Intelligent anchor rod system early warning method and threshold determination method

Abstract

The invention discloses an intelligent anchor rod system early warning method and a threshold determination method, and belongs to the field of anchor rod intelligent early warning, and the method comprises the steps: S1, monitoring data preprocessing, S2, data trend regression analysis, and S3, early warning method and index. The step S1 comprises the sub-steps of S11, singular point detection and correction, S12, equal interval data acquisition and S13, data noise reduction and smoothing; the step S2 comprises the substeps of S21, establishing a regression model and S22, determining a regression coefficient; the step S3 comprises the sub-steps of S31, determining a limit stress static threshold value of each monitoring point, S32, determining a stress change rate static threshold value, S33, determining a tip sudden change model and a section stability threshold value, and S34, performing a multi-factor combined early warning method. According to the method, the monitoring data is preprocessed, the data trend is subjected to regression analysis, and early warning methods and indexes are divided. Therefore, the overall trend and rule of the stress change of the surrounding rock of the roadway can be accurately analyzed, and the early warning threshold value of the anchor rod early warning system is accurately determined.

Description

Technical Field

[0001] The present invention relates to the field of intelligent early warning technology for anchor rods, and in particular to an intelligent early warning method for an anchor rod system and a threshold determination method. Background Art

[0002] The values ​​collected by the intelligent anchor system are mainly stress values ​​related to time, while the changes in stress and strain of the tunnel surrounding rock are a complex nonlinear dynamic system. At present, based on the stress value monitored by the intelligent anchor as the main indicator, the data curve obtained by analyzing the overall trend and law of stress changes in the tunnel surrounding rock has a strong personality. Affected by the prestress, the magnitude and change trend of the collected data are also different. It is mainly manifested in the following aspects: (1) The early data collection is unstable and is significantly affected by interference factors. It can only be used as a reference in the later stage and is not included in the early warning statistics. (2) Since the prestress applied at the tail end of the anchor cannot be controlled, the magnitude of each anchor monitoring is different, which makes it very difficult to study the fixed threshold. (3) In the early data, the data collection interval is inconsistent, which brings difficulties to the data homogenization processing. It should be unified in the later stage. (4) The data trends are different, and the overall change shows a nonlinear and complex change trend. Some data show obvious periodic characteristics, some data show violent fluctuations, and some data show multiple peaks and troughs. Summary of the invention

[0003] In order to solve the technical problems in the above-mentioned prior art such as unstable data collection, obvious influence of interference factors, difficulty in fixing thresholds and large data oscillation amplitude, the present invention provides an intelligent anchor system early warning method and a threshold determination method.

[0004] According to one aspect of the present invention, there is provided an intelligent anchor system early warning method, comprising the steps of:

[0005] S1: monitoring data preprocessing, said step S1 includes sub-steps, S11: detection and correction of singular points, S12:

[0006] Acquire equally spaced data and S13: data noise reduction and smoothing;

[0007] S2: data trend regression analysis, said step S2 comprises sub-steps, S21: regression model establishment and S22: determination of regression coefficients;

[0008] S3: Early warning method and indicators, the step S3 includes sub-steps, S31: a method for determining the static threshold of the ultimate stress of each monitoring point, S32: a method for determining the static threshold of the stress change rate, S33: a tip mutation model, a method for determining the section stability threshold and S34: a multi-factor joint early warning method.

[0009] Furthermore, in step S11, the collected data is set as a sequence {x i , i=1,2,3...,n} calculate the average value of two items before and after a certain distance of the collected data series

[0010]

[0011] Calculate the allowable value of deformation fluctuation range m i :

[0012]

[0013] Calculate x i The fluctuation value Δ i :

[0014]

[0015] When |Δ i |>m i When x i For outliers, the formula for adjusting them is:

[0016]

[0017] Further, in step S12, it is set as a sequence {x i ,i=1,2,3...,n} is the data collection sequence, x i-1 is the stress value of the i-1th item in the sequence, x i+1 is the i+1th stress value. The calculation formula for calculating the i-th stress value x based on interpolation is:

[0018]

[0019] Assuming the i-2th item and the i+1th deformation as the actual observed value, it is necessary to interpolate the i-1th item stress value and the i-th item deformation value, and the calculation formula is:

[0020]

[0021] Furthermore, in step S13, Fourier transform processing and noise reduction processing are performed on the data to make the data smooth and avoid the influence of complex fluctuations on trend analysis.

[0022] Further, in step S2,

[0023] S21: Regression model establishment: Perform polynomial regression simulation on the monitored time-stress curve:

[0024] yt=β0+β1t+…+β m t m +εt

[0025] S22: Determine the regression coefficient: When the regression equation is determined as a simple trend, m takes the value of 3; when the mutation model is constructed, m takes the value of 5. The least squares method is used to calculate the regression coefficients β0, β1, β2, β3, ...β m .

[0026] Further, step S31 includes,

[0027] Historical data extreme value statistics, use max and min functions to find the maximum and minimum values ​​in the monitored historical data, and use the interval between these two values ​​as the safety interval;

[0028] The newly collected data is compared with the threshold. If the newly collected data exceeds 1.2 times the extreme value, a single factor alarm is triggered. If the alarm is canceled and it is found that there is no problem, the new maximum value is used as the historical extreme value data to expand the warning range.

[0029] Furthermore, in step S32, a time interval of 30 minutes and four consecutive points lasting 2 hours are set. When the slope is continuously greater than the stress change rate threshold tan60, a single factor alarm is triggered:

[0030] Step 1: Fit the data of one day before the current time point; the prediction in a short time can be done by three times polynomial fitting to obtain the stress change rate:

[0031] Fitting curve:

[0032] P t =β0+β1t+β2t 2 +β3t 3

[0033] Stress change rate:

[0034]

[0035] Step 2: Statistic the stress change rate in the last two hours. If the stress change rate is too fast for two consecutive hours and the slope is greater than tan60 or less than -tan60, it means that the stress activity distribution inside the surrounding rock is obvious and the surrounding rock may be unstable, then a single factor alarm will be issued.

[0036] Furthermore, in step S33, the early warning threshold of the potential function is determined by using the mutation amount of the potential function mutation point in the cusp mutation model, eliminating the influence of subjective factors in the calculation process and making up for the deficiencies of step S31 and step S32:

[0037] Step 1: Construct a mathematical model of the mutation of the time-stress function, and regard the collected stress values ​​as a function of time t; perform regression analysis on the data two days before the current time point, and use a 5th-order polynomial to approximate the fitting, and obtain the following formula:

[0038] P t =β0+β1t+β2t 2 +β3t 3 +β4t 4 +β5t 5

[0039] Using multiple regression analysis to determine the coefficients, the above equation is derived:

[0040] V=β1+2β2t+3β3t 2 +3β4t 3 +3β4t 5

[0041] Let b0=β1, b1=β2, b2=β3, b3=β4, b4=β5, we get

[0042] V=b0+b1t+b2t 2 +b3t 3 +b4t 5

[0043] make t=yqSubstitute it into the above formula and eliminate the third-order term to obtain:

[0044] V=c0+c1y+c2y 2 +c4y 4

[0045] Where:

[0046]

[0047] make Substituting it into the above formula, we can get the potential function expression of the cusp mutation model of the stress-time curve:

[0048]

[0049] Derivative the potential function. When the first-order derivative of the potential function is zero, the system is in equilibrium, and at this time:

[0050] V'(Z)=Z 3 +uZ+v or V'(Z)=-Z 3 +uZ+v

[0051] The sum of the broken lines or cusps of this surface becomes an odd number of points, and the equation of the mutation point set consisting of all points with vertical tangents on the equilibrium surface is:

[0052] V”(Z)=3Z 2 +u or V”(Z)=-3Z 2 +u

[0053] Combining the above four equations and eliminating Z, we can obtain the equation of the bifurcation set formed by the projection of the singular point set on the control variable plane:

[0054] Δ=8u 3 +27v 2 =0

[0055] Step 2: Combine the latest 3-day data of each monitoring point and use the cusp mutation model to make a judgment: when Δ>0, the system is in a stable equilibrium state, the deformation of the system is continuous, and no mutation will occur; when Δ=0, and u and v are not 0 at the same time, the system is in a critical stable state, and the system state variables will mutate with just a slight disturbance; when Δ<0, the surrounding rock system is in an unstable state, and a single factor warning should be issued.

[0056] Furthermore, in step S34, in combination with steps S31, S32 and S33, and in combination with the correlation between different data collection points, a multi-parameter joint early warning is performed to predict the danger level of instability of the surrounding rock of the tunnel.

[0057] Furthermore, sensors are placed at the construction site and their data are analyzed and processed, and the method includes:

[0058] S41: Data preprocessing, comparing the original collected data with the preprocessed data, and removing data with obvious deviations;

[0059] S42: stress extreme value threshold, using the maximum and minimum values ​​of the sensor's historical data as the upper and lower limits of the threshold respectively;

[0060] S43: Stress change rate threshold, the monitoring data can be expressed by the stress-time relationship equation:

[0061] P(t)=21.525+(-0.0549)t+0.00236t 2 +(-0.000035)t 3

[0062] Then, the stress change rate is expressed as:

[0063]

[0064] The stress change rate in the last two hours was calculated and the stress change rate was small for two consecutive hours, with the slope between tan60 and -tan60, indicating that the stress activity inside the surrounding rock was not obvious, the warning value was not triggered, and the surface tunnel surrounding rock was stable;

[0065] S44: mutation model early warning, perform 5-time polynomial regression analysis on the data, determine the unknown coefficients, and obtain the fitted data relationship curve;

[0066] The fitting equation for 2 days of data is:

[0067] P(t)=21.21+0.21488t+(-0.00395)t 2 +(-0.0000857)t 3 +0.00000227t 4 +(-1.2557×10 -8 )t 5

[0068] According to the previous method, the mutation mathematical equation is:

[0069]

[0070] Tip mutation model discrimination threshold:

[0071] Δ=4u 3 +27v 2 =108.04

[0072] Judgment result: Δ>0, the system is in a stable state and the warning value is not triggered, indicating that the surrounding rock of the tunnel is stable.

[0073] Compared with the prior art, the present invention has the following beneficial effects:

[0074] The present invention provides an intelligent anchor system early warning method and a threshold determination method. The intelligent anchor system provided by the present invention collects mainly time-related stress values, and the changes in the stress and strain of the tunnel surrounding rock are a complex nonlinear dynamic system, in order to solve the cumbersome problem of the changes in the stress and strain of the tunnel surrounding rock. The present invention uses the stress value monitored by the intelligent anchor as the main indicator, pre-processes the monitoring data, performs regression analysis on the data trend, and divides the early warning methods and indicators. In this way, the overall trend and law of the stress change of the tunnel surrounding rock can be accurately analyzed, and the early warning threshold of the anchor early warning system can be accurately determined. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0076] Figure 1 Find graphs for extreme values ​​of historical data;

[0077] Figure 2 This is the installation location diagram of the 20080 sensor;

[0078] Figure 3 Data collection diagram for 20080 sensors;

[0079] Figure 4 This is the 20080 sensor data preprocessing diagram;

[0080] Figure 5 Find the graph for the 20080 sensor stress threshold;

[0081] Figure 6 It is the monitoring data and fitting curve graph;

[0082] Figure 7 This is the fitting graph for the last 2 days of data;

[0083] Figure 8 This is the fitting graph for the last 3 days of data;

[0084] Fig. 9 A flowchart of an intelligent anchor system early warning method and a threshold determination method provided by the present invention. DETAILED DESCRIPTION

[0085] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0086] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0087] Unless otherwise specifically stated, the relative arrangement of the parts and steps described in these embodiments, numerical expressions and numerical values ​​do not limit the scope of the present invention. At the same time, it should be understood that, for ease of description, the sizes of the various parts shown in the accompanying drawings are not drawn according to the actual proportional relationship. The technology, methods and equipment known to ordinary technicians in the relevant field may not be discussed in detail, but in appropriate cases, the technology, methods and equipment should be regarded as a part of the authorization specification. In all examples shown and discussed here, any specific value should be interpreted as being merely exemplary, rather than as a limitation. Therefore, other examples of exemplary embodiments may have different values. It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once a certain item is defined in an accompanying drawing, it does not need to be further discussed in subsequent drawings.

[0088] See also Figures 1 to 9 As shown, the present invention provides an intelligent anchor system early warning method and threshold determination method, including steps, S1: monitoring data preprocessing, S2: data trend regression analysis and S3: early warning method and indicators, wherein,

[0089] Step S1 includes sub-steps, S11: detection and correction of singular points, S12: acquisition of equally spaced data and S13: data denoising and smoothing;

[0090] Step S2 includes sub-steps S21: regression model establishment and S22: determining regression coefficients;

[0091] Step S3 includes sub-steps, S31: a method for determining a static threshold value of the ultimate stress of each monitoring point, S32: a method for determining a static threshold value of the stress change rate, S33: a method for determining a tip mutation model, a method for determining a section stability threshold value, and S34: a multi-factor joint early warning method.

[0092] S1: Monitoring data preprocessing

[0093] There are many problems with the data collected by the system. Before data analysis, the collected data must be preprocessed first. According to the typical characteristics of the monitored data, the following three problems are mainly preprocessed:

[0094] (1) During the monitoring process, the data may be disturbed by abnormal events or affected by errors, resulting in obvious abnormal trends and data in the collected data;

[0095] (2) The curve of each monitoring point is highly individual. Affected by prestress, the magnitude and change trend of the collected data are also different.

[0096] (3) During the monitoring process, the data frequencies collected by each system are not consistent, resulting in the monitoring data not being equally spaced;

[0097] According to the typical problems existing in the monitoring data, the original collected data must be pre-processed before it can be used for subsequent analysis. To address the existing problems, three steps of pre-processing are first implemented.

[0098] S11: Detection and correction of singular points

[0099] There are many reasons for the generation of singular data in monitoring data, which may be electromagnetic interference, human disturbance, system error and other factors. Combined with the previous typical waveform analysis, there are three main types of obvious interference data: data that suddenly jumps or sinks at a certain moment; high-frequency violent jitter data; data with particularly large or small values. These abnormal values ​​will have a great impact on the subsequent analysis process, affecting the accuracy of the prediction model itself and the final prediction results. Deleting singular data is crucial for early warning. The first step is to find such data and correct or delete it.

[0100] Singular point detection or correction can be performed by using the sliding average detection method to correct singular data points. The basic process is as follows:

[0101] Set the collected data as a sequence {x i , i=1,2,3...,n} calculate the average value of two items before and after a certain distance of the collected data series

[0102]

[0103] Calculate the allowable value of deformation fluctuation range m i :

[0104]

[0105] Calculate x i The fluctuation value Δ i :

[0106]

[0107] When |Δ i |>m i When x i It is an outlier and needs to be adjusted or deleted. The adjustment formula is:

[0108]

[0109] S12: Get equally spaced data

[0110] The data collected by the system, the deletion of singular points, the non-uniformity of the collection interval, and the monitoring data after the first step of processing are still not equally spaced data. Generally, the prediction model requires that the monitoring data is equally spaced. Non-equally spaced data will bring great inconvenience in prediction analysis. Therefore, before the prediction model is established, the original collected data needs to be homogenized in order to obtain an equally spaced data series.

[0111] Mathematical methods are used to interpolate according to the changing trend of the collected data sequence, and then the interpolation data is calculated.

[0112] Set to a sequence {x i ,i=1,2,3...,n} is the data collection sequence, x i-1 is the stress value of the i-1th item in the sequence, x i+1 is the i+1th stress value. The calculation formula for calculating the i-th stress value x based on interpolation is:

[0113]

[0114] Interpolate two observed values ​​between two measured values.

[0115] Assuming the i-2th item and the i+1th deformation as the actual observed value, it is necessary to interpolate the i-1th item stress value and the i-th item deformation value, and the calculation formula is:

[0116]

[0117] In other cases, the linear interpolation method can be used to arrange the data into evenly distributed and equally spaced data.

[0118] S13: Data noise reduction and smoothing

[0119] After the data processing in the previous two steps, the data still has some complex and changeable inflection points and burrs. Therefore, it is necessary to perform Fourier transform processing and noise reduction on the data to make the data smooth and avoid the influence of complex fluctuations on trend analysis. And the noise-reduced data can be output.

[0120] S2: Data trend regression analysis

[0121] The preprocessed data still represents the correspondence between stress and time. The monitored time-stress curve is represented by a mathematical equation and fitted using a multivariate linear regression model. Different polynomial equations are fitted according to different data lengths, time periods, and early warning indicator requirements.

[0122] The general approach to building a polynomial regression model is divided into two steps:

[0123] (1) Regression model establishment: the monitored stress value is taken as the variable y, and the time independent variable is the equally spaced acquisition cycle number t, t = 1, 2, ..., n. A polynomial regression simulation is performed on the monitored time-stress curve.

[0124] yt=β0+β1t+…+β m t m +ε t (8)

[0125] Where β0, β1, …β m is the regression coefficient; ε is the random error.

[0126] (2) Determine the regression coefficient. First, preliminarily determine the polynomial index m. According to relevant literature, it can be preliminarily determined that the value of m for the simple trend regression equation is 3. When constructing a mutation model, m is generally 5. Use the least squares method to calculate the regression coefficients β0, β1, β2, β3, ... β m .

[0127] S3: Early warning methods and indicators

[0128] After the data processing is completed, it is analyzed on the basis of statistics and regression analysis to determine the early warning methods and indicators.

[0129] The determination of traditional warning thresholds is mostly based on historical data and experience. The application time of smart anchors is relatively short, and there are no relevant scholars and experts to study their workmanship, and there is no established theory to refer to. Common warning methods can set a fixed threshold based on experience, and issue a warning when the threshold is exceeded. This method is relatively simple and easy to operate; it can also be based on data trend analysis. For relatively simple trend predictions, the trend inflection point and rate can be used for prediction and warning, but this method is only applicable to cases with simpler change rules; for highly nonlinear prediction and warning technologies, gray prediction, neural network and other methods can be used for prediction based on a large amount of data, but this method is based on a sufficiently large test database and a large amount of successful data; for another highly nonlinear prediction technology, mathematical methods, mutation theory, fractal theory, and random chaos theory can be used to find the internal connection between data factors through mathematical calculations, establish a suitable mathematical warning model, and output the finalized results by inputting variable parameters to issue warnings. This method has too high requirements for mathematics, and the construction of multi-factor mathematical models is too complex, which is not conducive to application.

[0130] The internal stress of the tunnel surrounding rock is a physical quantity in a special subjective sense. Due to the influence of various factors such as original rock stress, tunnel shape, structural influence, surrounding rock conditions, loose zone thickness, and monitoring position, it is difficult to measure the tunnel surrounding rock pressure truly and accurately. This special physical quantity must be transmitted through a medium, that is, a medium that can work with the tunnel and is easy to monitor to transmit the internal stress changes of the tunnel. The stress changes of the anchor bolts monitored by the intelligent anchor bolts have nothing to do with the monitored stress magnitude, but can reflect the changes in the surrounding rock stress.

[0131] First, set a single warning parameter to determine the safety, criticality and disaster status that may be caused by a single factor indicator, and then use the set warning method to conduct a joint warning with multiple parameters. The warning parameters for a single factor are as follows:

[0132] (1) Warning parameter 1: It is the static threshold value of the ultimate stress of each monitoring point;

[0133] (2) Warning parameter 2: static threshold of stress change rate;

[0134] (3) Warning parameter 3: tip mutation model, segment stability threshold.

[0135] Introduce the three warning parameters and specific application methods

[0136] S31: Method for determining warning parameter 1

[0137] At the beginning of anchor bolt installation, the anchor bolt anchoring force acts on the surrounding rock, which is first a balancing process. Generally, the stress monitoring data changes greatly and cannot be used for early warning. After 7 days, the anchor bolt and the surrounding rock begin to change in coordination, and the stress value begins to be collected. The data collected at the latest moment is compared with the maximum or minimum value of the historically collected data to analyze the stress change trend. According to experience, combined with the surrounding rock stress distribution and anchor bolt strength, when the stress value reaches 1.2 times the stress threshold (the previous statistical data, the historical maximum or minimum value), a single factor alarm is triggered. The basic process is 2 steps:

[0138] Step 1: If Figure 1 As shown, for the extreme value statistics of historical data, first use the max and min functions (please refer to the relevant functions when implementing the program) to perform a simple analysis of the historical data, find the maximum and minimum values ​​in the monitored historical data, and use the interval between these two values ​​as a safe interval.

[0139] Step 2: Compare the newly collected data with the threshold (historical extreme value). If the newly collected data exceeds 1.2 times the extreme value, a single factor alarm will be triggered. If the alarm is cancelled and it is found that there is no problem, the new maximum value will be used as the historical extreme value data to expand the warning range.

[0140] S32: Method for determining warning parameter 2

[0141] Trend analysis and early warning: predict tunnel instability based on the changing speed and acceleration of the measuring point data and the correlation analysis between the data of different substations.

[0142] Time-stress curve, the change of stress reflects the adjustment and distribution of stress inside the surrounding rock. When the stress change rate is large, it means that the scale of stress redistribution inside the surrounding rock is gradually increasing, and its stress change rate is accelerating. This often indicates that the stress change inside the surrounding rock is intensifying and may be dangerous. When the slope of 4 consecutive points (time interval 30min, lasting 2 hours) is continuously greater than the stress change rate threshold (value is tan60), a single factor alarm is triggered. The basic steps are also divided into 2 steps:

[0143] Step 1: At the current time point, the data of the previous day are fitted. According to relevant literature, short-term predictions can be made using three-times polynomial fitting to obtain the stress change rate.

[0144] Fitting curve:

[0145] P t =β0+β1t+β2t 2 +β3t 3 (9)

[0146] Stress change rate:

[0147]

[0148] Step 2: Statistics on stress change rate within the last two hours. If the stress change rate is too fast for two consecutive hours and the slope is greater than tan60 or less than -tan60, it means that the stress activity distribution inside the surrounding rock is obvious and the surrounding rock may be unstable, then a single factor alarm will be issued.

[0149] S33: Method for determining warning parameter 3

[0150] Cusp mutation, also known as Riena Hugonioc point mutation, is the most commonly used mutation model. Cusp mutation has the characteristics of sudden jump, hysteresis, divergence, uncontrollability, and bimodality. The cusp mutation model is suitable for situations where the data sequence changes regularly with a certain trend. The data collected by the intelligent anchor monitoring system has the characteristics described by the cusp mutation, such as hysteresis, divergence and sudden jump. Therefore, the use of cusp mutation theory in the study of intelligent anchor warning thresholds conforms to objective laws and the requirements of the system itself. The critical value of the potential function, that is, the warning threshold, is determined by the mutation amount of the potential function mutation point in the cusp mutation model, eliminating the influence of subjective factors in the calculation process and making up for the shortcomings of the previous two warning parameter methods. The basic steps are divided into 3 steps:

[0151] Step 1: Construct a mathematical model of the mutation of the time-stress function. The collected data has obvious timeliness, that is, the collected stress value can be regarded as a function of time t. Perform regression analysis on the data of the current time point and the previous two days. According to relevant literature, a 5th-order polynomial can be used for approximate fitting, and the following formula is obtained:

[0152] P t =β0+β1t+β2t 2 +β3t 3 +β4t 4 +β5t 5 (11)

[0153] Use multiple regression analysis to determine the coefficients and derive the above formula

[0154] V=β1+2β2t+3β3t 2 +3β4t 3 +3β4t 5 (12)

[0155] Let b0=β1, b1=β2, b2=β3, b3=β4, b4=β5, we get

[0156] V=b0+b1t+b2t 2 +b3t 3 +b4t 5 (13)

[0157] make t=yqSubstitute it into the above formula and eliminate the third-order term to obtain:

[0158] V=c0+c1y+c2y 2 +c4y 4 (14)

[0159] Where:

[0160]

[0161] Re-order

[0162] Substituting into the above formula, we can get the potential function expression of the cusp mutation model of the stress-time curve:

[0163]

[0164] Where:

[0165]

[0166] Derivative the potential function. When the first-order derivative of the potential function is zero, the system is in equilibrium. At this time,

[0167] V'(Z)=Z 3 +uZ+v or V'(Z)=-Z 3 +uZ+v (17)

[0168] The sum of the broken lines or cusps of this surface becomes an odd number of points, and the equation of the mutation point set consisting of all points with vertical tangents on the equilibrium surface is:

[0169] V”(Z)=3Z 2 +u or V”(Z)=-3Z 2 +u (18)

[0170] Combining the above four equations and eliminating Z, we can obtain the equation of the bifurcation set formed by the projection of the singular point set on the control variable plane:

[0171] Δ=8u 3 +27v 2 =0 (19)

[0172] Step 2: Combine the latest 3-day data of each monitoring point and use the cusp mutation model to make a judgment: when Δ>0, the system is in a stable equilibrium state, the deformation of the system is continuous, and no mutation will occur; when Δ=0, and u and v are not 0 at the same time, the system is in a critical stable state, and the system state variables will mutate with just a slight disturbance; when Δ<0, the surrounding rock system is in an unstable state, and a single factor warning should be issued.

[0173] S34: Multi-factor joint early warning method

[0174] The three single-factor early warning indicators mentioned above have their own advantages: the maximum stress value can reflect the degree of energy accumulation from the side; the stress change rate can reflect the degree of stress activity inside the surrounding rock mass; the cusp mutation model can judge the stability of the surrounding rock system over a period of time from nonlinear data. If each parameter is used for early warning separately, there are inevitable defects: the maximum stress value and stress strain rate values ​​are based on a large amount of empirical data, which are greatly affected by subjective factors, and can only judge the state of a certain point in the trend, and cannot strictly reflect the state of catastrophic evolution; and although the mutation theory can judge the state of the system based on the trend of change over a period of time, it is slightly lagging behind the first two indicators to a certain extent.

[0175] Therefore, the three separate warning parameters are combined with the correlation between different data collection points to carry out multi-parameter joint warning and predict the danger level of tunnel surrounding rock instability.

[0176] Warning level classification

[0177] Warning level Danger level Measures should be taken Level 3 warning Early warning Reporting to management personnel Second level warning Disaster Warning Oral report, hold meetings, make written reports and suggestions, analyze the reasons Level 1 warning Disaster Warning On-site investigation and research on emergency measures

[0178] Early warning method and standard setting

[0179]

[0180] In combination with the above-mentioned intelligent anchor system early warning method, a threshold determination method of the intelligent anchor system can be further obtained. The threshold determination method of the intelligent anchor system includes the following steps:

[0181] 1. Single parameter warning threshold verification;

[0182] like Figure 2 As shown, the set warning threshold is verified in combination with the specific data of the 20080 sensor; the 20080 sensor is located at the center of the arch of the monitoring section of the No. 2 auxiliary shaft.

[0183] 2. Collect data;

[0184] like Figure 3As shown, the data is read and statistics are made according to the stress-time relationship. This part of the data is collected by 20080 sensors before June 12, 2014, with an interval of 30 minutes.

[0185] 3. Data preprocessing;

[0186] like Figure 4 As shown, according to the data preprocessing method introduced above, the data are processed by singular point deletion, equally spaced data supplementation, and data noise reduction and smoothing processes; as a result of the processing, the data with obvious deviations are removed and the curve is smoothed.

[0187] 4. Determination of single parameter warning threshold

[0188] (1) Threshold 1: Stress extreme value threshold

[0189] like Figure 5 As shown, according to the previous introduction: find the maximum and minimum values ​​of the historical data of 20080 sensors, which are used as the upper and lower limits of the threshold respectively.

[0190] according to Figure 5 It can be seen that the upper and lower limits of the stress threshold of 20080 are 44KN to 16KN respectively. When the stress exceeds 1.2 times the stress amplitude, that is, when it exceeds 47KN or is lower than 13KN, a single parameter alarm is triggered. The data collected subsequently are all within the stress threshold range, indicating that the surrounding rock of the tunnel is stable and no alarm is triggered.

[0191] (2) Threshold 2: Stress change rate

[0192] Select the data collected by sensor 20080 on the last day of June 12, 2014:

[0193] 20080 sensors collected data in the last 24 hours

[0194]

[0195]

[0196] Perform a 3rd order polynomial regression analysis on the data to determine the unknown coefficients, and obtain the following: Figure 6 The relationship curve shown. Among them, the points are pre-processed data, and the curve is the fitting curve, which can be expressed by the stress-time relationship equation:

[0197] P(t)=21.525+(-0.0549)t+0.00236t 2 +(-0.000035)t 3 (20)

[0198] Then, the stress change rate is expressed as:

[0199]

[0200] The stress change rate was calculated within the last two hours. The stress change rate was small for two consecutive hours, and the slope was between tan60 and -tan60, indicating that the stress activity inside the surrounding rock was not obvious, the warning value was not triggered, and the surface tunnel surrounding rock was stable. The statistical results are shown in the table:

[0201] Statistical results of stress change rate

[0202] time Stress change rate Threshold Judgment results 2014 / 6 / 12 5:02 -0.267525496 -1.73 Safety 2014 / 6 / 12 5:32 -0.277080496 -1.73 Safety 2014 / 6 / 12 6:02 -0.286845496 -1.73 Safety 2014 / 6 / 12 6:32 -0.296820496 -1.73 Safety

[0203] (3) Threshold 3: Mutation model warning

[0204] Select the data collected by the 20080 sensor on the last two days of June 12, 2014: Perform a 5-fold polynomial regression analysis on the data to determine the unknown coefficients, and obtain the following Figure 7 , Figure 8 The relationship curve shown is a fitting curve with a time interval of 30 minutes.

[0205] The fitting degree of the data of 3 days is generally poor, the trend of change within 2 days is simpler, the fitting degree is higher, and the data of the latest two days meet the requirements of stability analysis, so the fitting equation of the data of 2 days is taken as:

[0206] P(t)=21.21+0.21488t+(-0.00395)t 2 +(-0.0000857)t 3 +0.00000227t 4 +(-1.2557×10 -8 )t 5

[0207] According to the previous method, the mutation mathematical equation is:

[0208]

[0209] Tip mutation model discrimination threshold:

[0210] Δ=4u 3 +27v 2 =108.04 (23)

[0211] Judgment result: Δ>0, the system is in a stable state and the warning value is not triggered, indicating that the surrounding rock of the tunnel is stable.

[0212] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An intelligent anchor system early warning method, characterized in that: Includes steps: S1: monitoring data preprocessing, step S1 includes sub-steps, S11: detection and correction of singular points, S12: acquisition of equally spaced data and S13: data denoising and smoothing; S2: data trend regression analysis, step S2 includes sub-steps, S21: regression model establishment and S22: determination of regression coefficients; S3: Early warning method and indicators, step S3 includes sub-steps, S31: method for determining the static threshold of the ultimate stress of each monitoring point, S32: method for determining the static threshold of the stress change rate, S33: tip mutation model, method for determining the section stability threshold and S34: multi-factor joint early warning method.

2. The intelligent anchor system early warning method according to claim 1 is characterized in that: In step S11, the collected data is set as a sequence {x i , i=1,2,3...,n} calculate the average value of two items before and after a certain distance of the collected data series Calculate the allowable value of deformation fluctuation range m i : Calculate x i The fluctuation value Δ i : When |Δ i |>m i When x i For outliers, the formula for adjusting them is:

3. The intelligent anchor system early warning method according to claim 2 is characterized in that: In step S12, a sequence {xi,i=1,2,3...,n} is set as the collected data sequence, x i-1 is the stress value of the i-1th item in the sequence, x i+1 is the i+1th stress value. The calculation formula for calculating the i-th stress value x based on interpolation is: Assuming the i-2th item and the i+1th deformation as the actual observed value, it is necessary to interpolate the i-1th item stress value and the i-th item deformation value, and the calculation formula is:

4. The intelligent anchor system early warning method according to claim 3 is characterized in that: In step S13, Fourier transform processing and noise reduction processing are performed on the data to make the data smooth and avoid the influence of complex fluctuations on trend analysis.

5. The intelligent anchor system early warning method according to claim 4 is characterized in that: In step S2, S21: Regression model establishment: Perform polynomial regression simulation on the monitored time-stress curve: yt=β0+β1t+…+β m t m +e t S22: Determine the regression coefficient: When the regression equation is determined as a simple trend, m takes the value of 3; when the mutation model is constructed, m takes the value of 5. The least squares method is used to calculate the regression coefficients β0, β1, β2, β3, ...β m .

6. The intelligent anchor system early warning method according to claim 5, characterized in that: Step S31 includes, Historical data extreme value statistics, use max and min functions to find the maximum and minimum values ​​in the monitored historical data, and use the interval between these two values ​​as the safety interval; The newly collected data is compared with the threshold. If the newly collected data exceeds 1.2 times the extreme value, a single factor alarm is triggered. If the alarm is canceled and it is found that there is no problem, the new maximum value is used as the historical extreme value data to expand the warning range.

7. The intelligent anchor system early warning method according to claim 6, characterized in that: In step S32, a time interval of 30 minutes is set for four consecutive points lasting 2 hours, so that when the slope is continuously greater than the stress change rate threshold tan60, a single factor alarm is triggered: Step 1: Fit the data of one day before the current time point; the prediction in a short time can be done by three times polynomial fitting to obtain the stress change rate: Fitting curve: P t =β0+β1t+β2t 2 +β3t 3 Stress change rate: Step 2: Statistic the stress change rate in the last two hours. If the stress change rate is too fast for two consecutive hours and the slope is greater than tan60 or less than -tan60, it means that the stress activity distribution inside the surrounding rock is obvious and the surrounding rock may be unstable, then a single factor alarm will be issued.

8. The intelligent anchor system early warning method according to claim 7, characterized in that: In step S33, the early warning threshold of the potential function is determined by using the mutation amount of the potential function mutation point in the cusp mutation model, eliminating the influence of subjective factors in the calculation process and making up for the deficiencies of step S31 and step S32: Step 1: Construct a mathematical model of the mutation of the time-stress function, and regard the collected stress values ​​as a function of time t; perform regression analysis on the data two days before the current time point, and use a 5th-order polynomial to approximate the fitting, and obtain the following formula: P t =β0+β1t+β2t 2 +β3t 3 +β4t 4 +β5t 5 Using multiple regression analysis to determine the coefficients, the above equation is derived: V=β1+2β2t+3β3t 2 +3β4t 3 +3β4t 5 Let b0=β1, b1=β2, b2=β3, b3=β4, b4=β5, we get <h2 style=";text-align:left;direction:ltr">V = b0 + b1t + b2t<h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> +b3t<h2 style=";text-align:left;direction:ltr"> 3 <h2 style=";text-align:left;direction:ltr"> +b4t<h2 style=";text-align:left;direction:ltr"> 5 make t=yqSubstitute it into the above formula and eliminate the third-order term to obtain: V=c0+c1y+c2y 2 +c4y 4 Where: make Substituting it into the above formula, we can get the potential function expression of the cusp mutation model of the stress-time curve: Derivative the potential function. When the first-order derivative of the potential function is zero, the system is in equilibrium, and at this time: V'(Z)=Z 3 +uZ+v or V'(Z)=-Z 3 +uZ+v The sum of the broken lines or cusps of this surface becomes an odd number of points, and the equation of the mutation point set consisting of all points with vertical tangents on the equilibrium surface is: <h2 style=";text-align:left;direction:ltr">V”(Z) = 3Z<h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> +uV”(Z)=-3Z<h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> +u Combining the above four equations and eliminating Z, we can obtain the equation of the bifurcation set formed by the projection of the singular point set on the control variable plane: Δ=8u 3 +27v 2 =0 Step 2: Combine the latest 3-day data of each monitoring point and use the cusp mutation model to make a judgment: when Δ>0, the system is in a stable equilibrium state, the deformation of the system is continuous, and no mutation will occur; when Δ=0, and u and v are not 0 at the same time, the system is in a critical stable state, and the system state variables will mutate with just a slight disturbance; when Δ<0, the surrounding rock system is in an unstable state, and a single factor warning should be issued.

9. The intelligent anchor system early warning method according to claim 8, characterized in that: In step S34, in combination with steps S31, S32 and S33, and in combination with the correlation between different data collection points, a multi-parameter joint early warning is performed to predict the danger level of tunnel surrounding rock instability.

10. A method for determining a threshold value of an intelligent anchor rod, comprising the intelligent anchor rod system early warning method according to any one of claims 1 to 9, characterized in that: The method of placing sensors at the construction site and analyzing and processing their data includes: S41: Data preprocessing, comparing the original collected data with the preprocessed data, and removing data with obvious deviations; S42: stress extreme value threshold, using the maximum and minimum values ​​of the sensor's historical data as the upper and lower limits of the threshold respectively; S43: Stress change rate threshold, the monitoring data can be expressed by the stress-time relationship equation: P(t)=21.525+(-0.0549)t+0.00236t 2 +(-0.000035)t 3 Then, the stress change rate is expressed as: The stress change rate in the last two hours was calculated and the stress change rate was small for two consecutive hours, with the slope between tan60 and -tan60, indicating that the stress activity inside the surrounding rock was not obvious, the warning value was not triggered, and the surface tunnel surrounding rock was stable; S44: mutation model early warning, perform 5-time polynomial regression analysis on the data, determine the unknown coefficients, and obtain the fitted data relationship curve; The fitting equation for 2 days of data is: P(t)=21.21+0.21488t+(-0.00395)t 2 +(-0.0000857)t 3 +0.00000227t 4 +(-1.2557×10 -8 )t 5 According to the previous method, the mutation mathematical equation is: Tip mutation model discrimination threshold: Δ=4u 3 +27v 2 =108.04 Judgment result: Δ>0, the system is in a stable state and the warning value is not triggered, indicating that the surrounding rock of the tunnel is stable.

Images