterior decision-making based on the working condition matching degree on the Pareto front solution set obtained by algorithm optimization to select the parameter group that is most suitable for the current and near-future predicted working conditions. In addition, a real-time safety correction and smoothing filtering module is introduced. This module comprehensively considers the physical constraints of the equipment, the coupling relationship between subsystems and the control command of the previous moment to generate a final safe, smooth and directly identifiable control command sequence.

[0080] In some embodiments, a posterior decision based on the working condition matching degree can be made on the frontier solution set obtained by the algorithm optimization based on the comprehensive performance criterion, the working condition matching degree criterion, and the output stationarity criterion to select the parameter set that best suits the current and near-future predicted working conditions; based on the parameter set, rate limit correction and coupling constraint correction are performed to generate a control command sequence.

[0081] This embodiment can select optimization parameters based on multi-criteria posterior decision-making: assuming that after the algorithm iterations are completed, an elite solution set is retained. and their corresponding fitness values Related operating condition weights and operating condition perception indicators Decisions should be made based on the following criteria: 1) Criterion 1 (Overall Performance): Ranking of Normalized Fitness The smaller the value, the better; in, Indicates a sorting operation. Minimum solution Second smallest And so on.

[0082] 2) Criterion 2 (Operating Condition Matching Degree): The matching degree between the predicted operating state and each typical operating condition under the solution parameters is calculated using operating condition-aware features. This is then used to predict the application through a lightweight forward model. Steady-state values ​​of post-critical operating data Then calculate its working condition weights. Further define the matching score A larger value indicates a better match with a typical working condition; in, This represents a lightweight forward prediction model, specifically implemented using an empirical regression model, used to quickly evaluate the steady-state response of key operating states of equipment under given control parameters; Indicates the predicted state For the The weights of each working condition segment are calculated using the working condition perception feature mapping function.

[0083] 3) Criterion 3 (Output Stationarity): Calculate the solution parameters and the parameter vector being executed at the current time. Differences: A smaller value indicates a smoother transition; in, This represents the vector of operating adjustment parameters currently being executed on the device, which is the output of the previous control cycle.

[0084] Furthermore, a weighted sum method is used for decision-making, but the weights are determined by the current real-time operating condition perception indicators. and load change rate Dynamically determined, expressed as:

[0085]

[0086]

[0087] In the formula, This represents the real-time dynamic weights of criterion 1. This represents the real-time dynamic weights of criterion 2. This represents the real-time dynamic weights of criterion 3; The normalized operating condition deviation represents the current real-time operating status, and is calculated as follows: ; This refers to the operational condition perception index calculated based on current real-time operational data. The load change rate characteristic represents the current real-time operating status, and its calculation method is the same as that of the real-time load change rate characteristic. The calculation method; This represents the first weighted sensitivity coefficient. This represents the second weighted sensitivity coefficient. This represents the third weight sensitivity coefficient, a preset positive constant used to adjust the weights of each criterion. and Sensitivity to change.

[0088] In practical implementation, when the operating conditions are stable ( When the focus is on performance (small) Large) and stability ( Large); when the operating conditions deviate ( When the value is large, the focus is on the matching degree. big).

[0089] Furthermore, the comprehensive decision score for each elite solution The calculation method is expressed as ; in, yes After min-max normalization to The value after the interval. In the specific implementation, choose... The minimum solution is used as the initial decision output. .

[0090] This embodiment can perform real-time security correction and smooth command generation: to ensure To ensure safe and reliable operation on equipment, a correction filter is defined to perform rate limiting correction and coupling constraint correction: 1) Rate limiting correction: for The Middle Parameters Its rate of change cannot exceed the actuator's maximum permissible rate of change. Define the first correction vector The d-th element The calculation method is expressed as ; in, The time interval of the control cycle, in seconds, determines the length of time between two control commands and is used to calculate the maximum allowable change in parameters within a single control cycle. Indicates the first The actual execution value of each parameter at the current moment; This represents a sign function used to determine the direction of parameter adjustment.

[0091] 2) Coupling constraint correction: Define a set of coupling relationship functions ,and ;For example, ; If the first correction vector If the constraints are violated, the out-of-bounds parameters are scaled proportionally until all constraints are satisfied. This process is an iterative fine-tuning process, which outputs a second correction vector. .

[0092] In practical implementation, the correction process can be specified as follows: if Violation of any constraint Then a scaling factor is introduced. And let the second correction vector ,from Start by gradually decreasing (For example, reduce by 0.1 each time), until for all All have Established.

[0093] Furthermore, a first-order low-pass filter is applied for condition-adaptive smoothing filtering, but its time constant... Working condition sensing indicators Changes are represented as: , , ; In the formula, This represents the filter time constant used in the current control cycle, and its value is dynamically adjusted according to the operating conditions. This represents the lower bound of the time constant. This represents the upper limit of the time constant and is a preset positive constant. ; The filter update coefficients are used to calculate the weighted average of the current control command and the control command at the previous moment to achieve a smooth transition. This represents the control command vector that is ultimately output to the equipment for execution. In one embodiment, taking a coal mining machine as an example, the control command vector that is ultimately output to the equipment for execution is the equipment operation adjustment parameters after safety correction and smoothing, including the traction speed setpoint, the cutting motor power limit, the hydraulic system pressure setpoint, the cooling water valve opening, and the spray dust suppression system flow rate.

[0094] Step S114: Control the operating status of the target mining equipment based on the control command sequence.

[0095] After obtaining safe, smooth, and adaptable control commands that are suitable for the current operating state through the aforementioned steps, the system enters the execution phase, realizing closed-loop equipment control and regulation based on adaptive perception of operating conditions. This process is a continuous cyclic adaptive closed loop.

[0096] At the beginning of each control cycle, the system collects the latest operating data of the equipment in real time and immediately standardizes it. Then, the standardized data is input into the pre-built working condition perception feature mapping function to calculate the similarity weight vector between the current operating state and each typical working condition. Then, based on the weight vector and the historical working condition center, the working condition perception index and real-time load change rate characteristics at the current moment are calculated. The adaptive working condition perception module dynamically generates the convergence factor and development factor adjustment factor of the Grey Wolf optimization algorithm suitable for the current state. After completing the quantitative perception of the current operating status, the system takes the status information, adjustment factor, and control command from the previous cycle as input to activate the equipment control and regulation optimization system based on the improved gray wolf optimization algorithm. The system uses a guided intelligent population initialization strategy to generate an initial solution and executes an iterative update mechanism that deeply integrates operating condition perception information and real-time load characteristics. At the same time, it uses a composite fitness function that integrates operating condition perception and algorithm search status to evaluate each set of candidate parameters. Finally, within a limited number of iterations, it finds a set of candidate parameters with excellent comprehensive performance that is suitable for the current operating conditions.

[0097] After the optimization process is completed, the system does not directly output the optimal solution. Instead, it enters the equipment control and regulation strategy output stage. It performs a post-hoc decision based on multiple criteria on the elite solution set, selects the parameter set that best matches the current and near-future operating conditions and takes into account both performance and stability. After rigorous real-time safety correction and operating condition adaptive smoothing filtering, it generates the final control command vector that can be directly and safely issued to each actuator of the equipment. This command acts on the equipment, causing its operating regulation parameters to change accordingly.

[0098] As the equipment operates under new parameters, its state changes accordingly, generating new operating data. This initiates the data acquisition, condition perception, optimization decision-making, and command output process for the next control cycle. This cycle repeats continuously, forming a dynamic closed loop of "perception-decision-execution-re-perception." This allows the control system to continuously track the time-varying characteristics of the equipment's operating conditions and constantly adjust the control strategy adaptively. Ultimately, this enables intelligent, refined, adaptive balance and optimization of equipment operating efficiency, energy consumption, and safety in the ever-changing and complex mining environment.

[0099] This invention provides a method for optimizing equipment control and regulation based on adaptive perception of operating conditions. This method can improve the control system's ability to perceive and adapt to complex time-varying operating conditions of mining equipment, enhance the search efficiency and accuracy of optimization algorithms, achieve intelligent dynamic balance of multi-objective performance of equipment operation, and ensure high reliability, smoothness, and safety of control commands.

[0100] In summary, the methods provided in the embodiments of the present invention mainly provide the following: 1. By adopting a dynamic operating condition segmentation and perception feature mapping method, the operating conditions are automatically divided through unsupervised clustering, and a weight vector is constructed to quantify the similarity between the current state and each typical operating condition, thereby realizing enhanced representation and quantitative perception of complex operating modes.

[0101] 2. An improved gray wolf optimization algorithm with deep integration of working condition perception is adopted. Working condition characteristics, indicators and adjustment factors are embedded into the entire process of population initialization, leader wolf reorganization, position update and diversity perturbation of the algorithm, so as to realize the adaptation of the optimization strategy to the operating state.

[0102] 3. A composite fitness function that combines dynamic weighting and exploration rewards is adopted. The weights of the sub-objectives are dynamically adjusted according to the deviation of the working conditions. An exploration reward item based on population distribution is introduced to guide the algorithm to balance multi-objective optimization and exploration-development behavior under specific working conditions from the evaluation level.

[0103] 4. A two-stage, safe and robust output mechanism is adopted, which includes posterior decision-making based on multi-criteria dynamic weights and instruction generation with integrated rate limiting, coupling constraint correction and operating condition adaptive smoothing filtering, to ensure that the output instructions have superior performance, operating condition matching, smooth switching and absolute safety.

[0104] The method provided in the embodiments of the present invention has the following main advantages: 1. Significantly improves the control system's ability to perceive and adapt to complex time-varying operating conditions of mining equipment, accurately identifies operating status and quantifies its deviation from typical patterns, providing a reliable basis for intelligent decision-making.

[0105] 2. The search efficiency and accuracy of the optimization algorithm have been enhanced. The guided initialization and working condition adaptive update mechanism enable the algorithm to quickly focus on potential solution space regions and dynamically adjust the search behavior under different working conditions, thereby finding high-quality solutions faster and more accurately.

[0106] 3. It realizes intelligent dynamic balancing of multi-objective performance of equipment operation. Through dynamic weighting of fitness function, it can prioritize improving efficiency and energy efficiency under stable operating conditions, and automatically strengthen the protection of safety and stable operation under abnormal or transitional operating conditions.

[0107] 4. It ensures the high reliability, smoothness and security of control commands. The post-processing mechanism effectively prevents parameter mutations, constraint violations and command jitter, so that the optimization results can be safely and stably applied to actual equipment and reduce operational risks.

[0108] Example 2: Corresponding to the above method embodiments, this embodiment provides a device control and adjustment optimization system based on adaptive working condition sensing. See [link to relevant documentation]. Figure 7 The diagram shown illustrates the structure of a device control and regulation optimization system based on adaptive working condition sensing. This system includes: Data layer 71 is used to continuously collect operating data of the target mining equipment under various working conditions from multiple sensor arrays deployed on the target mining equipment; The feature extraction layer 72 is used to construct a working condition-aware feature mapping function using a dynamic working condition segmentation strategy, and to enhance the representation of the running data based on the working condition-aware feature mapping function. The working condition perception layer 73 is used to calculate the working condition perception index based on the working condition perception feature mapping function, quantify the deviation between the current operating state and each working condition through the working condition perception index, and dynamically generate the parameter adjustment factor of the gray wolf optimization algorithm based on the deviation. The optimization algorithm layer 74 is used to adopt a guided intelligent population initialization strategy based on the gray wolf optimization algorithm. It uses the feature information of historical running data to generate the initial population position in a region that is closer to the high-quality solution, and ensures that the initial solution satisfies the constraints while introducing randomness, so as to obtain the updated population position. The optimization algorithm layer 74 is also used to construct a composite fitness function that integrates multi-dimensional performance indicators, working condition penalties, and the search state feedback of the gray wolf optimization algorithm. The control output layer 75 is used to make posterior decisions based on the working condition matching degree by optimizing the frontier solution set obtained by the algorithm, and to select the parameter set that best suits the current and near-future predicted working conditions; based on the parameter set, and combined with the physical constraints of the equipment, the coupling relationship between subsystems and the control command of the previous moment, a control command sequence is generated. The control output layer 75 is also used to control the operating status of the target mining equipment based on the sequence of control commands.

[0109] This invention provides an equipment control and adjustment optimization system based on adaptive working condition perception, which can improve the control system's ability to perceive and adapt to complex time-varying working conditions of mining equipment, enhance the search efficiency and accuracy of optimization algorithms, realize intelligent dynamic balance of multi-objective performance of equipment operation, and ensure high reliability, smoothness and safety of control commands.

[0110] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the equipment control and regulation optimization system based on adaptive working condition perception described above can be referred to the corresponding process in the aforementioned embodiments of the equipment control and regulation optimization method based on adaptive working condition perception, and will not be repeated here.

[0111] Example 3: This invention also provides an electronic device for running the above-described device control and adjustment optimization method based on adaptive working condition sensing; see also Figure 8 The diagram shows the structure of an electronic device, which includes a memory 100 and a processor 101. The memory 100 stores one or more computer instructions, which are executed by the processor 101 to implement the above-mentioned device control and regulation optimization method based on adaptive sensing of operating conditions.

[0112] Furthermore, Figure 8 The electronic device shown also includes a bus 102 and a communication interface 103, with the processor 101, the communication interface 103 and the memory 100 connected via the bus 102.

[0113] The memory 100 may include high-speed random access memory (RAM) and may also include non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 103 (which can be wired or wireless), such as the Internet, wide area network, local area network, metropolitan area network, etc. The bus 102 may be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 8 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0114] Processor 101 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of processor 101 or by instructions in software form. Processor 101 can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly manifested as execution by a hardware decoding processor, or execution by a combination of hardware and software modules in the decoding processor. The software module can reside in a readily available storage medium in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory 100, and processor 101 reads information from memory 100 and, in conjunction with its hardware, completes the steps of the method described in the foregoing embodiments.

[0115] This invention also provides a computer-readable storage medium storing computer-executable instructions. When these computer-executable instructions are called and executed by a processor, they cause the processor to implement the above-described device control and adjustment optimization method based on adaptive working condition perception. For specific implementation details, please refer to the method embodiments, which will not be repeated here.

[0116] The computer program product of the equipment control and adjustment optimization method and system based on adaptive working condition perception provided in the embodiments of the present invention includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods in the preceding method embodiments. For specific implementation, please refer to the method embodiments, which will not be repeated here.

[0117] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system and / or device described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0118] Furthermore, in the description of the embodiments of the present invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in the present invention based on the specific circumstances.

[0119] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0120] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0121] Finally, it should be noted that the above embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for optimizing equipment control and regulation based on adaptive sensing of operating conditions, characterized in that, The method includes: Continuously collect operational data of the target mining equipment under various working conditions from multiple sensor arrays deployed on the target mining equipment; A dynamic operating condition segmentation strategy is adopted to construct an operating condition perception feature mapping function, and the operating data is enhanced based on the operating condition perception feature mapping function. The operating condition perception index is calculated based on the operating condition perception feature mapping function. The deviation between the current operating state and each operating condition is quantified by the operating condition perception index. The parameter adjustment factor of the Grey Wolf optimization algorithm is dynamically generated based on the deviation. Based on the gray wolf optimization algorithm, a guided intelligent population initialization strategy is adopted. The feature information of historical running data is used to generate the initial population position in a region that is closer to the high-quality solution. While introducing randomness, the initial solution is guaranteed to meet the constraints, and the updated population position is obtained. Construct a composite fitness function that integrates multi-dimensional performance indicators, working condition penalties, and the search state feedback of the gray wolf optimization algorithm; By performing posterior decision-making based on the working condition matching degree on the frontier solution set obtained by the algorithm optimization, the parameter set that best suits the current and near-future predicted working conditions is selected; based on the parameter set, and combined with the physical constraints of the equipment, the coupling relationship between subsystems and the control command of the previous moment, a control command sequence is generated. The operating status of the target mining equipment is controlled based on the control command sequence.

2. The method according to claim 1, characterized in that, The sensor array is installed in key parts of the target mining equipment, including: motor, bearing, hydraulic system, and machine body; The operational data covers multiple physical dimensions, including electrical, mechanical, and thermal aspects. The operational data includes motor current and voltage, key component temperature, body vibration acceleration, and working traction force.

3. The method according to claim 1, characterized in that, The steps of constructing a condition-aware feature mapping function using a dynamic operating condition segmentation strategy, and enhancing the representation of the operating data based on the condition-aware feature mapping function, include: Collect multi-sensor historical time-series data of the target mining equipment during operation, and standardize each feature of the historical time-series data; use an unsupervised clustering algorithm to cluster the historical time-series data, and dynamically divide the operating status of the target mining equipment into multiple working condition segments; Construct a working condition perception feature mapping function based on the results of the working condition segmentation; Based on the operating condition-aware feature mapping function, the operating data is mapped into a weight vector that is similar to each operating condition, and the weight vector is used as the enhanced feature representation.

4. The method according to claim 1, characterized in that, The steps of calculating the operating condition perception index based on the operating condition perception feature mapping function, quantifying the deviation between the current operating state and various operating conditions through the operating condition perception index, and dynamically generating the parameter adjustment factor of the Grey Wolf optimization algorithm based on the deviation include: Based on the cluster centers of each working condition segment and the weight vector obtained by the working condition perception feature mapping function, the working condition perception index of the current sample is calculated. Using the aforementioned operating condition perception indicators, the degree of deviation between the current operating state and each operating condition is quantified by weighted average distance. Using the aforementioned operating condition perception index, the parameter adjustment factor of the Grey Wolf optimization algorithm is dynamically calculated through a smooth mapping function.

5. The method according to claim 1, characterized in that, The method further includes: The effect of adaptive sensing of operating conditions is analyzed by combining scatter plots and kernel density plots; wherein, the horizontal and vertical axes of the scatter plot represent two dimensions of operating characteristics after standardization, and the kernel density plot is used to display the distribution of the operating condition sensing indicators.

6. The method according to claim 1, characterized in that, Based on the gray wolf optimization algorithm, a guided intelligent population initialization strategy is adopted. This strategy utilizes the characteristic information of historical running data to generate the initial population position in a region closer to the optimal solution. While introducing randomness, it ensures that the initial solution satisfies the constraints. The steps to obtain the updated population position include: A device operation adjustment parameter vector is defined for each working condition segment. Based on the similarity weight between the sample and each working condition, multiple device operation adjustment parameter vectors are weighted and fused. Random perturbations modulated by the working condition perception state are introduced to generate the initial population position. The initial population position is subjected to boundary constraint processing. Through element-by-element comparison and pruning, the parameter values ​​of multiple dimensions of the initial population position are restricted to a preset feasible domain to obtain the updated population position.

7. The method according to claim 1, characterized in that, The method further includes: By integrating operating condition perception information, the exploration and development behavior of the Gray Wolf optimization algorithm is dynamically adjusted through adaptive parameter adjustment factors, operating condition perception feature weights, and real-time load characteristics. This allows the Gray Wolf optimization algorithm to intelligently balance the following of the leader wolf, random exploration, and guidance toward historical high-quality solution regions based on the stability and typicality of the current equipment operating status.

8. The method according to claim 1, characterized in that, The steps for constructing a composite fitness function that integrates multi-dimensional performance indicators, operating condition penalties, and the search state feedback of the Gray Wolf optimization algorithm include: Multiple sub-objective components are defined based on the equipment physical model, operating data, and optimization parameters; wherein, the sub-objective components include: productivity component, energy consumption component, safety risk component, and operating condition deviation penalty component; Multiple sub-target components are aggregated into a scalar fitness value, and the weights are dynamically adjusted according to the current sample's working condition perception index. By introducing an exploration reward based on the current population distribution and individual location, the fitness is modified to obtain a composite fitness function.

9. The method according to claim 1, characterized in that, By performing posterior decision-making based on the working condition matching degree on the frontier solution set obtained by the algorithm optimization, the parameter set that best fits the current and near-future predicted working conditions is selected. Based on the parameter set, and in conjunction with the physical constraints of the equipment, the coupling relationships between subsystems, and the control commands from the previous moment, the step of generating a control command sequence includes: Based on the comprehensive performance criterion, the working condition matching degree criterion, and the output stationarity criterion, the frontier solution set obtained by the algorithm optimization is subjected to a posterior decision based on the working condition matching degree to select the parameter set that best suits the current and near-future predicted working conditions. Based on the parameter set, rate limiting correction and coupling constraint correction are performed to generate a control command sequence.

10. A device control and regulation optimization system based on adaptive working condition sensing, characterized in that, The system includes: The data layer is used to continuously collect operational data of the target mining equipment under various working conditions from multiple sensor arrays deployed on the target mining equipment; The feature extraction layer is used to construct a working condition-aware feature mapping function using a dynamic working condition segmentation strategy, and to enhance the representation of the operating data based on the working condition-aware feature mapping function. The working condition perception layer is used to calculate the working condition perception index based on the working condition perception feature mapping function, quantify the deviation between the current operating state and each working condition through the working condition perception index, and dynamically generate the parameter adjustment factor of the gray wolf optimization algorithm based on the deviation. The optimization algorithm layer is used to adopt a guided intelligent population initialization strategy based on the gray wolf optimization algorithm. It uses the feature information of historical running data to generate the initial population position in a region that is closer to the high-quality solution, and ensures that the initial solution meets the constraints while introducing randomness, so as to obtain the updated population position. The optimization algorithm layer is also used to construct a composite fitness function that integrates multi-dimensional performance indicators, working condition penalties, and the search state feedback of the gray wolf optimization algorithm. The control output layer is used to make posterior decisions based on the working condition matching degree by optimizing the frontier solution set obtained by the algorithm, and to select the parameter set that best suits the current and near-future predicted working conditions; based on the parameter set, and combined with the physical constraints of the equipment, the coupling relationship between subsystems and the control command of the previous moment, a control command sequence is generated. The control output layer is also used to control the operating status of the target mining equipment based on the control command sequence.

Images