Recommended method for manual control strategy after hydraulic support automation

Abstract

The present invention relates to a method for recommending manual control strategies after the automation of hydraulic supports, and belongs to the technical field of intelligent mining working faces. It includes: collecting current data of comprehensive mining equipment in the current production state of the intelligent mining working face; preprocessing the current data of the comprehensive mining equipment to obtain the current operation data of the comprehensive mining equipment; extracting the characteristic values ​​of the current operation data; inputting the characteristic values ​​of the current operation data into a pre-trained manual intervention working condition identification decision tree model; judging whether the current working condition of the comprehensive mining equipment requires manual intervention behavior according to the identification result of the manual intervention working condition identification decision tree model; if it is determined that the current working condition of the comprehensive mining equipment requires manual intervention behavior, then recommending operation strategies for the manual intervention behavior after the automation of the hydraulic support cluster according to the preset operation strategy recommendation model. This method can discover the manual intervention working conditions in the actual production process of the intelligent mining working face and can recommend specific operation strategies.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent mining working faces, and in particular to a method for recommending a manual control strategy after automation of a hydraulic support. Background Art

[0002] With the rapid development of the coal mining industry, especially the vigorous promotion of intelligent coal mine construction in recent years, the intelligence level of coal mines has been significantly improved, and the production capacity has also been steadily increasing. At the same time, the death rate per million tons of coal has dropped sharply, which has supported the healthy development of the coal industry to a large extent. At present, most scholars have only proposed ideas and theoretical systems for the construction of intelligent mining working faces; although some scholars have achieved certain results in the research of autonomous follow-up of hydraulic supports, it is difficult to achieve the goal of completely "unmanned" intelligent mining working faces. Due to the different geological conditions of coal mines, for intelligent mining working faces with complex geological conditions, manual intervention after the automation of hydraulic support clusters is still indispensable.

[0003] However, the current automation system of the intelligent mining face is based on process control, which cannot make real-time decisions on the comprehensive mining equipment according to the real-time working conditions of the intelligent mining face, and lacks knowledge discovery of the working conditions of manual intervention in the actual production process, and lacks effective expression of the working condition information during manual intervention in production. Therefore, it is of great significance for safe and efficient mining to conduct knowledge discovery and logical reasoning on the working conditions of manual intervention to form an operation strategy that is instructive for underground workers. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a method for recommending manual control strategies after automation of hydraulic supports. The technical solution is as follows:

[0005] A method for recommending a manual control strategy after automation of a hydraulic support comprises the following steps:

[0006] S1, collects the current data of the fully mechanized mining equipment in the current production status of the intelligent mining working face;

[0007] S2, preprocessing the current data of the comprehensive mining equipment to obtain the current operation data of the comprehensive mining equipment;

[0008] S3, extracting the characteristic value of the current operation data;

[0009] S4, inputting the characteristic values ​​of the current operating data into a pre-trained decision tree model for identifying manual intervention conditions;

[0010] S5, judging whether the current working condition of the fully mechanized mining equipment requires manual intervention according to the recognition result of the manual intervention working condition recognition decision tree model;

[0011] S6: If it is determined that the current working condition of the comprehensive mining equipment requires manual intervention, an operation strategy recommendation is made for the manual intervention after the automation of the hydraulic support cluster according to a preset operation strategy recommendation model.

[0012] Optionally, before inputting the characteristic value of the current operating data into a pre-trained decision tree model for identifying manual intervention conditions, S4 further includes:

[0013] S41, obtaining historical data of comprehensive mining equipment of historical production status of intelligent mining working face;

[0014] S42, preprocessing the historical data of the fully mechanized mining equipment to obtain historical operation data of the fully mechanized mining equipment;

[0015] S43, extracting characteristic values ​​of the historical operation data;

[0016] S44, training a manual intervention condition identification decision tree model through the characteristic values ​​of the historical operation data.

[0017] Optionally, when preprocessing the historical data of the fully mechanized mining equipment, the step S42 includes:

[0018] S421, performing outlier processing and missing value processing on the historical data of the fully mechanized mining equipment;

[0019] S422, filtering the historical data of the fully mechanized mining equipment after outlier processing and missing value processing according to the column pressure value of the hydraulic support, the stroke of the push cylinder, the action data and the position data of the coal mining machine;

[0020] S423, sorting each type of filtered data by time to obtain historical operation data of the comprehensive mining equipment.

[0021] Optionally, the characteristic value of the historical operation data includes the change in the stroke of the automatic follower forward and backward push cylinder Δ X , Automatic following machine pulling distance X , the absolute value Δp of the difference between the position of the coal mining machine and the number of the hydraulic support to be judged, when extracting the characteristic value of the historical operation data, the S43 is implemented by the following formulas (1) to (3):

[0022]

[0023] In formula (1): x1 is the stroke of the push cylinder before automatic following, and x2 is the stroke of the push cylinder after automatic following;

[0024]

[0025] In formula (2): Assume that T (n-1) is the time point corresponding to the manual intervention of the pulling action, then T(n-2) is the time point corresponding to the automatic follow-up frame pulling action; x T(n-2)-10 x is the stroke of the hydraulic cylinder 10 seconds before the start of the automatic follow-up frame pulling action. T(n-2)+6 It is the travel of the push cylinder 6s after the automatic following and pulling action starts;

[0026]

[0027] In formula (3): N is the number of the hydraulic support to be judged, P (CMJ) It is the frame number of the hydraulic support where the coal mining machine is located.

[0028] Optionally, the characteristic value of the historical operation data includes the change in the stroke of the automatic follower forward and backward push cylinder Δ X , Automatic following machine pulling distance X , the absolute value Δp of the difference between the position of the coal mining machine and the hydraulic support number to be judged, and the pressure value Y of the column after the column is raised. When extracting the characteristic value of the historical operation data, S43 is implemented by the following formulas (1) to (4):

[0029]

[0030] In formula (1): x1 is the stroke of the push cylinder before automatic following, and x2 is the stroke of the push cylinder after automatic following;

[0031]

[0032] In formula (2): Assume that T (n-1) is the time point corresponding to the manual intervention of the pulling action, then T (n-2) is the time point corresponding to the automatic follow-up frame pulling action; x T(n-2)-10 x is the stroke of the hydraulic cylinder 10 seconds before the start of the automatic follow-up frame pulling action. T(n-2)+6 It is the travel of the push cylinder 6s after the automatic following and pulling action starts;

[0033]

[0034] In formula (3): N is the number of the hydraulic support to be judged, P (CMJ) The frame number of the hydraulic support where the coal mining machine is located;

[0035]

[0036] In formula (4), It is the column pressure value 20s after the automatic following frame pulling action starts.

[0037] Optionally, before training the manual intervention condition identification decision tree model by using the characteristic values ​​of the historical operation data, S44 further includes:

[0038] Performing data cleaning on the characteristic values ​​of the historical operation data to remove abnormal characteristic values;

[0039] The eigenvalues ​​after data cleaning are standardized.

[0040] Optionally, when recommending an operation strategy for the manual intervention behavior after the automation of the hydraulic support cluster according to the preset operation strategy recommendation model, S6 includes:

[0041] According to the preset operation strategy recommendation model, the manual intervention behavior after the automation of the hydraulic support cluster is recommended in terms of the manual frame pulling adjustment amount, manual column lowering time, manual column raising time and the target column pressure value achieved by raising the column.

[0042] Optionally, for the intelligent mining working face without inertial navigation installed, the stroke accumulation value calculation method is adopted to determine the manual pull adjustment amount; for the intelligent mining working face with inertial navigation installed, the inertial navigation is used to calculate the manual pull adjustment amount;

[0043] Recommend the duration of manual column lowering and manual column raising based on the statistical analysis results of historical data of fully mechanized mining equipment;

[0044] The target pressure value of the column to be reached by the lifting column is determined based on the initial support force of the hydraulic support used in each intelligent mining working face.

[0045] Optionally, the calculation formula of the trip cumulative value calculation method is:

[0046]

[0047] In formula (5): The manual pull adjustment amount, m is the number of work cycles completed on that day, s g The distance between the scaffolds is specified in the coal mine operating regulations. s 1, s 2, ..., s m For the bracket m The pulling distance completed in one working cycle.

[0048] The calculation formula for calculating the manual rack adjustment amount using inertial navigation is:

[0049]

[0050] In formula (6): is the manual pull adjustment, D is the displacement relative to the origin of the coordinate system, θ is the pitch angle of the coal mining machine, is the heading angle of the coal mining machine.

[0051] All the above optional technical solutions can be combined arbitrarily, and the present invention does not provide detailed descriptions of the structures after the combinations.

[0052] By means of the above scheme, the beneficial effects of the present invention are as follows:

[0053] The current operating data of the comprehensive mining equipment is obtained by collecting and preprocessing, the characteristic values ​​of the current operating data are extracted, and when it is determined that the current working condition of the comprehensive mining equipment requires manual intervention behavior based on the recognition results of the manual intervention working condition recognition decision tree model, the operation strategy recommendation model is used to recommend the manual intervention behavior after the automation of the hydraulic support cluster. This provides a method for making real-time decisions on the comprehensive mining equipment based on the real-time working conditions of the intelligent mining working face. This method can discover the manual intervention conditions in the actual production process of the intelligent mining working face and can recommend specific operation strategies, forming an operation strategy that is instructive for underground workers. It can not only reduce the workload of laborers, but also promote the construction of intelligent coal mines, and is of great significance for safe and efficient mining.

[0054] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a schematic diagram of the implementation environment of the present invention.

[0056] Figure 2 It is a flow chart of the present invention.

[0057] Figure 3 It is a frequency histogram of artificially adjusted descending column duration and ascending column duration.

[0058] Figure 4 The figure is a schematic diagram of an implementation process of an embodiment of the present invention.

[0059] Figure 5 The figure is a schematic diagram of the implementation process of another embodiment of the present invention. DETAILED DESCRIPTION

[0060] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0061] like Figure 1 As shown, it is a schematic diagram of the implementation environment of the recommended method for manual intervention behavior after the automation of the hydraulic support cluster provided by the present invention. Figure 1As shown, the implementation environment mainly includes the equipment layer, the perception layer, the transmission layer and the analysis application layer. Among them, the equipment layer is the comprehensive mining face equipment group, which is the object and data source for strategy recommendation. The perception layer uses the controller of the comprehensive mining face equipment group and the installed sensors to collect the comprehensive mining equipment data of the production status in real time, and transmits it to the chute control center through the underground ring network, playing the role of interactive communication and information transmission. The transmission layer is responsible for transmitting the data collected by the perception layer to the database, the dispatching room and the well data cloud center through the 10G industrial ring network and dedicated line and other equipment. The analysis application layer is the core component of the intelligent mining face hydraulic support behavior strategy recommendation framework. It trains the intelligent mining face hydraulic support behavior strategy recommendation model (in the embodiment of the present invention, the manual intervention working condition identification decision tree model and the preset operation strategy recommendation model) by preprocessing the historical operation data of the comprehensive mining equipment of the intelligent mining face in the database, extracting features and other steps. At the same time, the intelligent mining face hydraulic support behavior strategy recommendation model is driven by the current data of the comprehensive mining equipment collected in real time, and then the operation parameters are recommended. The specific content of the embodiment of the present invention is detailed in the following embodiment.

[0062] Combined with the above implementation environment diagram, Figure 2 As shown, the method for recommending manual control strategies after automation of hydraulic supports provided by an embodiment of the present invention comprises the following steps:

[0063] S1, collects the current data of the comprehensive mining equipment in the current production status of the intelligent mining working face.

[0064] This step is achieved through the controller of the comprehensive mining face equipment group and the installed sensors.

[0065] S2, preprocessing the current data of the comprehensive mining equipment to obtain the current operation data of the comprehensive mining equipment.

[0066] The preprocessing mainly includes outlier processing and missing value processing. The current data of the comprehensive mining equipment after outlier processing and missing value processing are filtered according to the column pressure value of the hydraulic support, the stroke of the push cylinder, the action data and the coal mining machine position data, and each type of filtered data is sorted by time to obtain the current operation data of the comprehensive mining equipment.

[0067] S3, extracting characteristic values ​​of the current running data.

[0068] The principle of this step is the same as that of S43 in the following content, and will not be described here for the time being.

[0069] S4, inputting the characteristic values ​​of the current operating data into a pre-trained decision tree model for identifying manual intervention conditions.

[0070] S5, judging whether the current working condition of the fully mechanized mining equipment requires manual intervention according to the recognition result of the manual intervention working condition recognition decision tree model.

[0071] The recognition results of the decision tree model for identifying manual intervention conditions include behaviors that require manual intervention and behaviors that do not require manual intervention.

[0072] S6: If it is determined that the current working condition of the comprehensive mining equipment requires manual intervention, an operation strategy recommendation is made for the manual intervention after the automation of the hydraulic support cluster according to a preset operation strategy recommendation model.

[0073] The preset operation strategy recommendation model includes the specific types of manual intervention behaviors and how to operate each type.

[0074] Wherein, before inputting the characteristic value of the current operation data into the pre-trained decision tree model for identifying the manual intervention condition, the S4 needs to first train the decision tree model for identifying the manual intervention condition. The method for training the decision tree model for identifying the manual intervention condition in the embodiment of the present invention comprises the following steps:

[0075] S41, obtain the historical data of the comprehensive mining equipment of the historical production status of the intelligent mining working face.

[0076] This step is achieved by collecting the comprehensive mining equipment data collected by the controllers of the comprehensive mining face equipment group and the installed sensors at each production status.

[0077] S42, pre-processing the historical data of the comprehensive mining equipment to obtain the historical operation data of the comprehensive mining equipment.

[0078] Specifically, when the S42 pre-processes the historical data of the fully mechanized mining equipment, it includes:

[0079] S421, performing outlier processing and missing value processing on the historical data of the comprehensive mining equipment.

[0080] Due to the complex production environment of the intelligent mining work surface, various sensors are often damaged, which leads to errors in the collected data. The following two types are common: data missing and data anomaly. Data missing refers to a circuit break in the data transmission path, which is common in damage to the sensor end and damage to the sensor signal line; data anomaly refers to an erroneous value transmitted by the sensor, which is generally an outlier, and is common in damage to the internal components of the sensor. In response to the problem of missing data, in order to ensure the integrity of the data, the method of filling adjacent values ​​in the time domain is adopted. For example, the column pressure value of a hydraulic support at a certain moment is missing, and it is filled with the pressure value of the support at the previous moment; in the case of data anomalies, the abnormal value is replaced by the average pressure of the five adjacent hydraulic supports at that moment.

[0081] S422, filtering the historical data of the fully mechanized mining equipment after outlier processing and missing value processing according to the column pressure value of the hydraulic support, the stroke of the push cylinder, the action data and the position data of the coal mining machine.

[0082] S423, sorting each type of filtered data by time to obtain historical operation data of the comprehensive mining equipment.

[0083] S43, extracting characteristic values ​​of the historical operation data.

[0084] In one embodiment, the characteristic value of the historical operation data includes the change in the stroke of the automatic follower forward and backward cylinder Δ X , Automatic following machine pulling distance X , the absolute value Δp of the difference between the coal mining machine position and the hydraulic support number being judged.

[0085] When calculating the above three characteristic values, the data of the state before and after the automatic following machine is needed. Considering that the specific adjustment operation of the worker cannot be judged when the model is applied, but the automatic following machine operation can be determined, the embodiment of the present invention needs to locate the manual intervention working condition first when using historical operation data for sample production, and then find the automatic following machine action before the manual intervention working condition, and then calculate according to the data of 10 minutes before and after the automatic following machine (this parameter is obtained by continuously calculating the accuracy during the model training process). It should also be noted that the data positioning is based on the hydraulic support action, but the hydraulic support action has a duration, so the starting point of each hydraulic support action must also be located. The method of locating the starting point in the embodiment of the present invention is: merge the hydraulic support action time points into an array, and make a difference between the adjacent elements in the array. If the time difference is less than 10s (because the operation duration of a hydraulic support action is generally less than 10s, so 10s is taken here), then remove the latter time point to get the start time of each hydraulic support action. In this way, n support action starting points can be obtained from the data of each day, recorded as T (1) , T (2) , T (3) ……T (n-1) , T (n) .

[0086] Based on the above content, when extracting the characteristic value of the historical operation data, S43 is implemented by the following formulas (1) to (3):

[0087]

[0088] In formula (1): x1 is the stroke of the push cylinder before automatic following, and x2 is the stroke of the push cylinder after automatic following;

[0089]

[0090] In formula (2): Assume that T (n-1) is the time point corresponding to the manual intervention of the pulling action, then T (n-2) is the time point corresponding to the automatic follow-up frame pulling action; x T(n-2)-10 x is the stroke of the hydraulic cylinder 10 seconds before the start of the automatic follow-up frame pulling action. T(n-2)+6 It is the travel of the push cylinder 6s after the automatic following and pulling action starts;

[0091]

[0092] In formula (3): N is the number of the hydraulic support to be judged, P (CMJ) It is the frame number of the hydraulic support where the coal mining machine is located.

[0093] In another embodiment, the characteristic values ​​of the historical operation data include the change in the stroke of the automatic follower forward and backward cylinder Δ X , Automatic following machine pulling distance X In addition to the absolute value Δp of the difference between the position of the coal mining machine and the hydraulic support number to be judged, it also includes the pressure value Y of the column after the column is raised. The pressure value Y of the column after the column is raised is realized by the following formula (4):

[0094]

[0095] In formula (4), It is the column pressure value 20s after the automatic following frame pulling action starts.

[0096] S44, training a manual intervention condition identification decision tree model through the characteristic values ​​of the historical operation data.

[0097] When training the decision tree model for identifying manual intervention conditions, the training process can include the following two methods, depending on the sample size, the type of feature value, the accuracy requirement, and the on-site application requirements:

[0098] The first method: When the characteristic value of the historical operation data includes the change in the stroke of the automatic follow-up cylinder Δ X , Automatic following machine pulling distance X , when the absolute value of the difference between the position of the coal mining machine and the judged hydraulic support number is Δp, the training process of S44 is as follows:

[0099] Since the decision tree belongs to supervised learning, the characteristic values ​​of the historical operation data need to be manually labeled first. After the manual labeling is completed, the data used in the embodiment of the present invention obtains 663 normal working condition samples and 256 samples of manual intervention after the automation of the hydraulic support cluster, which together constitute a sample set of size 4 (three columns of characteristic values ​​plus one column of manual labeling) * 919. Then, the sample set is divided into a training set and a test set by a stratified sampling method, of which the training set accounts for 80% and the test set accounts for 20%.

[0100] The purpose of model training is to optimize the decision tree parameters so that the model can achieve the best effect. The training model uses all samples. When the embodiment of the present invention divides the training set and the test set, n_splits=10, that is, the sample set is first divided into 10 groups, and then the sample set is divided according to the proportion. This can ensure the uniform distribution of samples and help improve the generalization ability of the model; the decision tree can be prevented from overfitting through pruning operations. Finally, the parameters of the decision tree model are selected as follows: max_depth=5, that is, the maximum depth of the decision tree is 5 layers. After experiments, it can be seen that the accuracy of the model training set trained by this method is 92.27%, and the accuracy of the test set is 93.75%, which is a good effect.

[0101] The second method: When the characteristic value of the historical operation data includes the change in the stroke of the cylinder before and after the automatic follow-up machine, Δ X , Automatic following machine pulling distance X , the absolute value Δp of the difference between the position of the coal mining machine and the hydraulic support number to be judged and the pressure value Y of the column after the column is raised, the training process of S44 is as follows:

[0102] Since the decision tree belongs to supervised learning, the characteristic values ​​of the historical operation data need to be manually labeled first. After the manual labeling is completed, the data used in the embodiment of the present invention obtains 663 normal working condition samples, 91 samples of manual pressure adjustment after the automation of the hydraulic support cluster, and 165 samples of manual lowering, shifting and raising adjustment after the automation of the hydraulic support cluster, which together constitute a sample set of size 5 (four columns of characteristic values ​​plus one column of manual labeling) * 919. Then, the sample set is divided into a training set and a test set by a stratified sampling method, of which the training set accounts for 80% and the test set accounts for 20%.

[0103] The purpose of model training is to optimize the decision tree parameters so that the model can achieve the best effect. The training model uses all samples. When the embodiment of the present invention divides the training set and the test set, n_splits=5, that is, the sample set is first divided into 5 groups, and then the sample set is divided according to the proportion. The uniform distribution of samples can be ensured, which is conducive to improving the generalization ability of the model; the decision tree can be prevented from overfitting through pruning operations. Finally, the parameters of the decision tree model are selected as follows: max_depth=6, min_samples_split=30, that is, the maximum depth of the decision tree is 6 layers, and the minimum number of samples required for segmentation is 30. After experiments, it can be seen that the training set accuracy of the model trained by this method is 85.57%, and the test set accuracy is 84.78%, which is a good effect.

[0104] Optionally, before training the decision tree model for identifying manual intervention conditions through the characteristic values ​​of the historical operation data, S44 may also: perform data cleaning on the characteristic values ​​of the historical operation data to remove abnormal characteristic values; and perform standardization on the characteristic values ​​after data cleaning. Standardization can scale data of different proportions and dimensions to the same data interval and range, thereby reducing the impact of proportion, characteristics and distribution differences on the model. The processing results are used as a data set for instantiating the intelligent mining working face hydraulic support behavior strategy recommendation model.

[0105] Further, referring to the automatic following process, the manual intervention behavior is also considered from the target value and action duration. The embodiment of the present invention mainly recommends strategies for the manual rack adjustment amount, the time for lowering the column, the time for raising the column, and the column pressure value reached by raising the column. The manual rack adjustment amount, the time for lowering the column, the time for raising the column, and the column pressure value reached by raising the column are recommended for lowering, moving and raising operations (lowering the column, moving the rack, raising the column), and the manual rack adjustment amount is recommended for pressure-carrying operations. That is to say, when the S6 recommends the operation strategy for the manual intervention behavior after the automation of the hydraulic support cluster according to the preset operation strategy recommendation model, the manual rack adjustment amount, the time for manual lowering the column, the time for manual raising the column, and the column target pressure value reached by raising the column are recommended for the manual intervention behavior after the automation of the hydraulic support cluster according to the preset operation strategy recommendation model.

[0106] Specifically, for the intelligent mining working face without inertial navigation installed, the stroke cumulative value calculation method is used to determine the manual pulling adjustment amount; for the intelligent mining working face with inertial navigation installed, the inertial navigation is used to calculate the manual pulling adjustment amount.

[0107] Furthermore, at the beginning of each production shift, the scraper straightness is adjusted to the optimal state, and the pulling distance of each hydraulic support after each working cycle is completed is counted, and then compared and calculated with the provisions of the coal mine operating procedures to obtain the distance that each hydraulic support should be manually adjusted after each working cycle. The calculation formula of the stroke cumulative value calculation method is:

[0108]

[0109] In formula (5): The manual pull adjustment amount, m is the number of work cycles completed on that day, s g The distance between the scaffolds is specified in the coal mine operating regulations. s 1, s 2, ..., s m For the bracket m The pulling distance completed in one working cycle.

[0110] At present, the inertial navigation system has become a standard equipment in the intelligent fully mechanized mining working face. It is the most common and accurate method for monitoring the straightness of the scraper. The calculation formula of the manual pull adjustment amount using the inertial navigation in the embodiment of the present invention is:

[0111]

[0112] In formula (6): is the manual pull adjustment, D is the displacement relative to the origin of the coordinate system, θ is the pitch angle of the coal mining machine, For the working face with installed inertial navigation, the calculation using inertial navigation will be more accurate.

[0113] The recommended values ​​of the working time (manual column lowering time and artificial column raising time) in the embodiment of the present invention are obtained by performing sample statistical analysis on the historical data of the fully mechanized mining equipment. Figure 3 It is a frequency histogram of manually adjusting the time for lowering the column and the time for raising the column. It can be found from the histogram that: the operation time of 1-5 seconds accounts for more than 60%, indicating that the operation time of manual adjustment is short, and most adjustments are made by inching operation. Therefore, the embodiment of the present invention recommends that the manual adjustment operation time parameter is: 2-5s, that is, the manual column lowering time and the manual column raising time are 2-5s.

[0114] Furthermore, the embodiment of the present invention determines the target pressure value of the column to be reached by the lifting column according to the initial support force of the hydraulic support used in each intelligent mining working face.

[0115] Figure 4 and Figure 5The following are schematic diagrams of the implementation process of two embodiments of the present invention. The embodiment of the present invention uses big data mining technology, based on the production data collected by the chute centralized control center, and uses intelligent algorithms to deeply mine the logic of manual intervention behavior after the automation of the hydraulic support cluster contained in the production process data of the intelligent mining working face, find the data law of manual intervention conditions after the automation of the hydraulic support cluster, and based on this law, intelligently judge whether manual intervention behavior is needed, and recommend operation strategies for working conditions that require manual intervention behavior, which can not only reduce the workload of workers, but also promote the construction of intelligent coal mines.

[0116] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the technical principles of the present invention, and these improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A method for recommending manual control strategies after automation of hydraulic supports, characterized in that: The steps include: S1, collects the current data of the fully mechanized mining equipment in the current production status of the intelligent mining working face; S2, preprocessing the current data of the comprehensive mining equipment to obtain the current operation data of the comprehensive mining equipment; S3, extracting the characteristic value of the current operation data; S4, inputting the characteristic values ​​of the current operating data into a pre-trained decision tree model for identifying manual intervention conditions; S5, judging whether the current working condition of the fully mechanized mining equipment requires manual intervention according to the recognition result of the manual intervention working condition recognition decision tree model; S6, if it is determined that the current working condition of the fully mechanized mining equipment requires manual intervention, an operation strategy is recommended for the manual intervention after the automation of the hydraulic support cluster according to a preset operation strategy recommendation model; Before inputting the characteristic value of the current operation data into the pre-trained decision tree model for identifying the manual intervention condition, S4 further includes: S41, obtaining historical data of comprehensive mining equipment of historical production status of intelligent mining working face; S42, preprocessing the historical data of the fully mechanized mining equipment to obtain historical operation data of the fully mechanized mining equipment; S43, extracting characteristic values ​​of the historical operation data; S44, training a decision tree model for identifying manual intervention conditions by using the characteristic values ​​of the historical operation data; The characteristic values ​​of the historical operation data include the change in the stroke of the automatic follower forward and backward cylinder ΔX, the automatic follower support distance X, and the absolute value Δp of the difference between the position of the coal mining machine and the hydraulic support number to be judged. When extracting the characteristic values ​​of the historical operation data, S43 is implemented by the following formulas (1) to (3): ΔX=MAX(x1)-MAX(x2) (1) In formula (1): x1 is the stroke of the push cylinder before automatic following, and x2 is the stroke of the push cylinder after automatic following; In formula (2): Assume that T (n-1) is the time point corresponding to the manual intervention of the pulling action, then T (n-2) is the time point corresponding to the automatic follow-up frame pulling action; x T(n-2)-10 x is the stroke of the hydraulic cylinder 10 seconds before the start of the automatic follow-up frame pulling action. T(n-2)+6 It is the travel of the push cylinder 6s after the automatic following and pulling action starts; Δp=|NP (CMJ) | (3) In formula (3): N is the number of the hydraulic support to be judged, P (CMJ) It is the frame number of the hydraulic support where the coal mining machine is located.

2. The method for recommending manual control strategies after automation of hydraulic support according to claim 1 is characterized in that: When the S42 pre-processes the historical data of the comprehensive mining equipment, it includes: S421, performing outlier processing and missing value processing on the historical data of the fully mechanized mining equipment; S422, filtering the historical data of the fully mechanized mining equipment after outlier processing and missing value processing according to the column pressure value of the hydraulic support, the stroke of the push cylinder, the action data and the position data of the coal mining machine; S423, sorting each type of filtered data by time to obtain historical operation data of the comprehensive mining equipment.

3. The method for recommending manual control strategies after automation of hydraulic support according to claim 1 is characterized in that: The characteristic value of the historical operation data also includes the column pressure value Y after the column is raised. When extracting the characteristic value of the historical operation data, S43 is also implemented by the following formula (4): And=and (T(n-2)+20) (4) In formula (4), y (T(n-2)+20) It is the column pressure value 20s after the automatic following frame pulling action starts.

4. The method for recommending manual control strategies after automation of hydraulic support according to claim 1 is characterized in that: Before training the manual intervention condition identification decision tree model by using the characteristic values ​​of the historical operation data, S44 further includes: Performing data cleaning on the characteristic values ​​of the historical operation data to remove abnormal characteristic values; The eigenvalues ​​after data cleaning are standardized.

5. The method for recommending manual control strategies after automation of hydraulic support according to claim 1, characterized in that: When the S6 recommends an operation strategy for the manual intervention behavior after the automation of the hydraulic support cluster according to the preset operation strategy recommendation model, it includes: According to the preset operation strategy recommendation model, the manual intervention behavior after the automation of the hydraulic support cluster is recommended in terms of the manual frame pulling adjustment amount, manual column lowering time, manual column raising time and the target column pressure value achieved by raising the column.

6. The method for recommending manual control strategies after automation of hydraulic support according to claim 5, characterized in that: For the intelligent mining working face without inertial navigation installed, the stroke accumulation value calculation method is used to determine the manual pull adjustment amount; for the intelligent mining working face with inertial navigation installed, the inertial navigation is used to calculate the manual pull adjustment amount; Recommend the duration of manual column lowering and manual column raising based on the statistical analysis results of historical data of fully mechanized mining equipment; The target pressure value of the column to be reached by the lifting column is determined based on the initial support force of the hydraulic support used in each intelligent mining working face.

7. The method for recommending manual control strategies after automation of hydraulic support according to claim 6, characterized in that: The calculation formula of the travel cumulative value calculation method is: X'=m·s g -(s1+s2+···+s m ) (5) In formula (5), X' is the manual pull adjustment amount, m is the number of work cycles completed on the day, s g The distance between the support frames specified in the coal mine operation regulations, s1, s2, ..., s m The pulling distance completed by the bracket in the mth working cycle; The calculation formula for calculating the manual rack adjustment amount using inertial navigation is: In formula (6), X' is the manual pull adjustment amount, D is the displacement relative to the coordinate origin, θ is the pitch angle of the coal mining machine, is the heading angle of the coal mining machine.

Images