Scraper S-bend position detection method based on support pushing and action

Abstract

The invention discloses a scraper S-bend position detection method based on support pushing and action. Comprising the steps of recognizing the S bend from pushing of a support, recognizing the S bend from action of the support, and finally conducting double verification analysis to determine the accurate S bend state of the scraper conveyor. According to the method, the problem of how to effectively combine bracket action and pushing data is solved, the recognition precision and the response speed of the system are improved by fusing multi-source data, and the accurate bending position of the scraper is calculated.

Description

Technical Field

[0001] The invention relates to the technical field of coal mine comprehensive excavation tunnel support, and in particular to a method for detecting the S-bend position of a scraper based on the displacement and movement of a support. Background Art

[0002] Longwall fully mechanized mining is one of the most efficient underground coal mining methods. The fully mechanized mining face consists of more than 300 pieces of equipment, including drum shearers, hydraulic supports, scraper conveyors, crushers, bridge transfer machines and belt conveyors. These equipment together form a complex system of mechanical equipment interacting with coal and rock. In order to improve production efficiency and reduce the workload of miners, automated and intelligent mining has become a common goal pursued by the global coal mining industry.

[0003] The intelligent comprehensive mining working face includes three points: reliable coal cutting, maintaining the geometric relationship of the working face, and reliable support of the roof. It aims to achieve reliable and automatic coal cutting of the coal mining machine by integrating advanced sensing technology and data analysis technology, while ensuring that the working face maintains a certain straightness and maintains a stable coupling relationship between the hydraulic support and the surrounding rock.

[0004] At present, the industry mainly calculates the position and inclination of the scraper by installing inertial navigation equipment on the coal mining machine, outputting the position and inclination of the coal mining machine at the current time point and other data. The specific implementation method is to first install the inertial navigation equipment on the coal mining machine, read the position information and attitude information, and transmit the collected position information and attitude information to the ground control center through a specific transmission protocol, and then filter the collected data for noise, obtain the accurate position and attitude of the coal mining machine, and use geometric calculation and dynamic model to calculate the position and attitude of the scraper.

[0005] The defects of the above-mentioned prior art are: 1. First, since the coal mining machine runs in a complex working face environment, the inertial navigation equipment is easily affected by vibration, temperature changes and electromagnetic interference, which leads to a decrease in data accuracy, thereby affecting the accuracy of scraper position calculation; 2. Secondly, there are delays in the data transmission and processing process, and the inertial navigation transmission data has a lag, which cannot realize real-time scraper position detection and cannot reflect the scraper pushing process of the physical working face in real time; 3. Finally, the inertial navigation equipment has a high maintenance cost and requires professional personnel to operate and maintain it. Moreover, because it is a single device, its ability to respond to emergencies is weak. Summary of the invention

[0006] The purpose of the present invention is to address the above-mentioned problems and to provide a method for detecting the S-bend position of a scraper based on the displacement and movement of the support. The method aims to intelligently analyze the movement and position of the support by utilizing the support displacement data, support movement data, and basic sensor measurement point information such as the position of the coal mining machine, and to accurately identify the S-bend position of the scraper through relevant algorithms, thereby achieving efficient and accurate results.

[0007] In order to achieve the above object, the technical solution of the present invention is:

[0008] A method for detecting the S-bend position of a scraper based on the displacement and movement of a bracket, characterized in that it includes identifying the S-bend from the displacement of the bracket, identifying the S-bend from the movement of the bracket, and finally performing a double verification analysis to determine the accurate S-bend state of the scraper.

[0009] As an improvement to the above technical solution, the steps of identifying the S-bend based on the displacement of the stent are:

[0010] S1, obtaining the first support travel data and infrared coal machine trajectory data;

[0011] S2, performing data integrity verification and data supplementation on the first bracket pushing stroke data and infrared coal machine trajectory data;

[0012] S3, S-bend feature recognition;

[0013] S4, S-bend direction is determined;

[0014] S5, S-bend reliability identification.

[0015] As an improvement on the above technical solution, in step S1, the first support's displacement travel data is acquired in a step-by-step manner according to historical time, and the displacement data set of each minute in the last 10 minutes is acquired in turn, and the most reliable first support's displacement travel data is acquired based on the obligation analysis; the infrared coal mining machine trajectory data refers to the acquisition of the last 20 non-stop data of the coal mining machine within 24 hours.

[0016] As an improvement to the above technical solution, in step S2, the first bracket pushing stroke data integrity check is performed to calculate the proportion of the number of empty values ​​in the data group, and the calculation result is compared with the first bracket pushing stroke data integrity threshold. If the proportion does not meet the data integrity threshold ≧65% requirement, the first bracket pushing stroke data cleaning operation is performed; the first bracket pushing stroke data supplementation must first supplement the first and last first bracket pushing stroke data, and then use linear interpolation to supplement the first bracket pushing stroke data in the middle area.

[0017] As an improvement to the above technical solution, in step S3, the S-bend feature recognition refers to performing two feature recognition operations on the first bracket displacement travel data of a plurality of time slices that meet the conditions:

[0018] 1. The difference between the front and rear values ​​of the first bracket's displacement data is greater than 50, and one of the values ​​is not less than 100, and the feature point takes the maximum value of the rise or fall, and this feature is recorded as N1;

[0019] 2. The value of the first bracket displacement data is less than 100, and this feature is recorded as N2;

[0020] 3. Calculate the number of feature N1 and feature N2 respectively, and then determine whether the number of N1 is greater than 4. If the condition is met, directly obtain the first and last frame numbers of the group and start the recognition of the next group; if not, determine whether the number of N1+N2 meets the requirement. If so, store N1 as 0 and not 0 respectively and start the recognition of the next group.

[0021] As an improvement to the above technical solution, the feature recognition value priority is: N1>N1+N2>N2.

[0022] As an improvement to the above technical solution, in step S4, determining the S-bend direction refers to identifying the S-bend direction after acquiring the S-bend feature recognition data:

[0023] Set the positive direction from small to large to 1, and the negative direction from large to small to -1; perform polynomial fitting on the S-bend result data set, the formula is as follows:

[0024]

[0025] Where m is specified as 1;

[0026] Use the least squares method to fit the polynomial, and after obtaining the coefficient list, judge the slope to determine whether the data set is in an upward or downward trend, thereby obtaining the direction of the S-bend.

[0027] As an improvement to the above technical solution, in step S5, the S-bend reliability identification refers to obtaining one or more groups of S-bends identified, and determining a group of the most reliable S-bends therefrom. The S-bend reliability determination steps are as follows:

[0028] 1. Obtain the infrared coal mining machine trajectory data of the middle support in each group of S-bend supports. If the infrared coal mining machine trajectory data exists, obtain the coal mining machine direction according to the latest N non-stop data of the infrared coal mining machine trajectory data, so as to determine the S-bend direction;

[0029] 2. If the infrared coal machine trajectory data does not exist (i.e., the latest N non-stop data cannot be obtained), determine whether there are continuous step S-bends. If continuous step S-bends are found, take the S-bend with the most step feature points in the continuous group. If the number of step feature points is the same, take the first group of S-bends after sorting.

[0030] 3. If no S-bend recognition with continuous steps is found, take the S-bend with the most step feature points in all the recognition groups. If the number of step feature points is the same, take the first group of S-bends after sorting.

[0031] As an improvement to the above technical solution, the steps of identifying the S-bend based on the bracket action are:

[0032] S6, acquiring motion data, respectively acquiring the second bracket moving stroke data, bracket manual motion data and bracket automatic working data; performing timeline merging processing on the same manual motion data and automatic motion data;

[0033] S7, identify the continuous action group, traverse the action data in sequence, add the first action directly to the temporary storage area, compare the time difference of adjacent actions, if the time difference of adjacent actions is less than 0.5s, it can be recorded and stored, and determine whether the number N of adjacent brackets in the temporary storage area is greater than the set value. If the condition is met, save the regrouped data;

[0034] Continuity identification is performed on each group of brackets. If the difference between adjacent brackets is greater than 2, the brackets are grouped and split, and finally the forward sliding action group and the backward sliding action group are obtained;

[0035] S8, identifying the S-bend, for the backward sliding action group, send it into the model for identification, and obtain the nearest group, which is the backward sliding S-bend;

[0036] For the forward sliding action group, determine how many brackets in each group have a displacement change greater than 100 through the second bracket displacement stroke data. If the number of brackets that meet the conditions is greater than or equal to X brackets, it is an S-bend that meets the conditions. Finally, take the group action closest to the current one as the S-bend.

[0037] In step S6, the second bracket shift travel data obtains all shift travel data of each bracket in the last two minutes, the bracket manual motion data obtains the bracket manual motion data in the last two minutes, and the bracket automatic motion data obtains the bracket automatic motion data in the last two minutes.

[0038] Compared with the prior art, the present invention has the following advantages and positive effects:

[0039] The present invention uses support displacement data, support motion data, and basic sensor measurement point information such as coal mining machine position to efficiently and accurately identify the S-bend position of the scraper. The method intelligently analyzes the support motion and position, and realizes accurate identification of the S-bend position of the scraper through relevant algorithms.

[0040] The present invention solves the problem of how to effectively combine the support action and displacement data, improve the recognition accuracy and response speed of the system by fusing multi-source data, and calculate the accurate bending position of the scraper.

[0041] Specific:

[0042] 1. Multi-source data fusion: Utilize the support movement data, support action data and coal mining machine position data to achieve multi-source data fusion, improve data reliability and strengthen risk resistance.

[0043] 2. Comprehensive analysis of the bracket's motion and position overcomes the problem of insufficient accuracy of single inertial navigation data.

[0044] 3. Real-time and high efficiency: It can quickly process and analyze large amounts of data, provide real-time detection results of the S-bend position of the scraper, and reduce the lag in data processing.

[0045] 4. Easy maintenance: The technology uses automatic data collection and processing to reduce manual intervention, thereby reducing the cost and difficulty of maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0047] Figure 1 A detailed schematic diagram for the push travel data;

[0048] Figure 2 It is the S-bend and top view of the scraper;

[0049] Figure 3 A flowchart of steps for identifying S-bends based on stent displacement;

[0050] Figure 4 A flowchart of the steps for identifying S-bends based on bracket movements. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work, any modifications, equivalent substitutions, improvements, etc., should be included in the protection scope of the present invention.

[0052] Figure 1 This is a schematic diagram for explaining the travel data. Figure 1As shown in the figure, the displacement data is the data measured by the displacement sensor at the support push rod. This data can be used to calculate the distance between the support and the scraper connected to it. During the mining process, the support and the scraper are slowly pushed forward along with the mining process, that is, pushing and pulling the frame. Figure 1 In the figure, 1 is the bracket, 2 is the push rod, and 3 is the scraper. The push stroke sensor can detect the extension and retraction distance of the push rod in real time.

[0053] The bracket action data is the data transmitted by the bracket controller when the bracket makes a specific action. This data can be used to obtain the real-time operation status of the bracket. This application mainly obtains the two actions of the bracket pushing forward and sliding backward, which include manual and automatic actions. Based on this action, batch movement can be determined, thereby determining the approximate range of the S-bend.

[0054] The position data of the shearer is the position data of the bracket number where the shearer is located. There are two ways to obtain it, namely, the shearer infrared sensor and the shearer shaft coding sensor. The main purpose of this application to obtain the shearer is to determine the positive and negative directions of the S-bend of the scraper according to the position of the shearer and the coal mining process.

[0055] S-bend is the bending of the scraper when viewed from above due to the support frame moving and sliding. Figure 2 As shown, during the normal coal cutting production process, as the coal mining machine moves, the passing supports will be pushed and moved in batches. The intersection of the frame movement and the frame movement that has not been executed is where the S-bend is generated. The purpose of this application is to obtain the frame number range of the scraper S-bend through the push data and the action data. After combined analysis of the two methods, the final S-bend recognition error is within 2 frames, realizing real-time monitoring of the S-bend position.

[0056] This application is divided into three modules: push recognition S-bend, action recognition S-bend, and double verification analysis.

[0057] Steps for identifying S-bends based on displacement:

[0058] The first step is to obtain data, and obtain the first support travel data and infrared coal machine trajectory data respectively. The acquisition of the first displacement data requires a step-by-step acquisition of the historical time, and obtains the displacement data set of each minute in the last 10 minutes in turn, and obtains the most reliable first displacement data based on the obligation analysis. The infrared coal machine trajectory obtains the latest 20 non-stop data within 24 hours.

[0059] The second step is data processing, which is divided into data integrity verification and data supplementation. Data integrity verification requires calculating the percentage of null values ​​in the data group and comparing it with the data integrity threshold. If the percentage does not meet the threshold requirement, data cleaning is performed. Data supplementation requires supplementing the data of the first bracket push stroke at the beginning and end, and then using linear interpolation to supplement the data of the first bracket push stroke in the middle area. The data integrity threshold requirement is ≧65%.

[0060] The third step is S-bend recognition. Two feature recognition operations are performed on the first bracket displacement data of several time slices that meet the conditions, and two features are identified: 1. The difference between the front and back values ​​of the first bracket displacement data is greater than 50, and one of the values ​​is not less than 100, and the feature point takes the maximum value of the rise or fall, and this feature is recorded as N1. 2. The value of the first bracket displacement data is less than 100, and this feature is recorded as N2. Calculate the number of feature one N1 and the number of feature two N2 respectively, and then determine whether N1 is greater than the condition threshold (the condition threshold is 4). If it meets the conditions, directly obtain the first and last frame numbers of the group and start the recognition of the next group; if it does not meet the conditions, determine whether the number N1+N2 meets the conditions. If it meets the conditions, store N1 as 0 and not 0 respectively and start the recognition of the next group. The priority of feature recognition values ​​is: N1>N1+N2>N2.

[0061] The fourth step is to determine the direction of the S-bend. After obtaining the S-bend identification data, the S-bend direction is identified, and the positive direction is set to 1 from small to large, and the negative direction is set to -1 from large to small. The S-bend result data set is fitted with a polynomial, and the formula is as follows, specifying m as 1.

[0062]

[0063] Use the least squares method to fit the polynomial, and after obtaining the coefficient list, judge the slope to determine whether the data set is in an upward or downward trend, thereby obtaining the direction of the S-bend.

[0064] The fifth step is the reliability identification of S-bends. After the S-bend identification, one or more groups of S-bends are obtained. This stage is to determine a group of the most reliable S-bends. If there are multiple groups of S-bends, the steps are as follows: obtain the infrared coal machine trajectory data of the middle-numbered bracket in each group of S-bend brackets. If the infrared coal machine trajectory data exists, obtain the direction of the coal mining machine based on the latest N non-stop data of the infrared coal machine trajectory data, thereby determining the direction of the S-bend. If the infrared coal machine trajectory data does not exist (that is, the latest N non-stop data cannot be obtained), determine whether there are continuous step S-bends. If the S-bend identification of continuous steps is found, take the S-bend with the most step feature points in the continuous group. If the step feature points are the same, take the first group of S-bends after sorting. If the S-bend identification of continuous steps is not found, take the S-bend with the most step feature points in all identification groups. If the step feature points are the same, take the first group of S-bends after sorting.

[0065] As shown in 3, the steps of S-bend recognition based on action are:

[0066] The first step is to acquire data. A total of three types of data need to be acquired, namely the second bracket push stroke data acquisition, bracket manual action acquisition and bracket automatic work acquisition. The second bracket push stroke data acquires short-term data, which refers to all the push stroke data of each bracket in the last two minutes. Manual action and automatic action acquire the bracket push forward and pull backward action data. After acquiring the action data, the same manual action and automatic action are merged on the timeline. The bracket manual action data acquires the bracket manual action data of the last two minutes, and the bracket automatic action data acquires the bracket automatic action data of the last two minutes.

[0067] The second step is to identify the continuous action group, traverse the action data in sequence, add the first action directly to the temporary storage area, compare the time difference of adjacent actions, and record and store if the time difference of adjacent actions is less than 0.5s. Determine whether the number of adjacent brackets N in the temporary storage area is greater than the set value (the set value is 2). If the condition is met, save the regrouped data. Finally, identify the continuity of each group of brackets. If the difference between adjacent brackets is too large, split them into groups, and finally get the front sliding action group and the back sliding action group.

[0068] The third step is to identify the S-bend. For the backward sliding action group, the model is sent for identification, and the nearest group is the backward sliding S-bend. For the forward sliding action group, the second bracket displacement data is used to determine how many brackets in each group have displacement changes greater than the specified threshold (the specified threshold is 100). If the number of brackets that meet the conditions is greater than or equal to X brackets, it is an S-bend that meets the conditions. Finally, the group action closest to the current one is taken as the S-bend.

[0069] The above is the S-bend identification process based on action and displacement respectively. Finally, the two are combined according to the specific business scenario to obtain the accurate S-bend state of the scraper conveyor.

[0070] The fitting methods include:

[0071] 1. Use polynomial interpolation to fit the missing values ​​of the shifted data.

[0072] 2. Use the piecewise cubic Hermite interpolation method to fit the missing values ​​of the shifted data.

[0073] 3. Use Gaussian filtering and Kalman filtering to eliminate data noise.

[0074] 4. Use XGBoost to build a deep learning model to identify S-bend feature points.

[0075] The above methods are all known in the prior art and will not be described in detail here.

[0076] The above technical solutions only reflect the preferred technical solutions of the technical solutions of the present invention. Some changes that may be made to certain parts thereof by technicians in this technical field all reflect the principles of the present invention and fall within the protection scope of the present invention.

Claims

1. A method for detecting the S-bend position of a scraper based on the displacement and movement of a support, characterized in that: It includes identifying the S-bend from the movement of the bracket, identifying the S-bend from the action of the bracket, and finally performing double verification analysis to determine the accurate S-bend state of the scraper conveyor.

2. The method for detecting the S-bend position of a scraper based on the displacement and movement of the support as claimed in claim 1, characterized in that: The steps for identifying S-bends based on stent displacement are: S1, obtaining the first support travel data and infrared coal machine trajectory data; S2, performing data integrity verification and data supplementation on the first bracket pushing stroke data and infrared coal machine trajectory data; S3, S-bend feature recognition; S4, S-bend direction is determined; S5, S-bend reliability identification.

3. The method for detecting the S-bend position of a scraper based on the displacement and movement of the support as claimed in claim 2, characterized in that: In step S1, the first support displacement travel data is obtained in a step-by-step manner according to historical time, and the displacement data set of each minute in the last 10 minutes is obtained in turn, and the most reliable first support displacement travel data is obtained according to the obligation analysis; the infrared coal mining machine trajectory data refers to obtaining the most recent 20 non-stop data of the coal mining machine within 24 hours.

4. The method for detecting the S-bend position of a scraper based on the displacement and movement of the support as claimed in claim 2, characterized in that: In the step S2, the first bracket pushing stroke data integrity check is performed to calculate the percentage of the number of empty values ​​in the data group, and the calculation result is compared with the first bracket pushing stroke data integrity threshold. If the percentage does not meet the data integrity threshold ≧65% requirement, the first bracket pushing stroke data cleaning operation is performed; the first bracket pushing stroke data supplementation must first supplement the first and last first bracket pushing stroke data, and then use linear interpolation to supplement the first bracket pushing stroke data in the middle area.

5. The method for detecting the S-bend position of a scraper based on the displacement and movement of the support as claimed in claim 2, characterized in that: In step S3, the S-bend feature recognition refers to performing two feature recognition operations on the first bracket displacement stroke data of a plurality of time slices that meet the conditions:

1. The difference between the front and rear values ​​of the first bracket's displacement data is greater than 50, and one of the values ​​is not less than 100, and the feature point takes the maximum value of the rise or fall, and this feature is recorded as N1; 2. The value of the first bracket displacement data is less than 100, and this feature is recorded as N2; 3. Calculate the number of feature N1 and feature N2 respectively, and then determine whether the number of N1 is greater than 4. If the condition is met, directly obtain the first and last frame numbers of the group and start the recognition of the next group; if not, determine whether the number of N1+N2 meets the requirement. If so, store N1 as 0 and not 0 respectively and start the recognition of the next group.

6. The method for detecting the S-bend position of a scraper based on the displacement and movement of the support as claimed in claim 5, characterized in that: The priority of feature recognition values ​​is: N1>N1+N2>N2.

7. The method for detecting the S-bend position of a scraper based on the displacement and movement of the support as claimed in claim 2, characterized in that: In step S4, determining the direction of the S-bend refers to identifying the direction of the S-bend after acquiring the S-bend feature recognition data: Set the positive direction from small to large to 1, and the negative direction from large to small to -1; perform polynomial fitting on the S-bend result data set, the formula is as follows: Where m is specified as 1; Use the least squares method to fit the polynomial, and after obtaining the coefficient list, judge the slope to determine whether the data set is in an upward or downward trend, thereby obtaining the direction of the S-bend.

8. The method for detecting the S-bend position of a scraper based on the displacement and movement of the support as claimed in claim 2, characterized in that: In step S5, the S-bend reliability identification refers to obtaining one or more groups of S-bends identified, and determining a group of the most reliable S-bends from them. The S-bend reliability determination steps are as follows:

1. Obtain the infrared coal mining machine trajectory data of the middle support in each group of S-bend supports. If the infrared coal mining machine trajectory data exists, obtain the coal mining machine direction according to the latest N non-stop data of the infrared coal mining machine trajectory data, so as to determine the S-bend direction; 2. If the infrared coal machine trajectory data does not exist (i.e., the latest N non-stop data cannot be obtained), determine whether there are continuous step S-bends. If continuous step S-bends are found, take the S-bend with the most step feature points in the continuous group. If the number of step feature points is the same, take the first group of S-bends after sorting.

3. If no S-bend recognition with continuous steps is found, take the S-bend with the most step feature points in all the recognition groups. If the number of step feature points is the same, take the first group of S-bends after sorting.

9. The method for detecting the S-bend position of a scraper based on the displacement and movement of the support as claimed in claim 2, characterized in that: The steps to identify S-bends based on bracket movements are: S6, acquiring motion data, respectively acquiring the second bracket moving stroke data, bracket manual motion data and bracket automatic working data; performing timeline merging processing on the same manual motion data and automatic motion data; S7, identify the continuous action group, traverse the action data in sequence, add the first action directly to the temporary storage area, compare the time difference of adjacent actions, if the time difference of adjacent actions is less than 0.5s, it can be recorded and stored, and determine whether the number N of adjacent brackets in the temporary storage area is greater than the set value. If the condition is met, save the regrouped data; Continuity identification is performed on each group of brackets. If the difference between adjacent brackets is greater than 2, the brackets are grouped and split, and finally the forward sliding action group and the backward sliding action group are obtained; S8, identifying the S-bend, for the backward sliding action group, send it into the model for identification, and obtain the nearest group, which is the backward sliding S-bend; For the forward sliding action group, determine how many brackets in each group have a displacement change greater than 100 through the second bracket displacement stroke data. If the number of brackets that meet the conditions is greater than or equal to X brackets, it is an S-bend that meets the conditions. Finally, take the group action closest to the current one as the S-bend.

10. The method for detecting the S-bend position of a scraper based on the displacement and movement of the support as claimed in claim 9, characterized in that: In step S6, the second bracket shift travel data obtains all shift travel data of each bracket in the last two minutes, the bracket manual motion data obtains the bracket manual motion data in the last two minutes, and the bracket automatic motion data obtains the bracket automatic motion data in the last two minutes.

Images