A method and system for load control of steeply inclined coal seam support groups based on digital twins

Abstract

The present application belongs to the technical field of intelligent control of hydraulic support in fully mechanized coal mining face, and discloses a load regulation and control method and system for support group in steeply inclined coal seam based on digital twinning. By establishing a digital twinning model containing gravity component force coefficient in inclined direction and contact interface friction coefficient, virtual units are divided and compressed information is extracted. After field collection of stress observation values of multiple types of supports, abnormal values are corrected, compressed supports are screened and candidate areas are aggregated. After bidirectional matching and marking of regulation and control areas by numbering and compression time, load regulation and control is performed. The method is suitable for steeply inclined working conditions, improves regulation and control accuracy and timeliness, enhances support stability, and realizes intelligent and efficient regulation and control.

Description

Technical Field

[0001] This invention belongs to the field of intelligent control technology for hydraulic supports in fully mechanized coal mining faces, specifically relating to a method and system for load regulation of support groups in steeply inclined coal seams based on digital twins. Background Technology

[0002] The working environment of steeply inclined, high-extraction coal seam working faces is unique: Firstly, the coal seam has a large inclination angle, and the component of gravity along the inclination direction is significant. Rocks, backfill, or loose coal falling from the roof will slide or press down along the inclination direction, causing the upper side area of ​​the working face to be pressured first, followed by the lower side area. Secondly, the coal seam is highly mined, and the working face space has obvious undulations. Due to some areas being too high or some areas having hard roofs, long strip-shaped areas of roof hardness change will be formed on the working face, resulting in inconsistent load-bearing capacity of the same row of supports. Furthermore, the hydraulic supports, scraper conveyors, and roof affect each other. The impact of the scraper conveyor tilting up or down on the posture angle and stress of the supports will be amplified on steeply inclined working faces.

[0003] Due to the unique working environment of steeply inclined and high-extraction coal seam working faces, the actual load transfer exhibits the following characteristics: First, the load moves faster along the inclined direction of the coal seam than the working face advances. The load's force front will cross or skip the preset support pressure sequence, resulting in patchy area movement rather than sequential advancement of each support. Second, the load values ​​collected by each sensor are easily distorted by coal block impacts, support posture angle deviations, or hydraulic cable tension. The distorted load values ​​have similar characteristics to the suddenly arriving real load values.

[0004] However, the working logic of existing hydraulic support control technology is usually as follows: first, the load value is monitored; then, based on the load value, it is determined that the roof will press down; finally, the load is increased according to the pre-divided support area. This single-value triggered control method is based on the premise that the load will be transmitted sequentially according to the preset support arrangement order as the working face advances. For working faces with gentle coal seams, small inclination angles, and stable roofs, this premise is basically valid. However, for emergency inclined working faces with high mining heights, this premise will have the following problems: on the one hand, the front edge of the actual load moves across the preset support pressing sequence to the current support, while the sensor of the current support still collects the load value according to the preset load transmission order, causing the load value collection to lag, resulting in a lag in control commands; on the other hand, the load value collected by the sensor is interfered with by coal block impacts or sudden changes in the support's attitude angle, leading to the mistaken belief that a certain section of the support has been pressed down, and group load increases, delayed pressure holding, or limit operations are performed on it, amplifying the erroneous adjustment.

[0005] In summary, in the working scenario of steeply inclined and high-extraction coal seams, the existing hydraulic support control technology causes the dynamic load control commands of the support group to be delayed or erroneously triggered, and even amplifies erroneous adjustments in uncompressed support areas. Summary of the Invention

[0006] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0007] Firstly, a load control method for a group of supports in steeply inclined coal seams based on digital twins is proposed, including the following steps: establishing a digital twin model of the working face; introducing the gravity component coefficient along the dipping direction of the coal seam and the friction coefficient of each contact interface into the digital twin model; dividing the support group into multiple virtual units in the digital twin model; each virtual unit contains multiple supports along the extension direction of the working face; extracting the number and time of the virtual unit that is first subjected to pressure from the operation results of the digital twin model; for each support: continuously acquiring multiple types of observations reflecting the stress state of the support on-site, filtering out outlier segments in each type of observation, and then... Outlier segments in each type of observation are corrected; each pressure support is selected based on the corrected multi-type observations; adjacent pressure supports are aggregated into candidate regions along the extension direction of the working face; the number and pressure time of the candidate region that was pressured first are obtained; if the number of the virtual unit that was pressured first is the same as the number of the candidate region that was pressured first, and the difference between the pressure time of the virtual unit that was pressured first and the pressure time of the candidate region that was pressured first is less than or equal to the stroke time of the coal mining machine, the candidate region that was pressured first is marked as a group control region; load control instructions for the group control region are generated, and load control is performed on the supports in the group control region according to the load control instructions.

[0008] Secondly, a load control system for steeply inclined coal seam support groups based on digital twins is proposed, including:

[0009] The model building module is used to create a digital twin model of the working face; the digital twin model incorporates the gravity component coefficient along the dipping direction of the coal seam and the friction coefficient of each contact interface.

[0010] The group division module is used to divide the support group into multiple virtual units; each virtual unit contains multiple supports along the extension direction of the working surface.

[0011] The data extraction module is used to extract the number and pressure time of the first virtual unit to be pressured from the running results of the digital twin model, as well as to obtain the number and pressure time of the candidate region to be pressured first.

[0012] The data acquisition module is used to continuously acquire various observation values ​​on site to reflect the stress state of the support.

[0013] The data filtering module is used to filter out outlier segments in each type of observation.

[0014] The data correction module is used to correct outlier segments in each type of observation;

[0015] The support screening module is used to select each pressure support based on the corrected multi-class observation values;

[0016] The support aggregation module is used to aggregate adjacent pressure supports into candidate regions along the extension direction of the working surface.

[0017] The area marking module is used to mark the first pressured candidate area as a group control area when the number of the first pressured virtual unit is consistent with the number of the first pressured candidate area, and the difference between the pressure time of the first pressured virtual unit and the pressure time of the first pressured candidate area is less than or equal to the stroke time of the coal mining machine.

[0018] The instruction generation module is used to generate load control instructions for the group control area;

[0019] The load control module is used to control the load on the support structure in the group control area according to the load control command.

[0020] Compared with existing technologies, this invention has the following advantages and beneficial effects: It establishes a digital twin model of the working face and introduces the gravity component coefficient along the coal seam's dip direction and the friction coefficient of each contact interface, enabling the model to accurately replicate the slip characteristics and contact transmission law of the load along the dip direction under steeply inclined conditions, providing a physical basis that fits the actual situation on site for subsequent stress prediction; it divides the support group into multiple virtual units containing multiple supports along the working face's extension direction, adapting to the "sheet-like transmission" characteristic of loads on steeply inclined working faces, reducing the complexity of model prediction, and achieving accurate positioning of the stress area, avoiding the control range deviation caused by traditional whole-row zoning; it extracts the number and pressure time of the first pressure-bearing virtual unit from the digital twin model, providing a clear prediction benchmark for determining the stress area on site, and building a comparison bridge between "virtual prediction and on-site verification"; it continuously acquires multiple observation values ​​reflecting the stress state of the supports on site, and by screening and correcting outlier segments, it effectively eliminates false interference signals such as coal block impact and cable disturbance, ensuring the accuracy of observation data. The authenticity and continuity of the data provide reliable data support for the accurate identification of pressure supports. Based on the corrected multi-class observations, pressure supports are screened and aggregated into candidate areas along the extension direction, which aligns with the spatial correlation of load transmission in steeply inclined working faces. This avoids the positioning deviation of the stress area caused by misjudgment of single-support signals, accurately locking the actual area that is first subjected to pressure on site. By marking the group control area with the bidirectional matching conditions of "consistent numbering" and "pressure time difference ≤ coal mining machine stroke time," it ensures a high degree of consistency between the control area and the stress area predicted by the digital twin model. This avoids both false triggering of control (controlling non-real stress areas) and control lag (missing the optimal support opportunity). Load control commands are generated to target the group control area, achieving coordinated support of the support group. This effectively resists the impact of sliding loads in steeply inclined working conditions, improving the stability and safety of the working face support. Simultaneously, it avoids energy waste and equipment wear caused by indiscriminate control, ultimately achieving intelligent, precise, and efficient load control of steeply inclined coal seam supports.

[0021] Thirdly, a computer device is proposed, comprising a memory, a processor, and a transceiver connected in sequence, wherein the memory is used to store a computer program, the transceiver is used to send and receive data, and the processor is used to read the computer program and execute a load control method for a steeply inclined coal seam support group based on digital twins as described in the first aspect.

[0022] Fourthly, a computer-readable storage medium is proposed, characterized in that the computer-readable storage medium stores instructions that, when executed on a computer, perform a load control method for a steeply inclined coal seam support group based on digital twins as described in the first aspect. Attached Figure Description

[0023] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0024] Figure 1 The present invention provides a method flow for load control of a steeply inclined coal seam support group based on digital twins, as described in Embodiment 1 of the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. The illustrative embodiments and descriptions of this invention are for illustrative purposes only and are not intended to limit the invention. The embodiments described below are some, but not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0026] In the following description, numerous specific details are set forth to provide a thorough understanding of the invention. However, it will be apparent to those skilled in the art that these specific details are not necessary to practice the invention. In other embodiments, well-known structures, materials, or methods are not specifically described to avoid obscuring the invention. Unless otherwise specified, the materials, instruments, and reagents used in the following embodiments are commercially available. Unless otherwise specified, the techniques used in the embodiments are conventional methods well known to those skilled in the art.

[0027] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0028] Example 1: In steeply inclined, high-extraction coal seams, the actual speed of load movement along the coal seam's dip direction is usually faster than the face's forward movement. The load's stress front may cross the preset support compression sequence. Simultaneously, field sensor values ​​are easily affected by factors such as coal block impacts, support posture deviations, and cable disturbances, leading to delayed execution or erroneous triggering of load dynamic control commands from the control system, and even amplifying erroneous adjustments in uncompressed support areas. To address these technical problems, this example provides a digital twin-based method for load control of support groups in steeply inclined coal seams. First, the predicted initial stress area of ​​the support is given on the digital twin side; then, the actual initial stress area of ​​the support is determined on the field side based on multi-source values ​​collected by sensors; finally, verification is performed between the "predicted initial stress area" and the "actual initial stress area." When both point to the same support area within a set spatial and temporal range, group control is triggered. This method verifies the consistency between the "prediction sequence of the digital twin model" and the "real-time measurement values ​​on site" in both time and space, and re-divides the stent groups online according to the actual position of the load, thereby accurately identifying the stent groups that truly need to be controlled and performing group control, avoiding delayed or erroneous adjustments triggered by the control commands of the stent groups.

[0029] Based on the above overall technical approach, the specific implementation steps of a digital twin-based method for load control of steeply inclined coal seam support groups are as follows: Figure 1 As shown, it includes:

[0030] Step 1: Establish a digital twin model of the steeply inclined, high-extraction working face.

[0031] The purpose of this step is to construct a virtual model of the working face that reflects changes in coal seam dip angle, mining height, and roof hardness bands, providing a model basis for subsequent prediction of the area of ​​the support that will be stressed first. The specific modeling method is as follows:

[0032] Step 1.1: Collect the actual inclination angle of the working face, the extension distance of the working face along the coal seam, and the mining height data at different locations (such as the two ends, middle, fault, fold zone and roof hardness change area of ​​the working face).

[0033] Step 1.2: Record the distribution range and location of soft-capped rock layers (such as shale and mudstone), hard-capped rock layers (such as sandstone and granite), and composite-capped rock layers (rock layers with alternating soft-capped and hard-capped rock layers).

[0034] Step 1.3: Collect the structural and performance parameters of the hydraulic support. The structural parameters include: top beam length, base dimensions, number and spacing of columns, and side guard plate width. The performance parameters include: rated support force, maximum column extension length, maximum top beam rotation angle, maximum extendable length of the top beam, side guard plate limit range, and the connection and support relationship between the hydraulic support and the scraper conveyor (the hydraulic support is connected to the scraper conveyor via a pushing jack, and simultaneously supports the scraper conveyor to prevent it from sliding along the inclined direction).

[0035] Step 1.4: Import the data collected in Steps 1.1 to 1.3 above into a digital twin development platform (such as Unity or Unreal Engine) or a digital twin system (such as GE Digital Twin Experience or Siemens MindSphere) to generate a digital twin model of the steeply inclined, high-extraction working face.

[0036] Furthermore, to adapt the digital twin model to complex conditions such as steeply inclined and high-extraction coal seams where the load moves faster along the dipping direction than the face advances, and the load's stress front may cross or skip the preset support compression sequence, resulting in patchy area movement rather than sequential support advancement, this embodiment introduces a gravity component coefficient along the dipping direction of the coal seam and a friction coefficient for each contact interface into the digital twin model. This allows the area where the load is first stressed to move faster than the face advances during the operation of the digital twin model, and also allows the load to move in leaps along the dipping direction, thus not being limited to a fixed mode of sequential support advancement. The gravity component coefficient is denoted as Gsinα, where α is the measured dipping angle of the coal seam; the contact surfaces include: the contact surfaces between each caving body and / or backfill and the roof, and the contact surfaces of the supports.

[0037] It should be noted that introducing the gravity component coefficient along the coal seam dip direction and the friction coefficient of each contact interface into the digital twin model is key to ensuring that the "predicted first stressed support area" conforms to the load transfer characteristics of a steeply inclined, high-extraction working face. In a steeply inclined, high-extraction working face, in addition to being compressed as the working face advances, the caving body and backfill are also subjected to a gravity component along the dip direction. This gravity component propels the load to move faster, making the load's movement speed faster than the working face's advance speed. This is because the support cluster in a steeply inclined, high-extraction working face is affected by a tangential force along the coal seam dip direction. This force is generated by the rolling of roof debris and the component of the coal and rock mass's own weight along the dip angle, and is a key force affecting load transfer. On the one hand, tangential force pushes individual supports down the inclined direction. As the support cluster is interconnected through anti-tipping and anti-slip devices and side guards, the downward trend transforms into inter-support pushing force, causing the load to transfer from the upper supports to the lower supports. This results in a phenomenon where "the load on the upper supports is transmitted downwards, while the lower supports bear the superimposed load," which is a key reason why the lower-side supports in steeply inclined working faces often bear several times the load of the upper-side supports. On the other hand, tangential force also exacerbates the rotational tendency of the supports, leading to uneven stress on the support columns. This, in turn, causes asymmetrical load transfer between different columns of the same support and between adjacent supports, such as causing inter-support biting and pushing, further altering the load transfer path.

[0038] Step 2: In the digital twin model, the scaffold group is divided into multiple virtual units.

[0039] The purpose of this step is to adapt to the special working conditions of "jumping transmission" and "sheet distribution" of loads in steeply inclined and high-extraction working faces. On the one hand, in steeply inclined and high-extraction working faces, loads may skip 1-2 supports and act directly on the area below (such as when a falling body slides across a soft roof area). If virtual units are not divided and prediction is still performed sequentially for each support, the load path will deviate from the actual load path. However, by dividing virtual units, the digital twin model can directly predict a virtual unit that is stressed first, covering the characteristics of load jump transmission. On the other hand, the actual pressure state is not that a single support is stressed in isolation, but rather that a "sheet-like stress zone" is formed along the inclined direction (multiple supports are stressed at the same time). By dividing virtual units, the sheet-like stress zone can be directly mapped to "adjacent virtual unit combinations", avoiding the problem of unstressed supports being mistakenly included in the control due to fixed partitioning.

[0040] In addition, it should be noted in advance that: 1. When performing dynamic group division in the subsequent process, the temporary group is formed by expanding from the virtual unit that is first subjected to pressure. If virtual units are not divided, control can only be performed according to fixed zones. This will result in some supports located at the center of the fixed zone being controlled even when the center of pressure is at the edge of the fixed zone. By subdividing the support group with virtual units as the smallest unit, the temporary group can be aligned with the actual center of pressure, avoiding the incorrect control of the pressure-bearing supports. 2. Based on the explanation in point 1, when performing model calibration in the subsequent process, it is necessary to compare the deviation between the "actual force-bearing unit" and the "predicted force-bearing unit". If virtual units are not divided, the deviation can only be located in a large area, and local parameter correction cannot be achieved.

[0041] The specific method for dividing the support group into multiple virtual units is as follows:

[0042] A virtual unit is configured to contain multiple supports extending along the working surface. Taking a virtual unit containing two supports as an example: along the extension direction of the working surface, the first and second supports are classified as virtual unit 1, the third and fourth supports as virtual unit 2, and so on. The total number of virtual units is calculated based on the total number of supports in the support group, where N ≈ n ÷ 2. The last support in the support group is incorporated into the last virtual unit. For example, when the support group contains 100 supports, it can be divided into 50 virtual units, each containing two supports. When the support group contains 101 supports, it can also be divided into 50 virtual units, with the 50th virtual unit containing three supports numbered 99-101. Similarly, each virtual unit can be configured to contain at least two supports, preferably two supports.

[0043] Step 3: Extract the number and timing of the first virtual unit to be stressed from the results of the digital twin model.

[0044] By using dynamic simulation and finite element analysis algorithms in a digital twin model, combined with a roof pressure prediction model (such as a statistical model based on mine pressure observation data), the stress process of each virtual unit is simulated. When the simulated stress of a virtual unit (or a combination of adjacent virtual units) reaches a preset pressure threshold, it is determined to be a "stressed area," and the time when its simulation is triggered is the "pressure time." Finally, the model data interface outputs the number of the earliest pressure-bearing virtual unit and the pressure time. The virtual unit number is represented by a support number, indicating the location of the support area that is first pressure-bearing. For example, the support number corresponding to virtual unit 1 is (1-2), the support number corresponding to virtual unit 2 is (3-4), and so on.

[0045] The virtual unit number and compression time predicted in this step will be used as one of the conditions for subsequent time and space bidirectional matching determination.

[0046] Step 4: For each support: continuously acquire multiple types of observations on site to reflect the stress state of the support, filter out outlier segments in each type of observation, and correct each outlier segment.

[0047] Based on the above overview of the overall technical approach, this method needs to be validated between the "predicted first stressed support area" and the "actual first stressed support area". It is necessary to ensure the accuracy of the "first stressed support area" actually detected. However, the field sensor values ​​are easily interfered with. Therefore, it is necessary to detect whether there are any abnormal values ​​in the various observation values ​​collected, and to repair any abnormal values ​​detected.

[0048] The multiple observation values ​​used to reflect the stress state of the support described in this embodiment include at least: column pressure value, column extension length value, and pressure difference value between adjacent supports. These three types of observation values ​​collectively reflect the stress state of the support, forming a "direct perception + dual verification" system. The load on a steeply inclined working face is transmitted in a sheet-like manner; the area initially stressed is the coordinated stress of multiple adjacent supports, rather than a single support being stressed in isolation. Column pressure is a direct quantitative indicator of the hydraulic support's resistance to roof load, falling material pressure, and backfill pressure, and is the primary basis for determining whether the support is under stress. However, column pressure values ​​are easily affected by coal block impacts, cable pulling, and support posture deviations, producing abrupt waveforms similar to "real pressure" and failing to reflect the stress correlation between adjacent supports. Column extension length and column pressure are mechanically coupled. When real pressure occurs, the roof load compresses the support's top beam, causing the column to be compressed and the extension length to shorten, thus verifying the authenticity of the pressure value and eliminating interference from posture deviations and impacts. The pressure difference between adjacent supports reflects the continuity of load transmission, thus verifying the continuity of the stress area and eliminating interference from isolated single supports. Only when the pressure value shows an upward trend and the synergistic value is consistent with the change in pressure value (such as the shortening of the expansion length and the pressure difference between adjacent values ​​conforming to the load transfer law) can it be confirmed that the force is real, thus avoiding the one-sidedness of a single value.

[0049] In addition to the column pressure value, column extension length value, and pressure difference value between adjacent supports, tilt angle value, attitude angle value, and navigation angle value can also be collected to participate in the value consistency verification and physical law verification, so as to ensure that the determined "first stress area" is true and accurate.

[0050] Furthermore, the method for filtering out outlier segments in each category of observations is as follows:

[0051] Perform steps 4.1.1 to 4.1.3 for each type of observation.

[0052] Step 4.1.1: Establish a physical model of the observations and obtain the predicted value for each observation time based on the physical model.

[0053] (1) According to Hooke's Law, the physical model expression for the column pressure value is: P = k ×( L 0- L ).in, P This indicates the pressure on the support column. k Indicates the hydraulic stiffness coefficient. L 0 represents the initial length of the support column. L This indicates the actual length of the support column.

[0054] For each observation time: the hydraulic stiffness coefficient k Initial length of the support column L 0. Actual length of the support column L (Observed values) Input the physical model of the column pressure value, and output the predicted column pressure value.

[0055] (2) Based on the physical model expression of the column pressure value, the physical model expression of the column expansion length value is derived by reverse derivation as follows: L = L 0- P / k .

[0056] For each observation time: the hydraulic stiffness coefficient k Initial length of the support column L 0. Column pressure P (Observed values) Input the physical model of the column extension length value, and output the predicted actual length of the column.

[0057] (3) In steeply inclined coal seams, the supports not only bear the vertical load of the roof but also need to balance their own gravitational component along the inclination direction (due to the coal seam dip angle α). The pressure difference between adjacent supports is essentially the result of the combined effect of the gravitational component and frictional resistance during load transfer. According to the law of load transfer along the inclination direction of the coal seam, the physical model expression for the pressure difference between adjacent supports is: △P = ±( m × G ×sinα); where, m This represents the coefficient of friction at the contact interface. G The weight of the support is represented by α, and the measured inclination angle of the coal seam is α.

[0058] For each observation time: the coefficient of friction at the contact interface. m , self-weight of the support G The physical model takes the measured inclination angle α of the coal seam as input and the pressure difference between adjacent supports as input, and outputs the predicted pressure difference between adjacent supports as output.

[0059] Step 4.1.2: Obtain the residual between the observed value and the predicted value at each observation time, and establish a residual sequence.

[0060] The residual of the column pressure value = measured column pressure value - predicted column pressure value; sort the residuals of the column pressure values ​​according to the order of observation time to obtain the residual sequence of column pressure values.

[0061] The residual of the column expansion / contraction length value = measured column expansion / contraction length value - predicted column expansion / contraction length value; sort the residuals of the column expansion / contraction length values ​​according to the order of observation time to obtain the residual sequence of the column expansion / contraction length values.

[0062] The residual of the pressure difference between adjacent stents = measured pressure difference between adjacent stents - predicted pressure difference between adjacent stents; sort the residuals of the pressure difference between adjacent stents according to the order of observation time to obtain the residual sequence of the pressure difference between adjacent stents.

[0063] Step 4.1.3: Use a sliding time window to filter out outlier segments from the residual sequence that meet the anomaly detection criteria.

[0064] The anomaly detection criteria described in this embodiment are: the absolute value of the residual mean is greater than a first threshold, the variance of the residual is greater than a second threshold, and the number of aberrations in the residual is not equal to zero. Here, the number of aberrations refers to the number of times the absolute value of the residual is greater than three times the standard deviation of the residual. The first and second thresholds are statistical thresholds based on historical data; the first threshold can be taken as ± three times the standard deviation of the residual mean (denoted as ). m ±3 s , m The mean of the residuals, s The first threshold is the standard deviation of the residuals. If the residuals conform to a normal distribution, it indicates that the deviation between the predicted and observed values ​​is within a reasonable range. The second threshold can be the maximum value of the residual variance.

[0065] Step 4.1.3 indicates that an observed value can only be considered a normal value (undisturbed value) if the residuals between the predicted and observed values ​​meet the following conditions: the absolute value of the mean of the residuals is greater than the first threshold, the variance of the residuals is greater than the second threshold, and the number of abrupt changes in the residuals is not equal to zero. Observations that do not meet any of the above conditions are considered outliers and need to be corrected to restore the true observed values, resulting in a continuous and reliable value sequence for subsequent "judgment of the first stressed support area" and "spatiotemporal bidirectional matching".

[0066] Furthermore, the method for correcting outlier segments in each type of observation is as follows:

[0067] (1) Correcting outlier segments in the column pressure values. This includes the following steps:

[0068] Step 4.2.1: Obtain the residual column pressure value at each observation time within the outlier segment.

[0069] Step 4.2.2: Mark column pressure values ​​whose absolute value of residual column pressure value is greater than the third threshold as abnormal column pressure values.

[0070] Based on the pressure fluctuation characteristics of the support hydraulic system and the measurement accuracy of the pressure sensor, a calculation formula for the third threshold can be established to ensure that the third threshold covers both normal fluctuations in column pressure and accurately identifies abnormal column pressure values. The calculation formula for the third threshold is: E P =3 s P + d P ,in, E P This represents the third threshold. s P This represents the standard deviation of the column pressure residuals under historical normal operating conditions. d P This indicates the maximum permissible error of the pressure sensor. For example, if the standard deviation of the column pressure residual under historical normal operating conditions... s P =1 MPa, the maximum permissible error of the pressure sensor d P =0.5 MPa (the actual maximum permissible error of the pressure sensor can be taken from the pressure sensor manual), then the third threshold E P =3×1+0.5=3.5Mpa.

[0071] Step 4.2.3: Mark the observation time corresponding to the abnormal column pressure value.

[0072] Step 4.2.4: Extract the column pressure values ​​corresponding to the marked observation times for the front and rear i-frame supports respectively.

[0073] 'i' represents the support frame number. For example, for the outlier segment in the column pressure value of the 50th support frame, the marked outlier column pressure value is the column pressure value P50-140ms collected at 140ms. Then, the column pressure value collected at 140ms for the 49th support frame along the inclined direction is taken. P 49-140ms The column pressure value collected at 140ms along the inclined direction and the 51st support. P 49-140ms .

[0074] Step 4.2.5: Calculate the average pressure value based on the column pressure value extracted in Step 4.2.4.

[0075] Step 4.2.6: Determine whether the average pressure value falls within the physical constraint boundary of the column pressure value. If yes, replace the abnormal column pressure value with the average pressure value. If no, execute i=i+1 and return to step 4.2.4.

[0076] The physical constraint boundary for the column pressure value is [0.2]. F max , F max ], F max This refers to the rated support force of the support frame. The minimum effective output value of the hydraulic support pressure sensor is typically no less than 20% of the rated support force (as per the standard parameters in the pressure sensor's instruction manual). This applies when the column pressure value is below 0.2. F max The time indicates that the sensor cable is broken, the interface is loose, or the sensor itself is malfunctioning, and the observed column pressure value is invalid data; the rated support force of the hydraulic support is the maximum safe load-bearing value designed. Under normal stress, the load is stable at 80%-100% of the rated value. However, the characteristics of impact interference such as coal block impact and support collision are "instantaneous jump and rapid fall", and the peak value often exceeds the rated support force. When the column pressure value is higher than the time, it means that the column pressure value was observed under instantaneous impact interference or dangerous overload conditions, neither of which conforms to the characteristics of normal stress. Therefore, [0.2] F max , F max This serves as a standard for determining whether the calculated average pressure value can be used to replace abnormal column pressure values.

[0077] If the calculated average pressure falls within [0.2] F max , F max Within the range of [0.2], the average pressure value can be directly used to replace the abnormal column pressure value; if the calculated average pressure value does not fall within [0.2]... F max , F max If the pressure falls within the range of [0.2], then the number of front and rear supports is increased, the average pressure is calculated, and physical constraint boundaries are determined until the average pressure falls within [0.2]. F max , F max Within the range.

[0078] It should be noted that steps 4.2.1 to 4.2.6 above utilize the continuity of load transfer between adjacent supports and directly generate a substitute value for the column pressure that conforms to physical laws using the mean value theory. However, after multiple iterations, the calculated mean value of the column pressure for all supports may not fall within the range of [0.2]. F max , F max This situation falls within the scope of the technical solution, but its essence is caused by "abnormal load on the entire working surface" or "system-level signal failure". It requires a full-range abnormality judgment and handling of the hydraulic system and is not within the scope of this technical solution.

[0079] (2) Correcting outlier segments of the column extension / retraction length values. This includes the following steps:

[0080] Step 4.3.1: Obtain the column extension length residual at each observation time within the outlier segment.

[0081] Step 4.3.2: Mark the column extension length values ​​whose absolute value is greater than the fourth threshold as abnormal column extension length values.

[0082] Based on the length fluctuation range of the column's hydraulic response and the measurement accuracy of the displacement sensor, a calculation formula for the fourth threshold can be established to ensure that the fourth threshold covers both normal fluctuations in the column's telescopic length and accurately identifies abnormal telescopic length values. The calculation formula for the fourth threshold is: E L =3 s L + d L ,in, E L This represents the fourth threshold. s L This represents the standard deviation of the residuals in the column expansion and contraction length under normal historical operating conditions. d L This indicates the maximum permissible error of the displacement sensor. For example, if the standard deviation of the column extension length residual under historical normal operating conditions is... s L =1mm, the maximum permissible error of the displacement sensor d L =0.3mm (the actual maximum permissible error of the displacement sensor can be taken from the displacement sensor manual), then the fourth threshold E P =3×1+0.3=3.3mm.

[0083] Step 4.3.3: Mark the observation time corresponding to the abnormal column extension length value.

[0084] Step 4.3.4: Determine whether the column pressure value corresponding to the marked observation time is an outlier. If so, establish a fitting function using the column expansion and contraction length values ​​corresponding to multiple observation times before the marked observation time, use the fitting function to obtain the column expansion and contraction length value corresponding to the marked observation time, and replace the outlier column expansion and contraction length value with the fitted column expansion and contraction length value. If not, substitute the pressure observation value corresponding to the marked observation time into the physical model of the column expansion and contraction length value, deduce the column expansion and contraction length value in reverse, and replace the outlier expansion and contraction length observation value with the calculated column expansion and contraction length value.

[0085] (3) Correcting outlier segments in the pressure difference between adjacent supports. This includes the following steps:

[0086] Step 4.4.1: Obtain the residual of the pressure difference between adjacent stents at each observation time within the outlier segment.

[0087] Step 4.4.2: Mark adjacent stent pressure differences where the absolute value of the residual difference is greater than the fifth threshold as abnormal adjacent stent pressure differences.

[0088] Based on the fluctuation range of load transfer differences and the measurement accuracy of pressure sensors, a formula for calculating the fifth threshold can be established to ensure that the fifth threshold covers both normal fluctuations in pressure differences between adjacent supports and accurately identifies abnormal pressure differences between adjacent supports. The formula for calculating the fifth threshold is: ,in, E △P This represents the fifth threshold. s △P This represents the standard deviation of the residuals of the pressure difference between adjacent supports under historical normal operating conditions. This represents the sum of the errors of the two pressure sensors. (When calculating the pressure difference between adjacent supports, the error of the difference is contributed by the errors of the two independent pressure sensors. The combined error needs to be calculated using the "Root Sum of Squares" (RSS method), and the final result is...) (Maximum permissible error of the pressure sensor). For example, if the residual standard deviation of the pressure difference between adjacent supports under historical normal operating conditions... s △P =0.8MPa, the sum of errors from the two pressure sensors =0.7MPa, then the fifth threshold E △P =3×0.8+0.7=3.1MPa.

[0089] Step 4.4.3: Mark the stent combinations corresponding to abnormal pressure differences between adjacent stents.

[0090] Step 4.4.4: Correct the column pressure value of the marked support assembly according to the methods described in steps 4.2.1 to 4.2.6.

[0091] Step 4.4.5: Use the column pressure value corrected in step 4.4.4 to obtain the new pressure difference value between adjacent supports.

[0092] Step 4.4.6: Determine whether the new adjacent stent pressure difference value meets the physical constraints of adjacent stent pressure difference value; if yes, replace the abnormal adjacent stent pressure difference value with the new adjacent stent pressure difference value; if no, extract the adjacent stent pressure difference values ​​of multiple stent combinations adjacent to the marked stent combination, take the median of the adjacent stent pressure difference values ​​of the marked stent combination and the extracted adjacent stent pressure difference values ​​of multiple stent combinations, and replace the abnormal adjacent stent pressure difference value with the median of the adjacent stent pressure difference values.

[0093] The physical constraint that the new pressure difference between adjacent supports must satisfy is Δ. P’ ≤ m × G ×sinα,△ P’ This represents the new pressure difference between adjacent supports. If the new pressure difference between adjacent supports does not meet the physical constraints, then the marked support combination... G j The pressure difference between adjacent supports is △ P j , j If the stent assembly number is specified, then the stent assembly number needs to be extracted. G j-1 Pressure difference between adjacent supports △ P j-1 , support assembly G j-2 Pressure difference between adjacent supports △ P j-2 , support assembly G j+1 Pressure difference between adjacent supports △ P j+1 Combination with bracket G j+2 Pressure difference between adjacent supports △ P j+2 Then for △ P j-2 , △ P j-1 , △ P j , △ P j+1 , △ P j+2 Take the median.

[0094] Step 5: Select each pressure support based on the corrected multi-class observation values.

[0095] Since the column pressure can directly reflect the force on the support, the support is compressed after being stressed, and the column extension length decreases. Therefore, in this embodiment, the column pressure is used as the primary basis for judging whether a single support is stressed, and the column extension length is used as an auxiliary basis for judging whether a single support is stressed. The screening conditions include: (1) the column pressure value collected in two consecutive sampling periods is ≥0.2×rated support force; (2) the rate of change of the column pressure value collected in two consecutive sampling periods is ≥ pressure change rate threshold; (3) the column extension length values ​​collected in two consecutive sampling periods are both shortening.

[0096] It should be noted that: (1) The purpose of collecting the column pressure value and the rate of change of the column pressure value for two consecutive sampling periods is to determine whether the load is continuously applied or fluctuates instantaneously; (2) The pressure change rate threshold can be determined based on the actual measured data on site. For example, if the actual measured column pressure value shows that the rise rate of the support pressure value is between 0.3 and 0.8 MPa / s when the pressure is actually applied, then the rate of change of the column pressure value can be taken as 0.5 MPa / s (intermediate value), which can cover most normal pressure application scenarios; (3) The purpose of collecting the shortening trend of the column extension length for two consecutive sampling periods is to keep in line with the rising trend of the column pressure.

[0097] Based on the above screening criteria, step 5 specifically involves performing steps 5.1 to 5.2 for each support frame.

[0098] Step 5.1: Collect the column pressure value and column extension length value within two consecutive usage cycles.

[0099] Step 5.2: Obtain the rate of change of column pressure based on the column pressure value collected in Step 5.1, and obtain the trend of change of column extension length based on the column extension length value collected in Step 5.1.

[0100] Step 5.3: If conditions A, B, and C are all met simultaneously, mark the support as a pressure support; if conditions A, B, and C are not met simultaneously, mark the support as a non-pressure support.

[0101] Condition A: The column pressure values ​​collected in two consecutive use cycles are greater than or equal to 0.2 times the rated support force; Condition B: The rate of change of column pressure in two consecutive use cycles is greater than or equal to 0.5 MPa / s; Condition C: The trend of column expansion and contraction length is shortening in two consecutive use cycles.

[0102] Step 6: Along the extension direction of the working surface, aggregate adjacent pressure supports into candidate areas.

[0103] Because the load moves in sheet-like areas rather than sequentially along individual supports on the steeply inclined working surface, it is necessary to aggregate adjacent individual load-bearing supports into candidate areas.

[0104] Step 7: Obtain the number and pressure time of the candidate region that was first subjected to pressure.

[0105] Includes the following steps:

[0106] Step 7.1: Add a number and stress time to each candidate region.

[0107] The candidate region is numbered according to the stent number. For example, if a candidate region contains the fifth, sixth, seventh, and eighth stents, then the candidate region is numbered 5-8. The compression time of the candidate region is the initial compression time of the earliest compressed stent in the region.

[0108] Step 7.2: Sort each candidate region in order of the time of pressure application, and take the first candidate region as the first pressure-apposed region.

[0109] Step 8: If the number of the first virtual unit to be compressed is the same as the number of the first candidate region to be compressed, and the difference between the compression time of the first virtual unit to be compressed and the compression time of the first candidate region to be compressed is less than or equal to the time difference threshold, mark the first candidate region to be compressed as the group control region.

[0110] The purpose of this step is to perform a two-way temporal and spatial matching determination between the "predicted first stressed support area" and the "actual first stressed support area," ensuring that group control is triggered only when both the prediction and the on-site measurement point to the same location. The number of the virtual unit that experiences the first stress is consistent with the number of the candidate area that experiences the first stress, including two cases: First, the number of the candidate area that experiences the first stress only includes the numbers of two adjacent supports; in this case, the support area pointed to by the candidate area that experiences the first stress is only the support area pointed to by one virtual unit that experiences the first stress. Second, the number of the candidate area that experiences the first stress includes variations of multiple adjacent supports; in this case, the support area pointed to by the candidate area that experiences the first stress is the support area pointed to by a combination of virtual units that experiences the first stress. This forms a temporary group control area based on virtual units or combinations of virtual units, achieving flexible zoning and avoiding the problem of unstressed supports being mistakenly included in the control due to fixed zoning.

[0111] Furthermore, setting a time difference threshold (coal mining machine stroke time) aims to address the time asynchrony between the "actually measured signal" and the "digital twin model predicted signal." By setting a time difference threshold, the "model-predicted initial stress time" and the "actually measured initial stress time" can be accurately compared, avoiding misjudgments of mismatch due to objective time deviations. Simultaneously, irrelevant signal interference is filtered out, ensuring the accuracy of the matching judgment. Further, using the coal mining machine stroke time as the time difference threshold is because the roof pressure on a steeply inclined working face is not a random event, but occurs gradually as the coal mining machine advances. After the coal mining machine cuts the coal, the roof is exposed and collapses, and the load is transferred along the inclined direction. This process exactly covers the time of one stroke. Setting the time difference threshold to the stroke time can fully cover the entire cycle from model-predicted stress to actual on-site stress, avoiding missing matching signals due to an excessively short window.

[0112] Step 9: Generate load control instructions for the group control area, and adjust the load on the support in the group control area according to the load control instructions.

[0113] A unified target load control command is issued to all supports within the group control area, including: initial support force control command, pressure holding time control command, lateral displacement control command, and top beam rotation angle control command. The initial support force control command is used to increase the support strength of the supports to the roof slab and resist incoming pressure loads; the pressure holding time control command is used to extend the pressure holding time, stabilize the roof slab, and prevent secondary collapse; the lateral displacement control command and the top beam rotation angle control command are used to maintain the stability of the support posture and avoid the support from tilting due to uneven force distribution in the tilt direction.

[0114] Furthermore, the increase in initial support force can be determined based on the measured roof pressure on site, or by referring to the relevant provisions of the "Technical Conditions for Hydraulic Supports in Coal Mines" (MT / T 1097-2021); the extension of pressure holding time can refer to the relevant requirements of the "Technical Specification for Support Quality and Dynamic Monitoring of Roof in Fully Mechanized Mining Faces" (AQ / T 2021-2019); the limitation of lateral displacement can refer to the relevant provisions of the "Coal Mine Safety Regulations"; the limitation of roof beam rotation angle can be found in the support design manual, and the roof beam rotation angle shall not exceed the maximum allowable angle of the hinge point.

[0115] Example 2: Corresponding to Example 1, this example provides a hydraulic support yaw angle correction system for a steeply inclined, high-extraction working face, comprising:

[0116] The model building module is used to create a digital twin model of the working face; the digital twin model incorporates the gravity component coefficient along the dipping direction of the coal seam and the friction coefficient of each contact interface.

[0117] The group division module is used to divide the support group into multiple virtual units; each virtual unit contains multiple supports along the extension direction of the working surface.

[0118] The data extraction module is used to extract the number and pressure time of the first virtual unit to be pressured from the running results of the digital twin model, as well as to obtain the number and pressure time of the candidate region to be pressured first.

[0119] The data acquisition module is used to continuously acquire various observation values ​​on site to reflect the stress state of the support.

[0120] The data filtering module is used to filter out outlier segments in each type of observation.

[0121] The data correction module is used to correct outlier segments in each type of observation;

[0122] The support screening module is used to select each pressure support based on the corrected multi-class observation values;

[0123] The support aggregation module is used to aggregate adjacent pressure supports into candidate regions along the extension direction of the working surface.

[0124] The area marking module is used to mark the first pressured candidate area as a group control area when the number of the first pressured virtual unit is consistent with the number of the first pressured candidate area, and the difference between the pressure time of the first pressured virtual unit and the pressure time of the first pressured candidate area is less than or equal to the stroke time of the coal mining machine.

[0125] The instruction generation module is used to generate load control instructions for the group control area;

[0126] The load control module is used to control the load on the support structure in the group control area according to the load control command.

[0127] The working process, working details and technical effects of the aforementioned system provided in this embodiment can be found in the method described in Embodiment 1 or any method that may involve the method described in Embodiment 1, and will not be repeated here.

[0128] Example 3: Based on the method provided in Example 1 and the system provided in Example 2, this example provides a computer device that executes the method described in Example 1 or any other method that may involve the method described in Example 1. The device includes a memory, a processor, and a transceiver connected in sequence. The memory stores a computer program, the transceiver sends and receives messages, and the processor reads the computer program and executes the method described in Example 1 or any other method that may involve the method described in Example 1. Specifically, the memory may include, but is not limited to, random-access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out (FIFO) memory, and / or last-in-first-out (FILO) memory, etc.; the processor may include, but is not limited to, a microprocessor of the STM32F105 series. Furthermore, the computer device may also include, but is not limited to, a power module, a display screen, and other necessary components.

[0129] The working process, working details and technical effects of the aforementioned computer device provided in this embodiment can be found in the method described in Embodiment 1 or any method that may involve the method described in Embodiment 1, and will not be repeated here.

[0130] Example 4: This example provides a computer-readable storage medium that stores instructions that include the method described in Example 1 or any other method that may involve the method described in Example 1. Specifically, the computer-readable storage medium stores instructions that, when executed on a computer, perform the method described in Example 1 or any other method that may involve the method described in Example 1. The computer-readable storage medium refers to a data storage medium, which may include, but is not limited to, floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or Memory Sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0131] The working process, working details and technical effects of the aforementioned computer-readable storage medium provided in this embodiment can be found in the method described in Embodiment 1 or any method that may be related to Embodiment 1, and will not be repeated here.

[0132] Example 5: This example provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform the method described in Example 1 or any method that may involve the method described in Example 1. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0133] It should be understood that the terms "system," "device," "unit," and / or "module" as used in this specification are a method of distinguishing different components, elements, parts, sections, or assemblies at different levels. However, if other terms can achieve the same purpose, they may be replaced by other expressions.

[0134] As indicated in this specification and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not specifically refer to the singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of expressly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0135] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0136] It should be noted that the structures, proportions, sizes, etc., illustrated in the accompanying drawings are merely for illustrative purposes to aid those skilled in the art and to facilitate understanding and reading. They are not intended to limit the scope of the invention and therefore have no substantial technical significance. Any modifications to the structure, changes in proportions, or adjustments to size, without affecting the effectiveness and purpose of the invention, should still fall within the scope of the technical content disclosed herein. Furthermore, terms such as "upper," "lower," "left," "right," and "middle" used in this specification are merely for clarity and not intended to limit the scope of the invention. Changes or adjustments to their relative relationships, without substantially altering the technical content, should also be considered within the scope of the invention's implementation.

Claims

1. A method for regulating the load of a group of supports in a steeply inclined coal seam based on digital twinning, characterized in that, The method comprises the following steps: A digital twin model of the working face is established; the digital twin model introduces a gravity component coefficient along the inclined direction of the coal seam and a friction coefficient of each contact interface; In the digital twin model, the support group is divided into multiple virtual units; each virtual unit contains multiple supports along the extension direction of the working face; The number and the time of the first compressed virtual unit are extracted from the running results of the digital twin model; For each support: continuously acquire multiple types of observation values reflecting the stress state of the support in the field, screen out abnormal value segments in each type of observation value, and correct the abnormal value segments in each type of observation value; Each compressed support is screened out according to the corrected multiple types of observation values; Adjacent compressed supports are aggregated into a candidate region along the extension direction of the working face; the number and the time of the first compressed candidate region are acquired; If the number of the first compressed virtual unit is consistent with the number of the first compressed candidate region, and the difference between the time of the first compressed virtual unit and the time of the first compressed candidate region is less than or equal to the stroke time of the coal mining machine, the first compressed candidate region is marked as a group control region; A load control instruction of the group control region is generated, and the supports in the group control region are controlled according to the load control instruction.

2. The digital twin-based load regulation method for a group of supports in a steeply inclined coal seam according to claim 1, characterized in that, The multiple types of observation values include: column pressure values, column telescopic length values, and pressure difference values of adjacent supports.

3. The digital-twin-based load regulation method for a group of supports in a steeply inclined coal seam according to claim 2, characterized in that, The method for screening out abnormal value segments in each type of observation value is: S1 to S3 are performed for each type of observation value; S1: a physical model of the observation value is established, and a predicted value at each observation time is acquired according to the physical model; S2: a residual error between the observation value and the predicted value at each observation time is acquired, and a residual error sequence is established; S3: abnormal value segments satisfying an abnormality judgment condition are screened out from the residual error sequence by using a sliding time window; the abnormality judgment condition is that the absolute value of the mean of the residual error is greater than a first threshold value, the variance of the residual error is greater than a second threshold value, or the number of mutations of the residual error is not equal to zero; the number of mutations is the number of times that the absolute value of the residual error is greater than 3 times the standard deviation.

4. The load regulation method for a group of supports in a steeply inclined coal seam based on digital twinning according to claim 3, characterized in that, The correction of the abnormal value segments in each type of observation value includes: For the abnormal value segments in the column pressure values, the abnormal column pressure values in the abnormal value segments are replaced by the mean value of the normal column pressure values of the adjacent supports; For the abnormal value segments of the column telescopic length values, the abnormal column telescopic length values are replaced by the fitting value of the normal column pressure values or the normal column telescopic length values; For the abnormal value segments of the pressure difference values of the adjacent supports, the abnormal pressure difference values are replaced by the corrected pressure difference values, or the median of the pressure difference values of a plurality of adjacent support combinations is taken, and the abnormal pressure difference values are replaced by the median of the pressure difference values.

5. The digital twin-based load regulation method for a group of supports in a steeply inclined coal seam according to claim 4, characterized in that, Replacing the abnormal column pressure values in the abnormal value segments by the mean value of the normal column pressure values of the adjacent supports comprises the following steps: Step A1: abnormal column pressure values and corresponding observation times are marked out; Step A2: column pressure values of the front and rear i supports at the marked observation times are extracted respectively, i being the support number; Step A3: the pressure mean value is calculated according to the column pressure values extracted in step A2; Step A4: judging whether the pressure mean value falls within the physical constraint boundary of the column pressure value, if yes, replacing the abnormal column pressure value with the pressure mean value, if no, executing i=i+1 and returning to step A3.

6. The digital twin-based load regulation method for a group of supports in a steeply inclined coal seam according to claim 5, characterized in that, Replacing the abnormal column telescopic length value with the fitted value of the normal column pressure value or the normal column telescopic length value, comprising the following steps: Step B1: marking the abnormal column telescopic length value and the corresponding observation time; Step B2: judging whether the column pressure value corresponding to the marked observation time is an abnormal value; if yes, establishing a fitting function with the column telescopic length values corresponding to multiple observation times before the marked observation time, obtaining the column telescopic length value corresponding to the marked observation time by using the fitting function, and replacing the abnormal column telescopic length value with the fitted column telescopic length value; if no, substituting the pressure observation value corresponding to the marked observation time into the physical model of the column telescopic length value to inversely deduce the column telescopic length value, and replacing the abnormal telescopic length observation value with the calculated column telescopic length value.

7. The digital-twin-based load regulation method for a group of supports in a steeply inclined coal seam according to claim 6, characterized in that, Replacing the abnormal pressure difference value with the corrected pressure difference value, or taking the median of the pressure difference values of the multiple adjacent support combinations before and after the abnormal pressure difference value, and replacing the abnormal pressure difference value with the median of the pressure difference values, comprising the following steps: Step C1: marking the abnormal pressure difference value and the corresponding support combination; Step C2: correcting the column pressure value of the marked support combination according to the method of steps A1 to A4; Step C3: obtaining a new pressure difference value by using the column pressure value corrected in step C2; Step C4: judging whether the new pressure difference value satisfies the physical constraint of the pressure difference value; if yes, replacing the abnormal pressure difference value with the new pressure difference value; if no, extracting the pressure difference values of multiple adjacent support combinations before and after the marked support combination, taking the median of the pressure difference value of the marked support combination and the extracted multiple support combinations, and replacing the abnormal pressure difference value with the median of the pressure difference values.

8. The support group load regulation method based on digital twinning according to claim 7, characterized in that, Marking the abnormal column pressure value, comprising the following steps: Obtaining the column pressure value residual of each observation time within the abnormal value segment; The column pressure value whose absolute value of column pressure value residual is greater than a third threshold value is marked as an abnormal column pressure value; the third threshold value is calculated by a formula: E P = 3 σ P + δ P , wherein, E P represents the third threshold value, σ P represents a standard deviation of column pressure residuals in a historical normal working condition, δ P represents a maximum allowable error of the pressure sensor; Marking the abnormal column telescopic length value, comprising the following steps: Obtaining the column telescopic length residual of each observation time within the abnormal value segment; mark the column telescopic length value whose absolute value of the column telescopic length residual is greater than a fourth threshold value as an abnormal column telescopic length value; the fourth threshold value is calculated according to the formula: E L = 3 σ L + δ L , wherein, E L the fourth threshold value is represented by, σ L the standard deviation of the column telescopic length residual under the historical normal working condition is represented by, δ L the maximum allowable error of the displacement sensor is represented by. Marking the abnormal pressure difference value, comprising the following steps: Obtaining the residual of the adjacent support pressure difference value of each observation time within the abnormal value segment; The adjacent support pressure difference value with a residual absolute value of the adjacent support pressure difference value greater than a fifth threshold value is marked as an abnormal adjacent support pressure difference value; the calculation formula of the fifth threshold value is: wherein, E △P represents the fifth threshold value, σ △P represents the standard deviation of the residual of the adjacent support pressure difference value under the historical normal working condition, represents the error superposition value of the two pressure sensors.

9. The digital-twin-based group load regulation method for steep-inclined seam supports according to any one of claims 2-8, characterized in that, The method for screening each pressure support according to the corrected multiple types of observation values is as follows: Collecting the column pressure value and the column telescopic length value in two consecutive sampling periods; Obtaining the change rate of the column pressure value in the two sampling periods and the change trend of the column telescopic length value in the two consecutive sampling periods; If conditions A, B and C are simultaneously satisfied, the support is marked as a pressure support; If conditions A, B and C are not met simultaneously, the support is marked as a non-pressurized support; wherein condition A: the column pressure value collected in two consecutive cycles is greater than or equal to 0.2 times the rated support force; condition B: in two consecutive cycles, the change rate of the column pressure is greater than or equal to 0.5 MPa / s; condition C: in two consecutive cycles, the change trend of the column extension length is a shortening trend. The load regulation instruction includes an initial support force regulation instruction, a pressure maintaining time regulation instruction, a lateral displacement regulation instruction, and a top beam rotation angle regulation instruction.

10. The digital-twin-based load regulation method for a group of supports in a steeply inclined coal seam according to claim 1, characterized in that, The method comprises:

11. A digital-twin-based group load regulation system for a steep-inclined seam support group, characterized in that, a model construction module for establishing a digital twin model of the working face; a gravity component coefficient along the inclination direction of the coal seam and a friction coefficient of each contact interface are introduced into the digital twin model; a group division module for dividing the support group into a plurality of virtual units; each virtual unit contains a plurality of supports along the extension direction of the working face; a data extraction module for extracting the number and pressure time of the first-pressurized virtual unit from the running results of the digital twin model, and obtaining the number and pressure time of the first-pressurized candidate area; a data acquisition module for continuously acquiring a plurality of observation values reflecting the stress state of the support in the field; a data screening module for screening abnormal value segments in each type of observation value; a data correction module for correcting the abnormal value segments in each type of observation value; a support screening module for screening each pressurized support according to the corrected plurality of observation values; a support aggregation module for aggregating adjacent pressurized supports into a candidate area along the extension direction of the working face; a region marking module for marking the first-pressurized candidate area as a group regulation region when the number of the first-pressurized virtual unit is consistent with the number of the first-pressurized candidate area, and the difference between the pressure time of the first-pressurized virtual unit and the pressure time of the first-pressurized candidate area is less than or equal to the stroke time of the coal mining machine; an instruction generation module for generating a load regulation instruction for the group regulation region; a load regulation module for regulating the load of the support in the group regulation region according to the load regulation instruction. The memory, the processor and the transceiver are sequentially communicatively connected, wherein the memory is used to store a computer program, the transceiver is used to transceive data, and the processor is used to read the computer program and execute the load dynamic regulation method for the support group in the steeply inclined coal seam based on the digital twin as claimed in any one of claims 1-10.

12. A computer device, comprising: The computer readable storage medium stores instructions, and when the instructions run on the computer, the load dynamic regulation method for the support group in the steeply inclined coal seam based on the digital twin as claimed in any one of claims 1-10 is executed.

13. A computer-readable storage medium, characterized in that, ​

Images