Method and system for determining the open state of a hydraulic support safety valve

Abstract

The application provides a method and system for determining the open state of a safety valve of a hydraulic support. The method comprises: obtaining original mine pressure monitoring data of the hydraulic support through a mine pressure monitoring system; manually marking the open state of the safety valve on the original mine pressure monitoring data, and generating a safety valve open marking time sequence according to the marking result; arranging the original mine pressure monitoring data, the safety valve open state time sequence and target data to be determined for the safety valve open state into equidistant data, and generating a Tsfresh time sequence according to the equidistant data; performing feature extraction and model training on the Tsfresh time sequence through a Tsfresh machine learning framework, and determining the safety valve open state result of each group of data in the target data through the trained machine learning model. The method combines manual marking and a supervised machine learning method to analyze the open state of the safety valve, and improves the timeliness and accuracy of the safety valve open state recognition.

Description

Technical Field

[0001] This application relates to the field of coal mine equipment technology, and in particular to a method and system for determining the opening state of a hydraulic support safety valve. Background Technology

[0002] Currently, safety in coal mining is attracting increasing attention. During the longwall mining process, hydraulic supports, as a crucial component of fully mechanized mining equipment, are vital for roof control and are essential equipment for ensuring mine safety.

[0003] The safety valve is a crucial component of the hydraulic support. Its malfunction and the appropriateness of its set pressure are critical to the reliable and effective support and control of the working face roof. Insufficient safety valve pressure reduces support strength and increases the risk of roof collapse; conversely, a malfunctioning safety valve or an excessively high set pressure can lead to cylinder explosion and injury. Therefore, precise control of the safety valve's opening is essential.

[0004] In related technologies, the opening control method for safety valves is relatively simple, generally relying solely on historical experience to manually determine the valve's opening state and control its opening. However, in practical applications, using only this single method of determining the opening state may fail to ensure timely opening of the safety valve in certain situations, or prevent it from opening at a more appropriate time. Summary of the Invention

[0005] This application aims to at least partially address one of the technical problems in the related art.

[0006] Therefore, the first objective of this application is to propose a method for determining the opening state of a hydraulic support safety valve. This method is based on the support pressure monitoring time series data obtained by the mine pressure monitoring system, and uses supervised machine learning to analyze the opening state of the safety valve, which can provide early warning of the support operation status for coal mine managers.

[0007] The second objective of this application is to provide a system for determining the opening state of a hydraulic support safety valve.

[0008] The third objective of this application is to provide a non-transitory computer-readable storage medium.

[0009] To achieve the above objectives, a first aspect of this application provides a method for determining the open state of a hydraulic support safety valve, comprising the following steps:

[0010] The raw mine pressure monitoring data of the hydraulic support is obtained through the mine pressure monitoring system;

[0011] The original mine pressure monitoring data is manually marked with the safety valve opening status, and a safety valve opening marking time series is generated based on the marking results;

[0012] The original mine pressure monitoring data, the safety valve opening status time series, and the target data of the safety valve opening status to be determined are organized into equally spaced data, and a Tsfresh time series is generated based on the equally spaced data.

[0013] The Tsfresh time series data is processed using the Tsfresh machine learning framework to extract features and train a model. The trained machine learning model is then used to determine the safety valve opening status of each data set in the target data.

[0014] Optionally, in one embodiment of this application, manually marking the safety valve opening status on the original mine pressure monitoring data includes: marking the starting point of each round of safety valve opening on the curve of the original mine pressure monitoring data; adding a preset time interval to the starting point to determine the ending point of each round of safety valve opening; and determining the safety valve opening pressure value for each round of safety valve opening based on the vertical axis of the curve.

[0015] Optionally, in one embodiment of this application, generating the safety valve opening state time sequence based on the marking results includes: generating multiple groups based on the start point, the end point, and the safety valve opening pressure value for each round of safety valve opening; and sorting the multiple groups in order from the earliest to the latest occurrence time of the start point to generate the safety valve opening state time sequence.

[0016] Optionally, in one embodiment of this application, the step of organizing the original mine pressure monitoring data, the safety valve opening status time series, and the target data of the safety valve opening status to be determined into equally spaced data includes: traversing the original mine pressure monitoring data and segmenting the original mine pressure monitoring data according to the marked safety valve opening time periods; truncating the non-safety valve opening time periods between two adjacent safety valve opening time periods according to the time interval, and removing data in the truncated non-safety valve opening time periods that do not meet the time interval; truncating the target data according to the time interval, and removing data in the truncated target data that does not meet the time interval.

[0017] Optionally, in one embodiment of this application, generating the Tsfresh time series based on the equally spaced data includes: determining the safety valve opening marker time series, the non-safety valve opening time series, and the safety valve opening status value of each group in the target data time series generated after equally spaced division; sorting each group containing the status value in order from farthest to nearest start time to generate the Tsfresh time series.

[0018] Optionally, in one embodiment of this application, the step of extracting features and training a model from the Tsfresh time series using the Tsfresh machine learning framework includes: setting the type of feature extraction using Tsfresh, and extracting features from the Tsfresh time series using a preset feature extraction function to generate a feature time series; dividing the feature time series into a training set, a validation set, and a test set using a preset segmentation function; and training a preset decision tree classifier using data from the training set to obtain a trained machine learning model for analyzing the opening state of a safety valve.

[0019] Optionally, in one embodiment of this application, the amount of data in the test set is the number of groups in the target data time series, and the step of determining the safety valve opening status result of each group of data in the target data through the trained machine learning model includes: predicting the data in the test set through the trained decision tree classifier to obtain the safety valve opening status result.

[0020] Optionally, in one embodiment of this application, after the safety valve opening status result of each group of data in the target data is determined by the machine learning model that has been trained, the method further includes: visually displaying the safety valve opening status result of each group of data in the support pressure curve of the hydraulic support.

[0021] To achieve the above objectives, a second aspect of this application provides a system for determining the opening state of a hydraulic support safety valve, comprising the following modules:

[0022] The acquisition module is used to acquire raw mine pressure monitoring data of hydraulic supports through the mine pressure monitoring system;

[0023] The marking module is used to manually mark the safety valve opening status on the raw mine pressure monitoring data, and generate a safety valve opening marking time series based on the marking results;

[0024] The generation module is used to organize the original mine pressure monitoring data, the safety valve opening status time series, and the target data of the safety valve opening status to be determined into equally spaced data, and generate a Tsfresh time series based on the equally spaced data;

[0025] The determination module is used to extract features and train models on the Tsfresh time series using the Tsfresh machine learning framework, and to determine the safety valve opening status of each group of data in the target data using the trained machine learning model.

[0026] To implement the above embodiments, a third aspect of this application also proposes a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method for determining the opening state of the hydraulic support safety valve in the above embodiments.

[0027] The technical solution provided by the embodiments of this application brings at least the following beneficial effects: This application first acquires time-series data of support pressure monitoring based on the mine pressure monitoring system, then manually marks the safety valve opening status of the data within a certain period, processes the raw data and the marked data results into the time series required by the machine learning framework tsfresh, and then automatically analyzes the safety valve opening status of the hydraulic support through machine learning methods, including steps such as feature extraction, feature filtering, training, and prediction, and realizes the safety valve opening status analysis through supervised machine learning methods. Therefore, this application enriches the methods of safety valve opening status analysis, can determine the time period of safety valve opening in a timely and accurate manner, thereby providing timely early warning information to coal mine workers, ensuring that the safety valve opens on time within the appropriate time period, improving the timeliness and accuracy of safety valve opening status identification, and contributing to improving the level of coal mine roof management and the safety of coal mining.

[0028] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0029] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein

[0030] Figure 1 This is a flowchart illustrating a method for determining the opening state of a hydraulic support safety valve according to an embodiment of this application.

[0031] Figure 2 This is a schematic diagram of the results of manual marking of the opening state of a safety valve according to an embodiment of this application;

[0032] Figure 3 This is a flowchart illustrating a method for generating equally spaced data according to an embodiment of this application;

[0033] Figure 4 This is a schematic diagram illustrating a data capture method proposed in an embodiment of this application;

[0034] Figure 5 This is a flowchart of a method for predicting the opening state of a safety valve according to an embodiment of this application;

[0035] Figure 6This is a schematic diagram of a system for determining the opening state of a hydraulic support safety valve, as proposed in an embodiment of this application. Detailed Implementation

[0036] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0037] The following description, with reference to the accompanying drawings, illustrates a method and system for determining the opening state of a hydraulic support safety valve according to an embodiment of the present invention.

[0038] Figure 1 This is a flowchart illustrating a method for determining the opening state of a hydraulic support safety valve according to an embodiment of this application, as shown below. Figure 1 The method includes the following steps:

[0039] Step S101: Obtain the raw mine pressure monitoring data of the hydraulic support through the mine pressure monitoring system.

[0040] It should be noted that hydraulic supports are structures used to control mine pressure at the coal face. The mine pressure acts on the hydraulic supports as an external load. Hydraulic supports consist of various components, including hydraulic cylinders, load-bearing structural parts, and pushing devices. The safety valve of the hydraulic support serves multiple purposes, including maintaining a constant working resistance of the support column and releasing fluid from the working chamber when the roof pressure exceeds the support's working resistance, thus providing overload protection. Therefore, it is necessary to accurately analyze the opening status of the safety valve and open it at the appropriate time.

[0041] Most coal mines currently have mine pressure monitoring systems installed, which can monitor the pressure on equipment such as hydraulic supports in real time and upload the monitored pressure data to the ground monitoring host. Therefore, this application utilizes existing mine pressure monitoring systems to obtain raw mine pressure monitoring data of hydraulic supports. The raw mine pressure monitoring data is unprocessed and directly collected by the mine pressure monitoring system, representing the pressure data experienced by the supports at different times; the raw mine pressure monitoring data is time-series data.

[0042] In practice, raw mine pressure monitoring data can be obtained through various methods. As a first example, data can be exchanged with existing mine pressure monitoring systems using various communication protocols, including FTP (File Transfer Protocol) and WebService. Taking WebService as an example, it is a remote invocation technology that is cross-programming language and cross-operating system platform. It allows remote invocation of data from the mine pressure monitoring system from the terminal device executing the safety valve opening status determination method of this application, without considering the different programming languages ​​of different equipment programs in the coal mine, making data acquisition relatively convenient.

[0043] As a second example, a database for storing mine pressure monitoring data from the mine pressure monitoring system can be set up in advance, and the data can be directly read from the database when data is needed.

[0044] In one embodiment of this application, for a longwall face with a non-electro-hydraulic control system, the installed mine pressure monitoring system typically installs one support pressure monitoring sensor every n supports. Therefore, the list of supports with installed sensors is L =<n,2n,…,m×n> Where m is the group number of the stent. In this embodiment, the raw monitoring data of each stent is sorted in ascending order of time value, resulting in the following set:

[0045] S m×n =<(t1,x1),(t2,x2)…,(t i ,x i )>,

[0046] This set represents the original mine pressure monitoring data of the m×n-th support group, (t i ,x i ) is the time t of the monitoring of the stent. i The corresponding raw data of stent pressure monitoring x i , where i is the data record number within the analysis period of the stent.

[0047] Step S102: Manually mark the safety valve opening status on the original mine pressure monitoring data, and generate a safety valve opening mark time series based on the marking results.

[0048] The safety valve opening status includes data such as the opening time, closing time, and opening pressure of the safety valve at the opening time.

[0049] Specifically, select any hydraulic support from all supports and manually mark the safety valve opening status based on the original mine pressure monitoring data for that support. In this application, the safety valve opening status can be determined sequentially for each support in the mine, following the methods described in this and subsequent steps. Specifically, when manually marking the safety valve opening status, various factors such as the support's historical operating data (including the actual opening status of the safety valve before), expert knowledge, and personal experience can be considered.

[0050] In one embodiment of this application, manual marking can be performed through the following steps: first, mark the starting point of each round of safety valve opening on the curve of the original mine pressure monitoring data; add a preset time interval to the starting point to determine the ending point of each round of safety valve opening; and determine the safety valve opening pressure value of each round of safety valve opening based on the vertical axis of the curve.

[0051] Specifically, in this embodiment, the raw mine pressure monitoring data is first used to generate data such as... Figure 2 The graph shows the change in column pressure of the support structure over time. Then, at predetermined time intervals, such as 30 minutes, the starting point t for the safety valve opening is selected from the historical data curve. s Marking the end point t of the safety valve opening indicates that the safety valve is open. e =t s +30, thereby obtaining the start and end times of all safety valve openings and the safety valve opening pressures within a certain period of time by manually marking the support.

[0052] like Figure 2 As shown in the figure, after marking, the horizontal lines (T1 to T8) at the curve positions in the figure represent the safety valve opening results obtained by manual marking. The start and end positions of the horizontal lines represent the start and end times of the safety valve opening, respectively, and the vertical axis scale value corresponding to the horizontal lines is the safety valve opening pressure value.

[0053] Furthermore, based on the results of manual labeling, namely the labeled safety valve opening status, a safety valve opening label time series is generated.

[0054] In one embodiment of this application, generating a safety valve opening state time sequence based on the marking results includes: generating multiple groups based on the start point, end point, and safety valve opening pressure value of each round of safety valve opening; sorting the multiple groups in order from the earliest to the latest occurrence time of the start point to generate a safety valve opening state time sequence.

[0055] Specifically, the results of the safety valve opening markings are compiled, including the markings for each round of safety valve opening. Figure 2The data corresponding to each horizontal line is recombined into grouped data in the form of (start time, end time, safety valve opening pressure). Each round of safety valve opening marks corresponds to one group of data. The groups are then arranged in order of the occurrence time of the starting point from the oldest to the newest, that is, in ascending order of the start time value, to obtain a new time series as shown below:

[0056] T = <(t) s1 ,t e1 ,P1),(t s2 ,t e2 ,P2)…,(t si ,t ei ,P i )>,

[0057] Wherein, the sequence represents the time t at which the safety valve opens each time. si Until the safety valve ends at time t ei The corresponding safety valve opening pressure P i The data represented by this sequence can also be organized into a table.

[0058] Step S103: Organize the original mine pressure monitoring data, the safety valve opening status time series, and the target data of the safety valve opening status to be determined into equally spaced data, and generate the Tsfresh time series based on the equally spaced data.

[0059] Specifically, this application uses the tsfresh machine learning framework to extract features, train, and predict the generated time series. Since the tsfresh time series processing framework requires that the processed time series be equally spaced data, the acquired raw mine pressure monitoring data N of the support structure, the generated safety valve opening marker time series T, and the data P for safety valve opening status analysis need to be processed into equally spaced data required by the tsfresh time series processing algorithm.

[0060] It should be noted that the target data for determining the opening status of the safety valve can be unmarked data from the acquired mine pressure monitoring data, or it can be the pressure monitoring data of the support currently acquired in real time. The target data can be determined according to the actual safety valve opening status analysis requirements.

[0061] In one embodiment of this application, to more clearly illustrate the specific implementation process of processing various data into the required equally spaced data, the following description uses a data processing method proposed in an embodiment of this application as an example. Figure 3 This is a flowchart of a method for generating equally spaced data according to an embodiment of this application, as shown below. Figure 3 As shown, the method includes the following steps:

[0062] Step S201: Traverse the original mine pressure monitoring data and segment the original mine pressure monitoring data according to the marked safety valve opening time period.

[0063] Specifically, the data in the original mine pressure monitoring data N is iterated through in a loop. The original mine pressure monitoring data is divided by the safety valve opening time period marked in step S102 (i.e., the time period between the start and end points of each round of safety valve opening). The divided original mine pressure monitoring data consists of multiple data segments, including the data of each safety valve opening time period and the data of the non-open safety valve opening time period between two adjacent safety valve opening time periods.

[0064] Step S202: Extract the non-safety valve opening time period between two adjacent safety valve opening time periods according to the time interval, and remove the data in the extracted non-safety valve opening time period that does not meet the time interval.

[0065] Specifically, according to the preset time interval used in the manual marking process, such as 30 minutes, the data for each non-open safety valve opening period is extracted in 30-minute intervals. Data remaining less than 30 minutes after the previous extraction is discarded. Figure 4 The method shown extracts data from the segmented non-safety valve opening periods. In the figure, N1 and N2 represent the data groups extracted from the non-safety valve opening periods.

[0066] Step S203: The target data is truncated according to time intervals, and data that does not meet the time intervals in the truncated target data is removed.

[0067] Specifically, following the extraction method in step S202 above, the data P that is about to be analyzed for the safety valve opening is also extracted at 30-minute intervals, and data less than 30 minutes is discarded, thus obtaining more than one time series group.

[0068] Therefore, the original mine pressure monitoring data N, the generated safety valve opening mark time series T, and the target data P to be analyzed for the safety valve opening status are processed into equally spaced data divided by the preset intervals when manually marked.

[0069] Furthermore, Tsfresh time series are generated based on the equally spaced data obtained from the segmentation.

[0070] In one embodiment of this application, generating a Tsfresh time series includes the following steps: determining the safety valve opening status value of each group in the safety valve opening marker time series, the non-safety valve opening time series and the target data time series generated after equal interval segmentation, and then sorting each group containing the status value in order from the earliest to the latest start time to generate a Tsfresh time series.

[0071] Specifically, for the equally spaced safety valve opening marker time series T, the corresponding safety valve opening state is set to True; for the non-safety opening marker time series N, the corresponding safety valve opening state is set to False; and for the target time series P, the safety valve opening state is temporarily assumed to be False. This allows us to determine the safety valve opening state value for each group in each time series. Then, each group is sorted according to its start time from oldest to youngest, and merged into a new time series Ts (i.e., the Tsfresh time series). Refer to the above embodiment. Figure 4 Given the data partitioning scenario, a possible Ts sequence is shown below:

[0072] Ts=<(T1,True),(N1,False),(N2,False)…(T2,True),…,(P1,False)>

[0073] Step S104: Use the Tsfresh machine learning framework to extract features and train the model for the Tsfresh time series, and use the trained machine learning model to determine the safety valve opening status of each group of data in the target data.

[0074] Tsfresh (TimeSeries Fresh) is a third-party Python toolkit that can automatically calculate features for large amounts of time series data and includes methods for feature importance assessment and feature selection.

[0075] Specifically, this application constructs a supervised machine learning framework based on Tsfresh, using Tsfresh to extract features, train, and predict new time series Ts. The analysis yields the safety valve opening status results for each set of data in the sequence P to be analyzed.

[0076] In one embodiment of this application, to more clearly illustrate the specific implementation process of feature extraction and model training for Tsfresh time series data, the following example uses a method for predicting the opening state of a safety valve proposed in this embodiment as an example. Figure 5 This is a flowchart of a method for predicting the opening state of a safety valve according to an embodiment of this application, as shown below. Figure 5 As shown, the method includes the following steps:

[0077] Step S301: Set the type of feature extraction using Tsfresh, and extract features from the Tsfresh time series using a preset feature extraction function to generate a feature time series.

[0078] Specifically, in practical applications, the amount of data extracted using Tsfresh is large, and the computational load for extracting time series features is significant, potentially leading to a long extraction time. Therefore, this application sets the formal parameters for feature extraction to determine the type of feature extraction. In this embodiment, the feature extraction setting type is defined as ComprehensiveFCParameters, which employs an exhaustive method to enumerate all possible parameters. Feature reduction can also be performed during feature extraction.

[0079] Furthermore, features are extracted from the Tsfresh time series using a preset feature extraction function. The feature extraction function can be selected as needed. As an example, the extract_relevant_features function can be used to extract features from the Ts time series to obtain a new feature time series X_filtered, which is the feature time series.

[0080] In practice, as a possible implementation method, it can be achieved through the following code:

[0081] "extract_settings=ComprehensiveFCParameters()

[0082] X_filtered=extract_relevant_features(TsX,TsY,column_id='id',

[0083] column_sort='time',default_fc_parameters=extract_settings)".

[0084] Step S302: The feature time series is divided into a training set, a validation set, and a test set using a preset segmentation function.

[0085] In this embodiment, the segmentation function can be the `train_test_split` function. The `train_test_split` function segments the feature time series X_filtered after the Ts time series is transformed, automatically dividing it into a training set, a validation set, and a test set. The data in the training set is used to train the machine learning model, while the data in the test set is the target data to be predicted; that is, the amount of data in the test set is equal to the number of groups in the target data time series.

[0086] In practice, as a possible implementation method, it can be achieved through the following code:

[0087] "X_train,X_test,X_filtered_train,X_filtered_test,y_train,

[0088] y_test=train_test_split(TsX,X_filtered,TsY,test_size=4)”.

[0089] The number of test_size parameters is the number of groups of the sequence P to be analyzed, which is 4 in this example.

[0090] Step S303: Train the preset decision tree classifier using data from the training set to obtain a trained machine learning model for analyzing the opening state of the safety valve.

[0091] Among them, the Decision Tree Classifier is a type of classifier that provides a set of attributes and classifies data by making a series of decisions based on these attributes. A decision tree is a tree-like structure where each branch represents a test output. Decision tree learning uses a top-down recursive method for training.

[0092] Specifically, the result type of the safety valve opening state analysis output by the decision tree classifier is determined to be (True and False). The data in the training set is used as training data to train the decision tree classifier. It is understood that since the training set includes manually labeled data, it meets the model training requirements. The specific training process of the decision tree classifier can be found in relevant technical documentation and will not be elaborated upon here.

[0093] Step S304: The trained decision tree classifier is used to predict the data in the test set to obtain the safety valve opening status result.

[0094] Specifically, after the decision tree classifier is trained, it predicts the safety valve opening state to obtain the analysis results. That is, through the trained decision tree classifier model, the output result type of the safety valve opening state analysis for each group of data in the target data of the test set includes True or False. At the same time, if it is True, the predicted safety valve opening pressure value can also be output.

[0095] In practice, as a possible implementation method, it can be achieved through the following code:

[0096] "cl=DecisionTreeClassifier()

[0097] cl.fit(X_train,y_train)

[0098] cl.predict(X_test)".

[0099] Therefore, this application can obtain the safety valve opening status of each group of data in the target data to be analyzed, so as to determine whether the safety valve needs to be opened.

[0100] Furthermore, in order to more clearly display the identification results of the safety valve opening status, in one embodiment of this application, after determining the safety valve opening status result of each group of data in the target data through the trained machine learning model, the method further includes: visually displaying the safety valve opening status result of each group of data on the support pressure curve of the hydraulic support.

[0101] Specifically, the stent pressure curve can be as follows: Figure 2 As shown, a graph with time on the horizontal axis and support pressure on the vertical axis displays the support pressure curve at different times. In this embodiment, the graph is generated based on the support pressure data at different times, and the predicted safety valve opening state is then displayed. Figure 2 The horizontal lines shown are displayed in the curve graph, thus providing a visual representation through chart drawing, making it easy for relevant staff to clearly and intuitively view the recognition results.

[0102] Furthermore, in one embodiment of this application, early warning information can be provided to relevant personnel in the coal mine based on the determined safety valve opening status. For example, the terminal device executing the safety valve opening status determination method of this application can pre-store the account information of the mobile terminals (e.g., mobile phones) of relevant legitimate personnel. In practical applications, after establishing a wireless connection with the mobile terminals of relevant personnel, their legitimacy is verified. After successful verification, early warning information can be sent to the relevant mobile terminals via a wireless network. For instance, when it is determined that the opening status of the hydraulic support safety valve is True at a certain moment, and the safety valve needs to be opened, early warning information can be sent to the mobile terminals in advance to remind personnel to open the safety valve in a timely manner. Thus, the more accurate safety valve opening status determined through machine learning assists personnel in controlling the safety valve, adding a new means to safety valve opening analysis and helping to improve the level of coal mine roof management.

[0103] In summary, the method for determining the opening status of the hydraulic support safety valve in this application first acquires time-series data of support pressure monitoring based on the mine pressure monitoring system. Then, the data within a certain period is manually marked with the safety valve opening status. The raw data and the marked data results are processed into the time series required by the machine learning framework tsfresh. Finally, the opening status of the hydraulic support safety valve is automatically analyzed using machine learning methods, including feature extraction, feature filtering, training, and prediction. Supervised machine learning is used to achieve the safety valve opening status analysis. Therefore, this method enriches the ways to analyze the safety valve opening status, enabling timely and accurate determination of the safety valve opening period. This allows for timely early warning information to coal mine workers, ensuring that the safety valve opens on time at the appropriate time, improving the timeliness and accuracy of safety valve opening status identification, and contributing to improved coal mine roof management and safety during coal mining.

[0104] To achieve the above embodiments, this application also proposes a system for determining the opening state of a hydraulic support safety valve. Figure 6 This is a schematic diagram of a system for determining the opening state of a hydraulic support safety valve, as proposed in an embodiment of this application.

[0105] like Figure 6 As shown, the system includes an acquisition module 100, a marking module 200, a generation module 300, and a determination module 400.

[0106] The acquisition module 100 is used to acquire the original mine pressure monitoring data of the hydraulic support through the mine pressure monitoring system.

[0107] The marking module 200 is used to manually mark the safety valve opening status on the raw mine pressure monitoring data, and generate a safety valve opening marking time series based on the marking results.

[0108] The generation module 300 is used to organize the original mine pressure monitoring data, the safety valve opening status time series, and the target data of the safety valve opening status to be determined into equally spaced data, and generate Tsfresh time series based on the equally spaced data.

[0109] The determination module 400 is used to extract features and train models on Tsfresh time series data using the Tsfresh machine learning framework, and to determine the safety valve opening status of each group of data in the target data using the trained machine learning model.

[0110] Optionally, in one embodiment of this application, the marking module 200 is specifically used to: mark the starting point of each round of safety valve opening in the curve graph of the original mine pressure monitoring data; add a preset time interval to the starting point to determine the ending point of each round of safety valve opening; and determine the safety valve opening pressure value of each round of safety valve opening based on the vertical axis of the curve graph.

[0111] Optionally, in one embodiment of this application, the marking module 200 is further configured to: generate multiple groups based on the start point, end point and safety valve opening pressure value of each round of safety valve opening; and sort the multiple groups in order from the earliest to the latest occurrence time of the start point to generate a safety valve opening state time sequence.

[0112] Optionally, in one embodiment of this application, the generation module 300 is specifically used for: traversing the original mine pressure monitoring data, segmenting the original mine pressure monitoring data according to the marked safety valve opening time period; truncating the non-safety valve opening time period between two adjacent safety valve opening time periods according to the time interval, and removing data in the truncated non-safety valve opening time period that does not meet the time interval; truncating the target data according to the time interval, and removing data in the truncated target data that does not meet the time interval.

[0113] Optionally, in one embodiment of this application, the generation module 300 is further configured to: determine the safety valve opening status value of each group in the safety valve opening marker time series, the non-safety valve opening time series and the target data time series generated after equal interval segmentation; and sort each group containing the status value in order of the earliest to the latest start time to generate the Tsfresh time series.

[0114] Optionally, in one embodiment of this application, the determining module 400 is specifically used to: set the type of feature extraction using Tsfresh, and extract features from the Tsfresh time series using a preset feature extraction function to generate a feature time series; divide the feature time series into a training set, a validation set, and a test set using a preset segmentation function; and train a preset decision tree classifier using data from the training set to obtain a trained machine learning model for analyzing the opening state of the safety valve.

[0115] Optionally, in one embodiment of this application, the amount of data in the test set is the number of groups in the target data time series, and the determination module 400 is further configured to: predict the data in the test set using the trained decision tree classifier to obtain the safety valve opening status result.

[0116] Optionally, in one embodiment of this application, the system further includes a display module for visually displaying the safety valve opening status results of each set of data on the support pressure curve of the hydraulic support.

[0117] It should be noted that the description of the aforementioned method for determining the opening state of the hydraulic support safety valve is also applicable to the system of this embodiment, and the implementation principle is the same, so it will not be repeated here.

[0118] In summary, the hydraulic support safety valve opening status determination system of this application embodiment first acquires support pressure monitoring time-series data based on the mine pressure monitoring system. Then, it manually marks the safety valve opening status of the data over a period of time. The raw data and marked data results are processed into the time series required by the machine learning framework tsfresh. Finally, the system automatically analyzes the hydraulic support safety valve opening status using machine learning methods, including feature extraction, feature filtering, training, and prediction. Supervised machine learning methods are used to achieve safety valve opening status analysis. Therefore, this system enriches the methods for safety valve opening status analysis, enabling timely and accurate determination of the safety valve opening period. This allows for timely early warning information to coal mine workers, ensuring that the safety valve opens on time at the appropriate time, improving the timeliness and accuracy of safety valve opening status identification, and contributing to improved coal mine roof management and mining safety.

[0119] To implement the above embodiments, the present invention also proposes a non-transitory computer-readable storage medium storing a computer program thereon, wherein the computer program, when executed by a processor, implements the method for determining the opening state of the hydraulic support safety valve as described in the first aspect embodiment of the present application.

[0120] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the use of illustrative expressions for the above terms in multiple embodiments or examples does not imply that these embodiments or examples are identical. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0121] 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 at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0122] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0123] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0124] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0125] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it includes one or a combination of the steps of the method embodiments.

[0126] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0127] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A method for determining the open state of a hydraulic support safety valve, characterized in that, Includes the following steps: The raw mine pressure monitoring data of the hydraulic support is obtained through the mine pressure monitoring system; Manually marking the safety valve opening status on the raw mine pressure monitoring data includes: marking the starting point of each round of safety valve opening on the curve of the raw mine pressure monitoring data; adding a preset time interval to the starting point to determine the ending point of each round of safety valve opening; and determining the safety valve opening pressure value for each round of safety valve opening based on the vertical axis of the curve. Generate a safety valve opening time sequence based on the marking results; Organizing the original mine pressure monitoring data, the safety valve opening status time series, and the target data of the safety valve opening status to be determined into equally spaced data includes: traversing the original mine pressure monitoring data, segmenting the original mine pressure monitoring data according to the marked safety valve opening time period, truncating the non-safety valve opening time period between two adjacent safety valve opening time periods according to the time interval, and removing data in the truncated non-safety valve opening time period that does not meet the time interval; truncating the target data according to the time interval, and removing data in the truncated target data that does not meet the time interval. Generate a Tsfresh time series based on the equally spaced data; The Tsfresh time series data is processed using the Tsfresh machine learning framework to extract features and train a model. The trained machine learning model is then used to determine the safety valve opening status of each data set in the target data.

2. The method for determining the opening state of the hydraulic support safety valve according to claim 1, characterized in that, The step of generating a safety valve opening state time series based on the marking results includes: Multiple groups are generated based on the starting point, the ending point, and the safety valve opening pressure value for each round of safety valve opening; The multiple groups are sorted in order of their occurrence time from oldest to youngest to generate the safety valve opening state time sequence.

3. The method for determining the opening state of the hydraulic support safety valve according to claim 1, characterized in that, The step of generating the Tsfresh time series based on the equally spaced data includes: Determine the safety valve opening status value of each group in the safety valve opening marker time series, the non-safety valve opening time series, and the target data time series generated after equal interval segmentation; Each group containing the state value is sorted in order of its start time from oldest to newest to generate the Tsfresh time series.

4. The method for determining the opening state of the hydraulic support safety valve according to claim 3, characterized in that, The step of using the Tsfresh machine learning framework to extract features and train a model for the Tsfresh time series includes: Set the type of feature extraction using Tsfresh, and extract features from the Tsfresh time series using a preset feature extraction function to generate a feature time series; The feature time series is divided into a training set, a validation set, and a test set using a preset segmentation function; The preset decision tree classifier is trained using data from the training set to obtain a trained machine learning model for analyzing the opening state of the safety valve.

5. The method for determining the opening state of the hydraulic support safety valve according to claim 4, characterized in that, The amount of data in the test set is the number of groups in the target data time series. The determination of the safety valve opening status of each group of data in the target data by the trained machine learning model includes: The safety valve opening status is obtained by predicting the data in the test set using a trained decision tree classifier.

6. The method for determining the opening state of the hydraulic support safety valve according to claim 1, characterized in that, After the machine learning model, which has been trained, determines the safety valve opening status of each group of data in the target data, the method further includes: The safety valve opening status results for each set of data are plotted on the support pressure curve of the hydraulic support for visualization.

7. A system for determining the opening state of a hydraulic support safety valve, characterized in that, Includes the following modules: The acquisition module is used to acquire raw mine pressure monitoring data of hydraulic supports through the mine pressure monitoring system; The marking module is used to manually mark the safety valve opening status on the raw mine pressure monitoring data, including: marking the starting point of each round of safety valve opening on the curve of the raw mine pressure monitoring data; adding a preset time interval to the starting point to determine the ending point of each round of safety valve opening; determining the safety valve opening pressure value for each round of safety valve opening based on the vertical axis of the curve; and generating a safety valve opening marking time sequence based on the marking results. The generation module is used to organize the original mine pressure monitoring data, the safety valve opening status time series, and the target data of the safety valve opening status to be determined into equally spaced data. This includes: traversing the original mine pressure monitoring data; segmenting the original mine pressure monitoring data according to the marked safety valve opening time periods; truncating the non-safety valve opening time periods between two adjacent safety valve opening time periods according to the time interval, and removing data in the truncated non-safety valve opening time periods that do not meet the time interval; truncating the target data according to the time interval, and removing data in the truncated target data that does not meet the time interval; and generating a Tsfresh time series based on the equally spaced data. The determination module is used to extract features and train models on the Tsfresh time series using the Tsfresh machine learning framework, and to determine the safety valve opening status of each group of data in the target data using the trained machine learning model.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for determining the opening state of the hydraulic support safety valve as described in any one of claims 1-6.

Images