A method of determining the production footage of a mining face

Abstract

The application belongs to the technical field of positioning and ranging, and specifically discloses a method for determining the production footage of a mining working face, which comprises the following steps: collecting positioning and ranging data of a mining machine to determine the mining working face; modeling the mining working face to identify a mining operation cycle and an abnormal mode; the mining operation cycle is a tunneling operation cycle or a stoping operation cycle; the abnormal mode is identified and verified by using a variable point detection and trend test algorithm to obtain an abnormal mode subsequence; coordinate values after modeling are obtained, the coordinate values being the farthest end coordinate values of the tunneling operation cycle or the nearest end coordinate values of the stoping operation cycle; then, the obtained coordinate values are corrected according to the obtained abnormal mode subsequence, and the daily footage is calculated according to the corrected coordinate values. The application can effectively eliminate the interference caused by abnormal data by using a time series analysis method to mine the footage mode and the abnormal mode from historical data and classifying and alarming the abnormal mode, so that the footage can be accurately calculated.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of positioning and ranging, and relates to a method for determining production footage of a mining working face, in particular to a method for determining production footage of a mining working face. BACKGROUND

[0002] Coal mine safety is an eternal theme in the production process of coal enterprises, and adhering to the principle of "safety first and prevention first" is the most basic requirement for the production of coal enterprises. With the increasing demand for coal in China, on the one hand, the production must be ensured to improve the economic benefits of enterprises, and on the other hand, the safety production of coal mines is under great pressure, and the safety production of coal mines is related to the stability and harmony of society.

[0003] With the progress of mining work, it is necessary to continuously complete the development roadway, preparation roadway and stoping roadway, that is, to prepare the mining amount or coal amount of development, preparation and stoping on time, so as to closely cooperate development, preparation and stoping, ensure mining balance and continuous production. For example, according to the different geological conditions, mining methods, degree of mining mechanization and management level, the various proportional relationships of the mine are determined, such as the proportional relationship between the footage and the production, the proportional relationship between the number of stoping working faces and the number of development working faces, etc. Therefore, the production footage data of mining working face is an important index reflecting the production situation of coal mine, and the accuracy of the data directly affects the production plan and production benefit of coal mine.

[0004] The existing method for determining footage of mining working face based on UWB technology positioning and ranging only considers normal conditions, while actual production is very complex. For example, equipment failure, base station change, signal reception anomaly, antenna jump, improper operation of workers and the like will interfere with data collection and produce abnormal data, which seriously affects the accuracy of production footage of mining working face. SUMMARY

[0005] In view of the above problems, the purpose of the present application is to provide a method for determining production footage of mining working face, which can eliminate the interference of abnormal data and improve the accuracy of production footage of mining working face.

[0006] The technical scheme of the present application is: a method for determining production footage of mining working face, comprising:

[0007] Collecting positioning and ranging data of the mining machine to determine the mining working face;

[0008] Modeling the mining working face to identify mining operation cycle and abnormal mode; the mining operation cycle is a development operation cycle or a stoping operation cycle;

[0009] Identifying and checking the abnormal mode by using a variable point detection and trend test algorithm to obtain an abnormal mode subsequence;

[0010] Obtain the modeling coordinate value, the coordinate value is the farthest end coordinate value of the excavation operation cycle or the nearest end coordinate value of the mining operation cycle; then correct the obtained coordinate value according to the obtained abnormal mode subsequence, and calculate the daily footage according to the corrected coordinate value.

[0011] Further, the mining working face is modeled by a LearnShapelets subsequence mining algorithm, and the steps are:

[0012] Collect the positioning and ranging data sequence of the tunneling machine of N working faces, denoted as Y i ; wherein each Y i sequence contains at least 10 complete working days of mining process;

[0013] Observe the change trend of Y i sequence by visual method, mark the start time T s and the end time T e of each mining operation cycle, and mark the start time T s , the end time T e and the type of abnormal data mode;

[0014] Statistical data points in each [T s , T e ] range, set the maximum value as L max , the minimum value as L min , the ratio R = L max / L min , when R is greater than 3, adjust the sampling interval of the original sequence where L max is located, so that R is less than 3;

[0015] Suppose there are M kinds of data modes, number them from 1, 1 is the mining operation cycle, and the following are various abnormal modes; cut the subsequence S i from N Y j sequences and mark the type y j ;

[0016] All the (S j , y j ) paired data are constructed into training set and validation set in the ratio of 8:2, when the number of data of a certain category is less than 10, it is merged into the adjacent abnormal mode, and the model automatic learning is completed;

[0017] Then, the LearnShapelets algorithm is used for model training, and the obtained model has automatically learned the optimal subsequence representation of the excavation operation cycle or the abnormal mode, and performs classification prediction.

[0018] Further, from N Yi Subsequence S in sequence j and label type y j The constraint conditions are: a) the subsequence S j The sequence length is L=5L min ; b) the subsequence S j includes an uninterrupted mining operation cycle; c) the starting time interval of each subsequence S j is at least L min ; d) each subsequence S j can only contain one abnormal mode; e) when the subsequence S j contains an abnormal mode, y i is marked as the abnormal mode number, otherwise it is marked as 1.

[0019] Further, in the mining working face, when the standard deviation of Y i,t is greater than the default standard deviation, that is, there is a jump interval; wherein the default standard deviation is 300; the mining working face is a driving working face or a stoping working face;

[0020] In the driving working face, when the driving direction trend is downward, and the jump difference value is greater than the default jump difference value, the occurrence is determined as the existence of base station change;

[0021] In the stoping working face, when the mining direction trend is upward, and the jump difference value is greater than the default jump difference value, the occurrence is determined as the existence of base station change;

[0022] Wherein, the jump difference value is obtained according to the standard jump point detection algorithm; the default jump difference value is 30.

[0023] Further, the base station change mode is recorded as an abnormal mode A1, and the specific process of identifying the base station change mode by the change point detection algorithm is:

[0024] For Y i sequence, take the sliding window size as w, for all t at time t>=w / 2, calculate the mean values of the current ranging positions in the intervals [t-w / 2, t] and [t, t+w / 2] and

[0025] Calculate

[0026] When D>D thr , mark Y i,t as an abnormal mode A1, and the abnormal interval is [t-w / 2, t+w / 2]; wherein, D thr represents the distance change value threshold allowed in adjacent time periods.

[0027] Further, the Mann-Kendall trend test algorithm is used to determine whether the time sequence has an upward or downward trend.

[0028] Further, when the base station changes the mode and the actual position of the ranging target is more than the moving position of the base station, the data trend changes, and the footage is calculated, the data mode needs to be identified, otherwise the calculation will be deviated; the abnormal data mode is A2, and the real-time process for identifying the trend anomaly is tested by using the Mann-Kendall trend test algorithm:

[0029] Y i The sequence is used to identify the base station change position by using the change point detection algorithm, if not identified, the sequence is skipped, otherwise the sequence is cut into two subsequences according to the base station change position, and the subsequence S j after the base station change is taken.

[0030] For all subsequences S j , the data length is L j . When L j <L0, the sequence is skipped.

[0031] For k∈[L0,L], the loop is performed according to the step size of 5, and the subsequence L j of L j,1...k is taken. The Mann-Kendall test algorithm is used to determine the trend value of the subsequence, which is recorded as Q j,k , and the trend sequence Q j,k , k∈[L0,L] is obtained.

[0032] The Q j sequence obtained by the above process is smoothed by using a sliding window size of 3, the trend value in the sliding window size is calculated, if the three values are the same, the trend value is set, otherwise there is no trend, and the final trend sequence is obtained.

[0033] If the subsequence S j is a tunneling working face sequence and there is a downward trend in Q , or if the subsequence S j is a back-drawing working face sequence and there is an upward trend in Q , the subsequence S j is marked as an abnormal mode A2.

[0034] Further, the tunneling operation cycle sequence is J, the back-drawing operation cycle sequence is H, and the abnormal mode sequence is U, when the daily shift start time T s0 and the end time T e0 , the tunneling daily footage or the back-drawing daily footage is calculated.

[0035] Further, the tunneling daily footage calculation steps are:

[0036] S8-1, take all the tunneling operation cycle sequence J i i∈[1,n1] and abnormal pattern sequence U i i∈[1,n2], respectively, in chronological order;

[0037] S8-2, determine whether there is an abnormal pattern A1, if not, all sequences J i are divided into the same group Gj;

[0038] S8-3, take the earliest sequence J1 and the latest sequence J n1 in the G group, respectively, take their maximum value Determine whether there is an abnormal pattern A2 in the time period where the J1 sequence is located, if there is an abnormal pattern A2, then the footage Otherwise, the footage

[0039] S8-4, if so, according to the start and end time of the A1 pattern sequence, divide J i sequence into independent groups G ji i∈[1,n3]; for each group G ji Repeat step S8-3 to get footage F ji i∈[1,n3], calculate the final tunneling daily footage

[0040] Further, the backbreaking daily footage calculation steps are:

[0041] S9-1, take all the backbreaking operation cycle sequence H i i∈[1,n1] and abnormal pattern sequence U i i∈[1,n2], respectively, in chronological order;

[0042] S9-2, determine whether there is an abnormal pattern A1, if not, all sequences H i are divided into the same group Gu;

[0043] S9-3, take the earliest sequence H1 and the latest sequence H n1 in the Gu group, respectively, take their minimum value Determine whether there is an abnormal pattern A2 in the time period where the H1 sequence is located, if there is an abnormal pattern A2, then the footage Otherwise, the footage

[0044] S9-4, if so, according to the start and end time of the A1 pattern sequence, divide H i sequence into independent groups Gui ,i∈[1,n3]; for each group G ui Repeat step S9-3 to obtain the advance F. ui ,i∈[1,n3], calculate the final daily advance of the recovery.

[0045] Compared with existing technologies, the advantages of this invention are as follows: Based on the positioning and ranging data of the tunneling machine collected by UWB technology at the mining face, this invention identifies and extracts the tunneling operation cycle, the mining operation cycle, and abnormal patterns through three algorithms: change point detection, subsequence mining, and trend verification. It calculates the daily tunneling footage and daily mining footage based on the daily tunneling and mining operation cycles, and classifies and alarms abnormal patterns. Through time series analysis, it mines footage patterns and abnormal patterns from historical data, eliminating interference from abnormal data and thus accurately calculating the footage. Attached Figure Description

[0046] Figure 1 This is a flowchart for calculating the daily advance of the mining face in this invention;

[0047] Figure 2 This is an example of the original data of the footage made on a certain day in the application of this invention;

[0048] Figure 3 This is an application example of the present invention, where the fitted data is obtained by fitting the original data.

[0049] Figure 4 The change point detection and analysis data are obtained by calculation in an application example of this invention;

[0050] Figure 5 The punctuation detection and analysis results are obtained by calculation in the application example of this invention. Detailed Implementation

[0051] The following is in conjunction with the appendix Figures 1-5 The specific embodiments of the present invention will be described in detail below.

[0052] Example

[0053] like Figure 1 A method for determining the production footage of a mining face, as shown, includes:

[0054] Collect positioning and ranging data of the tunneling machine to determine the mining face.

[0055] Model the mining face and identify the mining operation cycle and abnormal patterns; the mining operation cycle is either the tunneling operation cycle or the longwall mining operation cycle.

[0056] Anomaly patterns are identified and verified using change point detection and trend verification algorithms to obtain anomaly pattern subsequences.

[0057] The obtained coordinate values are corrected according to the obtained abnormal mode subsequences, and the daily footage is calculated according to the corrected coordinate values.

[0058] Preferably, the mining working face is modeled by a LearnShapelets subsequence mining algorithm, and the steps are as follows:

[0059] The positioning and ranging data sequences of the heading machines of N working faces are collected and denoted as Y i ; each Y i sequence contains at least 10 complete working days of mining processes.

[0060] The change trend of the Y i sequence is observed by a visual method, the start time T s and the end time T e of each mining operation cycle are marked, and the start time T s , the end time T e and the type of the abnormal data mode are also marked.

[0061] The number of data points in each [T s , T e ] range is counted, the maximum value is denoted as L max , the minimum value is denoted as L min , the ratio R = L max / L min , and when R is greater than 3, the sampling interval of the original sequence where L max is located is adjusted so that R is less than 3.

[0062] Suppose there are M kinds of data modes, which are numbered from 1, 1 is the mining operation cycle, and the subsequent are various abnormal modes; subsequences S i are intercepted from the N Y j sequences and the types y j are marked.

[0063] All the (S j , y j ) paired data are constructed into a training set and a validation set in a ratio of 8:2, when the number of data of a certain category is less than 10, it is merged into the adjacent abnormal mode, and the automatic learning of the model is completed.

[0064] Then, the LearnShapelets algorithm is used for model training, and the obtained model has automatically learned the optimal subsequence representation of the heading operation cycle or the abnormal mode and performs classification prediction.

[0065] Preferably, the subsequences S i are intercepted from the N Yj and label type y j The constraint conditions are: a) the sequence length of the subsequence S j is L=5L min ; b) the subsequence S j includes an uninterrupted mining operation cycle; c) the starting time interval of each subsequence S j is at least L min ; d) each subsequence S j can only contain one abnormal mode; e) when the subsequence S j contains an abnormal mode, y i is labeled as the abnormal mode number, otherwise it is labeled as 1.

[0066] Preferably, in the mining working face, when the standard deviation of Y i,t is greater than the default standard deviation, that is, there is a jump interval; wherein the default standard deviation is 300, and the mining working face is a driving working face or a stoping working face.

[0067] In the driving working face, when the driving direction trend is downward, and the jump difference value is greater than the default jump difference value, the occurrence is determined as the existence of base station change.

[0068] In the stoping working face, when the mining direction trend is upward, and the jump difference value is greater than the default jump difference value, the occurrence is determined as the existence of base station change.

[0069] Wherein, the jump difference value is obtained according to the standard jump point detection algorithm; the default jump difference value is 30.

[0070] Preferably, the base station change mode is recorded as an abnormal mode A1, and the specific process of identifying the base station change mode by the variable point detection algorithm is as follows:

[0071] For the Y i sequence, take the sliding window size as w, and for all t at time t>=w / 2, calculate the mean values of the current ranging positions in the intervals [t-w / 2, t] and [t, t+w / 2] and

[0072] Calculate

[0073] When D>D thr , mark Y i,t as an abnormal mode A1, and the abnormal interval is [t-w / 2, t+w / 2]; wherein, D thr represents the distance change value threshold allowed in adjacent time periods, and the value in this embodiment is 4000. It should be noted that the specific value of D thr can be adjusted in actual use according to the actual scene of judging the jump.

[0074] Preferably, the Mann-Kendall trend test algorithm is used to determine whether the time series has an upward or downward trend.

[0075] Preferably, based on the base station change pattern, when the base station moves beyond the actual position of the ranging target, causing a change in the data trend, this data pattern needs to be identified when calculating the footage; otherwise, the calculation will be biased. This abnormal data pattern is denoted as A2, and the Mann-Kendall trend detection algorithm is used to verify the real-time process of identifying trend anomalies.

[0076] For Y i The sequence uses a change-point detection algorithm to identify base station location changes. If no change is detected, the sequence is skipped; otherwise, the sequence is divided into two subsequences based on the changed base station location, and the subsequence S after the base station change is taken. j .

[0077] For all subsequences S j Let its data length be L. j When L j If <L0, skip the sequence.

[0078] For k∈[L0,L], loop with a step size of 5, taking L j subsequence L j,1...k The Mann-Kendall test algorithm is used to determine the trend value of this subsequence, denoted as Q. j,k The trend sequence Q is obtained. j,k ,k∈[L0,L].

[0079] Using a sliding window size of 3, the Q obtained from the above process j The sequence is smoothed, and the trend value within the sliding window is calculated. If all three values ​​are the same, the trend value is set; otherwise, there is no trend, resulting in the final trend sequence.

[0080] If the subsequence S j For the sequence of tunneling faces and There is a downward trend in the subsequence S, or if the subsequence S j For the sequence of longwall mining faces and If there is an upward trend in the subsequence S, then the subsequence S is labeled. j This is an abnormal mode A2.

[0081] Preferably, the tunneling operation cycle sequence is denoted as J, the mining operation cycle sequence as H, the abnormal mode sequence as U, and the shift start time for the day is T. s0 and end time T e0 Calculate the daily tunneling progress or mining progress.

[0082] Preferably, the steps for calculating daily tunneling footage are as follows:

[0083] S8-1, take all the tunneling operation cycle sequence J i , i ∈ [1, n1] and abnormal pattern sequence U i , i ∈ [1, n2], respectively, in chronological order.

[0084] S8-2, determine whether there is an abnormal pattern A1, if not, all sequences J i are divided into the same group Gj.

[0085] S8-3, take the earliest sequence J1 and the latest sequence J n1 in the group G, respectively, take their maximum value determine whether there is an abnormal pattern A2 in the time period where the sequence J1 is located, if there is an abnormal pattern A2, then the footage otherwise, footage

[0086] S8-4, if so, according to the start and end time of the A1 pattern sequence, divide J i sequence into independent groups G ji , i ∈ [1, n3]; for each group G ji repeat step S8-3 to get footage F ji , i ∈ [1, n3], calculate the final tunneling daily footage

[0087] Preferably, the step of calculating the recovery daily footage is:

[0088] S9-1, take all the recovery operation cycle sequence H i , i ∈ [1, n1] and abnormal pattern sequence U i , i ∈ [1, n2], respectively, in chronological order.

[0089] S9-2, determine whether there is an abnormal pattern A1, if not, all sequences H i are divided into the same group Gu.

[0090] S9-3, take the earliest sequence H1 and the latest sequence H n1 in the group Gu, respectively, take their minimum value determine whether there is an abnormal pattern A2 in the time period where the sequence H1 is located, if there is an abnormal pattern A2, then the footage otherwise, footage

[0091] S9-4, if so, according to the start and end time of the A1 pattern sequence, divide H i sequence into independent groups Gui i∈[1,n3]; for each group G ui Repeat step S9-3 to obtain footage F ui i∈[1,n3], calculate the final mining day footage

[0092] Application example

[0093] Using the above method for determining the production footage of a mining face, taking the daily footage calculation data of a certain mining face as an example, the original data is shown in Figure 2 , the fitting data is shown in Figure 3 , the change point detection analysis data is shown in Figure 4 , and the mark point detection analysis result is shown in Figure 5 . As shown in Figure 5 , the vertical line mark position is identified as a base station change, the first point of the subsequence is found according to the previous subsequence period, the difference between the two points is calculated, which is the distance of the base station change. The calculation shows that the change distance is 11114 cm, the maximum value after the base station change, i.e. the vertical line position right side, is 7147 cm, and the farthest mining position of the previous day is 14533 cm, so the daily footage is 7147+11114-14533=3728 cm.

[0094] The above specific embodiments further illustrate the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and does not limit the present application. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application is included in the protection scope of the present application.

Claims

1. A method for determining the production footage of a mining face, characterized in that, include: Collect positioning and ranging data of the tunneling machine to determine the mining face; The mining face is modeled to identify the mining operation cycle and abnormal patterns; the mining operation cycle is either a tunneling operation cycle or a longwall mining operation cycle. The abnormal patterns are identified and verified using change point detection and trend verification algorithms to obtain abnormal pattern subsequences; Obtain the modeled coordinate values, where the coordinate values ​​are either the farthest coordinate value of the tunneling operation cycle or the nearest coordinate value of the mining operation cycle; then, correct the obtained coordinate values ​​according to the obtained abnormal pattern subsequence, and calculate the daily progress based on the corrected coordinate values; The mining face is modeled using the LearnShapelets subsequence mining algorithm, and the steps are as follows: Collect positioning and ranging data sequences of tunneling machines at N working faces, denoted as Y. i ; where each Y i The sequence must contain at least 10 full working days of mining operations; Observe Y using visualization methods i The sequence shows the trend of change, and the start time T of each mining operation cycle is marked. s and end time T e At the same time, the start time T of the abnormal data pattern is marked. s and end time T e and type; Statistical analysis of each [T] s T e The number of data points within the range, let the maximum value be L. max The minimum value is L min The ratio R = L max / L min When R is greater than 3, adjust L. max The sampling interval of the original sequence is such that R is less than 3; Suppose there are a total of M data patterns, numbered starting from 1, where 1 represents the mining operation cycle, and the others represent various abnormal patterns; from N Y... i Extracting a subsequence S from a sequence j And mark the type y j ; All of the above (S) j y j Paired data are used to construct training and validation sets in an 8:2 ratio. When the number of data in a certain category is less than 10, it is merged into the nearest abnormal pattern to complete the automatic learning of the model. Then, the LearnShapelets algorithm is used to train the model. The resulting model has automatically learned the optimal subsequence representation of the tunneling operation cycle or abnormal pattern and performs classification and prediction.

2. The method for determining the production footage of a mining face as described in claim 1, characterized in that, From N Y i Extracting a subsequence S from a sequence j And mark the type y j The constraints are: a) Subsequence S j The sequence length is L = 5L min b) Subsequence S j Including uninterrupted mining operation cycles; c) Each subsequence S j The start time interval is at least L min d) Each subsequence S j It can only contain one anomalous pattern; e) when the subsequence S j When an abnormal pattern is included, y i Mark it as the abnormal mode number, otherwise mark it as 1.

3. The method for determining the production footage of a mining face as described in claim 1, characterized in that, At the mining face, when Y i,t When the standard deviation is greater than the default standard deviation, there is a jump interval; where the default standard deviation is 300; the mining face is a tunneling face or a longwall face; At the tunneling face, if the tunneling direction is downward and the jump difference is greater than the default jump difference, it is determined that a base station change has occurred. In the longwall mining face, when the mining direction is trending upward and the jump difference value is greater than the default jump difference value, it is determined that a base station change has occurred. The jump difference value is obtained based on the standard jump point detection algorithm; the default jump difference value is 30.

4. The method for determining the production footage of a mining face as described in claim 3, characterized in that, If the base station change mode is denoted as abnormal mode A1, then the specific process for identifying the base station change mode using the change point detection algorithm is as follows: For Y i For a sequence, taking a sliding window of size w, for all times t>=w / 2, calculate the mean of the current ranging position in the intervals [tw / 2, t] and [t, t+w / 2] respectively. and calculate When D>D thr When, mark Y i,t The abnormal mode is A1, and its abnormal interval is [tw / 2, t+w / 2]; where, D thr This represents the threshold value for the allowable distance variation within adjacent time periods.

5. The method for determining the production footage of a mining face as described in claim 4, characterized in that, The Mann-Kendall trend test algorithm is used to determine whether a time series has an upward or downward trend.

6. The method for determining the production footage of a mining face as described in claim 5, characterized in that, Based on the base station change mode, when the base station moves beyond the actual position of the ranging target, causing a change in the data trend, it is necessary to identify this data pattern when calculating the advance, otherwise the calculation will be biased. Let the abnormal data pattern be A2. The real-time process for identifying trend anomalies using the Mann-Kendall trend test algorithm is as follows: For Y i The sequence uses a change-point detection algorithm to identify base station location changes. If no change is detected, the sequence is skipped; otherwise, the sequence is divided into two subsequences based on the changed base station location, and the subsequence S after the base station change is taken. j ; For all subsequences S j Let its data length be L. j When L j If <L0, skip the sequence; For k∈[L0,L], loop with a step size of 5, taking L j subsequence L j,1...k The Mann-Kendall test algorithm is used to determine the trend value of this subsequence, denoted as Q. j,k The trend sequence Q is obtained. j,k ,k∈[L0,L]; Using a sliding window size of 3, the Q obtained from the above process j The sequence is smoothed, and the trend value within the sliding window is calculated. If all three values ​​are the same, the trend value is set; otherwise, there is no trend, resulting in the final trend sequence. If the subsequence S j For the sequence of tunneling faces and There is a downward trend in the subsequence S, or if the subsequence S j For the sequence of longwall mining faces and If there is an upward trend in the subsequence S, then the subsequence S is labeled. j This is an abnormal mode A2.

7. The method for determining the production footage of a mining face as described in claim 6, characterized in that, Let J be the tunneling operation cycle sequence, H be the longwall mining operation cycle sequence, U be the abnormal mode sequence, and T be the start time of the shift for the day. s0 and end time T e0 Calculate the daily advance of tunneling or the daily advance of mining.

8. The method for determining the production footage of a mining face as described in claim 7, characterized in that, The steps for calculating daily tunneling footage are as follows: S8-1, Take the sequence of all tunneling operation cycles for the day J i ,i∈[1,n1] and the abnormal pattern sequence U i ,i∈[1,n2], sorted in chronological order; S8-2, Determine if there is an abnormal pattern A1. If not, then process all sequences J. i They are grouped into the same group Gj; S8-3, Take the earliest sequence J1 and the latest sequence J2 within group G. n1 Take their maximum values ​​respectively Determine if there is an abnormal pattern A2 in the time period of sequence J1. If there is an abnormal pattern A2, proceed with the stepping. Otherwise advance S8-4, If applicable, then J will be determined based on the start and end times of the A1 pattern sequence. i The sequence is divided into independent groups G ji ,i∈[1,n3]; for each group G ji Repeat step S8-3 to obtain the advance F. ji ,i∈[1,n3], calculate the final daily tunneling progress.

9. The method for determining the production footage of a mining face as described in claim 7, characterized in that, The steps for calculating the daily advance in mining are as follows: S9-1, Take the sequence H of all mining operation cycles for the day. i ,i∈[1,n1] and the abnormal pattern sequence U i ,i∈[1,n2], sorted in chronological order; S9-2. Determine if there is an abnormal pattern A1. If not, then process all sequences H. i They were divided into the same group, Gu; S9-3, Take the earliest sequence H1 and the latest sequence H1 within group Gu. n1 Take their minimum values ​​respectively Determine if there is an anomalous pattern A2 in the time period of sequence H1. If there is an anomalous pattern A2, proceed with the stepping stone. Otherwise advance S9-4. If applicable, then H should be set according to the start and end times of the A1 pattern sequence. i The sequence is divided into independent groups G ui ,i∈[1,n3]; for each group G ui Repeat step S9-3 to obtain the advance F. ui ,i∈[1,n3], calculate the final daily advance of the recovery.

Images