Hydraulic support adaptive following system and method based on shadow learning for fully mechanized coal mining face

Abstract

The application provides a fully-mechanized coal mining face hydraulic support adaptive following machine system and method based on shadow learning, and belongs to the field of intelligent mining of fully-mechanized coal mining faces. The system solves the problems of lack of adaptability of standardized parameter mode and lack of optimization ability of memory mode in the existing automatic following machine control of hydraulic supports. The system comprises a shadow following machine module for taking a current process parameter set as input, synchronously running an automatic following machine simulation process, and generating a shadow simulation data stream; a double-track acquisition module for synchronously acquiring manual operation data and the shadow simulation data stream when a worker manually moves a support; a time alignment module; a multi-dimensional deviation extraction module for extracting multi-dimensional deviations of the double-track data stream after time alignment; a deviation attribution module; a parameter optimization module; an adaptive parameter library module; and an iterative control module. The application is applied to the adaptive following machine of the fully-mechanized coal mining face hydraulic support.

Description

Technical Field

[0001] This application relates to the field of intelligent mining technology for fully mechanized coal mining faces, and in particular to an adaptive following system and method for hydraulic supports in fully mechanized coal mining faces based on shadow learning. Background Technology

[0002] Automatic hydraulic support following control is one of the core components of intelligent mining in fully mechanized longwall faces. In automatic following mode, the hydraulic support automatically performs actions such as lowering the support, moving the support, raising the support, and pushing the conveyor according to the position and direction of the coal mining machine and preset following process parameters, so as to achieve following support for the coal mining machine.

[0003] Currently, the automatic follow-up control of hydraulic supports in working faces mainly adopts a standardized parameter mode, which involves configuring a standard parameterized process flow according to the operating procedures. This mode is based on ideal working conditions and lacks adaptability. In actual production, factors such as coal seam geological conditions, roof conditions, and coal mining machine speed are constantly changing. Fixed parameters often lead to incomplete actions or low efficiency, affecting follow-up efficiency and support safety. Summary of the Invention

[0004] To address the issues of lack of adaptability in standardized parameter modes and lack of optimization capability in memory modes in existing automatic hydraulic support following control systems, this application proposes an adaptive following system and method for hydraulic supports in fully mechanized mining faces based on shadow learning. This system can both draw on the essence of manual operation experience and systematically optimize the following process parameters.

[0005] The technical solution adopted in this application is: an adaptive following system for hydraulic supports in fully mechanized mining faces based on shadow learning, comprising:

[0006] Shadow Tracking Module: Deployed in the backend of the hydraulic support controller, used to track the current set of process parameters. As input, an automated simulation process is run synchronously to generate a shadow simulation data stream. No actual control commands are output.

[0007] Dual-track data acquisition module: Used to simultaneously acquire manual operation data streams when workers manually move the machine along with the support frame. and shadow simulation data stream ;

[0008] Time alignment module: Used to align manual operation data streams with the position of the coal mining machine as the time anchor point. and shadow simulation data stream Perform time alignment;

[0009] Multidimensional Deviation Extraction Module: Used to extract multidimensional deviations from time-aligned dual-track data streams on an action-by-action basis and construct a deviation vector;

[0010] Deviation Attribution Module: Used to perform attribution analysis on deviation vectors, distinguish between operating condition adaptive deviations and operational arbitrariness deviations, and selectively map operating condition adaptive deviations to the corresponding process parameter adjustment amounts;

[0011] Parameter optimization module: used to generate optimized process parameter sets based on selective parameter mapping. ;

[0012] Adaptive parameter library module: used to store and manage operating condition-parameter mapping tables. Parameter retrieval and inference are performed based on real-time operating condition characteristics;

[0013] Iterative control module: Used to control the number of iterations in shadow learning, determine convergence conditions, and output the final working condition-parameter mapping table. .

[0014] Furthermore, the shadow tracking module incorporates a virtual execution model of the hydraulic support, calculating and simulating motion parameters based on the valve-controlled cylinder dynamics equations.

[0015] Furthermore, the deviation dimensions in the deviation vector include timing deviation, duration deviation, travel deviation, pressure deviation, and sequence deviation. The number of columns in the deviation vector is the number of all extracted specific deviation sub-items in a follow-up action cycle.

[0016] Furthermore, the data collected by the dual-track acquisition module includes data from the hardware sensing layer composed of pressure sensors, stroke sensors, and attitude sensors, as well as data from the operation command log recording unit.

[0017] Furthermore, the deviation attribution module includes a deviation-operating condition correlation calculation unit and a deviation coefficient of variation calculation unit. The deviation-operating condition correlation calculation unit is used to attribute each deviation dimension based on correlation and significance to determine the source of deviation. The deviation coefficient of variation calculation unit is used to calculate the coefficient of variation for each deviation dimension.

[0018] Furthermore, the parameter optimization module includes a condition-parameter mapping function fitting unit and a parameter update unit; wherein the condition-parameter mapping function fitting unit is used to map the condition adaptation deviation into the adjustment amount of the corresponding process parameter, and the parameter update unit is used to generate an optimized process parameter set based on the selective parameter mapping result. .

[0019] Furthermore, the adaptive parameter library module includes a similarity matching unit and a real-time fine-tuning unit; wherein the similarity matching unit is used to calculate the similarity between the real-time operating condition feature vector and the operating condition feature vector of each entry in the operating condition-parameter mapping table, and select the parameter set corresponding to the entry with the highest similarity. As the current follow-up control parameter, the real-time fine-tuning unit is used to continuously monitor the stroke feedback and pressure feedback of each hydraulic support during automatic follow-up execution. When the actual execution deviation exceeds the preset tolerance, the parameters are fine-tuned online according to the real-time deviation.

[0020] An adaptive following method for hydraulic supports in fully mechanized mining faces based on shadow learning, employing the aforementioned system, includes the following steps:

[0021] S1: Shadow machine initialization;

[0022] Configuration standard on-site process parameter set ,in For the first Each following process parameter includes following distance, following direction, following action, action duration, and action target parameters;

[0023] Switch the hydraulic support control system to manual follow-up mode and simultaneously activate the shadow follow-up module. The shadow follow-up module uses the standard follow-up process parameter set. Using the initial parameters, an automatic simulation process runs synchronously in the background;

[0024] S2: Simultaneous acquisition of dual-track data and time alignment;

[0025] During the manual machine-assisted frame relocation process, the following two tracks of data are collected simultaneously:

[0026] Manually operate the data stream : Real-time recording of data generated by workers' actual operations;

[0027] Shadow Simulation Data Stream The synchronous recording shadow tracking module records the current process parameter set. Data generated during simulation;

[0028] S3: Multidimensional bias extraction;

[0029] For each follow-up action cycle of each hydraulic support, the time-aligned manual operation data stream and shadow simulation data stream Perform action-by-action multidimensional deviation extraction and construct a deviation vector:

[0030] S4: Deviation-driven parameter optimization;

[0031] Based on the deviation vector, an optimized set of follow-up process parameters is generated using a deviation-driven parameter optimization strategy. Specifically, it includes:

[0032] S41: Deviation Attribution Analysis: Attribute the deviation to each dimension and determine the source of the deviation;

[0033] When the deviation is significantly correlated with the characteristic variables of the operating conditions, it is judged as an operating condition adaptation deviation; when the deviation is not significantly correlated with the characteristic variables of the operating conditions and the variance exceeds the preset value, it is judged as an operational arbitrariness deviation.

[0034] S42: Selective parameter mapping:

[0035] For operating condition adaptation deviations, they are mapped to the corresponding adjustment amounts of process parameters. ,in For the working condition feature vector, This is a condition-parameter mapping function, representing the condition under which parameters are mapped. lower deviation For on-site process parameters The direction and magnitude of the correction;

[0036] For arbitrary operational deviations, no parameter mapping is performed; the original parameter values ​​are maintained.

[0037] S43: Parameter Update: Based on the selective parameter mapping results, generate an optimized set of process parameters.

[0038] ;

[0039] in For learning rate, The overall adjustment vector generated for selective parameter mapping;

[0040] S5: Adaptive parameter generation;

[0041] Using the optimized process parameter set Replace the current process parameter set This will serve as the benchmark parameter for the next round of shadow tracking.

[0042] Build the working condition-parameter mapping table , feature vectors of different working conditions Corresponding optimization parameter set Stored as a mapping entry;

[0043] In automatic follow-up mode, based on the real-time collected operating condition feature vectors... From the operating condition-parameter mapping table Search for the most matching parameter set As the current operating condition-adaptive operating control parameter, it enables adaptive operating condition control.

[0044] S6: Iterative convergence determination;

[0045] Repeat steps S2 to S5 until any convergence condition is met, then output the final working condition-parameter mapping table. As an adaptive following process parameter library.

[0046] Furthermore, manually manipulate the data stream. The recorded data includes the trigger time, completion time, actual action duration, actual stroke, actual pressure change curve, action interval, and action sequence for each action of each hydraulic support.

[0047] Shadow Simulation Data Stream The recorded data includes the simulated action trigger time, simulated completion time, simulated action duration, simulated stroke, simulated pressure change curve, simulated action interval, and simulated action sequence.

[0048] Furthermore, the convergence conditions in step S6 include:

[0049] continuous During the round of iteration, the norm of the deviation vector is less than the preset convergence threshold;

[0050] Operating Condition-Parameter Mapping Table Current operating condition items in the continuous The parameter change rate is less than the preset stability threshold during the next update. ;

[0051] Reaching the maximum number of iterations .

[0052] The advantages of this application over the prior art are as follows:

[0053] (1) Breaking through the adaptive bottleneck of standardized parameter mode: The existing standardized parameter mode configures fixed parameters according to the operation procedure, which cannot adapt to changes in working conditions. This application continuously accumulates experience in optimizing working condition-adaptive parameters through shadow learning, and constructs a working condition-parameter mapping table, so that the following parameters can be dynamically adjusted according to the working conditions, solving the two core problems of "inadequate action" and "excessive action time".

[0054] (2) Overcoming the shortcomings of memory patterns: Existing memory patterns directly reproduce manual operations and cannot distinguish between reasonable experience and arbitrary deviations in manual operations. This application selectively retains the experience contained in the working condition adaptation deviations and filters out the arbitrary deviations in operations through deviation attribution analysis, thus realizing the "taking the essence and discarding the dross" of manual experience, rather than simply copying it.

[0055] (3) Innovation of the shadow parallel architecture: This application is the first to create a dual-track parallel architecture of "manual operation + shadow simulation". While the worker is manually following the machine, the background shadow following module runs automatically to simulate the machine. No additional demonstration tool or special training stage is required, and the normal production rhythm is not affected. The learning process is naturally embedded in the production process.

[0056] (4) Deviation-driven parameter optimization paradigm: Unlike existing learning methods that focus on action sequence generation or conflict detection, this application uses the multidimensional deviation between manual operation and shadow simulation as the optimization driving force. The deviation is both the basis for problem diagnosis and the direction guide for parameter adjustment, forming a closed-loop optimization path of "deviation discovery → attribution analysis → selective correction". Attached Figure Description

[0057] The following description, in conjunction with the accompanying drawings, further illustrates this application:

[0058] Figure 1 A flowchart illustrating the method provided in this application embodiment;

[0059] Figure 2 This is a system block diagram provided for an embodiment of this application. Detailed Implementation

[0060] like Figure 1 and Figure 2 As shown, this application provides an adaptive following system for hydraulic supports in fully mechanized mining faces based on shadow learning, comprising:

[0061] (1) Shadow Follow-up Module: Deployed in the background of the hydraulic support controller, used to follow the current set of process parameters. As input, an automated simulation process is run synchronously to generate a shadow simulation data stream. It does not output actual control commands; the shadow following module has a built-in virtual execution model of hydraulic support, which calculates simulated motion parameters based on the valve-controlled cylinder dynamics equation;

[0062] (2) Dual-track acquisition module: used to simultaneously acquire manual operation data streams when workers manually move the frame along with the machine. and shadow simulation data stream It includes a hardware sensing layer consisting of pressure sensors, stroke sensors, and attitude sensors, as well as an operation command log recording unit;

[0063] (3) Time alignment module: used to align the manual operation data stream with the position of the coal mining machine as the time anchor point. and shadow simulation data stream Perform time alignment;

[0064] (4) Multidimensional Deviation Extraction Module: Used to extract multidimensional deviations from the time-aligned dual-track data streams on an action-by-action basis and construct a deviation vector. The deviation dimensions include timing deviation, duration deviation, travel deviation, pressure deviation, and sequence deviation;

[0065] (5) Deviation Attribution Module: Used to perform attribution analysis on deviation vectors, distinguish between working condition adaptive deviations and operational arbitrary deviations, including a deviation-working condition correlation calculation unit and a deviation variation coefficient calculation unit; wherein the deviation-working condition correlation calculation unit is used to calculate the correlation coefficient between the deviation dimension and the working condition characteristic variable, and to determine whether the deviation is a working condition adaptive deviation based on the correlation coefficient; the deviation variation coefficient calculation unit is used to calculate the variation coefficient of the deviation dimension in multiple iterations, and to determine whether the deviation is an operational arbitrary deviation based on the correlation coefficient and the variation coefficient.

[0066] (6) Parameter optimization module: used to generate an optimized set of process parameters based on selective parameter mapping. It includes a condition-parameter mapping function fitting unit and a parameter update unit; the condition-parameter mapping function fitting unit is used to map the condition adaptation deviation to the corresponding process parameter adjustment amount, and the parameter update unit is used to generate an optimized process parameter set based on the selective parameter mapping result. ;

[0067] (7) Adaptive parameter library module: used to store and manage operating condition-parameter mapping tables. The system performs parameter retrieval and inference based on real-time operating condition characteristics, including a similarity matching unit and a real-time fine-tuning unit. The similarity matching unit calculates the similarity between the real-time operating condition feature vector and the operating condition feature vectors of each entry in the operating condition-parameter mapping table, and selects the parameter set corresponding to the entry with the highest similarity. As the current follow-up control parameters, the real-time fine-tuning unit is used to continuously monitor the stroke feedback and pressure feedback of each hydraulic support during automatic follow-up execution. When the actual execution deviation exceeds the preset tolerance, the parameters are fine-tuned online according to the real-time deviation.

[0068] (8) Iterative Control Module: Used to control the iteration rounds of shadow learning, determine the convergence condition, and output the final working condition-parameter mapping table. .

[0069] This application also provides an adaptive following method for hydraulic supports in fully mechanized mining faces based on shadow learning, which includes the following steps based on the above system:

[0070] S1: Shadow machine initialization;

[0071] Configuration standard on-site process parameter set ,in For the first Each following process parameter includes following distance, following direction, following action, action duration, and action target parameters;

[0072] Switch the hydraulic support control system to manual follow-up mode and simultaneously activate the shadow follow-up module. The shadow follow-up module uses the standard follow-up process parameter set. Using the initial parameters, an automatic simulation process runs synchronously in the background;

[0073] The shadow tracking module does not output actual control commands to the hydraulic support actuator; it only performs simulation calculations and records simulation operation data.

[0074] S2: Dual-track synchronous data acquisition;

[0075] During the manual machine-assisted frame relocation process, the following two tracks of data are collected simultaneously:

[0076] Manually operate the data stream It records data generated by workers' actual operations in real time, including the trigger time, completion time, actual action duration, actual stroke, actual pressure change curve, action interval, and action sequence of each hydraulic support action.

[0077] Shadow Simulation Data Stream The synchronous recording shadow tracking module records the current process parameter set. The data generated by the simulation includes the simulated action trigger time, simulated completion time, simulated action duration, simulated stroke, simulated pressure change curve, simulated action interval, and simulated action sequence;

[0078] Using the position of the coal mining machine as the time anchor point, the manual operation data stream was analyzed. and shadow simulation data stream Perform time alignment to ensure comparability between the two under the same coal mining machine position reference.

[0079] S3: Multidimensional bias extraction;

[0080] For each follow-up action cycle of each hydraulic support, the time-aligned manual operation data stream and shadow simulation data stream Perform action-by-action multidimensional deviation extraction and construct a deviation vector:

[0081] In the formula This represents the number of all extracted specific deviation sub-items within a single follow-up action loop; where each deviation dimension includes:

[0082] Timing deviation : The time difference between manual operation and shadow simulation in triggering actions and the time difference in action completion;

[0083] Duration deviation : The difference in duration between each action phase of manual operation and shadow simulation;

[0084] travel deviation : Difference in the moving stroke and the pushing stroke between manual operation and shadow simulation;

[0085] Pressure deviation : Difference in column lowering pressure and initial support force between manual operation and shadow simulation;

[0086] sequence deviation The difference in the sequence of actions between manual operation and shadow simulation is quantized using edit distance.

[0087] For example, suppose each hydraulic support has four action stages (lowering the column, moving the support, raising the column, and pushing the conveyor):

[0088] Timing deviation: There is a trigger time difference for each stage → 4;

[0089] Duration deviation: There are →4 in each stage;

[0090] Travel deviation: Shifting frame + pushing slide → 2;

[0091] Pressure deviation: Lowering column + Raising column → 2;

[0092] Sequence bias: 1 overall;

[0093] The total is m = 4 + 4 + 2 + 2 + 1 = 13.

[0094] S4: Deviation-driven parameter optimization;

[0095] Based on deviation vector An optimized set of follow-up process parameters is generated using a deviation-driven parameter optimization strategy. Specifically, it includes:

[0096] S41: Deviation Attribution Analysis: Attribute the deviation to each dimension and determine the source of the deviation.

[0097] Working condition adaptive deviation: Deviations caused by workers actively adjusting their operations in response to actual working conditions (roof breakage, floor undulation, changes in coal mining machine speed, etc.). These deviations contain working condition adaptive experience.

[0098] Random operational deviations: These are deviations caused by worker operating habits, fatigue, or distraction, and they have no optimization value.

[0099] The attribution judgment method is as follows: when the deviation is significantly correlated with the working condition characteristic variables (coal mining machine speed, adjacent frame pressure, roof condition, etc.), it is judged as a working condition adaptation deviation; when the deviation is not significantly correlated with the working condition characteristic variables and the variance is large, it is judged as an operational arbitrariness deviation.

[0100] Specifically:

[0101] Constructing the deviation-operating condition correlation matrix ,element For the first The first deviation dimension and the second Correlation coefficients of individual working condition characteristic variables;

[0102] Set a significance threshold ,when At that time, the judgment of the first The first deviation dimension and the second The characteristics of each working condition are significantly correlated, and this deviation is a working condition adaptation deviation.

[0103] When all operating condition characteristic variables are related to the first The correlation coefficients of all deviation dimensions satisfy the following conditions. Furthermore, the coefficient of variation of this bias dimension across multiple iterations. When this occurs, it is determined to be an operational arbitrariness deviation; among which the coefficient of variation... This is the coefficient of variation in statistics, calculated using the following formula: Standard deviation Divide by the mean , is used to measure the degree of dispersion of data relative to the mean.

[0104] In this embodiment, a threshold for the coefficient of variation is set. .

[0105] S42: Selective parameter mapping:

[0106] For operating condition adaptation deviations, they are mapped to the corresponding adjustment amounts of process parameters. ,in For the working condition feature vector, This is a condition-parameter mapping function, representing the condition under which parameters are mapped. lower deviation For parameters The direction and magnitude of the correction;

[0107] For arbitrary operational deviations, no parameter mapping is performed; the original parameter values ​​are maintained.

[0108] S43: Parameter Update: Based on the selective parameter mapping results, generate an optimized set of process parameters.

[0109] ;

[0110] in For learning rate, The overall adjustment vector generated for selective parameter mapping.

[0111] S5: Adaptive parameter generation;

[0112] Using the optimized process parameter set Replace the current process parameter set This will serve as the benchmark parameter for the next round of shadow tracking.

[0113] Build a working condition-parameter mapping table , feature vectors of different working conditions Corresponding optimization parameter set Stored as a mapping entry;

[0114] In automatic follow-up mode, based on the real-time collected operating condition feature vectors... From the operating condition-parameter mapping table Search for the most matching parameter set As the current operating condition control parameter, it enables adaptive operation and follow-up.

[0115] Preferably, the operating condition-parameter mapping table The retrieval method is as follows:

[0116] Calculate the real-time operating condition feature vector The feature vectors of each entry in the working condition-parameter mapping table Similarity between them:

[0117] ;

[0118] Select the parameter set corresponding to the item with the highest similarity. As the current follow-up control parameter; when the highest similarity is lower than the matching threshold At that time, using the current standard parameter set Based on this, using the working condition-parameter mapping function Perform real-time inference to generate adaptation parameters.

[0119] Furthermore, the automatic follow-up mode also includes a real-time fine-tuning mechanism: during automatic follow-up execution, the stroke feedback and pressure feedback of each hydraulic support are continuously monitored, and when the actual execution deviation exceeds the preset tolerance, the parameters are fine-tuned online based on the real-time deviation.

[0120] ;

[0121] in To fine-tune the gain coefficient, For the target value, This is the actual feedback value.

[0122] S6: Iterative convergence determination;

[0123] Repeat steps S2 to S5 until one of the following convergence conditions is met:

[0124] continuous In rounds of iteration, the deviation vector norm Less than the preset convergence threshold ;

[0125] Operating Condition-Parameter Mapping Table Current operating condition items in the continuous The parameter change rate is less than the preset stability threshold during the next update. ;

[0126] Reaching the maximum number of iterations .

[0127] After convergence, output the final working condition-parameter mapping table. As an adaptive following process parameter library.

[0128] A specific embodiment of this application also provides a computer device comprising: a memory and a processor, wherein the memory stores instructions executable on the processor. When the processor executes the instructions, it implements the methods described in the above embodiments. The number of memories and processors can be one or more. This computer device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The computer device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown in this application, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application as described and / or claimed herein.

[0129] The computer device may also include a communication interface for communicating with external devices and exchanging data. The devices are interconnected using different buses and can be mounted on a common motherboard or otherwise installed as needed. The processor processes instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as a display device coupled to the interface). In other embodiments, multiple processors and / or multiple buses can be used with multiple memories, if desired. Similarly, multiple electronic devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). The bus can be divided into address buses, data buses, control buses, etc.

[0130] Optionally, in a specific implementation, if the memory, processor, and communication interface are integrated on a single chip, then the memory, processor, and communication interface can communicate with each other through an internal interface.

[0131] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. General-purpose processors can be microprocessors or any conventional processor. It is worth noting that the processor can be a processor supporting advanced RISC machines (ARM) architecture.

[0132] This application provides a computer-readable storage medium (such as the memory described above) that stores computer instructions, which, when executed by a processor, implement the method provided in this application.

[0133] Optionally, the memory may include a stored program area and a stored data area, wherein the stored program area may store the operating system and application programs required for at least one function; the stored data area may store data created based on the use of the computer device for mapping. Furthermore, the memory may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory may optionally include memory remotely located relative to the processor, which can be connected to the computer device for mapping via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0134] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. An adaptive following system for hydraulic supports in fully mechanized mining faces based on shadow learning, characterized in that: include: Shadow Tracking Module: Deployed in the backend of the hydraulic support controller, used to track the current set of process parameters. As input, an automated simulation process is run synchronously to generate a shadow simulation data stream. No actual control commands are output. Dual-track data acquisition module: Used to simultaneously acquire manual operation data streams when workers manually move the machine along with the support frame. and shadow simulation data stream ; Time alignment module: Used to align manual operation data streams with the position of the coal mining machine as the time anchor point. and shadow simulation data stream Perform time alignment; Multidimensional Deviation Extraction Module: Used to extract multidimensional deviations from time-aligned dual-track data streams on an action-by-action basis and construct a deviation vector; Deviation Attribution Module: Used to perform attribution analysis on deviation vectors, distinguish between operating condition adaptive deviations and operational arbitrariness deviations, and selectively map operating condition adaptive deviations to the corresponding process parameter adjustment amounts; Parameter optimization module: used to generate optimized process parameter sets based on selective parameter mapping. ; Adaptive parameter library module: used to store and manage operating condition-parameter mapping tables. Parameter retrieval and inference are performed based on real-time operating condition characteristics; Iterative control module: Used to control the number of iterations in shadow learning, determine convergence conditions, and output the final working condition-parameter mapping table. .

2. The adaptive following system for hydraulic supports in fully mechanized mining faces based on shadow learning according to claim 1, characterized in that: The shadow tracking module has a built-in virtual execution model of the hydraulic support, which calculates and simulates motion parameters based on the dynamic equations of the valve-controlled cylinder.

3. The adaptive following system for hydraulic supports in fully mechanized mining faces based on shadow learning according to claim 1, characterized in that: The deviation dimensions in the deviation vector include timing deviation, duration deviation, travel deviation, pressure deviation, and sequence deviation. The number of columns in the deviation vector is the number of all extracted specific deviation sub-items in a follow-up action cycle.

4. The adaptive following system for hydraulic supports in fully mechanized mining faces based on shadow learning according to claim 1, characterized in that: The data collected by the dual-track acquisition module includes data from the hardware sensing layer consisting of pressure sensors, stroke sensors, and attitude sensors, as well as data from the operation command log recording unit.

5. The adaptive following system for hydraulic supports in fully mechanized mining faces based on shadow learning according to claim 1, characterized in that: The deviation attribution module includes a deviation-operating condition correlation calculation unit and a deviation coefficient of variation calculation unit. The deviation-operating condition correlation calculation unit is used to attribute each deviation dimension based on correlation and significance to determine the source of deviation. The deviation coefficient of variation calculation unit is used to calculate the coefficient of variation for each deviation dimension.

6. The adaptive following system for hydraulic supports in fully mechanized mining faces based on shadow learning according to claim 1, characterized in that: The parameter optimization module includes a condition-parameter mapping function fitting unit and a parameter update unit. The condition-parameter mapping function fitting unit maps the condition adaptation deviation to the corresponding process parameter adjustment amount, and the parameter update unit generates an optimized process parameter set based on the selective parameter mapping results. .

7. The adaptive following system for hydraulic supports in fully mechanized mining faces based on shadow learning according to claim 1, characterized in that: The adaptive parameter library module includes a similarity matching unit and a real-time fine-tuning unit. The similarity matching unit calculates the similarity between the real-time operating condition feature vector and the operating condition feature vector of each entry in the operating condition-parameter mapping table, and selects the parameter set corresponding to the entry with the highest similarity. As the current follow-up control parameter, the real-time fine-tuning unit is used to continuously monitor the stroke feedback and pressure feedback of each hydraulic support during automatic follow-up execution. When the actual execution deviation exceeds the preset tolerance, the parameters are fine-tuned online according to the real-time deviation.

8. An adaptive following method for hydraulic supports in fully mechanized mining faces based on shadow learning, characterized in that: Using the system as described in any one of claims 1-7, the method includes the following steps: S1: Shadow machine initialization; Configuration standard on-site process parameter set ,in For the first Each following process parameter includes following distance, following direction, following action, action duration, and action target parameters; Switch the hydraulic support control system to manual follow-up mode and simultaneously activate the shadow follow-up module. The shadow follow-up module uses the standard follow-up process parameter set. Using the initial parameters, an automatic simulation process runs synchronously in the background; S2: Simultaneous acquisition of dual-track data and time alignment; During the manual machine-assisted frame relocation process, the following two tracks of data are collected simultaneously: Manually operate the data stream : Real-time recording of data generated by workers' actual operations; Shadow Simulation Data Stream The synchronous recording shadow tracking module records the current process parameter set. Data generated during simulation; S3: Multidimensional bias extraction; For each follow-up action cycle of each hydraulic support, the time-aligned manual operation data stream and shadow simulation data stream Perform action-by-action multidimensional deviation extraction and construct a deviation vector: S4: Deviation-driven parameter optimization; Based on the deviation vector, an optimized set of follow-up process parameters is generated using a deviation-driven parameter optimization strategy. Specifically, it includes: S41: Deviation Attribution Analysis: Attribute the deviation to each dimension and determine the source of the deviation; When the deviation is significantly correlated with the characteristic variables of the operating conditions, it is judged as an operating condition adaptation deviation; when the deviation is not significantly correlated with the characteristic variables of the operating conditions and the variance exceeds the preset value, it is judged as an operational arbitrariness deviation. S42: Selective parameter mapping: For operating condition adaptation deviations, they are mapped to the corresponding adjustment amounts of process parameters. ,in For the working condition feature vector, This is a condition-parameter mapping function, representing the condition under which parameters are mapped. lower deviation For on-site process parameters The direction and magnitude of the correction; For arbitrary operational deviations, no parameter mapping is performed; the original parameter values ​​are maintained. S43: Parameter Update: Based on the selective parameter mapping results, generate an optimized set of process parameters. ； in For learning rate, The overall adjustment vector generated for selective parameter mapping; S5: Adaptive parameter generation; Using the optimized process parameter set Replace the current process parameter set This will serve as the benchmark parameter for the next round of shadow tracking. Build a working condition-parameter mapping table , feature vectors of different working conditions Corresponding optimization parameter set Stored as a mapping entry; In automatic follow-up mode, based on the real-time collected operating condition feature vectors... From the operating condition-parameter mapping table Search for the most matching parameter set As the current operating condition-adaptive operating control parameter, it enables adaptive operating condition control. S6: Iterative convergence determination; Repeat steps S2 to S5 until any convergence condition is met, then output the final working condition-parameter mapping table. As an adaptive following process parameter library.

9. The adaptive following method for hydraulic supports in fully mechanized mining faces based on shadow learning according to claim 8, characterized in that: Manually operate the data stream The recorded data includes the trigger time, completion time, actual action duration, actual stroke, actual pressure change curve, action interval, and action sequence for each action of each hydraulic support. Shadow Simulation Data Stream The recorded data includes the simulated action trigger time, simulated completion time, simulated action duration, simulated stroke, simulated pressure change curve, simulated action interval, and simulated action sequence.

10. The adaptive following method for hydraulic supports in fully mechanized mining faces based on shadow learning according to claim 8, characterized in that: The convergence conditions in step S6 include: continuous During the round of iteration, the norm of the deviation vector is less than the preset convergence threshold; Operating Condition-Parameter Mapping Table Current operating condition items in the continuous The parameter change rate is less than the preset stability threshold during the next update. ; Reaching the maximum number of iterations .

Images