Decoupled mining system and method for thin coal seams

Abstract

The present application discloses a decoupled mining system and method for thin coal seams, the system comprising: a mobile module, a conveying module, a cutting module, a control module and a neural network model; wherein the neural network model comprises a perception layer, a processing layer and a control layer which are sequentially connected in communication, the perception layer is used to collect state parameters and action parameters in the mining process as perception data, the processing layer is used to pre-process the perception data; the neural network model decouples the pre-processed perception data based on the decoupling module carried, the control layer is used to generate a control signal for adjusting the action parameter according to the decoupled perception data; the control module is used to generate a control instruction according to the control signal, and send the control instruction to the corresponding module to execute the instruction. The system decouples the mechanical coupling between the state parameters and the action parameters in the mining system through the neural network model, thereby improving the accuracy and stability of the thin coal seam mining control.

Description

Technical Field

[0001] The present application relates to the technical field of coal mining, and in particular to a decoupled mining system and method for thin coal seams. Background Art

[0002] Thin coal seam coal resources are widely distributed and have abundant reserves. However, due to the harsh underground mining environment, great safety hazards, low mining benefits and poor mining efficiency of thin coal seams, the current output of thin coal seams accounts for a low proportion of the total output, resulting in a large amount of thin coal seam resources being wasted.

[0003] In the relevant technologies, the degree of mechanization is generally not high for the mining of thinner coal seams such as 1.3m to 1.5m, especially the supporting technology of fully mechanized mining machinery automation equipment is still very immature. Taking the existing fully mechanized mining machinery as an example, the cutting part where the mining system contacts the thin coal seam is easily affected by the mechanical coupling of its own components, which is specifically manifested in the dynamic fluctuation of parameters such as the speed and torque of its own components, which leads to the mining of thin coal seams cannot be effectively controlled.

[0004] Therefore, how to avoid the influence of mechanical coupling on thin coal seam mining equipment and achieve effective control of thin coal seam mining has become an urgent problem to be solved. Summary of the invention

[0005] The purpose of this application is to solve one of the above-mentioned technical problems at least to a certain extent.

[0006] To this end, the first purpose of this application is to propose a decoupled mining system for thin coal seams. The system decouples the mechanical coupling between the state parameters and the action parameters in the mining system through a neural network model, thereby improving the accuracy and stability of thin coal seam mining control and solving the problem that the mechanical coupling of the cutting module in the existing mining system causes dynamic fluctuations in its own speed and torque.

[0007] The second object of this application is to propose a decoupled mining method for thin coal seams.

[0008] A third object of the present application is to provide a non-transitory computer-readable storage medium.

[0009] To achieve the above-mentioned purpose, the first embodiment of the present application proposes a decoupled mining system for thin coal seams, the system comprising: a mobile module, a conveying module and a cutting module, the conveying module and the cutting module are arranged on the mobile module, the system further comprises: a control module and a neural network model; wherein,

[0010] The neural network model comprises a perception layer, a processing layer and a control layer which are sequentially connected in communication, the control layer is also in communication with the control module, the perception layer is used to collect state parameters and action parameters of the mining system during the mining process as perception data, and the processing layer is used to pre-process the perception data;

[0011] The neural network model decouples the preprocessed perception data based on the decoupling module to separate and adjust the relationship between the state parameters and the action parameters. The control layer is used to generate a control signal for adjusting the action parameters of the mining system during subsequent mining according to the decoupled perception data;

[0012] The control module is respectively connected to the moving module, the conveying module and the cutting module for communication, and is used to generate a control instruction according to the control signal, and send the control instruction to a corresponding module to execute the control instruction.

[0013] In addition, the decoupled mining system for thin coal seams of the embodiment of the present application also has the following additional technical features:

[0014] Optionally, in some embodiments, the neural network model is based on a feedforward neural network, and the neural network model is represented by the following formula:

[0015] ,

[0016] in, represents the neural network model, A and B represents different time-invariant matrices, x Represents the model state quantity, u Indicates the control input quantity of the control module input, Indicates about time t Nonlinear coupling interference.

[0017] Optionally, in some embodiments, the neural network model includes a plurality of channels for transmitting data, and the neural network model is specifically used for:

[0018] The model i The input of the first channel j The coupling influence of the state parameters of the channels is taken as a measurable interference, and the measurable interference is eliminated by a feedforward method, wherein: i and j is any positive integer less than the number of channels;

[0019] Eliminate the first i The status parameters of the channel are j The influence of the state parameters of each channel;

[0020] Based on the elimination process of the feedforward method and the feedback method, the formula satisfied by the neural network model is deformed, and based on the deformation result, an auxiliary input quantity is introduced into the decoupling module to obtain a decoupling controller, and the decoupling between the input quantity and the state quantity is performed by the decoupling controller, wherein the decoupling controller satisfies the following formula:

[0021] S 1 u ( t )+ S 2 u ( t ) =- Kx ( t )+ Lv ( t ),

[0022] in, S 1 and L is the diagonal weighted matrix of different dimensions, S 1 Reversible, S 2 and K are weight matrices with different main diagonal elements being zero, u ( t ) is the input quantity of any channel, v ( t ) is the auxiliary input quantity, x ( t ) is about time t The state quantity.

[0023] Optionally, in some embodiments, the hardware system corresponding to the perception layer includes a three-dimensional force sensor, and the three-dimensional force sensor includes: a first force-bearing body, a second force-bearing body and an elastic body; wherein the elastic body is arranged between the first force-bearing body and the second force-bearing body, and a plurality of layers of slots are arranged at equal intervals on the periphery of the elastic body, and there is a fixed angle phase difference between the slots in different layers; strain gauges are arranged in the thin-walled area between each of the slots, and each of the strain gauges forms an electric bridge for force signal collection.

[0024] Optionally, in some embodiments, the cutting module includes: a rocker unit, a cutting unit, a spray unit and a hydraulic unit; wherein, the rocker unit is movably arranged on the moving module, and the rocker unit is rotationally controlled under the drive of the hydraulic unit; the cutting unit is arranged at the free end of the rocker unit, and the spray unit is arranged inside the rocker unit.

[0025] Optionally, in some embodiments, the cutting unit comprises: a plurality of coaxially arranged cutting drums, each of which is fixedly provided with a plurality of cutting teeth and a plurality of cams; the rocker arm unit comprises: a rocker arm seat and a transmission rod, wherein the transmission rod is arranged in the rocker arm seat, a return spring is arranged on the outer periphery of the transmission rod, one end of the transmission rod is in contact with the cam, and the other end of the transmission rod is connected to the spray unit to serve as the power input of the hydraulic spray.

[0026] Optionally, in some embodiments, the conveying module includes: a collecting hopper, which is arranged below the cutting module; a conveyor belt, which is arranged on the moving module, and the inlet of the conveyor belt is connected to the collecting hopper; a turntable, which is rotatably arranged on the collecting hopper, and the turntable is driven by a motor arranged inside the collecting hopper.

[0027] To achieve the above object, a second embodiment of the present invention provides a decoupled mining method for thin coal seams, which is applied to the decoupled mining system for thin coal seams of the first aspect. The method comprises:

[0028] Move the decoupled mining system to the coal mining opening of the thin coal seam to be mined, dock the loader behind the decoupled mining system, and obtain the mining conditions corresponding to the coal mining opening;

[0029] Acquire historical mining data of the thin coal seam as a sample data set, train a random forest model based on the sample data set by a grid search method and a cross-validation method, input the mining conditions into the trained random forest model, and acquire mining face path data output by the model;

[0030] The mining face path data is sent to the control module of the decoupled mining system, and the control module optimizes the action parameters of the decoupled mining system according to the mining face path data, and performs thin coal seam mining operations according to the initial action parameters;

[0031] During the thin coal seam mining process, the real-time state parameters and action parameters of the decoupled mining system are collected, and the collected state parameters and action parameters are decoupled through the neural network model in the system. A control signal for adjusting the action parameters of the system during subsequent mining is generated based on the decoupled data, and a control module generates a control instruction based on the control signal to adjust the action parameters of the corresponding module.

[0032] To achieve the above objectives, the third aspect of the present invention proposes a non-temporary computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, it implements the decoupled mining method of thin coal seams as described in any of the second aspect embodiments above.

[0033] The technical solution provided by the embodiments of the present application brings at least the following beneficial effects:

[0034] The present application uses a neural network model to decouple the mechanical coupling between the state parameters and the action parameters in the mining system, separate and adjust the relationship between the state parameters and the action parameters, and the neural network model and decoupling controller in the system can be regarded as a number of single variable systems that do not affect each other, achieving complete decoupling between input and state, thereby effectively reducing the impact of mechanical coupling in the system and automatically and accurately controlling the relevant mining equipment of the system. In addition, the present application can also adjust the action parameters of the system according to the real-time collected data, adapt to different mining conditions and coal seam characteristics, and adjust and optimize the system control method according to the continuously accumulated data and experience, so as to improve the adaptability and mining performance of the mining system. In addition, the information acquisition equipment such as the three-dimensional force sensor of the present application can provide high-precision force measurement data, including the magnitude and direction of the force; the structural design of the elastic body and the strain gauge makes the sensor have high stability and anti-interference ability, and can work stably in harsh environments; the sensor adopts a complex structural design and a precise calibration process, which can provide reliable force measurement data and provide reliable data support for the stable operation of the system. Therefore, the present application effectively controls the mining of thin coal seams through the decoupling process, and improves the accuracy, stability and intelligence of the mining control of thin coal seams.

[0035] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0037] Figure 1 A schematic diagram of the structure of a decoupled mining system for thin coal seams proposed in an embodiment of the present application;

[0038] Figure 2 A schematic diagram of the structure of a three-dimensional force sensor proposed in an embodiment of the present application;

[0039] Figure 3 A schematic diagram of the structure of a specific decoupled mining system for thin coal seams proposed in an embodiment of the present application;

[0040] Figure 4 A schematic diagram of the structure of an enlarged local area of ​​a cutting module proposed in an embodiment of the present application;

[0041] Figure 5 This is a flow chart of a decoupled mining method for thin coal seams proposed in an embodiment of the present application.

[0042] Wherein, the reference numerals represent:

[0043] 1-mobile module, 2-transport module, 3-cutting module, 4-three-dimensional force sensor;

[0044] 21-collecting hopper, 22-conveyor belt, 23-turntable;

[0045] 31- rocker unit, 32- cutting unit, 33- spray unit, 34- hydraulic unit;

[0046] 311- rocker arm seat, 312- transmission rod, 313- return spring;

[0047] 321-cutting drum, 322-cutting teeth, 323-cam;

[0048] 41 - first force-bearing body, 42 - second force-bearing body, 43 - elastic body, 44 - slot. DETAILED DESCRIPTION

[0049] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0050] In order to facilitate understanding of the technical solution of the decoupled mining system of the present application, some technical terms involved in the present application are explained below:

[0051] The term "state parameter" in this application refers to the state data of each component in the decoupled mining system of this application during the thin coal seam mining process, such as the temperature and speed of the cutting drum 321, the position of the mobile module 1, etc.

[0052] The term "action parameter" in this application refers to the relevant data of the actions performed by various components in the decoupled mining system of this application during the thin coal seam mining process, such as the movement time of the mobile module 1, the lifting or lowering action of the rocker arm unit by the hydraulic unit, etc.

[0053] The term "sensory data" in this application refers to raw data or information collected from a system or environment, which may be data obtained by sensors, monitoring equipment, etc. In this application, the sensory data includes data collected from three-dimensional force sensors 4, coal seam thickness measuring instruments, and other equipment;

[0054] The term "control signal" in this application refers to a signal used to adjust the behavior or operation of the system of this application, which is usually generated by the control layer of the neural network model. After the control layer is connected to the control module signal, the control signal is sent to the control module, and the control module generates a corresponding control instruction that can be recognized by the device according to the control signal and sends it to the actuator or device. For example, in the decoupled mining system of this application, the control signal can be used to adjust the speed of the cutting module, the injection pressure of the spray unit, etc.;

[0055] The term "decoupling process" in this application refers to separating and adjusting the relationship between state parameters and action parameters through a neural network model and a decoupling module, thereby achieving more precise control.

[0056] The term "channel" in this application refers to the path of information transmission or the path of signal transmission. In this application, it refers to different input or output paths in a neural network model, which are used to transmit perception data, control signals and other information.

[0057] The term "feedforward method" in this application refers to adjusting the behavior of a system by directly using the current state of the system as input without considering the past behavior or errors of the system. In the decoupling process, the feedforward method is used to eliminate the influence of the input on the state parameters.

[0058] The term "feedback method" in this application refers to reducing the error and achieving closed-loop control by measuring the output of the system and comparing it with the expected output, and then adjusting the input or parameters of the system. In the decoupling process, the feedback method is used to eliminate the mutual influence between state parameters.

[0059] The term "strain gauge" in this application refers to a sensor used to measure the strain or deformation of the elastic body 43. In this application, the strain gauge is used to sense the force condition of the cutting module.

[0060] The term "spray unit" in this application refers to a device or component for spraying a liquid. In this application, the spray unit is used to spray water or other liquids to reduce coal dust and reduce friction.

[0061] The term "historical mining data" in this application refers to the data collected during the previous mining process of thin coal seams, including information such as coal seam thickness, hardness and mining speed, which is used to build models or analyze mining effects.

[0062] The term "grid search method" in this application refers to a method for hyperparameter optimization, which finds the best parameter combination by performing a grid search in a given parameter space and evaluating the performance of each parameter combination.

[0063] The term "cross-validation method" in this application refers to a method for evaluating model performance by dividing the dataset into multiple subsets, then training the model multiple times and testing it on different subsets to evaluate the generalization ability of the model.

[0064] The term "hydraulic unit" in the present application refers to a device or system for generating and controlling hydraulic pressure, which may preferably be a device such as a hydraulic cylinder, and is used to control the movement of the rocker unit in the cutting module.

[0065] The following describes the decoupled mining system and method for thin coal seams according to an embodiment of the present application with reference to the accompanying drawings.

[0066] Figure 1 This is a schematic diagram of the structure of a decoupled mining system for thin coal seams proposed in an embodiment of the present application, such as Figure 1 As shown, the system includes: a mobile module 1, a conveying module 2 and a cutting module 3, wherein the conveying module 2 and the cutting module 3 are arranged on the mobile module 1, and the system also includes a control module and a neural network model ( Figure 1 (not shown in the figure), the control module and the neural network model can be set at the corresponding position of the mining system according to the actual situation of the equipment and the decoupling control requirements, etc., and this application does not impose any restrictions on this.

[0067] Among them, the neural network model includes a perception layer, a processing layer and a control layer which are connected in sequence. The control layer is also connected in communication with the control module. The perception layer is used to collect the state parameters and action parameters of the mining system during the mining process as perception data, and the processing layer is used to pre-process the perception data.

[0068] The neural network model decouples the preprocessed perception data based on the onboard decoupling module to separate and adjust the relationship between state parameters and action parameters. The control layer is used to generate control signals for adjusting the action parameters of the mining system during subsequent mining based on the decoupled perception data.

[0069] The control module is respectively connected to the moving module, the conveying module and the cutting module for communication. The control module is used to generate a control instruction according to a control signal and send the control instruction to a corresponding module to execute the control instruction.

[0070] It should be noted that the current thin coal seam comprehensive mining technology, especially the supporting technology of comprehensive mining machinery automation equipment, is still very immature and is in the exploratory stage. Taking the existing automated comprehensive mining machinery as an example, it was found in the loading experiment before it left the factory that the cutting part in contact with the thin coal seam is easily affected by the mechanical coupling of its own components, specifically the dynamic fluctuation of its own speed and torque, which leads to the mining of the thin coal seam cannot be effectively controlled. Based on the above problems, the present application proposes a decoupled mining system for thin coal seams, which is configured to move the position of the mobile module 1 during the mining process, and to set the control module to perform intelligent control during the mining process of the thin coal seam, and to collect and process data through the neural network model, and can decouple the mechanical coupling between the state parameters and the action parameters in the mining system, so that the automation process of the mining system is more accurate.

[0071] Specifically, the neural network model realizes data collection, preprocessing and control through the perception layer, processing layer and control layer respectively; the perception layer is responsible for collecting the state parameters and action parameters during mining as perception data, and the processing layer preprocesses these perception data; finally, the control layer is connected to the control module signal to generate a control signal for adjusting the action parameters during mining, thereby realizing automatic control of the system.

[0072] Among them, the control instructions sent by the control module can control various related modules such as the mobile module 1, the conveying module 2 and the cutting module 3 that perform actions during the thin coal seam mining process, and perform actions according to the parameters set in the instructions. Specifically, the control layer of the neural network model is responsible for generating control signals. The control layer processes the decoupled state parameters and action parameters through the decoupling process of the decoupling module to generate signals for adjusting the action parameters during mining. Since the control layer is connected to the control module in communication, these signals are sent to the control module, and then the control module generates precise control instructions based on the action parameters adjusted by the control signal, and then the control module sends the control instructions to the corresponding related modules that can execute the current instructions, and the modules that receive the instructions execute these instructions.

[0073] For example, the main function of the control signal is to adjust the action parameters of the system, such as the moving time of the moving module 1, the lifting or lowering action of the rocker unit 31 controlled by the hydraulic unit 34, etc. It can be understood that by adjusting these action parameters, the state parameters of the mining system will also be indirectly affected. For example, the speed of the cutting drum 321 is a state parameter, and the control signal can indirectly change the speed of the cutting drum 321 by adjusting the relevant action parameters (such as the rotation speed of the cutting unit 32).

[0074] In the above process, it is necessary to realize the decoupling process through the onboard decoupling module, that is, to separate and adjust the relationship between the state parameters and the action parameters, so as to reduce the influence of mechanical coupling in the system and make the mining process more accurate and stable.

[0075] In specific implementation, the processing layer of the neural network model in this application preprocesses the perception data as follows: first classify the original data of various influencing parameters, then convert the classified data into digital labels that can be calculated by computers, and then undergo SMOTE sampling and normalization processing. Then, the neural network model decouples the preprocessed data through the decoupling module it carries.

[0076] The following is a detailed description of the process of decoupling the neural network model of the present application in combination with the decoupling module. In one embodiment of the present application, the neural network model is based on a feedforward neural network, and the neural network model is represented by the following formula:

[0077] ,

[0078] in, represents the neural network model, A and B represents different time-invariant matrices, x Represents the model state quantity, u Indicates the control input quantity of the control module input, Indicates about time t Nonlinear coupling interference.

[0079] Among them, the model state refers to a set of variables used in the model to describe the current state of the system, such as the displacement, velocity, acceleration, temperature and pressure of the system; the control input refers to the external input that can be directly controlled by the control module to affect the system state. It can be a signal or force applied to the system through a control device or algorithm. For example, in a motor control system, the control input may be the voltage or current of the motor; in a hydraulic system, the control input may be the pressure or flow of the hydraulic oil.

[0080] In this embodiment, the neural network model includes multiple channels for transmitting data. When decoupling, the neural network model is specifically used to: i The input of the first channel j The coupling influence of the state parameters of the channels is taken as a measurable interference, and the measurable interference is eliminated by a feedforward method, wherein: i and j is any positive integer less than the number of channels; the first i The status parameters of the channel are j The influence of the state parameters of each channel; based on the elimination process of the feedforward method and the feedback method, the formula satisfied by the neural network model is deformed, and based on the deformation result, an auxiliary input quantity is introduced into the decoupling module to obtain a decoupling controller, and the decoupling between the input quantity and the state quantity is performed through the decoupling controller, wherein the decoupling controller satisfies the following formula:

[0081] S 1 u ( t )+ S 2 u ( t ) =- Kx ( t )+ Lv ( t ),

[0082] in, S 1 andL is the diagonal weighted matrix of different dimensions, S 1 Reversible, S 2 and K are weight matrices with different main diagonal elements being zero, u ( t ) is the input quantity of any channel, v ( t ) is the auxiliary input quantity, x ( t ) is about time t The state quantity.

[0083] Among them, the input quantity of the channel refers to the operating variable or instruction that directly affects the output of the system. For example, the signal that controls the action of the hydraulic unit can directly adjust the action of the rocker unit. The input quantity is the direct result of the control operation and can be regarded as the signal corresponding to the action parameter, while the state parameter is the internal expression of the system's response to these input quantities. That is, the input quantity is the specific control content, and the state parameter is the response result of the system.

[0084] In the neural network model, one channel corresponds to a specific input quantity or state parameter. At the same time, if there are multiple input quantities that need to be controlled in the system, each input quantity can be processed through an independent channel to achieve more precise decoupled control.

[0085] In the above decoupling process, the feedforward method is first used to eliminate i The input of the first channel j The coupling effect of the state parameters of the first channel is eliminated by the feedback method. i The state parameters are j The influence of the state parameters is eliminated; the auxiliary input quantity is then introduced to obtain the decoupling controller, and finally the decoupling between the input quantity and the state quantity is completed through the decoupling controller.

[0086] Among them, the purpose of feedforward elimination is to deal with the coupling effect directly at the input stage, by predicting the potential impact of the input on the system state parameters and adjusting the input signal before it is transmitted to the system. Feedback elimination is a control strategy based on system output, which measures the actual output of the system and compares it with the expected output, and then adjusts the input or system parameters according to the error signal to reduce or eliminate the mutual influence between the state parameters.

[0087] The decoupling controller is the last step of the entire decoupling process, which integrates the results of feedforward elimination and feedback elimination and further optimizes the control of the mining system. In this application, the design of the decoupling controller is based on a deep understanding of the dynamics of the mining system, including the state space model or transfer function of the system. The goal of the decoupling controller is to achieve complete decoupling between input quantities and state quantities, so that each input quantity can independently control the corresponding output quantity without being affected by other input quantities or state quantities.

[0088] Therefore, in this application, feedforward elimination algorithm and feedback elimination algorithm are used to eliminate coupling effects, but their timing and mode of action are different. Feedforward elimination is preventive, and it adjusts the input signal before it enters the system; while feedback elimination is corrective, and it adjusts according to the actual output of the system after the coupling effect occurs.

[0089] It should be noted that when establishing the above model, for the theoretical basis of the model, the present application is to establish an accurate mathematical model to describe the input-output relationship of the system, and the model is constructed based on physical principles, data analysis or machine learning methods to ensure accurate description of the system behavior. For the overall process of decoupled mining of thin coal seams, since the present application finally outputs a signal of modulated action parameters, for the mining system, the state parameters of the system are collected by sensors and other equipment as input data, and the action parameters are used as output data. These collected data may be affected by noise, interference, etc., so preprocessing is required, such as filtering, normalization, etc., to ensure the quality and reliability of the data; and then a decoupling model is designed to completely decouple the input and output of the system, with the goal of eliminating or minimizing the influence of input parameters on output parameters, and realizing independent control of input and output. More preferably, in one embodiment of the present application, the decoupling process is also verified and adjusted to ensure that it can effectively realize complete decoupling of input and output in practical applications.

[0090] The following is a detailed description of the process of establishing the neural network model and the process of decoupling the input quantity and the state quantity by the decoupling controller.

[0091] In this embodiment, during the establishment of the neural network model, the coupled state quantity and action quantity in the mining system can be regarded as a nonlinear unknown disturbance about time. Due to the good self-learning ability and good robustness of the feedforward neural network, the embodiment of the present application selects the feedforward neural network as the basis of the neural network model, and establishes its dynamic model as follows: ,

[0092] In the process of eliminating the coupling effect by the above feedforward method and feedback method, the above dynamic model can be transformed into the following formula:

[0093] ,

[0094] in, A 1 is a diagonal matrix whose diagonal elements are equal to the matrix A Elements on the main diagonal, A 2 = A - A 1 , A 2 is a matrix with zero elements on the main diagonal, and represents the interaction between different variables in the mining system of this application, B 1 is a diagonal matrix whose diagonal elements are equal to the matrix B Elements on the main diagonal, B 2 = B - B 1 , B 2 is a matrix with zero elements on the main diagonal, which represents the coupling between different channels.

[0095] Furthermore, in order to eliminate the interference of coupling terms, several auxiliary parameters are introduced: S 1 , S 2 and K .in, S 1 , S 2 and K The following relations are satisfied:

[0096]

[0097] Then, based on the relationship between the above parameters, the formula satisfied by the above neural network model can be transformed into the following formula:

[0098] ,

[0099] Then, based on the deformation result, after introducing the auxiliary input, the above-mentioned decoupling controller can be derived to satisfy the formula. From the above deformation result formula, it can be seen that the formula can be regarded as a number of single variable systems that do not affect each other, thereby achieving complete decoupling between input and state.

[0100] Therefore, unlike the traditional thin coal seam mining system which is affected by mechanical coupling and leads to unstable mining control, the mining system in the present application can effectively reduce the impact of mechanical coupling and improve the accuracy and stability of mining through the decoupling process of the neural network model; it realizes intelligent mining control and adjusts the system's action parameters according to real-time collected data to adapt to different mining conditions and coal seam characteristics; it also adjusts and optimizes according to continuously accumulated data and experience to improve the adaptability and performance of the system.

[0101] The other modules in the decoupled mining system of the present application are described in detail below.

[0102] In one embodiment of the present application, the hardware system corresponding to the perception layer includes a three-dimensional force sensor, and the three-dimensional force sensor includes: a first force-bearing body, a second force-bearing body and an elastic body; wherein the elastic body is arranged between the first force-bearing body and the second force-bearing body, and a plurality of layers of slots are arranged at equal intervals on the periphery of the elastic body, and there is a fixed angle phase difference between the slots in different layers; strain gauges are arranged in the thin-walled area between each slot, and each strain gauge forms an electric bridge to collect force signals.

[0103] Specifically, in the hardware system supporting the neural network model, the hardware subsystem corresponding to the perception layer includes a plurality of sensors for detecting the action parameters and state parameters of the mining system, including a three-dimensional force sensor 4. For example, Figure 2 As shown, the three-dimensional force sensor 4 includes a first force-bearing body 41, a second force-bearing body 42 and an elastic body 43. The elastic body 43 is arranged between the first force-bearing body 41 and the second force-bearing body 42. A plurality of layers of slots 44 are evenly spaced on the periphery of the elastic body 43, and there is a 45° phase difference between the slots 44 between the layers. Strain gauges are arranged in the thin-walled areas between the slots 44, and the strain gauges form an electric bridge to collect force signals.

[0104] Therefore, based on the above structure, the elastic body 43 converts the force exerted on the measured object into its deformation and then measures the magnitude and direction of the strain through the strain gauge, that is, converts the force signal into an electrical signal, and determines the correspondence between the force signal and the electrical signal through precise calibration to obtain the force condition of the measured object. Due to the fine design of the sensor structure and the sensitivity of the bridge, the above structure can provide high-precision force measurement, including the magnitude and direction of the force; the structural design of the elastic body 43 and the strain gauge makes the sensor have high stability and anti-interference ability, and can work stably in harsh environments. The sensor adopts a complex structural design and a precise calibration process, which can provide reliable force measurement data and provide reliable support for the stable operation of the system.

[0105] In one embodiment of the present application, the cutting module includes: a rocker unit, a cutting unit, a spray unit and a hydraulic unit; wherein the rocker unit is movably arranged on the moving module, and the rocker unit is controlled to rotate under the drive of the hydraulic unit;

[0106] The cutting unit is arranged at the free end of the rocker unit, and the spray unit is arranged inside the rocker unit.

[0107] Specifically, if Figure 3 As shown, the cutting module 3 includes a rocker unit 31, a cutting unit 32 and a spray unit 33. The rocker unit 31 is movably arranged with the mobile module 1 and is rotationally controlled by a hydraulic unit 34. The cutting unit 32 is arranged at the free end of the rocker unit 31, and the spray unit 33 is arranged inside the rocker unit 31. It should also be noted that the rocker unit 31 is rotationally controlled by the hydraulic unit 34, so that the cutting unit 32 and the spray unit 33 can achieve the required actions during the working process; the cutting unit is arranged at the free end of the rocker unit 31, and the cutting unit 32 is controlled to cut the thin coal seam through the rotation of the rocker unit 31; the spray unit 33 is arranged inside the rocker unit 31, and the spray unit 33 is controlled to spray and reduce dust in the cutting area through the spray unit 33.

[0108] Based on the above structure, the cutting module 3 integrates the rocker arm unit 31, the cutting unit 32 and the spray unit 33 to form a complete cutting system. Compared with the traditional decentralized design, it has the advantages of more compact structure and more convenient operation. Through hydraulic control, the rocker arm unit 31 can achieve precise rotation control, thereby realizing precise adjustment and control of the cutting unit 32 and the spray unit 33, improving the accuracy and efficiency of the cutting operation.

[0109] Further, such as Figure 3 As shown, in this embodiment, the cutting unit 32 includes: a plurality of coaxially arranged cutting drums 321. In order to more clearly and meticulously show the specific structure and connection mode of the rocker unit 31 and the cutting unit 32, Figure 3 A region A (i.e. Figure 3 The reference symbol A in the figure represents a selected local area to be enlarged), Figure 4 This is a schematic diagram of the structure after the area is enlarged. Figure 4As shown, a plurality of cutting teeth 322 and a plurality of cams 323 are fixedly arranged on each cutting drum 321. The rocker arm unit 31 comprises: a rocker arm seat 311 and a transmission rod 312, wherein the transmission rod 312 is arranged inside the rocker arm seat 311, that is, a part of the transmission rod 312 is inside the rocker arm seat 311, a return spring 313 is sleeved on the outer periphery of the transmission rod 312, one end of the transmission rod 312 abuts against the cam 323, and the other end of the transmission rod 312 is connected to the spray unit 33 to serve as the power input of the hydraulic spray.

[0110] It should be noted that, in this embodiment, the design of the transmission rod 312 enables synchronous control of the cutting and spraying processes, ensures the coordinated progress of cutting and spraying, and improves the work efficiency and accuracy; the reset spring 313 sleeved on the outer periphery of the movable rod can restore the rocker unit 31 to its initial position after the spraying and cutting processes are completed, so as to prepare for the next operation, while ensuring the stability and reliability of the system. When the hydraulic system is started, the rocker unit 31 starts to work, and the rotation of the cam 323 causes the transmission rod 312 to follow the movement, thereby driving the cutting drum 321 to rotate, and the cutting teeth 322 on the cutting drum 321 cut the thin coal seam during the rotation. At the same time, the other end of the transmission rod 312 is connected to the spray unit 33 as its power input. During the cutting process, the spray unit 33 performs hydraulic spraying according to preset conditions to achieve the spray dust reduction effect on the cutting area.

[0111] In one embodiment of the present application, the conveying module includes: a collecting hopper, which is arranged below the cutting module; a conveyor belt, which is arranged on the moving module, and the inlet of the conveyor belt is connected to the collecting hopper; a turntable, which is rotatably arranged on the collecting hopper, and the turntable is driven by a motor arranged inside the collecting hopper.

[0112] Specifically, if Figure 3 As shown, the conveying module 2 includes a hopper 21 arranged below the cutting module 3, and also includes a conveyor belt 22 arranged on the mobile module 1. The inlet of the conveyor belt 22 is connected to the hopper 21. A turntable 23 is rotatably arranged on the hopper 21, and the turntable 23 is driven by a motor arranged in the hopper 21. Based on the above structure, the combined design of the hopper 21 and the conveyor belt 22 enables the cut materials to be quickly and efficiently transported to the designated location, thereby improving production efficiency and material utilization. The turntable 23 in the hopper 21 is driven by a motor, and can realize the function of automatic adjustment and distribution of materials, thereby reducing manual intervention and management costs. The hopper 21 is arranged below the cutting module 3, which utilizes space, makes the entire system structure more compact, and saves space and resources. The integrated design of the hopper 21 and the conveyor belt 22 makes the material transportation process smoother, reduces intermediate links and energy loss, and improves the stability and reliability of the system.

[0113] In summary, the decoupled mining system for thin coal seams implemented in this application decouples the mechanical coupling between the state parameters and the action parameters in the mining system through the neural network model, separates and adjusts the relationship between the state parameters and the action parameters, and the neural network model and decoupling controller in the system can be regarded as a number of single variable systems that do not affect each other, realizing the complete decoupling between input and state, thereby effectively reducing the impact of mechanical coupling in the system, and automatically and accurately controlling the relevant mining equipment of the system. In addition, the system can also adjust the action parameters of the system according to the real-time collected data, can adapt to different mining conditions and coal seam characteristics, and adjust and optimize the system control method according to the continuously accumulated data and experience, so as to improve the adaptability and mining performance of the mining system. In addition, the information acquisition equipment such as the three-dimensional force sensor of the system can provide high-precision force measurement data, including the magnitude and direction of the force; the structural design of the elastic body and the strain gauge makes the sensor have high stability and anti-interference ability, and can work stably in harsh environments; the sensor adopts a complex structural design and a precise calibration process, which can provide reliable force measurement data and provide reliable data support for the stable operation of the system. Therefore, the system effectively controls thin coal seam mining through the decoupling process, improving the accuracy, stability and intelligence of thin coal seam mining control.

[0114] In order to more clearly illustrate the specific implementation process of mining thin coal seams through the decoupled mining system of thin coal seams, a decoupled mining method for thin coal seams proposed in the embodiment of the present application is described in detail below. The method is applied to the decoupled mining system for thin coal seams in the above embodiment, that is, the mining method of the present embodiment is realized by performing relevant control on the mining system in the above embodiment. The various components involved in the method are as described in the above embodiment, and will not be repeated here.

[0115] Figure 5 This is a flow chart of a decoupled mining method for thin coal seams proposed in an embodiment of the present application, such as Figure 5 As shown, the method comprises the following steps:

[0116] Step S101: Move the decoupled mining system to the coal mining opening of the thin coal seam to be mined, dock the transfer machine at the rear of the decoupled mining system, and obtain the mining conditions corresponding to the coal mining opening.

[0117] Specifically, the mining conditions corresponding to the coal mining opening include the current mining environment parameters of the coal mining opening, the initial state parameters of the equipment, and the target planning data that has been formulated for this mining task.

[0118] Step S102: Obtain historical mining data of thin coal seams as a sample data set, train a random forest model based on the sample data set through grid search and cross-validation methods, input mining conditions into the trained random forest model, and obtain mining face path data output by the model.

[0119] Specifically, after step S101 is completed, the historical mining data of the thin coal seam is used as a sample data set, and the sample data set is preprocessed and divided proportionally to establish a training set and a data set respectively, and the optimal random forest model is found according to the grid search method and cross-validation method. Finally, the collected current mining condition data is input into the trained random forest model, and the mining face path of the mining system is output according to the random forest model.

[0120] In one embodiment of the present application, the random forest model can be set in the mining system or as an independent software component. It is necessary to ensure that the output of the random forest model, that is, the prediction of the mining surface path, will be sent to the control module of the mining system to guide the actual mining operation.

[0121] In the process of training the random forest model, as a possible implementation method, the specific steps of model training are as follows:

[0122] The first step is data preprocessing: cleaning and standardizing historical mining data, which may include denoising, feature scaling, etc. The second step is feature selection: selecting key features that are highly correlated with the mining face path, such as coal seam thickness, hardness, water content, etc. The third step is data set segmentation: split the preprocessed data set into training sets and test sets according to a certain ratio, such as 70% training sets and 30% test sets. The fourth step is model training: using the training set, create multiple different sample sets through sampling with replacement; for each sample set, randomly select a subset of features and train a decision tree. The fifth step is parameter optimization: use grid search and cross-validation to find the optimal model parameters, such as the number of decision trees, the number of features considered by each decision tree, etc.

[0123] When performing path prediction using a trained model, the specific steps are as follows: using the trained random forest model, input new mining conditions to obtain the prediction results of the mining surface path; model evaluation: using the test set to evaluate the prediction accuracy of the model to ensure the reliability of the mining path prediction; real-time adjustment: during the mining process, the model can be updated and adjusted based on real-time monitoring data to adapt to changes in mining conditions.

[0124] Step S103: Send the mining face path data to the control module of the decoupled mining system, optimize the action parameters of the decoupled mining system according to the mining face path data through the control module, and perform thin coal seam mining operations according to the initial action parameters.

[0125] Specifically, the present application indirectly links the path prediction through the random forest model with the decoupling control in the above embodiment. The random forest model provides a prediction of the mining path, and the decoupling control uses this prediction information to optimize the action parameters in the mining process, reduce the impact of mechanical coupling, and improve the accuracy and stability of mining. When generating control instructions, the control module can combine the predicted mining path and adaptively adjust the action parameters of the execution module according to the specific working conditions of the mining path to perform mining operations that are more suitable for the current mining conditions.

[0126] Furthermore, the control module generates initial action parameters at the beginning of mining based on the current mining path prediction data and mining condition data, and controls related modules to perform thin coal seam mining operations according to the initial action parameters.

[0127] Step S104: During the thin coal seam mining process, the real-time state parameters and action parameters of the decoupled mining system are collected, and the collected state parameters and action parameters are decoupled through the neural network model in the system. A control signal for adjusting the action parameters of the system during subsequent mining is generated based on the decoupled data, and a control module generates a control instruction based on the control signal to adjust the action parameters of the corresponding module.

[0128] Specifically, during the mining operation of the mining system, the real-time state parameters and action parameters of the decoupled mining system are continuously collected, and the collected state parameters and action parameters are decoupled through the neural network model within the system in accordance with the decoupling method and control method described in the above system embodiment, and the action parameters of the relevant modules are adjusted according to the decoupled data, such as controlling the rotation speed and torque of the cutting module 3 during the mining process.

[0129] In one embodiment of the present application, a thin coal seam mining operation is performed, including: when the cutting drum rotates, the transmission rod is driven to move back and forth through the synchronous rotation of the cam, and the spray unit is controlled to spray synchronously.

[0130] Furthermore, in the subsequent mining process, data collection and decoupling control can be repeated in the above manner, and the planned mining path can be adjusted according to the real-time mining conditions, so that the action parameters of the mining system can be adjusted according to the current mining needs, thereby improving the real-time performance of the decoupled mining control.

[0131] Therefore, the present application uses a machine learning algorithm to predict the mining surface path, and combines the neural network model to control the cutting module 3 and the spray unit 33 in real time, thereby realizing intelligent control of the mining process and improving mining efficiency and accuracy. The method can adjust the model in real time according to real-time monitoring data to adapt to changes in mining conditions, and has better adaptability and flexibility. Through real-time control of the cutting module 3 and the spray unit 33, the accuracy and stability of the mining process can be ensured, and the risks and losses in the mining process can be reduced.

[0132] In summary, the decoupled mining method for thin coal seams implemented in the present application decouples the mechanical coupling between the state parameters and the action parameters in the mining system through a neural network model, separates and adjusts the relationship between the state parameters and the action parameters, and realizes complete decoupling between the input and the state, thereby effectively reducing the impact of the mechanical coupling in the system and performing automatic and precise control of the relevant mining equipment of the system. In addition, the method can also adjust the action parameters of the system according to the data collected in real time, can adapt to different mining conditions and coal seam characteristics, and adjust and optimize the system control method according to the continuously accumulated data and experience, so as to improve the adaptability and mining performance of the mining system. Therefore, the method effectively controls the mining of thin coal seams through the decoupling process, thereby improving the accuracy, stability and intelligence of the control of thin coal seams.

[0133] In order to implement the above embodiments, the present application also proposes a non-temporary computer-readable storage medium storing a computer program, which, when executed by a processor, implements the decoupled mining method for thin coal seams proposed in the aforementioned embodiments of the present application.

[0134] It should be noted that it should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0135] In addition, in the description of the present application, the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc., indicating the orientation or position relationship are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0136] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0137] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0138] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0139] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in the field can change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A decoupled mining system for thin coal seams, comprising: A mobile module, a conveying module and a cutting module, wherein the conveying module and the cutting module are arranged on the mobile module, and characterized in that it also includes: a control module and a neural network model; wherein, The neural network model comprises a perception layer, a processing layer and a control layer which are sequentially connected in communication, the control layer is also connected in communication with the control module, the perception layer is used to collect state parameters and action parameters of the mining system during the mining process as perception data, and the processing layer is used to pre-process the perception data; The neural network model decouples the preprocessed perception data based on the decoupling module to separate and adjust the relationship between the state parameters and the action parameters. The control layer is used to generate a control signal for adjusting the action parameters of the mining system during subsequent mining according to the decoupled perception data. The control module is respectively connected to the moving module, the conveying module and the cutting module for communication, and is used to generate a control instruction according to the control signal, and send the control instruction to the corresponding module to execute the control instruction; The neural network model is based on a feedforward neural network and is represented by the following formula: ， in, represents the neural network model, A and B represents different time-invariant matrices, x Represents the model state quantity, u Indicates the control input quantity of the control module input, Indicates about time t Nonlinear coupling interference; The neural network model includes a plurality of channels for transmitting data, and the neural network model is specifically used for: The model i The input of the first channel j The coupling influence of the state parameters of the channels is taken as a measurable interference, and the measurable interference is eliminated by a feedforward method, wherein: i and j is any positive integer less than the number of channels; Eliminate the first i The status parameters of the channel are j The influence of the state parameters of each channel; Based on the elimination process of the feedforward method and the feedback method, the formula satisfied by the neural network model is deformed, and based on the deformation result, an auxiliary input quantity is introduced into the decoupling module to obtain a decoupling controller, and the decoupling between the input quantity and the state quantity is performed by the decoupling controller, wherein the decoupling controller satisfies the following formula: S 1 u （ t ）+ S 2 u （ t ）=- Kx （ t ）+ Lv （ t ）， in, S 1 and L is the diagonal weighted matrix of different dimensions, S 1 reversible, S 2 and K are weight matrices with different main diagonal elements being zero, u ( t ) is the input quantity of any channel, v ( t ) is the auxiliary input quantity, x ( t ) is about time t The state quantity.

2. The decoupled mining system for thin coal seams according to claim 1, characterized in that: The hardware system corresponding to the perception layer includes a three-dimensional force sensor, which includes: a first force-bearing body, a second force-bearing body and an elastic body; wherein, The elastic body is arranged between the first force-bearing body and the second force-bearing body, and multiple layers of slots are arranged at equal intervals on the periphery of the elastic body, and the slots in different layers have a fixed angle phase difference; The thin-walled areas between the slots are provided with strain gauges, and the strain gauges form an electric bridge for collecting force signals.

3. The decoupled mining system for thin coal seams according to claim 1, characterized in that: The cutting module comprises: a rocker unit, a cutting unit, a spray unit and a hydraulic unit; wherein, The rocker arm unit is movably arranged on the moving module, and the rocker arm unit is rotationally controlled under the drive of the hydraulic unit; The cutting unit is arranged at the free end of the rocker unit, and the spray unit is arranged inside the rocker unit.

4. The decoupled mining system for thin coal seams according to claim 3, characterized in that: The cutting unit comprises: A plurality of coaxially arranged cutting drums, each of which is fixedly provided with a plurality of cutting teeth and a plurality of cams; The rocker arm unit comprises: a rocker arm seat and a transmission rod, wherein the transmission rod is arranged in the rocker arm seat, a return spring is sleeved on the outer periphery of the transmission rod, one end of the transmission rod abuts against the cam, and the other end of the transmission rod is connected to the spray unit to serve as the power input of the hydraulic spray.

5. The decoupled mining system for thin coal seams according to claim 1, characterized in that: The conveying module comprises: A collecting hopper, the collecting hopper is arranged below the cutting module; A conveyor belt, wherein the conveyor belt is arranged on the moving module, and an inlet of the conveyor belt is connected to the collecting hopper; A turntable is rotatably arranged on the collecting hopper, and the turntable is driven by a motor arranged inside the collecting hopper.

6. A decoupled mining method for thin coal seams, characterized in that: A decoupled mining system for thin coal seams as described in any one of claims 1 to 5, the method comprising the following steps: Move the decoupled mining system to the coal mining opening of the thin coal seam to be mined, dock the loader behind the decoupled mining system, and obtain the mining conditions corresponding to the coal mining opening; Acquire historical mining data of the thin coal seam as a sample data set, train a random forest model based on the sample data set by a grid search method and a cross-validation method, input the mining conditions into the trained random forest model, and acquire mining face path data output by the model; The mining face path data is sent to the control module of the decoupled mining system, and the control module optimizes the action parameters of the decoupled mining system according to the mining face path data, and performs thin coal seam mining operations according to the initial action parameters; During the thin coal seam mining process, the real-time state parameters and action parameters of the decoupled mining system are collected, and the collected state parameters and action parameters are decoupled through the neural network model in the system. A control signal for adjusting the action parameters of the system during subsequent mining is generated based on the decoupled data, and a control module generates a control instruction based on the control signal to adjust the action parameters of the corresponding module.

7. The decoupled mining method for thin coal seams according to claim 6, characterized in that: The thin coal seam mining operation comprises: When the cutting drum rotates, the transmission rod is driven to move back and forth through the synchronous rotation of the cam, and the spray unit is controlled to spray synchronously.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the decoupled mining method for thin coal seams as described in any one of claims 6-7 is implemented.

Images