A coking coal preparation full-process intelligent collaborative optimization and control system

Abstract

The present application relates to the technical field of coking coal preparation automation control, and discloses a coking coal preparation whole-process intelligent collaborative optimization and control system, which comprises a perception and execution equipment layer, an edge computing node and a cloud server. The edge computing node discretizes materials into micro-element to calculate absolute physical displacement, constructs a logistics mapping matrix, identifies a physical mixing node combined state parameter, gives a dynamic confidence degree, generates a fixed mechanism boundary, uses the dynamic confidence degree for contraction modulation, generates a final execution instruction within a constraint range, and calls an instruction to issue when a calculated expected remaining time meets a feedforward advance trigger time; the cloud server extracts a traceability feature slice to update an inherent diffusion resistance coefficient when a quality deviation exceeds a judgment threshold. The present application constructs a bottom-layer logistics mapping, quantifies a node mixing action, modulates equipment control boundaries, eliminates data distortion risks, makes up for physical control action time lag, and establishes a bottom-layer parameter closed-loop correction channel driven by terminal quality.

Description

Technical Field

[0001] This invention relates to the field of automated control technology for coking coal preparation, specifically to an intelligent collaborative optimization and control system for the entire coking coal preparation process. Background Technology

[0002] Coking coal preparation is a preliminary step in coke production. The logic involves mixing and crushing different types of coal according to process ratios. In existing coking coal preparation control processes, the system uses a macroscopic flow setting mode for regulation, failing to accurately track the continuously conveyed material on the belt conveyor at the underlying spatial level. As the continuously flowing coal is transported and passes through various levels of coal storage bins, it faces complex physical spatial transformations and slippage from the mechanical belts. Relying solely on theoretical kinematic positioning often results in significant tracking and positioning deviations.

[0003] When coal passes through the physical mixing nodes in the storage area, the internal physical mixing and tumbling effects interfere with the original physicochemical properties of the coal. Traditional control systems do not quantify the disturbance attenuation phenomenon caused by the physical mixing at these nodes on the material's own physicochemical properties. Even when the reliability of the underlying sensing data decreases, they still rely on fixed equipment operating constraints for optimization. This monotonous boundary constraint mechanism makes the optimization algorithm prone to outputting control commands that deviate from actual operating conditions under the guidance of distorted data, and may even trigger high-risk control actions.

[0004] Existing control systems rely on passive adjustments based on post-processing terminal quality detection results, failing to comprehensively calculate electrical network communication delays and mechanical transmission response delays. A significant physical time misalignment exists between the issuance of control commands and the actual arrival of materials at the target processing equipment, resulting in physical time lag in control intervention. Furthermore, existing optimized control models lack a closed-loop iterative mechanism for reverse correction of underlying key parameters based on end-product quality deviations. In long-term continuous production, calculation errors caused by changes in operating conditions and equipment wear accumulate, preventing the system from achieving long-term autonomous correction. This singular open-loop operation ultimately leads to large fluctuations in coal preparation quality, lag in control response, and a lack of adaptive adjustment capabilities across the entire system. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an intelligent collaborative optimization and control system for the entire coking coal preparation process. It solves the problems of traditional control processes, such as the lack of continuous material bottom-level spatial tracking and physical mixing interference quantification mechanisms, the susceptibility of fixed constraint models to being misled by distorted data and resulting in high-risk control actions, physical time lag in control intervention, and the lack of a closed-loop mechanism for reverse correction of bottom-level parameters based on terminal quality deviations, which lead to large fluctuations in the coal preparation quality of the entire system and a lack of adaptive adjustment capabilities.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] This invention provides an intelligent collaborative optimization and control system for the entire coking coal preparation process, comprising a sensing and execution device layer, edge computing nodes, and a cloud server. The edge computing nodes include:

[0008] The logistics mapping module is used to discretize continuously transported materials into data micro-elements and calculate the absolute physical displacement of the data micro-elements to construct a logistics mapping matrix.

[0009] The confidence calculation module is used to identify the physical mixed nodes through which the data micro-element passes based on the logistics mapping matrix, and to calculate and assign the dynamic confidence level to the data micro-element in combination with real-time status parameters.

[0010] The boundary decision module is used to generate a fixed mechanism boundary, use the dynamic confidence to shrink and modulate the fixed mechanism boundary, and generate the final execution instruction within the modulated boundary constraint range;

[0011] The feedforward control module is used to calculate the estimated remaining time for the data micro-element to reach the target processing equipment, and when it is determined that the feedforward advance trigger time is met, it calls the final execution instruction to be sent to the sensing execution device layer.

[0012] Preferably, the cloud server is equipped with a source tracing iteration module. When the quality deviation exceeds the judgment threshold, the source tracing iteration module reversely extracts the source tracing feature slices, updates the inherent diffusion resistance coefficient of the physical hybrid node, and sends it to the edge computing node.

[0013] Preferably, the logistics mapping module receives the instantaneous moisture detection signal, instantaneous ash detection signal, and instantaneous mass flow signal output by the sensing and execution device layer, performs a truncation operation on the continuously flowing material on the belt conveyor according to the sampling period time to generate independent data micro-elements, and calculates the average moisture measurement value, average ash measurement value, and absolute mass within the sampling period time based on the instantaneous moisture detection signal, the instantaneous ash detection signal, and the instantaneous mass flow signal.

[0014] Preferably, the logistics mapping module calls the Hardgrove Grindability Index as a hardness indicator based on the coal type code issued by the production management system, packages the average moisture measurement value, the average ash content measurement value, the absolute mass, the hardness indicator, and a preset absolute hardware clock, and assigns an initial attribute vector and a generation timestamp to the data micro-element.

[0015] Preferably, the logistics mapping module calculates the theoretical displacement of the data micro-element based on the operating speed collected by the sensing and execution device layer, compares the difference between the instantaneous stator current and the reference current, extracts the current difference exceeding the reference current, combines it with the empirical slippage coefficient, performs integration calculation in the time dimension to obtain the slippage error compensation amount, obtains the absolute physical displacement by subtracting the slippage error compensation amount from the theoretical displacement, and updates the logistics mapping matrix.

[0016] Preferably, the confidence calculation module compares the absolute physical displacement with the pre-entered equipment space topology coordinate dictionary to determine whether the material passes through the physical mixing node;

[0017] The sensing and execution device layer outputs the real-time material level height of the internal material. The confidence calculation module extracts the material level historical records within the observation time window to calculate the material level fluctuation variance. Based on the real-time material level height and the material level fluctuation variance, fluctuation influence weight coefficient and drop influence weight coefficient are introduced to construct a real-time mixing intensity characterizing the degree of mixing of materials inside the node.

[0018] Preferably, the confidence calculation module is used for:

[0019] An initial confidence benchmark value is assigned to the data micro-element generated in the early stage, and the inherent diffusion resistance coefficient corresponding to the physical mixing node is extracted based on the identified node index number;

[0020] When the absolute physical displacement is detected to have moved out of the spatial coordinate boundary of the physical mixing node, a multivariate exponential decay law is adopted to mathematically combine the real-time mixing intensity and the inherent diffusion resistance coefficient to form the exponential term of the decay function, and the dynamic confidence level is output.

[0021] Preferably, the boundary decision module is used for:

[0022] The highest output frequency of the inverter and the maximum allowable speed of the motor are read and set as fixed upper limits, and the lowest operating torque or speed is extracted and set as fixed lower limits. The fixed mechanism boundary is generated based on the physical limit parameters of the inverter and the motor and the production process constraints.

[0023] Read the instruction records under stable operation, calculate the arithmetic mean, and generate a safety baseline control value;

[0024] The dynamic confidence level is mapped onto the allowable range of the control action, and the fixed mechanism boundary is forced to shrink towards the safety benchmark control quantity, thereby generating a dynamic modulation boundary that follows the changes in the material state.

[0025] Preferably, the boundary decision module is used for:

[0026] The average moisture content, average ash content, absolute mass, and hardness indices associated with the data micro-elements are extracted and input into a predefined data-driven prediction algorithm to calculate and output the unconstrained optimal control quantity.

[0027] The dynamic modulation boundary is used as a hard constraint condition. The unconstrained optimal control quantity is projected into the limited solution space. When the unconstrained optimal control quantity exceeds the dynamic modulation boundary, it is truncated to the nearest boundary extremum to generate the final execution instruction.

[0028] Preferably, the feedforward control module is used for:

[0029] Extract the absolute installation coordinates of the center position of the feed inlet of the target processing equipment and convert them into linear unfolded coordinates along the running direction of the belt conveyor. Calculate the difference between the linear unfolded coordinates and the absolute physical displacement to obtain the remaining conveying distance.

[0030] The instantaneous linear velocity of the belt conveyor is read, and the estimated remaining time is calculated by dividing the remaining conveying distance by the instantaneous linear velocity.

[0031] The electrical network communication delay time and the mechanical transmission response time are extracted and summed on the time axis to generate the feedforward advance trigger time.

[0032] Preferably, the feedforward control module calls the final execution instruction to parse the target operating frequency of the main drive motor and the target feeding rate of the upstream associated feeder, and sends it down to the sensing and execution device layer;

[0033] After receiving multi-dimensional hardware control settings, the sensing and execution device layer adjusts the output state execution parameters smoothly according to the preset device frequency change rate, and continuously monitors the actual operating load of the main drive motor and the upstream associated feeder.

[0034] Preferably, the intelligent collaborative optimization and control system for the entire coking coal preparation process also includes a cloud server. The cloud server is equipped with a traceability iteration module. The traceability iteration module performs multi-dimensional distance calculations between the actual output material quality detection vector and the target setting vector to obtain the absolute deviation. When the absolute deviation exceeds the judgment threshold, a traceability trigger signal with an output timestamp is generated.

[0035] The output timestamp is used as the retrieval key to initiate a cross-table query to the edge computing node, and the associated data is extracted and spliced ​​into the source tracing feature slice;

[0036] A comprehensive residual energy function is constructed using the actual output material quality detection vector and the target setting vector. The gradient descent method is then used to calculate the partial derivative gradient of the inherent diffusion resistance coefficient corresponding to the physical mixing node to obtain the correction increment. The inherent diffusion resistance coefficient is then updated and sent to the edge computing node.

[0037] This invention provides an intelligent collaborative optimization and control system for the entire coking coal preparation process, which has the following beneficial effects:

[0038] This invention discretizes continuously conveyed materials to construct a logistics mapping matrix for physical spatial tracking, and quantifies the interference encountered by materials passing through physical mixing nodes, assigning dynamic confidence levels. The system uses dynamic confidence levels to contract and modulate fixed mechanism boundaries to generate final execution instructions. When the expected remaining time is determined to meet the feedforward early trigger time, the instructions are issued. When the output quality deviates from the judgment threshold, the traceability feature slices are extracted in reverse to update the inherent diffusion resistance coefficient. The collaborative operation mechanism of the entire system not only eliminates the control risks caused by the distortion of the underlying detection data and compensates for the time lag of physical actions, but also establishes a closed-loop correction channel for underlying parameters driven by the quality of the final product, thereby solving the problems of large fluctuations in coal preparation quality and lack of adaptive adjustment capability in traditional control.

[0039] This invention introduces dynamic compensation and follow-up constraint mechanisms in the underlying feature extraction and safety control stages. It uses the difference between the instantaneous stator current and the reference current, combined with an empirical slippage coefficient, to calculate the slippage error compensation amount, thus restoring the true absolute physical displacement of the material. Combined with a predefined data-driven prediction algorithm, it outputs an unconstrained optimal control quantity, which is then truncated within the boundary extreme values ​​modulated by dynamic confidence. The objective data processing and boundary decision-making methods can eliminate tracking and positioning offsets caused by mechanical slippage and ensure that the adjustment commands output by the optimization algorithm fall entirely within the equipment's safe allowable range, avoiding dangerous aggressive control actions caused by external interference.

[0040] This invention introduces a time compensation and reverse optimization architecture in the control issuance and feedback stages. By superimposing the electrical network communication delay time and the mechanical transmission response time, a feedforward advance trigger time is generated. The target operating frequency and target feeding rate are issued according to the preset equipment frequency change rate to perform a smooth transition. At the same time, a comprehensive residual energy function is constructed using the actual output material quality detection vector and the target setting vector. The gradient descent method is called to obtain the correction increment for the underlying attenuation characteristics. The iterative architecture of spatiotemporal compensation combined with gradient optimization eliminates the time misalignment caused by the inherent delay of software and hardware, ensuring the stable adjustment of production control parameters and enabling the long-term autonomous correction capability of the underlying node characteristic parameters, thus maintaining the optimization accuracy of the control system. Attached Figure Description

[0041] Figure 1This is a schematic diagram of the architecture of an intelligent collaborative optimization and control system for the entire coking coal preparation process according to the present invention.

[0042] Figure 2 This is a flowchart illustrating the overall process of intelligent collaborative optimization and control of the entire coking coal preparation process according to the present invention.

[0043] Figure 3 This is a flowchart of the continuous material discretization, initial data encapsulation, absolute displacement tracking, and slippage compensation of the present invention.

[0044] Figure 4 This is a flowchart of the physical mixing node feature extraction, mixing intensity modeling, and dynamic decay calculation of logistics diffusion confidence in this invention;

[0045] Figure 5 This is a flowchart illustrating the mechanism boundary generation, cross-dimensional dynamic modulation, and data-driven optimal search within the constrained solution space of this invention.

[0046] Figure 6 This is a flowchart illustrating the calculation of remaining conveying distance, determination of feedforward triggering mechanism, and pre-control execution of multivariable equipment operation compensation parameters in this invention.

[0047] Figure 7 This is a flowchart of the reverse timing logic retrieval, comprehensive residual calculation, and physical node diffusion parameter iteration triggered by quality deviation in this invention.

[0048] Figure 8 This is a projection trajectory diagram of the dynamic boundary modulation and optimal control quantity of the present invention;

[0049] Figure 9 This is a time-series comparison diagram of the operating load of the crushing unit of the present invention.

[0050] Among them, 100 is the perception and execution device layer; 200 is the edge computing node; 210 is the logistics mapping module; 220 is the confidence calculation module; 230 is the boundary decision module; 240 is the feedforward control module; 300 is the cloud server; and 310 is the traceability iteration module. Detailed Implementation

[0051] The technical solutions in the embodiments 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, and 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.

[0052] See Figure 1This invention provides an intelligent collaborative optimization and control system for the entire process of coking coal preparation. The system adopts a cloud-edge-device collaborative computing architecture, including a sensing and execution device layer 100, an edge computing node 200, and a cloud server 300.

[0053] The sensing and execution device layer 100 is configured in the industrial production site, encompassing belt scales, encoders, online microwave moisture meters, online ash analyzers, radar level gauges, as well as frequency converters and regulating valve actuators. Edge computing nodes 200 are deployed on the production workshop side, establishing a bidirectional data connection with the sensing and execution device layer 100 via industrial communication protocols. A cloud server 300 is deployed in a remote data center, connecting to the edge computing nodes 200 via a wide area network. The system includes a logistics mapping module 210, a confidence calculation module 220, a boundary decision module 230, and a feedforward control module 240 within the edge computing nodes 200. The system also includes a traceability iteration module 310 within the cloud server 300.

[0054] The logistics mapping module 210 communicates with the sensing and execution device layer 100 and receives the underlying detection data. It discretizes the continuously conveyed material into data micro-elements containing initial attribute vectors and generation timestamps. It calculates the absolute physical displacement of each data micro-element on the conveyor line by combining the belt running speed and current data, and then constructs a logistics mapping matrix in memory that reflects the real-time spatial coordinates of the material.

[0055] The confidence calculation module 220 communicates with the logistics mapping module 210 and the sensing and execution device layer 100, and is used to parse the spatial displacement data of data micro-elements to identify the physical mixing nodes through which each data micro-element passes. This module obtains the real-time status parameters of the mixing nodes transmitted by the sensing and execution device layer 100, calculates and assigns a dynamic confidence level to each data micro-element that decreases dynamically with the intensity of physical mixing.

[0056] The boundary decision module 230 is communicatively connected to the confidence calculation module 220 and the sensing and execution device layer 100. It is responsible for generating a fixed mechanism boundary for feedforward control actions based on the physical limit parameters of the controlled device and production process constraints. It uses the dynamic confidence output by the confidence calculation module 220 to dynamically shrink and modulate the fixed mechanism boundary. Within the modulated boundary constraint range, this module executes an optimization algorithm to calculate the unconstrained optimal control quantity, generates the final execution instruction, and waits for triggering.

[0057] The feedforward control module 240 is communicatively connected to the logistics mapping module 210, the boundary decision module 230, and the sensing and execution device layer 100. It primarily determines the feedforward triggering timing by continuously monitoring the micro-element displacement data in the logistics mapping matrix. Combining the estimated remaining time for the micro-element to reach the target processing equipment based on the conveyor belt's dynamic calculation data, when the estimated remaining time is determined to be less than or equal to the feedforward advance triggering time, the final execution instruction generated by the boundary decision module 230 is invoked and parsed into multi-dimensional hardware control setpoints. These setpoints are then sent to the sensing and execution device layer 100 to adjust the target equipment's operating state in advance.

[0058] The traceability iteration module 310 communicates with the edge computing node 200. When the actual output material quality detection vector deviates from the target set vector by more than a judgment threshold, it uses the output timestamp to perform a reverse retrieval to extract traceability feature slices of the corresponding batch data micro-elements. Based on this, a system-wide comprehensive residual energy function is constructed to update the inherent diffusion resistance coefficient of the physical mixing node. The updated parameter set is then distributed and overlaid on the underlying control logic at the edge.

[0059] See Figure 2 This invention provides an intelligent collaborative optimization and control method for the entire process of coking coal preparation, comprising the following steps:

[0060] S100, the sensing and execution device layer 100 collects the physical and chemical parameter detection values ​​and physical state measurement values ​​of the bulk material at the starting point of the conveying process. The logistics mapping module 210 receives the detection data, discretizes the continuously conveyed material into data micro-elements containing initial attribute vectors and generation timestamps, and calculates the absolute physical displacement of the data micro-elements in combination with the belt running status data to construct the logistics mapping matrix.

[0061] S200, the confidence calculation module 220 identifies the physical mixing nodes through which each data micro-element passes based on the spatial displacement data of the data micro-element, and calculates and assigns a dynamic confidence level to each data micro-element based on the real-time status parameters of the mixing nodes transmitted by the sensing and execution device layer 100, which decreases dynamically with the intensity of physical mixing.

[0062] S300, Boundary Decision Module 230 generates a fixed mechanism boundary based on the physical limit parameters of the controlled equipment and the constraints of the production process. It uses dynamic confidence to dynamically shrink and modulate the fixed mechanism boundary. Within the modulated boundary constraint range, it executes the optimization solution algorithm to calculate the unconstrained optimal control quantity and generates the final execution instruction.

[0063] S400, the feedforward control module 240 monitors the micro-element displacement data in the logistics mapping matrix, and calculates the estimated remaining time in conjunction with the dynamic calculation of the conveyor belt operation. When it is determined that the feedforward advance trigger time is met, the final execution instruction is parsed into multi-dimensional hardware control settings and sent to the sensing and execution device layer 100 the pre-control parameters for smooth transition in order to adjust the target equipment operation status in advance.

[0064] S500, when the actual output material quality detection vector deviates from the target set vector by more than the judgment threshold, the traceability iteration module 310 uses the output timestamp to perform reverse retrieval to extract the traceability feature slices of the corresponding batch data micro-element, calculates the system comprehensive residual energy function to update the inherent diffusion resistance coefficient of the physical mixing node, and sends down the updated parameter set to cover the edge end underlying control logic.

[0065] To enable those skilled in the art to better understand the technical essence of the present invention, the following will elaborate on the various technical details and specific implementation logic in the operation and execution of the above-mentioned system and method.

[0066] See Figure 3 In specific implementation, the method provided by this invention includes the following steps in step S100 regarding the discretization of continuous materials and initial data encapsulation:

[0067] S101, the sensing and execution device layer 100 synchronously collects the physicochemical parameter detection values ​​and physical state measurement values ​​of the continuously conveyed materials according to a unified sampling cycle.

[0068] The online microwave moisture meter, online ash meter, and belt scale configured in the sensing and execution device layer 100 have physically different output signal refresh frequencies due to their inherent hardware conversion times. The logistics mapping module 210 has a built-in hardware timer that triggers data reading operations on the underlying communication bus using a preset sampling period as a global time reference. The specific value of the sampling period is calculated and determined by the highest polling period of the underlying control bus and the rated linear speed of the belt conveyor, and is limited to a fixed range of 0.5s to 2.0s to prevent data communication blockage.

[0069] The logistics mapping module 210 synchronously latches the instantaneous moisture detection signal, instantaneous ash detection signal, and instantaneous mass flow rate signal output by the sensing execution device layer 100 at the end of each sampling cycle. For the underlying hardware measurement process of the online microwave moisture meter emitting microwaves to detect the dielectric constant of the material to calculate moisture, and the underlying signal acquisition mechanism of the online ash meter using the principle of X-ray absorption to determine the ash content of the material, those skilled in the art can refer to the working principles of existing industrial measuring instruments for configuration and implementation. Their internal mechanisms are well-known technologies in the field and will not be elaborated upon here.

[0070] S102, the logistics mapping module 210 performs a truncation operation on the continuously flowing material on the belt conveyor in the time dimension to generate a set of data micro-elements with fixed logical boundaries.

[0071] The bulk materials on the on-site belt conveyor exhibit a continuous distribution without a fixed shape. The material mapping module 210 logically segments the material passing through the physical cross-section of the online detection equipment according to a set sampling period. All bulk materials passing through the detection cross-section within a complete sampling period are defined by the system as an independent data element. Due to slight fluctuations in the linear speed of the belt conveyor and variations in the material feed rate during actual production, each generated data element... In physical space, this corresponds to a segment of solid material with an indefinite absolute length and continuously changing mass.

[0072] Through time-series-based segmentation calculations, the system transforms macroscopic continuous logistics entities into a sequence of independently addressable and logically traceable objects in the workshop server's memory. The system continuously generates new data micro-elements during equipment operation. This allows for the construction of a set of data micro-elements in memory that reflect the material distribution characteristics of conveyor belts in the real physical world. In this set, This represents the total number of data micro-elements generated during the continuous operation of the system, and is a positive integer.

[0073] S103, the logistics mapping module 210 is for newly generated data micro-elements Construct initial attribute vectors for multidimensional independent states to complete the digital spatial encapsulation of the physical parameters of the corresponding material segment.

[0074] For data micro-elements that have just been truncated The logistics mapping module 210 extracts the various physicochemical parameters latched in step S101 and calculates the cumulative physical quantity or average detection parameter within the corresponding time slice. The logistics mapping module 210 uses the instantaneous mass flow rate output by the belt scale to perform an integral operation over the entire sampling period, thereby obtaining the data micro-element. The absolute quality. Simultaneously, the logistics mapping module 210, based on the coal type code synchronously issued by the underlying production management system, calls the Hadgrove grindability index pre-stored in the local database as the data micro-element. The hardness index.

[0075] The logistics mapping module 210 structures and packages the parameters obtained from the above calculations with the system's absolute hardware clock to generate the data micro-element. initial attribute vector Initial attribute vector The mathematical expression structure is constructed as follows:

[0076] ;

[0077] in, Representing data micro-elements Average ash content measured within the corresponding time slice; Representing data micro-elements Average moisture content measured within the corresponding time slice; Representing data micro-elements Related hardness index; Representing data micro-elements The absolute mass; Representing data micro-elements Complete the truncation and generate the timestamp. as initial attribute vector A unique timing identifier used to support the tracking and reverse retrieval logic of the absolute displacement of materials in subsequent control processes; The index number represents the data element, and its value is greater than or equal to 1 and less than or equal to 1. Positive integers; The symbol for transpose of a matrix.

[0078] In specific execution, the method provided by this invention includes the following steps in step S100 regarding absolute displacement tracking and slippage compensation based on the state-space analytical algorithm:

[0079] S104, the sensing and execution device layer 100 collects the operating speed of the belt conveyor, and the logistics mapping module 210 calculates the theoretical displacement of the data micro-element based on the speed.

[0080] The encoder configured in the sensing and execution device layer 100 is installed on the driven roller shaft end of the belt conveyor to output pulse signals reflecting the roller's rotation angle in real time. After reading this pulse signal, the logistics mapping module 210 calculates the instantaneous linear velocity of the belt conveyor based on the preset physical circumference of the roller, and displays it as a data micro-element. Generation timestamp Using this as the starting point for integration, a time-domain integration operation is performed on the instantaneous linear velocity to obtain the data element. The theoretical conveying distance generated by the belt running.

[0081] For the data processing of encoder acquiring pulse signals and calculating the instantaneous linear velocity of the device, those skilled in the art can use conventional pulse frequency measurement and digital filtering algorithms. The conversion mechanism between rotational speed and linear velocity is a well-known technology in this field and will not be elaborated here.

[0082] S105, the sensing and execution device layer 100 monitors the operating load of the drive motor, and the logistics mapping module 210 calculates the slippage error compensation amount based on the load change characteristics.

[0083] During bulk material conveying in industrial settings, relative slippage can occur between the conveyor belt and the drive roller due to transient surges in material load or localized belt tension decay. This physical slippage causes the theoretical displacement calculated based on the driven roller to deviate from the actual movement coordinates of the material, accumulating into spatial positioning errors over long conveying distances. When slippage occurs, the resistance of the belt conveyor's transmission system increases, leading to an abnormally high output torque of the drive motor, manifested as a transient change in the motor's stator current.

[0084] To this end, the sensing and execution device layer 100 continuously collects instantaneous stator current data of the drive motor and transmits it to the logistics mapping module 210. Upon receiving the data, the logistics mapping module 210 compares the instantaneous stator current with a preset reference current. When the instantaneous stator current exceeds the reference current, the system extracts the current difference and multiplies it by an empirical slippage coefficient, then performs integration over time to calculate the slippage error compensation amount for the corresponding time interval.

[0085] S106, the logistics mapping module 210 obtains the absolute physical displacement of the data micro-element by subtracting the slippage error compensation amount from the theoretical displacement, and updates the logistics mapping matrix synchronously.

[0086] By data micro elements The theoretical displacement is calculated by differential calculation with the slippage error compensation amount obtained above, and the logistics mapping module 210 outputs the data micro-element. The absolute physical displacement on the current conveyor line. The specific state-space displacement update and slippage compensation calculation logic is expressed by the following mathematical formula:

[0087] ;

[0088] ;

[0089] in, Representing data micro-elements At the current observation time The absolute physical displacement; This represents the data micro-element determined in the aforementioned steps. The generated timestamp; Represented by data micro-element Generation timestamp As the lower limit, based on the current observation time The definite integral operator with an upper limit is used to calculate the cumulative physical quantity within the time interval; Indicates the sliding moment The differential symbol represents an infinitesimal increment in the time dimension; Indicates the sliding moment of the belt conveyor The instantaneous linear velocity function; Representing data micro-elements Up to the current observation time Cumulative slippage error compensation; Indicates the empirical slip coefficient; This represents the maximum value function, used to extract current differences greater than zero. When the instantaneous stator current is less than or equal to the reference current, this term takes the value of zero. Indicates the moment when the drive motor is sliding. The instantaneous stator current; This indicates the reference current.

[0090] In the parameter settings of the above formula, the reference current is... The specific values ​​are calibrated based on the average stator rated current of the belt conveyor when it is running at a constant speed under rated stable load; while the empirical slippage coefficient... This coefficient is used to characterize the physical mapping ratio between the current load deviation and the actual belt slip distance. The specific value range of this coefficient is set between 0.005 and 0.025. Those skilled in the art can assign values ​​to it during system implementation by combining the friction coefficient of the conveyor belt, the counterweight parameters of the tensioning device, and the wrap angle structure of the drive roller, through no-load and full-load trial runs. This yields the absolute physical displacement. Subsequently, the logistics mapping module 210 continuously updates the logistics mapping matrix in memory to ensure that the coordinates of the data micro-elements in the digital space always correspond to the actual positions of the materials on site.

[0091] In practical implementation, the logistics mapping matrix is ​​a two-dimensional numerical state table built in the workshop server's runtime memory, used to accurately bind discrete material entities in digital space to their physical locations. The rows of the logistics mapping matrix correspond to the sequence of data micro-elements currently in the transportation state, with each row representing an independent data micro-element. The columns of the logistics mapping matrix correspond to the real-time state dimensions of each data micro-element. The logistics mapping matrix at any given observation time... The mathematical expression is constructed as follows:

[0092] ;

[0093] in, The logistics mapping matrix represents the current observation time; This represents the total number of data elements currently located on the transport link; logistics mapping matrix. The first column represents the index number of the data element. Logistics mapping matrix The second column represents the generation timestamp of the data micro-element. Logistics mapping matrix The third column represents the absolute physical displacement calculated in real time based on the slippage compensation formula mentioned above. Logistics mapping matrix The fourth column represents the dynamic confidence level of the data element at the current observation time. Logistics mapping matrix Columns 5 through 8 represent the average ash content measurements of the solidified components in the initial attribute vector of the data element. Average moisture content Hardness index With absolute mass .

[0094] The dynamic update mapping mechanism between data micro-elements and the logistics mapping matrix includes synchronization operation logic in three dimensions:

[0095] The first dimension is matrix expansion. When the sensing and execution device layer 100 continuously collects and truncates new data micro-elements, the logistics mapping module automatically adds a new row at the bottom of the matrix, fills in the index number of the new data micro-element, the generation timestamp and the multi-dimensional initial attribute parameters, and initializes the absolute physical displacement corresponding to the new row to zero and the dynamic confidence to the baseline value of 1.0.

[0096] The second dimension is state refresh. In each underlying control cycle of the system, the logistics mapping module 210 calculates and refreshes the displacement value column of all rows in the matrix based on encoder pulses and motor current feedback; the confidence calculation module 220 refreshes the confidence value column of the corresponding row in the matrix based on the attributes of the physical mixed nodes passed through.

[0097] The third dimension is matrix reduction. When the absolute physical displacement value of a record in a row of the matrix is ​​greater than or equal to the absolute physical reference coordinates of the target processing equipment, the system determines that the corresponding physical material has completely left the conveyor link and entered the processing stage. At this time, the system immediately removes the current row of data from the logistics mapping matrix and persists the removed data to the edge historical database for subsequent reverse tracing of quality anomalies, thereby freeing up memory and maintaining the dynamic balance of the matrix dimensions.

[0098] See Figure 4 In specific implementation, the method provided by this invention includes the following steps in step S200 regarding the extraction of physical hybrid node features and modeling of hybrid intensity:

[0099] S201, the confidence calculation module 220 reads the data element displacement parameters in the logistics mapping matrix and identifies physical mixed nodes by combining them with the preset conveyor link spatial topology data table. The data table contains the three-dimensional mapping coordinates of the relative positions of all conversion equipment and conveyor belts on site.

[0100] As bulk materials move along the conveyor belt, they inevitably pass through various conversion devices. The confidence calculation module 220 continuously monitors the logistics mapping matrix, converting data elements... The absolute physical displacement is matched against the pre-entered equipment spatial topology coordinate dictionary in the system. When the absolute physical displacement enters the coordinate range of a specific device, the system determines that the micro-element has passed through the corresponding physical mixing node. In the coking coal preparation process, these nodes include the coal drop pipe of the transfer station, the material buffer bin below the coal blending disc, and the crushing chamber of the crusher. When materials pass through these physical mixing nodes, they are constrained by physical factors such as gravity falling, mechanical dispersal, and lateral cross-sectional contraction. The data micro-elements that were originally relatively independent on the conveyor belt will exchange spatial positions and mix states with the adjacent materials, resulting in a decrease in the accuracy of the initial attribute vector representation of the micro-element.

[0101] S202, the sensing and execution device layer 100 collects the working status parameters of the physical mixing node in real time, and the confidence calculation module 220 extracts the material level fluctuation characteristics based on the parameters.

[0102] To quantify the actual mixing state of materials within the nodes, the sensing and execution device layer 100 synchronously monitors each physical mixing node using instruments. Specifically, radar level gauges installed on the top of the silo or chute are responsible for outputting the real-time material level height inside. After receiving the raw measurement data, the confidence calculation module 220 extracts the historical material level data within a preset observation time window prior to the current observation time, and then calculates the material level fluctuation variance within that window. The specific length of this preset observation time window is set between 30s and 120s depending on the silo volume. The material level fluctuation variance directly reflects the instability of the bulk material accumulation state inside the container. The larger the variance value, the more violent the disordered movements such as local collapse and tumbling of the material within the container, and the more obvious the corresponding material mixing effect.

[0103] For the data acquisition process of radar level gauges emitting microwave pulses and receiving echoes to calculate the level height, those skilled in the art can use conventional electromagnetic wave time-of-flight ranging algorithms. The internal mechanism is a well-known technology in the field and will not be elaborated here.

[0104] S203, the confidence calculation module 220 uses the extracted state feature parameters to construct a real-time mathematical formula for calculating the mixing intensity, which characterizes the intensity of material mixing within the node.

[0105] Based on the acquired material level height and the calculated fluctuation variance, the confidence calculation module 220 targets the identified first... Each physical mixing node is used to calculate its current dynamic mixing index. This calculation process comprehensively considers the impact mixing caused by material drop and the dynamic backmixing effect caused by retention within the container. The specific real-time mixing intensity is calculated using the following mathematical formula:

[0106] ;

[0107] in, Indicates the first At the current observation time, the physical hybrid nodes are... Real-time mixing intensity; The index number of the physical hybrid node is a positive integer; This indicates the weighting coefficient for the impact of fluctuations; Indicates the first At the current observation time, the physical hybrid nodes are... The variance of material level fluctuation within the associated observation time window; This indicates the weighting coefficient for the impact of the drop; Indicates the first The maximum material level height is designed for each physical mixing node; Indicates the first At the current observation time, the physical hybrid nodes are... Real-time material level height.

[0108] In the parameter settings of the above formula, the fluctuation impact weighting coefficient is... and the weighting coefficient of the impact of falls The values ​​are all set between 0.1 and 0.9, and the system mandates that the sum of the two values ​​equals 1. Those skilled in the art can assign values ​​based on the specific structural form of various hybrid nodes during the system deployment phase. For coal blending bins that emphasize material buffering and storage, the fluctuation impact weighting coefficient is increased. The value of is chosen to highlight the impact of retention and backmixing on the material state; for transfer chutes with large drops, the weighting coefficient of the drop effect is increased accordingly. The value of is chosen to highlight the forced mixing effect caused by the impact of gravity drop. The confidence calculation module 220 uses this formula to continuously output the dynamic indicators of each node, providing a quantitative benchmark for subsequent data micro-attribute modulation.

[0109] In specific implementation, the dynamic decay calculation process of the logistics diffusion confidence level in step S200 of the method provided by this invention includes the following steps:

[0110] S204, the confidence calculation module 220 assigns an initial confidence benchmark value to the newly generated data micro-element, establishing the starting point for attenuation calculation.

[0111] When the data micro-element completes the reading of various physicochemical parameters and time-domain discretization at the physical section of the detection equipment, the corresponding physical material has not yet undergone any long-distance transportation or inter-equipment transfer process. At this time, the initial attribute vector within the digital space of the data micro-element can accurately map the actual physical distribution state of the material on site. Based on this high parameter matching, the confidence calculation module 220 calculates the data micro-element... The initial confidence baseline value is uniformly set to 1.0. This baseline value is stored in the system's running memory as the data element moves with the conveyor belt, serving as the basic base for subsequent calculations of various dynamic attenuation effects.

[0112] S205, the confidence calculation module 220 calls the inherent diffusion resistance coefficient of the physical hybrid node to quantify the structural resistance effect inside the equipment.

[0113] In the transport link of bulk materials, each physical mixing node, constrained by its specific internal mechanical structure, the inclination angle of the guide baffle, and the effective flow cross-sectional area, imposes varying degrees of structural restrictions on the spatial diffusion motion of the flowing material. This objectively existing physical obstacle effect is abstracted by the system and defined as the inherent diffusion resistance coefficient. Before performing specific attenuation calculations, the confidence calculation module 220 retrieves and extracts the inherent diffusion resistance coefficient corresponding to the identified node from the local database based on the node index number. The initial value range of the inherent diffusion resistance coefficient is set between 0.5 and 2.5. In the initial stage of system deployment, technicians statically assign values ​​to it based on the internal volume and guide structure dimensions of the mixing equipment. After the system enters closed-loop operation, this coefficient will be dynamically corrected based on the reverse feedback of the output quality data.

[0114] S206, the confidence calculation module 220 is based on the multivariate exponential decay algorithm to calculate the dynamic confidence of the data micro-element after being affected by the physical mixed nodes.

[0115] The confidence calculation module 220 combines the parameters collected in the above steps to detect data micro-elements. When the absolute physical displacement of a component moves beyond the spatial coordinate boundary of a physical hybrid node, the confidence decay logic for that micro-element is triggered. The system employs a multivariate exponential decay law, mathematically combining the real-time hybrid state within the node with inherent equipment resistance characteristics. This causes the state and characteristics to together constitute the exponential term of the decay function. The specific steps for executing the decay algorithm to obtain the index are as follows:

[0116] ;

[0117] in, Representing data micro-elements At the current observation time Dynamic confidence level; This represents the initial confidence baseline value for a data element; Represented by natural constant An exponential function with base 0; Representing data micro-elements up to the current observation time The set of all physically hybrid nodes that have already left; Indicates the first The inherent diffusion resistance coefficient of each physical mixing node; Representing data micro-elements Actual departure The crossing time of the boundary of a physical hybrid node; Indicates the first A physical hybrid node at the time of travel The real-time mixing intensity.

[0118] The confidence calculation module 220 continuously updates the parameters of the data element using the above formula. As the data element... Moving on the conveyor belt and continuously passing through newly added physical mixing nodes, the collection The number of elements gradually increases. This causes the absolute value of the exponent term in the formula to accumulate and increase, thus affecting the dynamic confidence level. It exhibits a smooth exponential downward trend. The calculated dynamic confidence level is synchronously written into the attribute vector of the micro-element, providing a direct quantitative basis for subsequent control steps to determine the reliability of the internal physicochemical parameters of the micro-element.

[0119] See Figure 5 In specific implementation, the method provided by this invention includes the following steps in step S300 regarding the generation of the mechanism boundary and the cross-dimensional dynamic modulation mechanism:

[0120] S301, Boundary Decision Module 230 extracts the physical limit parameters and production process constraints of the controlled equipment to generate a fixed mechanism boundary for feedforward control actions.

[0121] In industrial production environments, the output commands of the control system are limited by mechanical hardware conditions and process specifications. The boundary decision module 230 reads the highest output frequency of the underlying frequency converter and the maximum allowable speed of the motor, using these as the fixed upper limit of the control system's feedforward output. Simultaneously, the boundary decision module 230 calls the process control database to extract the minimum operating torque or speed required to maintain a basic production rhythm and prevent material accumulation and blockage in pipelines, setting these as the fixed lower limit. For obtaining the operating limit parameters of electromechanical equipment and setting process protection thresholds, those skilled in the art can directly configure them based on the equipment nameplates and basic operation manuals. The parameter mapping mechanism is a well-known technology in this field and will not be elaborated upon here.

[0122] S302, Boundary Decision Module 230 extracts the safety benchmark control quantity of the controlled equipment from the integrated historical operating data and establishes the conservative control contraction benchmark point.

[0123] During continuous conveying and processing, when the system faces extreme conditions where input data completely fails, the control loop requires a conservative instruction that balances equipment safety and basic output. The boundary decision module 230 reads instruction records from the production management system during a specific historical period of stable operation and calculates the arithmetic mean of this historical steady-state data to generate a safety baseline control quantity for the equipment. In the actual deployment phase of the system, this specific time period is set to 30 days. To simplify engineering implementation, this safety baseline control quantity can also be set empirically to a fixed constant between 60% and 75% of the equipment's rated design load. Physically, the safety baseline control quantity represents the most reliable output state for the system, maintaining operation solely based on the basic load without relying on any advanced material sensing data.

[0124] S303, the boundary decision module 230 introduces the dynamic confidence of the data micro-element into the control space and implements cross-dimensional dynamic contraction modulation on the fixed mechanism boundary.

[0125] The accuracy of feedforward control actions is highly dependent on the reliability of the input material information. The boundary decision module 230 reads the dynamic confidence level output by the confidence level calculation module 220 and directly maps this indicator, which reflects the risk of data distortion, to the allowable range of the control action. When the dynamic confidence level remains high, close to 1.0, it indicates that the attribute vector of the data element can truly reflect the physical material state, and the system allows the feedforward algorithm to make significant adjustments within a wide range close to the fixed mechanism boundary. As the dynamic confidence level decreases due to the material passing through multiple physical mixing nodes, the boundary decision module 230 forces the control boundary to contract towards the safe baseline control value, thereby limiting the absolute magnitude of the feedforward adjustment and preventing overly aggressive control responses caused by data deviations. This cross-dimensional boundary modulation logic is implemented through the following mathematical formula:

[0126] ;

[0127] ;

[0128] in, Indicates the data microelement At the current observation time The upper limit of the feedforward control quantity generated after dynamic modulation; The index number representing the data element; This indicates the safety baseline control parameters of the equipment; This represents the fixed upper limit within the fixed mechanism boundary; This represents the fixed lower limit within the fixed mechanism boundary; Representing data micro-elements At the current observation time Dynamic confidence level; Indicates the data microelement At the current observation time The lower limit of the feedforward control quantity generated after dynamic modulation.

[0129] Through the aforementioned computational logic, the boundary decision module 230 continuously calculates and outputs a dynamically modulated boundary that follows the real-time changes in the material state. The first layer of this dual-layer boundary mechanism not only ensures that all control commands will not exceed the physical safety limits of the equipment, but also effectively isolates the potential interference caused by material data distortion to the feedforward control system using a dynamic contraction mechanism.

[0130] In specific implementation, the method provided by this invention includes the following steps in step S300 regarding the data-driven optimal search process within the constrained solution space:

[0131] S304, the boundary decision module 230 extracts the multidimensional attribute features of the data micro-elements and uses a predefined data-driven prediction algorithm to calculate the unconstrained optimal control quantity.

[0132] As the data micro-element moves along the transmission link, the boundary decision module 230 reads the corresponding data micro-element. The initial attribute vector is generated. This attribute vector contains physicochemical parameters such as the material's average ash content, average moisture content, absolute mass, and hardness index. The boundary decision module 230 uses these parameters as input features and inputs them into the built-in machine learning regression algorithm. This regression algorithm is constructed during the system's offline phase and is essentially a multivariate linear regression mapping matrix or a multinomial fitting mapping matrix from multidimensional feature inputs to control command outputs. The mapping matrix is ​​based on a historical production database, and its offline training dataset selects material detection parameters and corresponding optimal equipment operating control parameters from the field equipment during the past 6 to 12 months when it was in a steady-state high-yield period. Through forward propagation calculation, the algorithm outputs an unconstrained optimal control quantity that matches the current material characteristics. This control quantity represents the optimal equipment driving parameters under ideal conditions without considering any external hardware limitations or data deviation risks.

[0133] For the internal optimization logic and matrix parameter update process of the data-driven prediction algorithm that uses historical data to construct a mapping matrix for prediction, those skilled in the art can conduct offline supervised learning training by combining the sample data collected on-site. Its internal feature mapping and weight optimization mechanism are well-known technologies in this field and will not be elaborated here.

[0134] S305, the boundary decision module 230 projects the unconstrained optimal control quantity into the constrained solution space formed by the double-layer boundary, and generates the final execution instruction.

[0135] After obtaining the unconstrained optimal control quantity, the system needs to prevent command jumps beyond the limits caused by the generalization error of the data-driven algorithm itself or unforeseen operating conditions. The boundary decision module 230 uses the dynamic modulation upper and lower limits generated in the aforementioned steps as hard constraints to forcibly project the unconstrained optimal control quantity into this limited solution space. Under this projection mechanism, if the optimal solution output by the algorithm falls within the dynamic boundary range, the system directly adopts the result; if the algorithm output exceeds the boundary, the system truncates it to the nearest boundary extremum. The specific control command generation logic is expressed by the following mathematical formula:

[0136] ;

[0137] in, Indicates the data microelement At the current observation time The final execution instructions generated; Indicates the current observation time; The index number representing the data element; This represents a function that takes the minimum value. This represents the function that takes the maximum value. Indicates the data microelement At the current observation time The upper limit of the feedforward control quantity generated after dynamic modulation; Indicates the data microelement At the current observation time The lower limit of the feedforward control quantity generated after dynamic modulation; Indicates data based on micro-elements The unconstrained optimal control quantity is obtained from the calculation of physicochemical properties.

[0138] S306, the boundary decision module 230 triggers the issuance of the final execution command at the time node based on the material displacement tracking status.

[0139] The final execution command does not immediately take effect on the field equipment. The boundary decision module 230 continuously monitors this data element. The absolute physical displacement in the logistics mapping matrix. When the value of the absolute physical displacement reaches the preset physical coordinate trigger range of the target processing equipment, it indicates that the physical material corresponding to the data element is about to enter the effective processing area of ​​the equipment. At this time, the boundary decision module 230 will execute the final instruction through the underlying industrial communication bus. The command is sent to the target frequency converter or servo drive in the sensing and execution device layer 100. Upon receiving the command, the execution hardware adjusts its output frequency or drive torque to achieve precise spatiotemporal matching between the feedforward control action and the actual material flow. The lead time of this physical coordinate triggering interval is determined based on the electrical response delay and mechanical transmission inertia of the equipment, and is set to correspond to the belt conveyor running distance between 1.5s and 3.0s before the material arrives at the inlet of the processing equipment.

[0140] See Figure 6 In specific execution, the method provided by this invention includes the following steps in step S400 regarding the calculation of the remaining conveying distance and the determination of the feedforward triggering mechanism:

[0141] S401, the feedforward control module 240 obtains the physical coordinate reference of the target processing equipment and calculates the remaining conveying distance of the data micro-element.

[0142] When implementing precise material tracking, the system needs to map the equipment layout in three-dimensional space into a one-dimensional linear mathematical coordinate system. The feedforward control module 240 reads the system spatial topology data table, retrieves the three-dimensional mapped coordinates recorded in the table, extracts the absolute installation coordinates of the center position of the target processing equipment's feed inlet, and converts them into linear unfolded coordinates along the belt conveyor's operating direction. These unfolded coordinates serve as fixed reference points on the physical material transport chain. The feedforward control module 240 continuously reads the data elements output by the logistics mapping module 210. The absolute physical displacement at the current observation moment is used to calculate the difference between the coordinates of the fixed reference point and the absolute physical displacement, thereby obtaining the real-time spatial span of the corresponding physical material from the target processing equipment, i.e. the remaining conveying distance.

[0143] S402, the feedforward control module 240 combines the dynamic operation of the conveyor belt to predict the time node when the data micro-element arrives at the target processing equipment.

[0144] After obtaining the remaining conveying distance in the spatial dimension, the system needs to convert it into a control reference quantity in the time dimension. The feedforward control module 240 reads the instantaneous linear velocity of the belt conveyor at the current observation moment. By dividing the remaining conveying distance by the instantaneous linear velocity, the feedforward control module 240 dynamically calculates the data micro-element. The estimated remaining time to reach the target processing equipment. This estimated remaining time decreases as the belt conveyor continues to operate and the material approaches the equipment. The calculation logic for this time dimension is expressed by the following mathematical formula:

[0145] ;

[0146] in, Representing data micro-elements At the current observation time Estimated remaining time to reach the target processing equipment; The index number representing the data element; This represents the absolute physical reference coordinates of the target processing equipment on the transport link; Representing data micro-elements At the current observation time The absolute physical displacement; This indicates the belt conveyor at the current observation time. The instantaneous linear velocity.

[0147] S403, the feedforward control module 240 extracts the response delay characteristics of the target processing equipment and determines the feedforward advance trigger time.

[0148] Heavy machinery in industrial settings cannot instantly reach the target operating state after receiving drive commands from the control system. From the initial data packet parsing in the underlying communication network and the output of variable frequency current from the inverter circuit, to the motor rotor overcoming its own rotational inertia to establish effective torque, this series of processes constitutes the inherent response dead zone of the equipment. The feedforward control module 240 calls the equipment parameter database, extracts the pre-set electrical network communication delay time and mechanical transmission response time, and sums them on the time axis to generate the feedforward advance trigger time for the target processing equipment.

[0149] For determining the delay characteristics of inverter communication parsing and motor torque establishment processes, those skilled in the art can refer to the factory test report of the relevant electrical control equipment, or use the step response method for on-site operational calibration. The internal electrical hysteresis mechanism is a well-known technology in this field and will not be elaborated upon here. The feedforward advance trigger time reflects the absolute time required for the equipment to take effect from receiving the signal. Its value ranges from 1.0s to 5.0s, depending on the on-site motor power and reducer speed ratio.

[0150] S404, the feedforward control module 240 executes the time-domain comparison logic, determines that the triggering condition is met, and activates the instruction issuance mechanism.

[0151] To ensure a high degree of synchronization between equipment movement adjustments and the timing of material entering the processing chamber, the feedforward control module 240 employs a cyclic monitoring and comparison mechanism. The system compares the real-time updated estimated remaining time with the previously determined feedforward advance trigger time. When the estimated remaining time is greater than the feedforward advance trigger time, the system maintains the command interception and suspension state. When the estimated remaining time is less than or equal to the feedforward advance trigger time, the feedforward control module 240 determines that the feedforward trigger condition is met. At this point, the system immediately invokes the final execution command generated for this data element in the boundary decision module 230 and sends it to the underlying driver corresponding to the target processing equipment via industrial Ethernet. This time-prediction-based backward triggering mechanism effectively compensates for the physical response dead zone of the electromechanical equipment and eliminates the phase deviation between control actions and material flow.

[0152] In specific implementation, the method provided by this invention includes the following steps in step S400 regarding the pre-control execution of multivariable equipment operation compensation parameters:

[0153] S405, after receiving the trigger signal, the feedforward control module 240 parses the final execution instruction generated for the data micro-element into multi-dimensional hardware control settings.

[0154] The bulk material processing equipment on site comprises multiple coordinating physical actuators. Taking the crushing unit in the process as an example, the control quantities in the final execution command need to be specifically analyzed to determine the target operating frequency of the main drive motor and the target feeding rate of the upstream feeder. The feedforward control module 240 performs differential parameter decomposition based on the physicochemical properties encapsulated within the data micro-element. When the data micro-element... When the attribute vector indicates that the moisture content and hardness of the current material slice are high, the instruction decomposition logic will simultaneously increase the target operating frequency of the main motor and appropriately reduce the feeding rate. Through this multi-variable joint adjustment mechanism, the system can effectively prevent the adhesion and blockage problems caused by highly viscous and hard materials in the crushing chamber at the physical level.

[0155] S406, the sensing and execution device layer 100 receives the above-mentioned multi-dimensional setting values ​​and calls the ramp control function of the underlying driver to perform a smooth parameter transition.

[0156] Sudden changes in operating parameters of heavy-duty industrial equipment can easily trigger grid excitation shocks and torque fatigue in mechanical structures. Upon receiving a new target operating frequency from the feedforward control module 240, the underlying frequency converter does not immediately respond in a step-like manner, but rather gradually adjusts its output state according to its internal program settings. The specific process of adjusting the actual operating frequency of the equipment is expressed by the following mathematical formula:

[0157] ;

[0158] in, This indicates the main motor of the target processing equipment at the current observation time. The actual operating frequency; Indicates the system at the absolute trigger moment Record the initial operating frequency of the main motor; This indicates the absolute triggering moment at which the feedforward triggering condition is determined to be met in the aforementioned steps; Indicates the absolute trigger time As the lower limit, based on the current observation time The definite integral operator with an upper limit; This indicates the preset rate of change of the device frequency; The sign function is used to determine the direction of frequency adjustment. When the input parameter is greater than zero, it outputs a positive 1; when the input parameter is less than zero, it outputs a negative 1; and when it is equal to zero, it outputs zero. This indicates the target operating frequency of the main motor, as parsed from the final executed instruction; Indicates the sliding moment The differential symbol represents an infinitesimal increment in the time dimension. It should be noted that the above integration operation continues until the actual operating frequency is reached. Achieve the target operating frequency of the main motor The frequency should be stopped immediately and kept constant to prevent the equipment from exceeding the operating limit.

[0159] For the parameter values ​​in the above formula, the preset equipment frequency change rate is... The specific value is limited by the rated operating capacity of the frequency converter and the rotational inertia of the motor rotor. To ensure that the stator current fluctuation during acceleration or deceleration does not exceed the safe allowable range of the equipment, the value of this frequency change rate is set between 0.5Hz / s and 2.0Hz / s. Those skilled in the art can precisely optimize this value by observing the load start-stop curve during the system commissioning phase.

[0160] S407, the feedforward control module 240 continuously monitors the actual operating load of the equipment after the pre-control parameters are executed, and completes the execution closed loop of a single feedforward action.

[0161] When data micro elements When the corresponding physical material arrives and falls into the target processing equipment at the expected time, the physical equipment has already adjusted its operating parameters to the optimal processing state matching the properties of the batch of material, relying on the aforementioned feedforward triggering and smooth transition mechanism. During the processing of the physical material, the sensing and execution device layer 100 collects the output torque of the main motor and the feedback values ​​of the three-phase operating current in real time and uploads them to the feedforward control module 240. The system confirms the physical execution effect of this multivariate compensation action by comparing the actual operating load with the safe operating threshold. This pre-control action based on spatiotemporal tracking effectively eliminates the system lag caused by waiting for load fluctuations caused by material in the traditional control loop before feedback adjustment, and realizes the pre-alignment of the equipment's mechanical work capacity with the dynamic properties of the material.

[0162] For the underlying electrical execution mechanism of pulse width modulation signal generation and motor stator magnetic field rotation frequency control inside the frequency converter, those skilled in the art can refer to the variable frequency speed regulation theory of standard AC asynchronous motor for conventional configuration and implementation. Its internal microscopic electromagnetic mechanism is a well-known technology in this field and will not be elaborated here.

[0163] See Figure 7 In specific execution, the method provided by this invention includes the following steps in step S500 regarding the reverse timing logic retrieval process for quality deviation triggering:

[0164] S501, cloud server 300 sets quality deviation trigger judgment threshold and implements operation status monitoring.

[0165] The traceability iteration module 310, internally configured in the cloud server 300, continuously receives the actual output material quality detection vector uploaded by the end-product quality inspection device. The traceability iteration module 310 performs multi-dimensional distance calculations between this actual output quality detection vector and a preset target vector to obtain the absolute deviation between the two in the quality feature space. When this absolute deviation exceeds a set trigger threshold, the system determines that the currently produced end-product has experienced a substantial quality deviation and generates a traceability trigger signal with an output timestamp.

[0166] For the physicochemical analysis process of sampling physical materials and generating multidimensional quality detection vectors by end-of-line quality testing equipment, those skilled in the art can achieve this using conventional online near-infrared spectroscopy scanning or automated mechanical sampling and testing mechanisms. The internal component analysis mechanism is well-known in the field and will not be elaborated upon here. Regarding the specific value of the trigger threshold, those skilled in the art will set it based on the process tolerance of the specific industrial product before the system is put into operation. In common application scenarios such as coking coal blending or ore mixing, this threshold is set between 2% and 5% of the target set vector magnitude.

[0167] S502, the source tracing iteration module 310 performs cross-table reverse retrieval based on the output timestamp.

[0168] Because there is a mapping relationship between the flow of materials in the physical corridor and the micro-element records in the digital space, the quality anomalies detected at the end originate from early control input deviations at the front-end processing nodes. After the trigger signal is generated, the traceability iteration module 310 extracts the output timestamp carried by the signal and uses it as the primary retrieval key to initiate a cross-table query request to the edge historical database. Based on the output timestamp and the preset operating delay parameters of each level of conveying equipment, the traceability iteration module 310 performs backtracking and deduction to locate the target batch data micro-element that caused this end-of-line quality deviation.

[0169] S503, the traceability iteration module 310 extracts the full life cycle feature data of micro-elements to construct a data chain.

[0170] After identifying the target data micro-element, the tracing and iteration module 310 deeply extracts multi-dimensional historical operation records associated with that micro-element from the historical database at the edge. The extracted content is divided into three levels: the system reads the initial attribute vector assigned to the target batch data micro-element at the initial stage of entering the transport link; the system retrieves the historical dynamic confidence decay data recorded when the micro-element passes through various physical mixing nodes according to the time series; and the system extracts the historical control curves issued and executed by the target processing equipment within the corresponding processing time period. Through the above reverse cross-table extraction mechanism, the tracing and iteration module 310 stitches together the data fragments originally scattered across different time nodes and control network segments into a complete tracing feature slice, providing complete information support for subsequent closed-loop parameter optimization.

[0171] In specific implementation, the method provided by this invention includes the following steps in step S500 regarding the calculation of the comprehensive residual and the iteration of the physical node diffusion parameters:

[0172] S504, the traceability iteration module 310 constructs the system comprehensive residual energy function based on the actual quality deviation.

[0173] After acquiring the traceability feature slices, the cloud server 300 needs to transform the material quality deviation phenomena on the physical production line into a numerical objective function that can be optimized by the optimization algorithm. The traceability iteration module 310 uses the extracted actual output quality detection vector and the target setting vector under ideal conditions to construct a quadratic form of the comprehensive residual energy function. This function is used to quantify the overall divergence between the actual processing results and the expected process theory. The specific comprehensive residual energy function is expressed by the following mathematical formula:

[0174] ;

[0175] in, This represents the comprehensive residual energy value calculated by the system for the target batch of materials; This represents the total number of quality feature dimensions contained in the quality detection vector, and is a positive integer. Indicates the index number of the quality feature dimension; The summation operator represents the summation of the characteristic deviations across all quality dimensions. This represents the first element in the actual output quality detection vector. Physical measurement values ​​for each dimension of quality characteristics; Indicating the first term in the target setting vector The expected values ​​of the process for each quality characteristic dimension.

[0176] S505, the source tracing iteration module 310 uses the gradient descent algorithm to calculate the correction increment of the inherent diffusion resistance coefficient.

[0177] Overall residual energy value The abnormally high value reflects a static calibration deviation in the inherent diffusion resistance coefficient of the physical mixing nodes set within the underlying control logic at the edge. This deviation causes a disconnect between the confidence decay trajectory of the system's output and the actual physical mixing state of the material during transport. The traceability iteration module 310 uses minimizing this comprehensive residual energy value as the global optimization objective, and calls the gradient descent method to calculate the partial derivative gradient of the inherent diffusion resistance coefficient of all nodes involved in the traceability link. Since the system's comprehensive residual energy value depends on the actual quality inspection results of the physical production line, the partial derivative gradient cannot be directly derived analytically. In practice, the system uses the finite difference method combined with historical operating data for numerical approximation. The traceability iteration module 310 extracts historical operating data from the historical database of adjacent batches with similar initial physicochemical properties to the target batch of material, but with slight perturbation differences in the inherent diffusion resistance coefficient setting. By calculating the ratio of the difference in comprehensive residual energy value between adjacent batches to the perturbation of the inherent diffusion resistance coefficient, the ratio is used as the current partial derivative gradient vector. The specific logic for iteratively updating the inherent diffusion resistance coefficient parameter is implemented through the following mathematical formula:

[0178] ;

[0179] in, This indicates the number generated after this cloud-based optimization calculation. The updated intrinsic diffusion resistance coefficient of each physical hybrid node; This indicates that the edge confidence calculation module 220 is currently resident and in use as the first... The historical inherent diffusion resistance coefficient of each physical hybrid node; This indicates the preset iterative learning rate; Indicates the overall residual energy value for the first... The partial derivative gradient vector calculated from the intrinsic diffusion resistance coefficient of each physical mixing node.

[0180] Regarding the iterative learning rate in the above formula The value of this step size determines the range of the single step size for the system parameters to self-correct. To prevent excessively large update step sizes from causing repeated overshooting or even system divergence in the feedforward control loop, the value of this iterative learning rate is constrained to be between 0.001 and 0.01. Those skilled in the art can match its basic values ​​through playback tests of historical samples in the offline phase, and the system can also use a momentum mechanism to adaptively scale and adjust it during operation.

[0181] S506, the cloud server 300 sends updated parameter files and overwrites the underlying control logic of the edge terminal.

[0182] After completing multi-dimensional gradient calculations, the source tracing iteration module 310 formats and encapsulates all the obtained updated inherent diffusion resistance coefficients to generate a standard configuration file. The cloud server 300 distributes this configuration file to edge computing devices in various areas of the field via industrial Ethernet. After receiving and verifying the integrity of the file, the edge computing devices erase the historical parameters running in memory within the next control heartbeat cycle of the system and synchronously replace them with the latest parameters. This cloud-edge collaborative configuration mechanism enables the underlying control strategy to autonomously adapt to long-term drift phenomena in operating conditions caused by factors such as wear of the conveyor chute and seasonal changes in material humidity, constructing an iterative closed loop where the entire process-driven data and physical mechanical entities continuously evolve synchronously.

[0183] To help those skilled in the art better understand the logic of the present invention, a specific application example based on a real-world scenario in a coking plant's coal blending workshop is provided below.

[0184] The rated linear speed of the belt conveyor in the coal blending workshop of a coking plant is 2.0 m / s. The target processing equipment is a crushing unit, and the physical distance between its feed inlet and the detection section of the current microwave moisture analyzer and ash analyzer is 100 m. The system's sampling period is set to 1.0 s. At any given moment, the sensing and execution device layer 100 truncates and generates data micro-elements. The average moisture content of this batch of coal was measured. Average ash content measurement value absolute quality The database retrieves its hardness index. (This belongs to a high-moisture, hard, and difficult-to-grind coal type). The dynamic confidence level at this point is... .

[0185] As the belt moves, data micro-element When the physical displacement reaches 40m, it passes through a coal blending buffer silo (physical mixing node). At this time, the radar level gauge measured the maximum designed material level height of the silo. Real-time material level height Variance of material level fluctuation over the past 60 seconds Fluctuations affect weighting coefficients The impact of the drop on the weighting coefficient Substitute into the formula to calculate the mixing intensity: Retrieve the inherent diffusion resistance coefficient of the container. Substitute the values ​​into the multivariate exponential decay formula to calculate the dynamic confidence level: Due to the drop and remixing within the storage chamber, the representativeness of the initial properties of this infinitesimal element has decreased to approximately 54.88%.

[0186] The fixed mechanism boundary of the main motor inverter of the crusher is set as: the fixed upper limit in the fixed mechanism boundary. Fixed lower limit in fixed mechanism boundary Safety baseline control quantity Dynamic boundary shrinkage is calculated based on the decayed confidence level, and the upper bound shrinkage is: The lower limit contraction is: Meanwhile, the data-driven prediction algorithm receives the characteristics of high humidity (10%) and high hardness (65) and outputs the unconstrained optimal control quantity. (The system determines that a higher rotational speed is needed for pulverization). Because... Greater than the upper limit of feedforward control quantity This triggers the truncation mechanism. Ultimately, the executed instructions are confined to spatial projection within a safe boundary: .

[0187] When the belt travels to a distance of 4m from the micro-element crusher, the absolute residual displacement Current instantaneous linear velocity The estimated remaining time is calculated as follows: It is known that the feedforward advance trigger time for communication and torque establishment in the crusher is exactly 2.0s. At this time, the following conditions are met: The system immediately issued a 45.488Hz control command, and the frequency converter... The slope begins to accelerate smoothly. When the material actually falls into the crushing chamber after 2 seconds, the motor rises smoothly and stabilizes at the optimal speed, avoiding material accumulation and equipment jamming caused by a sudden increase in material.

[0188] To verify the effectiveness of this invention, a 7-day operational test was conducted at the aforementioned site, and the changes in key system parameters and the electrical response characteristics of the controlled equipment were recorded. The specific analysis is as follows:

[0189] See Figure 8The figure illustrates the variation of the control frequency (vertical axis) under constraints as the data element sequentially passes through different logistics nodes (horizontal axis 0 to 3) in the above embodiment. The dashed lines in the figure represent the safety baseline control quantity. It maintains a constant 40Hz level throughout the entire process, serving as a reference for conservative control of the system. The solid line with a circle represents the unconstrained optimal control quantity. Since the physicochemical properties of the data micro-elements are determined in the early stage of generation, the optimal crushing frequency output by the machine learning regression algorithm remains at 48.5Hz.

[0190] As the node number increases, the degree of physical mixing experienced by the micro-element deepens, and the dynamic confidence continuously decays. Therefore, this represents the upper limit of the feedforward control quantity. The dashed line gradually shrinks downwards from the initial 50Hz, representing the lower limit of the feedforward control. The dotted line gradually shrinks upwards from the initial 30Hz, with both moving closer to the safety reference.

[0191] When the position is between node 0 and node 1, the dynamic confidence level is relatively high, and the upper limit of the feedforward control quantity is [not specified]. Above 48.5Hz, unconstrained optimal control quantity Located within the allowable boundaries. This indicates the final execution instruction. solid lines with asterisks and When the curves coincide, the system directly outputs the optimal value calculated by the machine learning regression algorithm.

[0192] When in the stage from node 1 to node 3, because As the confidence level drops below 48.5Hz, the system triggers a truncation mechanism. The final instruction is then executed. Limited to the upper limit, marked with an asterisk The curve then bends downwards, and the upper limit of the feedforward control quantity is reached. The overlap indicates that the dynamic contraction mechanism restricts control commands that exceed the boundary.

[0193] See Figure 9 This figure compares the stator current response of the crusher motor when faced with the same batch of coal using two schemes: traditional load feedback control (PID, dashed line) and the feedforward pre-control of this invention (solid line). The vertical dashed lines in the figure represent... This refers to the moment when physical materials enter the equipment.

[0194] During the 0s to 8s phase, the material has not yet arrived at the processing equipment, and the equipment is in an unloaded state. At this time, the traditional control dotted line and the invention's control solid line coincide, both maintaining a basic operating current level of 150A.

[0195] From 8s to 10s, the two curves diverged. The traditional control scheme, employing a simple feedback mechanism, kept the dashed line at 150A. However, the control scheme of this invention calculated at 8s that there were still 2 seconds until the material arrived, satisfying the feedforward early trigger condition. After the system issued the command, the solid line sloped upwards, smoothly increasing the current to approximately 185A before 10s, thus increasing the motor's output torque ahead of schedule.

[0196] At 10 seconds, after the material entered the equipment, the dotted line of the traditional control system began to fluctuate, and the current jumped from 150A to nearly 220A. This was because the motor's load suddenly increased under its basic operating conditions, and subsequently, due to the lag in PID regulation, the current curve exhibited a decaying oscillation state, only stabilizing after more than ten seconds.

[0197] In contrast, the actual line of this invention showed only a current increase of about 5A (maximum about 190A) at 10s because the equipment had already adjusted its operating parameters in advance, and then quickly recovered and stabilized at the operating level of 185A. Test data shows that the feedforward pre-control mechanism reduces the current surge caused by sudden load changes and reduces the fluctuation amplitude during equipment adjustment.

[0198] Furthermore, during the test run, the end-point analysis detected quality deviations in some products, with ash content fluctuations exceeding the set threshold. The traceability iteration module 310 used timestamps to extract historical records and calculated that when the material level was low, the actual mixing degree of the material was greater than the predicted value of the original edge-end underlying control logic.

[0199] Subsequently, the system uses the comprehensive residual energy function to optimize parameters, thereby determining the inherent diffusion resistance coefficient of the corresponding buffer chamber. The version was revised from 1.2 to 1.65. The updated parameters were distributed to the edge, reducing the system's boundary contraction adjustment range under similar operating conditions and decreasing control command deviation. After the system parameters were updated, the quality pass rate of the produced materials increased from 88.5% in the initial testing phase to 97.2%.

[0200] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A smart collaborative optimization and control system for the entire coking coal preparation process, characterized in that, It includes a sensing and execution device layer (100), an edge computing node (200), and a cloud server (300), wherein the edge computing node (200) includes: The logistics mapping module (210) is used to discretize the continuously conveyed materials into data micro-elements and calculate the absolute physical displacement of the data micro-elements to construct a logistics mapping matrix. The confidence calculation module (220) is used to identify the physical mixed nodes through which the data micro-element passes based on the logistics mapping matrix, and to calculate and assign the dynamic confidence level corresponding to the data micro-element in combination with real-time status parameters. The boundary decision module (230) is used to generate a fixed mechanism boundary, use the dynamic confidence to shrink and modulate the fixed mechanism boundary, and generate the final execution instruction within the modulated boundary constraint range; The feedforward control module (240) is used to calculate the estimated remaining time for the data micro-element to reach the target processing equipment, and when it is determined that the feedforward advance trigger time is met, it calls the final execution instruction to be sent to the sensing execution device layer (100).

2. The intelligent collaborative optimization and control system for the entire coking coal preparation process according to claim 1, characterized in that, The logistics mapping module (210) receives the instantaneous moisture detection signal, instantaneous ash detection signal and instantaneous mass flow signal output by the sensing and execution device layer (100), performs a truncation operation on the continuously flowing material on the belt conveyor according to the sampling period time to generate independent data micro-elements, and calculates the average moisture measurement value, average ash measurement value and absolute mass within the sampling period time based on the instantaneous moisture detection signal, the instantaneous ash detection signal and the instantaneous mass flow signal.

3. The intelligent collaborative optimization and control system for the entire coking coal preparation process according to claim 2, characterized in that, The logistics mapping module (210) calls the Hardgrove Grindability Index as a hardness index according to the coal type code issued by the production management system, packages the average moisture measurement value, the average ash content measurement value, the absolute mass, the hardness index and the preset absolute hardware clock, and assigns the data micro-element an initial attribute vector and generates a timestamp.

4. The intelligent collaborative optimization and control system for the entire coking coal preparation process according to claim 1, characterized in that, The logistics mapping module (210) calculates the theoretical displacement of the data micro-element based on the operating speed collected by the sensing and execution device layer (100), compares the difference between the instantaneous stator current and the reference current, extracts the current difference exceeding the reference current, combines it with the empirical slippage coefficient, performs integral calculation in the time dimension to obtain the slippage error compensation amount, obtains the absolute physical displacement by subtracting the slippage error compensation amount from the theoretical displacement, and updates the logistics mapping matrix.

5. The intelligent collaborative optimization and control system for the entire coking coal preparation process according to claim 1, characterized in that, The confidence calculation module (220) compares the absolute physical displacement with the pre-entered equipment space topology coordinate dictionary to determine whether the material passes through the physical mixing node. The sensing and execution device layer (100) outputs the real-time material level height of the internal material. The confidence calculation module (220) extracts the material level history records within the observation time window to calculate the material level fluctuation variance. Based on the real-time material level height and the material level fluctuation variance, the fluctuation influence weight coefficient and the drop influence weight coefficient are introduced to construct a real-time mixing intensity that characterizes the intensity of material mixing within the node.

6. The intelligent collaborative optimization and control system for the entire coking coal preparation process according to claim 5, characterized in that, The confidence calculation module (220) is used for: An initial confidence benchmark value is assigned to the data micro-element generated in the early stage, and the inherent diffusion resistance coefficient corresponding to the physical mixing node is extracted based on the identified node index number; When the absolute physical displacement is detected to have moved out of the spatial coordinate boundary of the physical mixing node, a multivariate exponential decay law is adopted to mathematically combine the real-time mixing intensity and the inherent diffusion resistance coefficient to form the exponential term of the decay function, and the dynamic confidence level is output.

7. The intelligent collaborative optimization and control system for the entire coking coal preparation process according to claim 2, characterized in that, The boundary decision module (230) is used for: The highest output frequency of the inverter and the maximum allowable speed of the motor are read and set as fixed upper limits, and the lowest operating torque or speed is extracted and set as fixed lower limits. The fixed mechanism boundary is generated based on the physical limit parameters of the inverter and the motor and the production process constraints. Read the instruction records under stable operation, calculate the arithmetic mean, and generate a safety baseline control value; The dynamic confidence level is mapped onto the allowable range of the control action, and the fixed mechanism boundary is forced to shrink towards the safety benchmark control quantity, thereby generating a dynamic modulation boundary that follows the changes in the material state.

8. The intelligent collaborative optimization and control system for the entire coking coal preparation process according to claim 7, characterized in that, The boundary decision module (230) is used for: The average moisture content, average ash content, absolute mass, and hardness indices associated with the data micro-elements are extracted and input into a predefined data-driven prediction algorithm to calculate and output the unconstrained optimal control quantity. The dynamic modulation boundary is used as a hard constraint condition. The unconstrained optimal control quantity is projected into the limited solution space. When the unconstrained optimal control quantity exceeds the dynamic modulation boundary, it is truncated to the nearest boundary extremum to generate the final execution instruction.

9. The intelligent collaborative optimization and control system for the entire coking coal preparation process according to claim 1, characterized in that, The feedforward control module (240) is used for: Extract the absolute installation coordinates of the center position of the feed inlet of the target processing equipment and convert them into linear unfolded coordinates along the running direction of the belt conveyor. Calculate the difference between the linear unfolded coordinates and the absolute physical displacement to obtain the remaining conveying distance. The instantaneous linear velocity of the belt conveyor is read, and the estimated remaining time is calculated by dividing the remaining conveying distance by the instantaneous linear velocity. The electrical network communication delay time and the mechanical transmission response time are extracted and summed on the time axis to generate the feedforward advance trigger time.

10. The intelligent collaborative optimization and control system for the entire coking coal preparation process according to claim 1, characterized in that, The feedforward control module (240) calls the final execution instruction to parse the target operating frequency of the main drive motor and the target feeding rate of the upstream associated feeder, and sends it to the sensing and execution device layer (100); After receiving multi-dimensional hardware control settings, the sensing and execution device layer (100) adjusts the output state execution parameters smoothly according to the preset device frequency change rate, and continuously monitors the actual operating load of the main drive motor and the upstream associated feeder.

Images