A method for dividing hard rock TBM excavation cycles based on inflection point detection

Abstract

The present invention relates to a hard rock TBM excavation cycle division method based on inflection point detection, comprising the following steps: collecting hard rock TBM excavation parameters and corresponding time identification parameters; pre-processing the hard rock TBM excavation parameters to obtain the hard rock TBM excavation cycle; dividing the idling section from the hard rock TBM excavation cycle; after dividing the idling section, based on the remaining data in the hard rock TBM excavation cycle, finding all its inflection points and arranging all the inflection points in ascending order according to the size of the time identification parameter; based on the inflection points, sequentially detecting the time division points of the idling section and the ascending section, the time division points of the ascending section and the stable section, and the time division points of the stable section and the descending section. The present invention divides the excavation cycle into five excavation stages of the idling section, the idling section, the ascending section, the stable section and the descending section through the inflection point detection method, thereby realizing batch, rapid and accurate division of the excavation cycle.

Description

Technical Field

[0001] The present invention relates to the field of TBM (Tunnel boring machine) big data technology, and in particular to a hard rock TBM excavation cycle division method based on inflection point detection. Background Art

[0002] During the TBM construction process, a large amount of tunneling parameters are recorded and stored in real time. With the rapid development of the Internet of Things, big data, and artificial intelligence technologies, TBM tunneling parameters have become valuable data resources. TBM big data mining has become an industry hotspot and is widely used in TBM performance optimization, surrounding rock prediction, and poor geological diagnosis. The tunneling process of a hard rock TBM from startup to shutdown is called a tunneling cycle, which is the basic unit of hard rock TBM big data mining. According to the working principle of the hard rock TBM and the data characteristics of the main tunneling parameters (total thrust, cutterhead torque, and tunneling speed), a complete tunneling cycle can be divided into several tunneling stages. The precise division of the tunneling cycle is of great significance to improving the efficiency and accuracy of data mining.

[0003] At present, there are few methods for dividing the excavation cycle of hard rock TBMs, and there are two main problems. First, the excavation cycle is usually divided into four excavation stages: empty push section, ascending section, stable section, and descending section. The empty push section can be further divided into an idle section and an empty push section, which is conducive to extracting the TBM rotation resistance and propulsion resistance from the two stages respectively. Secondly, the excavation cycle is usually divided according to the duration of each stage, and the division accuracy is low, which affects the data quality and has limited promotion effect on the research of TBM intelligent excavation. Summary of the invention

[0004] A complete hard rock TBM excavation cycle can be divided into five excavation stages: idling stage, idle pushing stage, ascending stage, stable stage and descending stage. The data characteristics of different excavation stages are significantly different, so the segmentation points of each stage can be regarded as inflection points. In view of the low accuracy of previous excavation cycle division methods, the present invention proposes a hard rock TBM excavation cycle division method based on inflection point detection, based on which the batch, rapid and accurate division of excavation cycles can be achieved.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is as follows:

[0006] A hard rock TBM excavation cycle division method based on inflection point detection comprises the following steps:

[0007] S1, collect hard rock TBM excavation parameters and corresponding time mark parameters to build a hard rock TBM big database;

[0008] The hard rock TBM excavation parameters include total thrust, cutter head torque and excavation speed;

[0009] S2, pre-processing the collected hard rock TBM big data, eliminating the zero-value data recorded during the downtime and the noise data caused by sensor errors, and obtaining the hard rock TBM excavation cycle;

[0010] The starting time point of the hard rock TBM excavation cycle is recorded as point A, and the ending time point is recorded as point B. Each excavation cycle includes three excavation parameter curves, namely, the total thrust-time curve, the cutterhead torque-time curve and the excavation speed-time curve;

[0011] S3, taking point A as the starting time point of the idle running segment, finding the dividing time point C of the idle running segment and the idle pushing segment;

[0012] S4, based on the tunneling parameter curve between time points C and B, find all the inflection points on the curve, and arrange all the inflection points in ascending order according to the size of the time identification parameter;

[0013] S5, taking point C as the starting time point of the empty push segment, sequentially detecting whether the inflection point in step S4 is the dividing time point D of the empty push segment and the rising segment, after obtaining point D, eliminating the inflection points less than or equal to point D, and retaining the remaining inflection points;

[0014] S6, taking point D as the starting time point of the rising segment, sequentially checking whether the inflection point in step S5 is the dividing time point E of the rising segment and the stable segment, after obtaining point E, eliminating the inflection points that are less than or equal to point E, and retaining the remaining inflection points;

[0015] S7, taking point E as the starting time point of the stable segment, and sequentially detecting whether the inflection point in step S6 is the dividing time point F of the stable segment and the descending segment;

[0016] S8, point F is used as the starting time point of the descending segment, and point B is used as the ending time point of the descending segment.

[0017] Further, in step S2, the hard rock TBM big data preprocessing includes the following steps:

[0018] a. Construct the state discriminant function TDF to obtain the hard rock TBM excavation cycle from the hard rock TBM big data, retain the data corresponding to when TDF is equal to 1, and eliminate the data corresponding to when TDF is equal to 0:

[0019] TDF=f(T)

[0020]

[0021] Where, T represents the cutter head torque;

[0022] b, The isolation forest algorithm is used to remove noise data in the hard rock TBM excavation cycle.

[0023] Furthermore, in step S3, the method for finding the time point C for dividing the idle running section and the idle pushing section is to construct a state discriminant function FDF, and the time point when the FDF changes from 0 to 1 is taken as point C:

[0024] FDF=f(F)·f(Pr)

[0025] Where F represents the total thrust and Pr represents the excavation speed.

[0026] Furthermore, in step S4, the step of finding the inflection point of the CB segment excavation parameter curve includes the following steps:

[0027] a. Use the moving average method to smooth the CB section excavation parameter curve to eliminate local sharp points caused by small fluctuations in data, making it easier to find inflection points;

[0028] b. Based on the smoothed CB section excavation parameter curve, the Kneedle algorithm is used to find all its inflection points, and all inflection points are arranged in ascending order according to the size of the time marker parameter.

[0029] Furthermore, in step S5, the method for detecting the time point D of the empty push segment and the rising segment is to detect the time point D of the inflection point and the time point after the inflection point. 上 Data within the time period, total (n 上 +1) data points are scored for their monotonically increasing nature and growth rate, and the inflection point with the highest score is selected as point D:

[0030]

[0031]

[0032]

[0033]

[0034] In the formula, Indicates the score obtained by the inflection point P when detecting the segmentation time point D. The value of the inflection point P is the time identification parameter value corresponding to the inflection point. score1 para and score2 para They represent the monotonically increasing score and the growth rate score of the tunneling parameter para curve, score1 i,para Indicates whether the change trend of the tunneling parameter para at time i increases or decreases, para i represents the excavation parameter para at time i,l 上 Indicates the statistical duration of the rising segment. Its unit is consistent with the time mark parameter unit.上 Indicates l 上 The number of data points collected during the time period, f represents the frequency of data collection, and n represents the frequency of data collection. 上 = l 上 f, F is the total thrust, T is the cutter head torque, and Pr is the excavation speed.

[0035] Furthermore, in step S6, the method for detecting the time point E for dividing the rising segment and the stable segment is to score the slope difference between the rising segment and the stable segment at the inflection point, and select the inflection point with the highest score as point E:

[0036]

[0037] In the formula, represents the score obtained by the inflection point P when detecting the segmentation time point E, l 稳 Indicates the statistical duration of the stable segment.

[0038] Furthermore, in step S7, the method for detecting the time point F for dividing the stable segment and the descending segment is to score the slope difference between the descending segment and the stable segment at the inflection point, and select the inflection point with the highest score as point F:

[0039]

[0040] In the formula, It represents the score obtained by the inflection point P when detecting the segmentation time point F, and L represents the time identification parameter corresponding to the last data point of the hard rock TBM excavation cycle.

[0041] Furthermore, steps S1 to S8 can be quickly implemented through computer programming.

[0042] The present invention has the following beneficial effects:

[0043] The present invention integrates hard rock TBM big data preprocessing and tunneling cycle division. First, the hard rock TBM tunneling parameters and corresponding time identification parameters are collected to build a hard rock TBM big database; secondly, the state discriminant function and isolation forest algorithm are used to preprocess the data to obtain the hard rock TBM tunneling cycle; finally, the segmentation point of the tunneling cycle is regarded as the inflection point, and the tunneling cycle is divided into five tunneling stages: idling section, idle pushing section, rising section, stable section and descending section through the inflection point detection method, realizing batch, fast and accurate division of the tunneling cycle. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a flow chart of a hard rock TBM excavation cycle division method based on inflection point detection of the present invention;

[0045] Figure 2 Schematic diagram of hard rock TBM data collected in an embodiment of the present invention;

[0046] Figure 3 Schematic diagram of the hard rock TBM excavation cycle after denoising in an embodiment of the present invention;

[0047] Figure 4 Schematic diagram of a hard rock TBM excavation cycle divided into five stages in an embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to better explain the present invention, the present invention will be described in detail below through specific implementation modes in conjunction with the accompanying drawings. Obviously, the described embodiments are only partial embodiments of the present invention, rather than all embodiments.

[0049] A hard rock TBM excavation cycle division method based on inflection point detection, such as Figure 1 As shown, the following steps are included:

[0050] S1, collect hard rock TBM excavation parameters and corresponding time stamp parameters to build a hard rock TBM big database.

[0051] In this embodiment, the data acquisition frequency is 1 Hz, and the time mark parameter is in seconds, that is, the hard rock TBM excavation parameters are recorded once per second, including total thrust, cutter head torque and excavation speed. The collected data are as follows: Figure 2 As shown, Figure 2 The data in the box are zero value data.

[0052] S2, pre-processing the hard rock TBM excavation parameters, eliminating zero-value data recorded during the shutdown period and noise data caused by sensor errors, and obtaining a hard rock TBM excavation cycle.

[0053] In this embodiment, the hard rock TBM excavation parameter preprocessing includes the following steps:

[0054] a. Construct the state discriminant function TDF to remove the zero-value data recorded during the downtime, retain the data corresponding to when TDF is equal to 1, and remove the data corresponding to when TDF is equal to 0:

[0055] TDF=f(T)

[0056]

[0057] Where T represents the cutter head torque.

[0058] The time point when the cutter head torque suddenly changes from zero or a negative number to a positive number is taken as the starting point, and the time point when the cutter head torque suddenly changes from a positive number to zero or a negative number is taken as the end point. The data from a starting point to an adjacent end point belongs to a hard rock TBM excavation cycle.

[0059] b. The isolation forest algorithm is used to remove noise data in the hard rock TBM excavation cycle.

[0060] In this embodiment, the noise data caused by sensor errors often appear in the excavation speed. The number of noise points in one excavation cycle is small, accounting for less than 3% of the total data in the excavation cycle. Taking the excavation speed as an example, the denoising process of one excavation cycle is explained:

[0061] 1) Draw a scatter plot of tunneling speed-time in two-dimensional space;

[0062] 2) Build an isolation forest consisting of 100 isolated trees, each of which independently searches for noisy data;

[0063] 3) The working principle of the isolation tree is to randomly generate a hyperplane to divide the entire data space into two subspaces, repeat this process until all subspaces have only one data point, and record the number of hyperplanes required to divide each data point. The fewer the number of hyperplanes required, the more likely the corresponding data point is a noise point;

[0064] 4) Count the working results of 100 isolated trees and calculate the average number of hyperplanes required to separate each data point;

[0065] 5) Arrange the average number of hyperplanes in ascending order, regard the first 3% of data points as noise points, remove the noise point data, and retain the remaining data.

[0066] The starting time point of the denoised hard rock TBM excavation cycle is recorded as point A, the ending time point is recorded as point B, and the excavation cycle duration is recorded as L. In this embodiment, the hard rock TBM records the excavation parameters once per second, so the total thrust after denoising is recorded as F = {F 1 , F 2 , …, F L}, the cutter head torque is recorded as T = {T 1 , T 2 , …, T L}, the excavation speed is recorded as Pr = {Pr 1 , Pr 2 , …, Pr L}, the time marker parameter is recorded as t = {t 1 , t 2 ,…,t L}={1,2,…,L}.

[0067] The above data are plotted into total thrust-time curve, cutterhead torque-time curve and excavation speed-time curve to obtain the pre-processed hard rock TBM excavation cycle, as shown in Figure 3 shown.

[0068] S3, taking point A as the starting time point of the idle running segment, finding the dividing time point C of the idle running segment and the idle pushing segment.

[0069] In this embodiment, a state discriminant function FDF is constructed, and the time point when FDF changes from 0 to 1 is taken as point C:

[0070] FDF=f(F)·f(Pr)

[0071]

[0072] S4, based on the excavation parameter curve between time points C and B, find all the inflection points on the curve, and arrange all the inflection points in ascending order according to the size of the time identification parameter.

[0073] In this embodiment, the step of finding the inflection point of the CB section excavation parameter curve includes the following steps:

[0074] a. Use the moving average method to smooth the CB section excavation parameter curve:

[0075]

[0076] In the formula, A′ i represents the data point at the i-th second after smoothing, A i represents the data point at the i-th second before smoothing, and n represents the smoothing window.

[0077] In this embodiment, the smoothing window n is taken as one sixth of the duration of the excavation cycle, that is, L / 6 (rounded up), and the data before n seconds and after n seconds of the excavation cycle are not smoothed.

[0078] b. Based on the smoothed CB section excavation parameter curve, the Kneedle algorithm is used to find all its inflection points, and all inflection points are arranged in ascending order according to the size of the time marker parameter.

[0079] Take the tunneling speed-time curve as an example to explain the process of finding the inflection point:

[0080] 1) Data input: The smoothed CB section excavation speed is expressed as Pr' = {Pr' C , Pr' C+1 , ..., Pr' L}, and its corresponding time identification parameter is expressed as t = {C, C+1, ..., L};

[0081] 2) Data transformation: A series of transformations are performed on Pr' to change its initial trend into an increasing trend and the area above the curve into a concave set. The transformed tunneling speed is expressed as transform_Pr' = {transform_Pr' C ,transform_Pr′ C+1, ..., transform_Pr′ L};

[0082] The Pr' has the following situations:

[0083] ① The initial trend of Pr' itself is increasing, and the area above the curve is a concave set, so Pr' is not transformed;

[0084] ② The initial trend of Pr' itself is increasing, and the area above the curve is a convex set. The following transformation is performed: transform_Pr′ i =max(Pr′)-Pr′ L-i+C

[0085] Where max(Pr′) represents the maximum value of the smoothed CB section excavation speed;

[0086] ③ The initial trend of Pr' itself is decreasing, and the area above the curve is a concave set. The following transformation is performed: transform_Pr′ i =Pr′ L-i+C

[0087] ④ The initial trend of Pr' itself is decreasing, and the area above the curve is a convex set. The following transformation is performed: transform_Pr′ i =max(Pr′)-Pr′ i

[0088] 3) Calculate the difference curve: Subtract t from transform_Pr′ to get the difference_Pr′: difference_Pr′ i =transform_Pr′ i -i

[0089] 4) Calculate the threshold: Based on the t, calculate the absolute value of its slope as the threshold, denoted as Tmx; in this embodiment, Tmx=1;

[0090] 5) Finding local extreme values: Based on the difference_Pr′, find all local maximum values ​​and local minimum values, record the position indexes of all local extreme values, and arrange them in ascending order, represented by index={index 1 , index 2 , …, index m};

[0091] 6) Find the inflection point: If the difference between the local extreme point and the next data point is greater than the threshold Tmx, then the local extreme point is the inflection point of the curve, that is, it satisfies the following formula:

[0092] |difference_Pr′ index-difference_Pr′ index+1 |>Tmx

[0093] In this embodiment, the above formula can be simplified as follows:

[0094] |Pr′ index -Pr′ index+1 |>0

[0095] Obviously, all local extreme values ​​satisfy this formula, that is, all local extreme values ​​of difference_Pr′ are inflection points of the tunneling speed-time curve, with a total of m inflection points, denoted by P = {P 1 , P 2 , ..., P m};

[0096] The values ​​of the inflection points are recorded as the corresponding time identification parameter values, arranged in ascending order. In this embodiment, P = {index 1 , index 2 , …, index m}.

[0097] S5, taking point C as the starting time point of the empty push segment, and sequentially detecting whether the turning point in step S4 is the dividing time point D of the empty push segment and the rising segment.

[0098] In this embodiment, the statistical results show that the duration of the rising section is usually between 100 and 300 seconds, and the excavation parameters of the rising section continue to rise rapidly. Therefore, the data within 100 seconds after the inflection point P is initially regarded as the rising section, which is recorded as {F P , F P+1 , …, F P+100}, {T P , T P+1 , …, T P+100},{Pr P , Pr P+1 , …, Pr p+100}, score the monotonically increasing nature and growth rate of the three data series:

[0099]

[0100]

[0101]

[0102] score2 para =(para P+100 -para P ) / 100

[0103] In the formula, Indicates the score obtained by the inflection point P when detecting the segmentation time point D. The value of the inflection point P is the time identification parameter value corresponding to the inflection point. score1 para and score2 para They represent the monotonically increasing score and the growth rate score of the tunneling parameter para curve, score1 i，para Indicates the change trend (increase or decrease) of the tunneling parameter para at time i, para i represents the excavation parameter para at time i, l 上 Indicates the statistical duration of the rising segment. Its unit is consistent with the time mark parameter unit. 上 Indicates l 上 The number of data points collected during the time period, f represents the frequency of data collection, and n represents the frequency of data collection. 上 = l 上 f, F represents the total thrust, T represents the cutter head torque, and Pr represents the excavation speed

[0104] All the inflection points in step S4 are scored in turn, and the inflection point with the highest score is selected as point D. After point D is obtained, the inflection points that are less than or equal to point D are eliminated, and the remaining inflection points are retained.

[0105] S6, taking point D as the starting time point of the rising segment, and sequentially detecting whether the inflection point in step S5 is the dividing time point E of the rising segment and the stable segment.

[0106] In this embodiment, the statistical results show that the duration of the stable segment is usually more than 100 seconds. If the inflection point P is the dividing point between the rising segment and the stable segment, the rising segment is initially recorded as {para D ,para D+1 ,…,para P}, and the stable segment is denoted as {para P ,para P+1 ,…,para P+100 The slopes of the rising segment and the stable segment are significantly different, so the difference between the slopes of the rising segment and the stable segment is scored:

[0107]

[0108] In the formula, represents the score obtained when detecting the inflection point P at the segmentation time point E.

[0109] All the inflection points in step S5 are scored in turn, and the inflection point with the highest score is selected as point E. After point E is obtained, the inflection points that are less than or equal to point E are eliminated, and the remaining inflection points are retained.

[0110] S7, taking point E as the starting time point of the stable segment, and sequentially detecting whether the inflection point in step S6 is the dividing time point F of the stable segment and the descending segment.

[0111] In this embodiment, if the inflection point P is the dividing point between the stable segment and the descending segment, the stable segment is initially recorded as {para E ,para E+1 ,…,para P}, and denote the descending segment as {para P ,para P+1 ,…,para L The difference in slope between the stable segment and the declining segment is also very significant, so the difference in slope between the stable segment and the declining segment is scored:

[0112]

[0113] In the formula, represents the score obtained when detecting the inflection point P at the segmentation time point F.

[0114] All the inflection points in step S6 are scored in turn, and the inflection point with the highest score is selected as point F.

[0115] S8, point F is used as the starting time point of the descending segment, and point B is used as the ending time point of the descending segment.

[0116] In Windows 10 environment, with a computer configuration of Intel(R) Core(TM) i5-10500 CPU@3.10GHz, steps S1 to S8 were implemented by programming in Python. 500 hard rock TBM excavation cycles were divided in batches, quickly and accurately. The time taken to divide a excavation cycle was less than 0.3 seconds, and the overall division accuracy was 91.6%. Compared with the commonly used division method based on the duration of each stage, the division method based on inflection point detection proposed in the present invention improves the accuracy by 17.4 percentage points.

[0117] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

[0118] The specific embodiments described herein are merely examples of the spirit of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in similar ways without departing from the spirit of the present invention or exceeding the scope defined by the appended claims.

Claims

1. A method for dividing the hard rock TBM excavation cycle based on inflection point detection. It is characterized in that The following steps are involved: S1, collecting hard rock TBM excavation parameters and corresponding time stamp parameters to build a hard rock TBM big database; the hard rock TBM excavation parameters include total thrust, cutter head torque and excavation speed; S2, pre-processing the collected hard rock TBM big data, eliminating the zero-value data recorded during the downtime and the noise data caused by sensor errors, and obtaining the hard rock TBM excavation cycle; The starting time point of the hard rock TBM excavation cycle is recorded as point A, and the ending time point is recorded as point B. Each excavation cycle includes three excavation parameter curves, namely, the total thrust-time curve, the cutterhead torque-time curve and the excavation speed-time curve; S3, taking point A as the starting time point of the idling section, finding the time point C for dividing the idling section and the idle push section; the method for finding the time point C for dividing the idling section and the idle push section is to construct a state discriminant function FDF, and the time point when FDF changes from 0 to 1 is taken as point C: In the formula, F represents the total thrust, Pr Indicates the excavation speed; S4, based on the tunneling parameter curve between time points C and B, find all the inflection points on the curve, and arrange all the inflection points in ascending order according to the size of the time identification parameter; S5, taking point C as the starting time point of the empty push segment, sequentially detecting whether the inflection point in step S4 is the dividing time point D of the empty push segment and the rising segment, after obtaining point D, eliminating the inflection points less than or equal to point D, and retaining the remaining inflection points; S6, taking point D as the starting time point of the rising segment, sequentially checking whether the inflection point in step S5 is the dividing time point E of the rising segment and the stable segment, after obtaining point E, eliminating the inflection points that are less than or equal to point E, and retaining the remaining inflection points; S7, taking point E as the starting time point of the stable segment, and sequentially detecting whether the inflection point in step S6 is the dividing time point F of the stable segment and the descending segment; S8, point F is used as the starting time point of the descending segment, and point B is used as the ending time point of the descending segment.

2. The method according to claim 1, It is characterized in that In step S2, the hard rock TBM big data preprocessing step includes: a. Construct the state discriminant function TDF to obtain the hard rock TBM excavation cycle from the hard rock TBM big data, retain the data corresponding to when TDF is equal to 1, and eliminate the data corresponding to when TDF is equal to 0: In the formula, T Indicates the cutter head torque; b, The isolation forest algorithm is used to remove noise data in the hard rock TBM excavation cycle.

3. The method according to claim 1, It is characterized in that In step S4, the step of finding the inflection point of the CB segment excavation parameter curve includes the following steps: a. The three tunneling parameter curves of the CB section are smoothed using the moving average method to eliminate local sharp points caused by small fluctuations in the data, making it easier to find inflection points; b. Based on the smoothed CB section excavation parameter curve, the Kneedle algorithm is used to find all the inflection points on the three excavation parameter curves, and all the inflection points are arranged in ascending order according to the size of the time marker parameter.

4. The method according to claim 1, It is characterized in that In step S5, the method for detecting the time point D of the empty push segment and the rising segment is to detect the time point D of the inflection point and the time point after the inflection point. Data within the time period, a total of ( ) data points are scored for their monotonically increasing nature and growth rate, and the inflection point with the highest score is selected as point D: In the formula, Indicates the turning point when detecting the segmentation time point D P Ratings obtained, turning point P The value of is the time identification parameter value corresponding to the inflection point. and Represents the excavation parameters para The monotonicity score and growth rate score of the curve, express i Time excavation parameters para The trend of change increases or decreases. express i Excavation parameters at the time para , Indicates the statistical duration of the rising segment. Its unit is consistent with the time mark parameter unit. express The number of data points collected during the duration, f Indicates the frequency of data collection, , F represents the total thrust, T Indicates the cutter torque, Pr Indicates the excavation speed.

5. The method according to claim 1, It is characterized in that In step S6, the method for detecting the time point E for dividing the rising segment and the stable segment is to score the slope difference between the rising segment and the stable segment at the inflection point, and select the inflection point with the highest score as point E: In the formula, Indicates the turning point when detecting the segmentation time point E P The ratings obtained, Indicates the statistical duration of the stable segment.

6. The method according to claim 1, It is characterized in that In step S7, the method for detecting the time point F for dividing the stable segment and the descending segment is to score the slope difference between the descending segment and the stable segment at the inflection point, and select the inflection point with the highest score as point F: In the formula, Indicates the turning point when detecting the segmentation time point F P The ratings obtained, L Indicates the time stamp parameter corresponding to the last data point of the hard rock TBM excavation cycle.

7. The method according to claim 1, It is characterized in that The steps S1 to S8 can be quickly implemented through computer programming.

Images