Intelligent detection method and system for deformation of deep foundation pit of pipe gallery

Abstract

The invention relates to the technical field of signal data processing, in particular to an intelligent detection method and system for deformation of a deep foundation pit of a pipe gallery, and the method comprises the steps: collecting a vibration signal and an IMF component signal in each short-time window; acquiring a vibration propagation sequence and a first-stage signal corresponding to each IMF; according to the similarity between the IMF component signals and the corresponding vibration signals, a first-stage correction coefficient is obtained by combining the similarity between the first-stage signals corresponding to the adjacent IMFs; obtaining a second-stage signal and a second-stage correction coefficient corresponding to each IMF according to the vibration propagation sequence and the IMF component signal; signals in the vibration propagation sequence are corrected, and a vibration source signal segment corresponding to each monitoring point is obtained according to a correction result and the similar situation of each vibration signal between IMF component signals in each short-time window; and obtaining a deformation monitoring result of the pipe gallery deep foundation pit. According to the invention, accurate separation and positioning of the independent vibration source are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of signal data processing, and particularly to an intelligent detection method and system for the deformation of deep foundation pits of utility tunnels. Background Art

[0002] With the development of urbanization, underground utility tunnels, as important infrastructure, have been widely constructed. However, due to the complex geological conditions, variable surrounding environments, and uncertainties during the construction process of deep foundation pits, deformation problems of utility tunnel deep foundation pits often occur. To ensure the stability of the foundation pit and the safety of surrounding buildings and underground facilities, it is crucial to monitor the deformation of the foundation pit in a timely manner.

[0003] The existing method is to install sensors at different monitoring points to collect vibration data of the utility tunnel deep foundation pit, and then use array sensor technology to estimate the positions of deformation sources and vibration sources. However, considering the relatively complex internal environment of the utility tunnel deep foundation pit, the situations of generating vibration sources are also complex, and echoes may be generated in the corridor after vibration occurs. This method is difficult to separate signals of different vibration sources, and the boundaries between vibration signal segments of a single vibration source are not clear, resulting in that the original vibration signals cannot be directly used to estimate the positions of vibration sources, affecting the detection results of the deformation of the utility tunnel deep foundation pit. Summary of the Invention

[0004] In order to solve the technical problem that the existing method is difficult to separate signals of different vibration sources and the original vibration signals cannot be directly used to estimate the positions of vibration sources, the purpose of the present invention is to provide an intelligent detection method and system for the deformation of utility tunnel deep foundation pits. The specific technical solutions adopted are as follows: In a first aspect, the present invention provides an intelligent detection method for the deformation of a utility tunnel deep foundation pit, including: Obtaining vibration signals at each monitoring point in the utility tunnel deep foundation pit, and performing signal feature decomposition on the vibration signals at each monitoring point to obtain IMF component signals of each vibration signal in each short-time window; Analyzing mutation signal points according to the change trends of each IMF component signal of each vibration signal in each short-time window, and obtaining the vibration propagation sequence of each vibration signal in each short-time window and the first-stage signals corresponding to each IMF; According to the similarity between each IMF component signal of each vibration signal in each short-time window and the corresponding vibration signal, and combining the similarity between the first-stage signals corresponding to adjacent IMFs in the short-time window, obtaining a first-stage correction coefficient; Obtaining second-stage signals corresponding to each IMF according to the vibration propagation sequence and the IMF component signals of each vibration signal in each short-time window; obtaining a second-stage correction coefficient according to the signal correlation relationship between the first-stage signals and the second-stage signals; The signals in the vibration propagation sequence are corrected using the first-stage correction coefficient and the second-stage correction coefficient, and according to the correction results and the similarity between each vibration signal and the IMF component signals within each short-time window, the vibration source signal segments corresponding to each monitoring point are obtained. Based on the vibration signal segments corresponding to each monitoring point and combined with the point cloud data inside the deep foundation pit of the utility tunnel, the deformation monitoring results of the deep foundation pit of the utility tunnel are obtained.

[0005] Preferably, the method for analyzing the mutation signal points according to the change trend of each IMF component signal of each vibration signal within each short-time window, obtaining the vibration propagation sequence of each vibration signal within each short-time window and the first-stage signals corresponding to each IMF specifically includes: For any short-time window of any vibration signal; The second-order difference values of each IMF component signal are respectively obtained, the moments corresponding to the second-order difference values greater than the preset threshold are recorded as mutation points, and the signals corresponding to the first mutation points in each IMF component signal are recorded as initial mutation signals; The initial mutation signals of all IMF component signals within the short-time window form the vibration propagation sequence of the vibration signal within the short-time window; the signal segments corresponding to the time periods corresponding to the initial mutation signals in the vibration propagation sequence are obtained on each IMF component signal within the short-time window as the first-stage signals corresponding to each IMF within the short-time window.

[0006] Preferably, the method for obtaining the first-stage correction coefficient according to the similarity between each IMF component signal of each vibration signal and the corresponding vibration signal within each short-time window and combined with the similarity between the first-stage signals corresponding to adjacent IMFs within the short-time window specifically includes: For any short-time window of any vibration signal, any IMF component signal is denoted as the target component signal; Based on the similarity between the first-stage signal corresponding to the target component signal and the original vibration signal, the first characteristic factor is determined; The number of repetitions between all frequency components in the target component signal and all frequency components in the adjacent next IMF component signal is obtained as the second characteristic factor; Based on the negative correlation coefficient of the ratio between the first characteristic factor and the second characteristic factor, the first-stage correction coefficient of the target component signal is determined.

[0007] Preferably, the method for obtaining the second-stage signals corresponding to each IMF according to the vibration propagation sequence and each IMF component signal of each vibration signal within each short-time window specifically includes: The time period after the last initial mutation point of all IMF component signals within the short-time window is obtained as the second stage; Take the signal corresponding to each IMF component signal in the second stage within the short-time window as the second-stage signal corresponding to each IMF within the short-time window.

[0008] Preferably, obtaining the second-stage correction coefficient according to the signal correlation relationship between the first-stage signal and the second-stage signal specifically includes: For any short-time window of any vibration signal; Obtain the first-stage signal function and the second-stage signal function corresponding to each IMF; calculate the integral value of the product of each first-stage signal function and the second-stage signal function within the short-time window at the corresponding time, and use it as the cross-correlation eigenvalue between each first-stage signal and each second-stage signal. Denote any IMF component signal as the selected component signal, and obtain the maximum value of the cross-correlation eigenvalues between the selected component signal and each first-stage signal, which is denoted as the maximum eigenvalue of the selected component signal. Obtain the time length between the first-stage signal corresponding to the maximum eigenvalue of the selected component signal and the second-stage signal corresponding to the selected component signal, and perform normalization processing on the ratio of the time length to the maximum eigenvalue to obtain the second-stage correction coefficient of the selected component signal within the short-time window.

[0009] Preferably, using the first-stage correction coefficient and the second-stage correction coefficient to correct the signals in the vibration propagation sequence specifically includes: For any short-time window of any vibration signal; Obtain the time interval length between the initial mutation point of each corresponding IMF component signal in the vibration propagation sequence and the initial mutation point of the adjacent next IMF component signal, and obtain the characteristic time length of each IMF component signal. Calculate the product of the mean value between the first-stage correction coefficient and the second-stage correction coefficient of each IMF component signal and the characteristic time length to obtain the moving time length of each IMF component signal. Translate the initial mutation point of each corresponding IMF component signal in the vibration propagation sequence to the right by the moving time length to obtain the moving signal, which is denoted as the corrected propagation signal of each IMF component signal. Take the sequence composed of the corrected propagation signals of all IMF component signals within the short-time window as the corrected propagation sequence, and the corrected propagation sequence is the correction result.

[0010] Preferably, obtaining the vibration source signal segment corresponding to each monitoring point according to the correction result and the similarity between each vibration signal and the IMF component signals within each short-time window specifically includes: Arrange the average phases corresponding to all IMF component signals within a short-time window in ascending order of the average phase to form a propagation phase sequence; Based on the time length between the vibration propagation signals of two adjacent IMF component signals within each short-time window, combined with the difference between two adjacent average phases in the propagation phase sequence and the time sequence length of the vibration propagation sequence, obtain the similarity index between each short-time window and the corrected propagation sequence; Based on the change of the similarity index between each short-time window of each vibration signal and the corrected propagation sequence, obtain the vibration source signal segment of the monitoring point corresponding to each vibration signal.

[0011] Preferably, the calculation formula for the similarity index between each short-time window and the corrected propagation sequence is: Wherein, represents the similarity index between the u-th short-time window of any vibration signal and the corrected propagation sequence, represents the time interval length between the initial mutation point of the i-th IMF component signal and the next adjacent IMF component signal in the corrected propagation sequence within the u-th short-time window, represents the total number of IMF component signals included in the short-time window, represents the time sequence length of the corrected propagation sequence, represents the absolute value of the difference between the average phases of the i-th IMF component signal and the next adjacent IMF component signal in the propagation phase sequence of the u-th short-time window of the vibration signal, that is, the phase delay amount, represents the maximum value of all phase delay amounts within the u-th short-time window of the vibration signal, represents the exponential function with the natural constant e as the base, is the normalization function.

[0012] Preferably, the step of obtaining the vibration source signal segment of the monitoring point corresponding to each vibration signal based on the change of the similarity index between each short-time window of each vibration signal and the corrected propagation sequence specifically includes: For any vibration signal, calculate the first-order difference value of the similarity index corresponding to each short-time window in the vibration signal, and use the short-time window adjacent to the previous one of the short-time windows corresponding to the minimum value of all first-order difference values as the first vibration source signal segment; According to the same method as the first vibration source signal segment, re-obtain the mutation point in the short-time window after the first vibration source signal segment, determine the second vibration source signal segment, and so on, to obtain all vibration source signal segments of the monitoring point corresponding to the vibration signal.

[0013] In a second aspect, the present invention provides an intelligent detection system for the deformation of a deep foundation pit of a utility tunnel, including a memory, a processor, and a computer program stored on the memory and running on the processor. When the computer program is executed by the processor, it implements the steps of an intelligent detection method for the deformation of a deep foundation pit of a utility tunnel.

[0014] The embodiments of the present invention at least have the following beneficial effects: The present invention first collects vibration signals at different monitoring points and performs data preprocessing to obtain signal decomposition results, providing a data basis for subsequent feature analysis. Then, in the first aspect, considering the signal propagation characteristics of an independent vibration source in the first stage, the mutation situation in the IMF component signal is analyzed to determine the initial vibration propagation sequence and the signal in the first stage of the corresponding propagation characteristics. Furthermore, the similarity between the signal in the first stage and the original signal is analyzed to quantify whether the current IMF component signal has the signal propagation characteristics of the first stage, and a first-stage correction coefficient is obtained. Secondly, in the second aspect, considering the signal propagation characteristics of an independent vibration source in the second stage, the signal in the second stage is determined, and the correlation between the signal in the first stage and the signal in the second stage is considered to quantify whether the current IMF component conforms to the signal propagation characteristics of the second stage to obtain a second-stage correction coefficient. Further, combining the correction feature analysis results of these two aspects, the initial vibration propagation sequence is corrected, and then by analyzing the similarity between the corrected vibration propagation situation and the IMF signal components within each short-time window, the vibration signal that most conforms to the propagation characteristics of the vibration source can be determined to screen out the vibration source signal segment, and finally the deformation monitoring result of the deep foundation pit of the utility tunnel is obtained. The present invention solves the problem of vibration source positioning in complex environments such as corridor foundation pits and realizes the precise separation and positioning of independent vibration sources. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings without creative efforts based on these drawings.

[0016] Figure 1 is a flowchart of the steps of an intelligent detection method for the deformation of a deep foundation pit of a utility tunnel provided by the present invention; Figure 2 is a schematic diagram of the three-dimensional point cloud data of the foundation pit of the utility tunnel without pipelines provided by the present invention; Figure 3 is a schematic diagram of the ITD decomposition result corresponding to the short-time window provided by the present invention; Figure 4 is a schematic diagram of the signal structure of the vibration propagation chain within the short-time window provided by the present invention; Figure 5 is a flowchart of the steps of the method for obtaining the first-stage correction coefficient provided by the present invention; Figure 6 is a flowchart of the steps of the method for obtaining the second-stage correction coefficient provided by the present invention; Figure 7 is a flowchart of the steps of the method for obtaining the vibration source signal segment corresponding to the monitoring point provided by the present invention. Detailed implementation manners

[0017] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following combines the accompanying drawings and preferred embodiments to detail the specific implementation manners, structures, features, and effects of an intelligent detection method and system for the deformation of a deep foundation pit of a pipe gallery according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0019] The following specifically describes the specific solutions of an intelligent detection method and system for the deformation of a deep foundation pit of a pipe gallery provided by the present invention with reference to the accompanying drawings.

[0020] Please refer to Figure 1 , which shows a flowchart of the steps of an intelligent detection method for the deformation of a deep foundation pit of a pipe gallery provided by an embodiment of the present invention. The method includes the following steps: Step S100, obtain the vibration signals at each monitoring point in the deep foundation pit of the pipe gallery, and perform signal feature decomposition on the vibration signals at each monitoring point to obtain the IMF component signals of each vibration signal in each short-time window.

[0021] In this embodiment, a plurality of monitoring points are evenly arranged along the extending length direction of the corridor at the bottom of the foundation pit, and vibration sensors are installed at the positions corresponding to the monitoring points to monitor the vibration data at the bottom of the corridor.

[0022] Further, in order to be able to detect the deformation of the foundation pit based on the vibration source signal subsequently, it is also necessary to obtain the three-dimensional point cloud data of the corridor foundation pit, as Figure 2 shown. Specifically, three-dimensional laser scanning of the foundation pit is performed. The three-dimensional laser scanning is realized by means of grid scanning, and usually the laser pulse ranging method is adopted, and then the point cloud data inside the corridor foundation pit can be obtained. This method is a well-known technology and will not be introduced in detail here.

[0023] Before the deformation occurs in the foundation pit of the corridor, there will also be vibration phenomena, and the vibration sensor can collect the signal data when these vibrations occur in real time. However, the deformation location in the foundation pit of the corridor may appear randomly, and the installation positions of the vibration sensors used to monitor the vibration conditions can only be set at a limited number of fixed positions and cannot cover all the positions where the vibration sources occur.

[0024] During the propagation process of the vibration signals generated by the vibration sources towards the vibration sensors, they may be affected by different positions, different paths, and different structural layers. These position differences make the spatio-temporal dimension of the data collected by the vibration sensors more complex, resulting in difficulties in data fusion between the vibration data and the point cloud data, that is, the positions between the vibration sources and the deformation areas cannot be corresponding, and an accurate foundation pit deformation model cannot be obtained. In order to accurately locate the deformation source, it is necessary to perform time and space calibration on the vibration data in the point cloud data.

[0025] When deformation phenomena such as cracking, collapse, uplift, and displacement occur in some structural layers in the deep foundation pit of the utility tunnel, the vibration signals when the anomalies occur can be collected by the vibration sensors, but it is difficult to determine the more accurate deformation source location from the monitoring data. Especially, the vibration signals generated in the deep foundation pit are multi-segment and superimposed, and it is difficult to independently separate the vibration signals of each independent vibration source to determine whether each independent vibration source may have deformation phenomena.

[0026] Based on this, in this embodiment, first, the vibration signals collected by the vibration sensors at each monitoring point position are subjected to characteristic signal decomposition, providing a data basis for further analyzing the signal source information corresponding to each decomposed characteristic signal. Among them, when performing characteristic decomposition on the signals, this embodiment uses Intrinsic Time Decomposition (ITD) decomposition and Fourier transform.

[0027] Specifically, each monitoring point location corresponds to a group of vibration signals, and ITD decomposition is performed on each group of vibration signals. It should be noted that the ITD algorithm has higher adaptability to irregular data. During its decomposition process, it will first separate short-time windows of different lengths, called intrinsic short-time windows, and then separately separate different component signals within each short-time window. Taking each intrinsic short-time window in the ITD decomposition result directly as a short-time window, performing short-time Fourier transform on each component signal of the vibration signal can obtain the amplitude spectrum and phase spectrum of each IMF component signal of each vibration signal within each short-time window.

[0028] It can be understood that each group of vibration signals corresponds to multiple short-time windows, and each short-time window contains a fixed number of IMF component signals. Each IMF component signal represents a type of signal component in the vibration signal. Such as Figure 3As shown, the ITD decomposition result corresponding to a short-time window in this embodiment contains a total of 5 IMF component signals. Figure 3 In the order from top to bottom, they are IMF1, IMF2, IMF3, IMF4, and IMF5.

[0029] The processing of short time series is to judge the starting to ending positions of a single vibration signal source through local signal analysis. However, since the energy of the vibration signal decays during propagation, the amplitude of the vibration signal shows a decreasing trend, and the attenuation speeds of different component signals of the same vibration source signal are also different. Therefore, the ITD decomposition algorithm can only separate the aliased components of the signal, and further feature analysis needs to be carried out on the decomposition result.

[0030] Step S200: Analyze the mutation signal points according to the change trend of each IMF component signal of each vibration signal in each short-time window, and obtain the vibration propagation sequence of each vibration signal in each short-time window and the first-stage signal corresponding to each IMF.

[0031] In the corridor foundation pit, a large change in stress occurs instantaneously, resulting in the deformation of the foundation pit, which is generally manifested as the mutability of the vibration signal on the vibration signal. Since ITD decomposition can retain components of different time scales, each IMF component signal can basically show the dynamic change characteristics of the signal from gentle to abrupt. Based on this, feature analysis can be carried out respectively on the signal change trend of each IMF component signal to screen out the data points with mutability in each IMF component signal, so as to initially extract the time series signal characteristics of vibration propagation.

[0032] Specifically, in this embodiment, taking the 5 IMF component signals in any short-time window of any vibration signal as an example, the second-order difference values of each IMF component signal are obtained respectively, and the moment corresponding to the second-order difference value greater than the preset threshold is recorded as the mutation point, and the signal corresponding to the first mutation point in each IMF component signal is recorded as the initial mutation signal.

[0033] Among them, for each IMF component signal, except that the second-order difference values cannot be calculated at the first moment and the second moment, the second-order difference values of each moment can be calculated and obtained at other moments. For any IMF component signal, the mean value of all second-order difference values is used as the preset threshold corresponding to this IMF component signal. That is, under this IMF component signal, when the second-order difference value is greater than the mean value of the corresponding second-order difference values, the moment corresponding to the second-order difference value is recorded as the mutation point under this IMF component signal, and the signal corresponding to the mutation point is the initial mutation signal. It can be understood that the method of second-order difference is very sensitive to the relationship between gentle signals and abrupt signals, and is convenient for detecting the moment when the signal undergoes a mutation phenomenon.

[0034] Further, the initial mutation signals of all IMF component signals within a short-time window form a vibration propagation sequence of the vibration signal within the short-time window; on each IMF component signal within the short-time window, a signal segment corresponding to the time period corresponding to the initial mutation signal in the vibration propagation sequence is obtained as the first-stage signal corresponding to each IMF within the short-time window.

[0035] It can be understood that the initial mutation signals in the vibration propagation sequence within the short-time window have a certain time sequence. As Figure 4 shown, after re-ordering the IMF component signals according to the time sequence of the initial mutation signals and connecting them with straight lines, the performance of the vibration propagation chain can be more intuitively displayed. According to the order of implementation, the IMF component signals corresponding to the vibration propagation chain are IMF3, IMF4, IMF5, IMF2, and IMF1 in sequence.

[0036] The vibration propagation chain characterizes the estimated result of the sequence in which all component signals in the vibration signal generated by a certain vibration source are sequentially transmitted to each monitoring point position based on the signal mutation in each IMF component signal. That is, each initial mutation signal in the vibration propagation chain corresponding to the vibration propagation sequence characterizes the position performance in each IMF component signal within the short-time window corresponding to a certain vibration source at the current monitoring point.

[0037] Based on this, the time period from the moment of the first initial mutation point to the moment of the last mutation point in the vibration propagation sequence can be understood as the process of the vibration signal propagating to the current monitoring point. Furthermore, the signals of each IMF component signal within the short-time window during this time period form the first-stage signal corresponding to each IMF component signal.

[0038] Step S300, according to the similarity between each IMF component signal of each vibration signal within each short-time window and the corresponding vibration signal, and combining the similarity between the first-stage signals corresponding to adjacent IMFs within the short-time window, a first-stage correction coefficient is obtained.

[0039] Considering that in a closed space, there is an echo situation in the vibration conduction, so the vibration signal collected by the vibration sensor may exhibit modal aliasing phenomenon and a certain degree of periodicity. By analyzing the characteristics of the vibration signal in these two aspects, the vibration propagation chain of the signal source can be corrected to a certain extent.

[0040] If there is an independent vibration source, the vibration signal generated by the vibration source can be divided into two propagation stages. If the signal characteristics in the signal segment corresponding to the vibration propagation chain include not only the characteristics of the two propagation stages, then it is necessary to correct the vibration propagation chain according to the characteristic distribution of the vibration signal. First, analyze the signal characteristics of the first stage exhibited by the independent vibration source to quantify the degree of correction required for the vibration propagation chain in the first stage.

[0041] It should be noted that the first stage specifically refers to the signal segment of all different signal components between the first initial mutation point (the initial mutation point with the earliest time) and the last initial mutation point (the initial mutation point with the latest time) in the vibration propagation chain. The propagation speeds and propagation paths (due to reflection and refraction) of different signal components may be different. Therefore, there is no aliasing relationship between all component signals in the first-stage signal.

[0042] In this embodiment, still taking the 5 IMF component signals of any vibration signal in any short-time window as an example for illustration, any one IMF component signal is denoted as the target component signal. In this embodiment, the nth IMF component signal can be used as the target component signal. As Figure 5 shown, the method for obtaining the first-stage correction coefficient can be implemented by steps S301 to S303.

[0043] Step S301: Determine the first characteristic factor based on the similarity between the first-stage signal corresponding to the target component signal and the original vibration signal.

[0044] If the characteristics of each component signal can be observed in the original signal waveform, it means that different component signals do not overlap with each other and the characteristics of the component signals are not covered. Therefore, the similarity between each IMF component signal and the original vibration signal can be analyzed to quantify the component signal characteristics of each IMF component signal.

[0045] Specifically, in this embodiment, the Pearson correlation coefficient is used to calculate the similarity between two data. More specifically, this embodiment takes the target component signal of any vibration signal in any short-time window as an example for illustration. In this vibration signal, the signal corresponding to the first-stage signal of the target component signal is intercepted in the corresponding time period and denoted as the original signal segment; the Pearson correlation coefficient between the first-stage signal of the target component signal and the original signal segment is calculated to obtain the first characteristic factor corresponding to the target component signal. The larger this value is, the clearer the characteristics of the target component signal are.

[0046] Step S302: Obtain the number of repetitions between all frequency components in the target component signal and all frequency components in the next adjacent IMF component signal as the second characteristic factor.

[0047] In the ITD decomposition result, the fewer the repeated signal features between the component signals of two adjacent components, the more independent the feature information of the component signals of each component is, and the more difficult it is to exhibit signal aliasing. Based on this, obtain the number of repetitions between all frequency components in the target component signal and all frequency components in the next adjacent IMF component signal.

[0048] Specifically, the target component signal is the nth IMF component signal within any short-time window of any vibration signal. Then, obtain the number of the same frequency values between all different frequency values in the nth IMF component signal and all different frequency values in the (n + 1)th IMF signal component, which is the second characteristic factor corresponding to the target component signal, that is, the nth IMF signal component. The more the number of the same frequency component values, the greater the possibility of aliasing between the component signals of two adjacent different components.

[0049] It should be noted that for the last IMF component signal, the next adjacent IMF component signal cannot be obtained, and the above calculation is not performed in this embodiment. In other embodiments, the implementer can set according to the specific implementation scenario.

[0050] Step S303, determine the first-stage correction coefficient of the target component signal based on the negative correlation coefficient between the ratio of the first characteristic factor and the second characteristic factor.

[0051] The first characteristic factor reflects the similarity between the target component signal and the original vibration signal. The larger this value, the clearer the signal feature information of the corresponding target component signal. The second characteristic factor reflects the signal repetition degree between the target component signal and the adjacent component signal. The larger this value, the greater the overlap between the target component signal and other signals, and the greater the possibility of signal aliasing.

[0052] Combining the characteristic performances of the two aspects, perform negative correlation normalization processing on the ratio of the first characteristic factor and the second characteristic factor of the target component signal to obtain the first-stage correction coefficient of the target component signal. Its calculation formula can be specifically expressed as , represents the first-stage correction coefficient of the target component signal, n represents the nth IMF component signal, represents the first characteristic factor, represents the second characteristic factor, is the normalization function.

[0053] For the ratio between the two Among them, the larger the numerator and the smaller the denominator indicate that the nth IMF component signal has clear signal component characteristics in the original vibration signal, and the number of repeated frequency types between the nth IMF component signal and the adjacent component signals is small. Its component signal components are more independent and non-overlapping. At this time, it more conforms to the propagation characteristics of the first stage of an independent vibration source. The corresponding component signal components require less correction, that is, the value of the first stage correction coefficient is smaller at this time.

[0054] Step S400, obtain the second stage signal corresponding to each IMF according to the vibration propagation sequence and the IMF component signals of each vibration signal in each short-time window; obtain the second stage correction coefficient according to the signal correlation relationship between the first stage signal and the second stage signal.

[0055] Considering that the vibration signals generated by an independent vibration source are specifically divided into two propagation stages. If the signal characteristics in the signal segment corresponding to the vibration propagation chain not only include the characteristics of the two propagation stages, then it is necessary to correct the vibration propagation chain according to the characteristic distribution of the vibration signals. Second, through two steps, analyze the signal characteristics of the second stage shown by an independent vibration source to quantify the degree of correction required for the vibration propagation chain in the second stage.

[0056] Among them, it should be noted that the second stage specifically refers to all the signals after the last initial mutation point (the initial mutation point with the latest time) in the vibration propagation chain, indicating the echo phenomenon that may occur after all the single vibration sources are propagated to the monitoring point. That is, there may be echoes of the signal components in the first stage in all the vibration signals after the last initial mutation point.

[0057] The first step is to obtain the signal manifestation form of the second stage.

[0058] Specifically, the time period after the last initial mutation point of all the IMF component signals in the short-time window is recorded as the second stage; the signals corresponding to each IMF component signal in the second stage in the short-time window are used as the second stage signals corresponding to each IMF in the short-time window. Among them, it can be understood that the last initial mutation point refers to the last signal point in time sequence, that is, the initial mutation point with the latest time.

[0059] The second step is to perform feature analysis on the first stage and the second stage to quantify the degree of correction required for each component under the feature manifestation of the second stage. In this embodiment, take any short-time window of any vibration signal as an example for illustration. As Figure 6 shown, the method for obtaining the second stage correction coefficient can be realized by steps S401 to S403.

[0060] Step S401: Obtain the cross - correlation eigenvalue between each first - stage signal and the second - stage signal within a short - time window.

[0061] Specifically, first obtain the first - stage signal function and the second - stage signal function corresponding to each IMF; calculate the integral value of the product of each first - stage signal function and the second - stage signal function within the short - time window at the corresponding time, and use it as the cross - correlation eigenvalue between each first - stage signal and each second - stage signal.

[0062] As a specific example, the calculation formula of the cross - correlation eigenvalue can be expressed as: Where, represents the cross - correlation eigenvalue between any first - stage signal and any second - stage signal, t represents the time sequence, represents any first - stage signal function, represents any second - stage signal function.

[0063] It should be noted that the method for obtaining the signal function can be obtained by fitting. When calculating the integral value, the signal values corresponding to the stage signals can be actually obtained for calculation. These processes are all well - known technologies and will not be introduced in detail here.

[0064] Step S402: Denote any IMF component signal as the selected component signal, and obtain the maximum value of the cross - correlation eigenvalues between the selected component signal and each first - stage signal, which is denoted as the maximum eigenvalue of the selected component signal.

[0065] When there is a signal segment in the second - stage signal whose product with the integral of the first - stage signal segment is the largest, it represents that and have the maximum cross - correlation.

[0066] In this embodiment, the m - th IMF component signal within the short - time window can be used as the selected component signal. A cross - correlation eigenvalue can be calculated between the selected component signal and the first - stage signal of each IMF component signal within the current short - time window, and the maximum value among them is the maximum eigenvalue of the selected component signal, which represents the degree of characteristic manifestation of the maximum cross - correlation relationship with the selected component signal.

[0067] Step S403: According to the first - stage signal corresponding to the maximum eigenvalue of the selected component signal and the second - stage signal corresponding to the selected component signal, combined with the maximum eigenvalue, obtain the second - stage correction coefficient of the selected component signal within the short - time window.

[0068] Specifically, obtain the time length between the first-stage signal corresponding to the maximum eigenvalue of the selected component signal and the second-stage signal corresponding to the selected component signal, and normalize the ratio of the time length to the maximum eigenvalue to obtain the second-stage correction coefficient of the selected component signal within the short-time window.

[0069] As a specific example, the calculation formula for the second-stage correction coefficient of the selected component signal can be expressed as: , where represents the second-stage correction coefficient of the selected component signal, that is, the second-stage correction coefficient of the m-th IMF component signal within the short-time window, and m represents the m-th IMF component signal. represents the time length between the first-stage signal corresponding to the maximum eigenvalue of the selected component signal and the second-stage signal corresponding to the selected component signal; represents the maximum eigenvalue of the selected component signal, and t represents the time sequence. is the normalization function.

[0070] The time length refers to the time length between the first moment in the first-stage signal corresponding to the maximum eigenvalue of the selected component signal and the first moment in the second-stage signal corresponding to the selected component signal, which reflects the time difference between the first-stage signals with the maximum cross-correlation relationship with the selected component signal, and also represents the delay time between the original vibration signal and the possible echo signal.

[0071] In the ratio , the larger the numerator and the smaller the denominator, that is, the larger the value of the time length , and the smaller the maximum eigenvalue, it indicates that the echo delay time is longer, the credibility of the selected component signal having echo characteristics is lower, and the corresponding component composition does not conform to the signal propagation characteristics of the second stage. Therefore, the greater the need for correction, and at this time, the value of the second-stage correction coefficient is larger, that is, the greater the degree of correction required.

[0072] Step S500, use the first-stage correction coefficient and the second-stage correction coefficient to correct the signals in the vibration propagation sequence, and obtain the vibration source signal segment corresponding to each monitoring point according to the correction result and the similarity between each vibration signal and the IMF component signals within each short-time window.

[0073] The first-stage correction coefficient reflects the degree to which the corresponding IMF component signal needs to be corrected in terms of the propagation characteristic performance of each IMF component signal in the first stage. The second-stage correction coefficient reflects the degree to which the corresponding IMF component signal needs to be corrected in terms of the propagation characteristic performance of each IMF component signal in the second stage. Based on this, first, combine the analysis results of the characteristics that need to be corrected in the two aspects of the component signal to correct the initially obtained vibration propagation chain to obtain a more accurate representation of the vibration propagation signal of the vibration source. Then, analyze the characteristics of the vibration source for the corrected vibration propagation situation to accurately determine the signal information of a single vibration source.

[0074] First step, use the first-stage correction coefficient and the second-stage correction coefficient to correct the signals in the vibration propagation sequence.

[0075] Specifically, the larger the value of the first-stage correction coefficient of each IMF component signal within the short-time window, the more chaotic the characteristic information of the corresponding IMF component signal, and the greater the possibility of aliasing. The larger the value of the second-stage correction coefficient of each IMF component signal within the short-time window, it indicates that there is no echo phenomenon in the subsequent corresponding IMF component signal, that is, the corresponding IMF component signal does not have the characteristics of the vibration wave signal propagating in the corridor foundation pit, and may not even be a vibration component.

[0076] It should be noted that the direction of correcting the signal in this embodiment is to further reduce the delay amount between adjacent signals. For example, within the short-time window, move the first initial mutation point in the vibration propagation chain closer to the second initial mutation point to reduce the time interval between the two initial mutation points.

[0077] More specifically, taking any short-time window in any vibration signal as an example, in the vibration propagation chain of the short-time window, as Figure 4 shown, obtain the time interval length between the initial mutation point of each IMF component signal and the initial mutation point of the adjacent next IMF component signal to obtain the characteristic time length of each IMF component signal. Calculate the product of the mean value between the first-stage correction coefficient and the second-stage correction coefficient of each IMF component signal and the characteristic time length to obtain the moving time length of each IMF component signal.

[0078] As a specific example, the calculation formula for the moving time length of the i-th IMF component signal in the vibration propagation sequence within the short-time window can be expressed as: , where represents the moving time length of the i-th IMF component signal in the vibration propagation sequence within the short-time window, represents the characteristic time length of the i-th IMF component signal in the vibration propagation sequence within the short-time window, represents the first-stage correction coefficient of the i-th IMF component signal in the vibration propagation sequence within the short-time window, represents the second-stage correction coefficient of the i-th IMF component signal in the vibration propagation sequence within the short-time window.

[0079] Then, translate the initial mutation point of each corresponding IMF component signal in the vibration propagation sequence to the right by the moving time length to obtain the moving signal, which is denoted as the corrected propagation signal of each IMF component signal. The sequence formed by the corrected propagation signals of all IMF component signals within the short-time window is used as the corrected propagation sequence, and the corrected propagation sequence is the correction result.

[0080] It can be understood that the corrected propagation sequence is also arranged in the chronological order of time like the initial vibration propagation sequence. This correction process is to move each initial mutation point in the vibration propagation chain in the direction of increasing time sequence by the corresponding moving time length to complete the correction process of the vibration propagation chain. The corrected vibration propagation chain can more accurately express the propagation characteristics of the first vibration signal component (the first vibration source).

[0081] It should be noted that for the same reason as in step 302, for the last IMF component signal in the vibration propagation chain, the adjacent next IMF component signal cannot be obtained, and the above calculation is not performed in this embodiment. In other embodiments, the implementer can set according to the specific implementation scenario.

[0082] Second step, according to the correction result and the similarity between each vibration signal and the IMF component signals within each short-time window, obtain the vibration source signal segment corresponding to each monitoring point.

[0083] The corrected vibration sequence characterizes the relative relationship of vibration propagation between the component signals of the vibration signal. Furthermore, by analyzing the signal characteristics within each short-time window to extract the information representation of the propagation characteristics, analyzing the similarity between the propagation characteristics of each short-time window and the corrected vibration sequence, and then screening out the short-time window where the vibration signal that best conforms to the vibration propagation characteristics in the current vibration signal is located, and then determining the complete vibration source signal by analyzing the signal segment range of the continuously propagated vibration source. Based on this, as Figure 7 shown, the method for obtaining the vibration source signal segment corresponding to the monitoring point can be implemented by steps S501 to S503.

[0084] Step S501, arrange the average phases corresponding to all IMF component signals within the short-time window in ascending order of the average phase to form a propagation phase sequence.

[0085] First, analyze the signal characteristics within each short-time window of the vibration signal, extract the propagation characteristic information corresponding to the vibration signal, and an absolute value of the difference between each average phase and the next adjacent average phase in each propagation phase sequence can be used to calculate a phase delay amount. .

[0086] Step S502: Based on the time length between the vibration propagation signals of two adjacent IMF component signals within each short-time window, combined with the difference between two adjacent average phases in the propagation phase sequence and the time sequence length of the vibration propagation sequence, obtain the similarity index between each short-time window and the corrected propagation sequence.

[0087] Since the initial mutation point represents the initial position when the vibration undergoes a mutation, and the vibration propagation chain composed of the initial mutation points represents the relative relationship of vibration propagation between each component signal of the vibration signal, therefore, if the vibration signals in subsequent short-time windows are generated by continuous independent vibration sources, the relative relationship of propagation between each component signal within these short-time windows is basically unchanged. Unless after the vibration source is interrupted and the vibration signal of a new vibration source is generated, this propagation relationship will change. Therefore, the short-time window with a greater similarity is more likely to be within the signal segment range of continuous propagation of this vibration source. Based on this, it is necessary to quantify the similarity index between the vibration signal and the corrected propagation sequence under each short-time window.

[0088] As a specific example, compare the signal propagation information within the short-time window with the corrected vibration propagation chain to obtain all short-time windows from the propagation to dissipation of this vibration source. In this embodiment, taking the u-th short-time window of any vibration signal as an example for illustration, the calculation formula for the similarity index between the u-th short-time window and the corrected propagation sequence can be expressed as: where, represents the similarity index between the u-th short-time window of any vibration signal and the corrected propagation sequence, represents the time interval length between the initial mutation point of the i-th IMF component signal and the initial mutation point of the next adjacent IMF component signal in the corrected propagation sequence within the u-th short-time window, represents the total number of IMF component signals included within the short-time window, represents the time sequence length of the corrected propagation sequence, represents the absolute value of the difference between the average phases of the i-th IMF component signal and the next adjacent IMF component signal in the propagation phase sequence of the u-th short-time window of the vibration signal, that is, the phase delay amount, represents the maximum value of all phase delay amounts within the u-th short-time window of the vibration signal, represents the exponential function with the natural constant e as the base, is a normalization function.

[0089] reflects the rate of change of phase delay at the i-th IMF component signal in the vibration propagation information within the u-th short-time window, reflects the slope value corresponding to the initial mutation point of the i-th IMF component signal in the corrected vibration propagation chain. Further, represents the error situation between the slope of each initial mutation point in the corrected vibration propagation chain and the rate of change of phase delay of all component signals within the current short-time window. The smaller the difference between the two, the greater the corresponding similarity degree, that is, the larger the value of the similarity index.

[0090] If the time distance between the initial mutation points in the corrected vibration propagation chain is large, it may not be different vibration signal components generated by the same vibration source. Therefore, it is necessary to assign a calculation doubt factor to the large time delay amount, that is, the time delay amount is larger, the corresponding value of the doubt factor is smaller, and it is used as a weight to weight the aforementioned difference.

[0091] Step S503, obtain the vibration source signal segment corresponding to each monitoring point of each vibration signal according to the change situation of the similarity index between each short-time window of each vibration signal and the corrected propagation sequence.

[0092] For any vibration signal, the larger the value of the similarity index between each short-time window and the corrected propagation sequence, the more the propagation characteristics of the vibration signal within the short-time window conform to the propagation characteristics of the corrected vibration propagation chain, that is, the greater the possibility that the short-time window is within the signal segment range of the continuous propagation of the vibration source.

[0093] Based on this, determine the vibration source in each vibration signal through the change trend of the similarity index of each short-time window. Specifically, for any vibration signal, calculate the first-order difference value of the similarity index corresponding to each short-time window in the vibration signal, and take the short-time window adjacent to the previous short-time window corresponding to the minimum value of all first-order difference values as the first vibration source signal segment.

[0094] It should be noted that since the signal propagation characteristics may be about to change in the time window with the smallest first-order difference value of this similarity degree. The previous time window is considered to be the last window before the significant change in signal propagation characteristics, so it is used as the cut-off position of the first vibration signal source, so that the complete signal segment of the signal source in the relatively stable propagation stage can be obtained, so as to better analyze and extract the characteristics of the signal source itself.

[0095] It can be understood that the first vibration source signal is the result of feature analysis for the first set of mutation points extracted from the original set of vibration signals. That is, one set of mutation points corresponds to a possible vibration source. Therefore, after obtaining the first vibration source signal segment, the first vibration source signal segment is removed from the original vibration signals, that is, starting from the next moment adjacent to the last signal of this vibration source signal segment, the feature analysis process of the new vibration source signal segment is carried out. According to the same method as the first vibration source signal segment, that is, the foregoing steps of this embodiment, and so on, until all independent vibration sources are extracted, and finally all vibration source signal segments corresponding to the monitoring points of the vibration signals can be obtained.

[0096] The vibration signal data collected by the vibration sensors at each monitoring point position may be separated into different independent vibration source signal segments.

[0097] Step S600, according to the vibration signal segments corresponding to each monitoring point, combined with the point cloud data inside the deep foundation pit of the utility tunnel, obtain the deformation monitoring results of the deep foundation pit of the utility tunnel.

[0098] Specifically, for each vibration source signal segment at every four adjacent monitoring points, the time difference of the vibration signal arriving at different vibration sensors can be calculated by the cross-correlation method, and then the beamforming method is used to deduce the incident direction of the vibration source relative to the array. According to the direction estimation result and the known layout of the sensor array, the triangulation method is used to calculate the position of the vibration source. Combining the geometric positioning method, the estimated values in different directions are intersected to determine the three-dimensional coordinate position of each independent vibration source. Among them, the cross-correlation algorithm, the beamforming method, and the geometric positioning method are commonly used in the fields of signal processing and vibration detection, and are all well-known technologies, so no more details will be given.

[0099] It should be noted that the method for signal source localization for the vibration source signals already extracted by the vibration sensors corresponding to each monitoring point is a well-known technology to those skilled in the art, and only a simple introduction is made here, and no more details will be given. Further, after determining the position where the signal source is located, relevant staff can detect whether there is a deformation phenomenon at the position where the vibration signal source is located. In other embodiments, the implementer can select a suitable method for processing according to the specific implementation scenario.

[0100] In summary, in this embodiment, through ITD decomposition and short-time Fourier transform, the initial mutation points are obtained according to the mutation of the vibration source, the vibration propagation chain is obtained and corrected according to the propagation characteristics of the vibration signal, and then the propagation phase sequence within each short-time window is established. According to the similarity between the propagation phase sequence and the corrected propagation chain, the signal cut-off points of each independent vibration source are obtained, and finally each independent vibration source signal is independently separated, which can greatly improve the accuracy. The problem of vibration source positioning in complex environments such as corridor foundation pits is solved, and the precise separation and positioning of independent vibration sources are realized.

[0101] The present invention also provides an intelligent detection system for the deformation of the deep foundation pit of a pipe gallery, including a memory, a processor, and a computer program stored on the memory and running on the processor. When the computer program is executed by the processor, the steps of an intelligent detection method for the deformation of the deep foundation pit of a pipe gallery are implemented. Since the embodiments of an intelligent detection method for the deformation of the deep foundation pit of a pipe gallery have been elaborated in detail, no further introduction will be made here.

[0102] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. An intelligent detection method for deformation of deep foundation pit of pipe gallery, characterized in that: The method comprises the following steps: Obtain the vibration signal at each monitoring point in the deep foundation pit of the pipe gallery, perform signal feature decomposition on the vibration signal at each monitoring point to obtain the IMF component signal of each vibration signal in each short-time window; According to the change trend of each IMF component signal of each vibration signal in each short-time window, the mutation signal point is analyzed to obtain the vibration propagation sequence of each vibration signal in each short-time window and the first-stage signal corresponding to each IMF; According to the similarity between the IMF component signal and the corresponding vibration signal of each vibration signal in each short-time window, combined with the similarity between the first-stage signals corresponding to adjacent IMFs in the short-time window, the first-stage correction coefficient is obtained; According to the vibration propagation sequence and the IMF component signal of each vibration signal in each short-time window, a second-stage signal corresponding to each IMF is obtained; according to the signal correlation relationship between the first-stage signal and the second-stage signal, a second-stage correction coefficient is obtained; The signal in the vibration propagation sequence is corrected using the first-stage correction coefficient and the second-stage correction coefficient, and the vibration source signal segment corresponding to each monitoring point is obtained according to the correction result and the similarity between the IMF component signals of each vibration signal in each short-time window; According to the vibration signal segment corresponding to each monitoring point and combined with the point cloud data inside the deep foundation pit of the pipeline gallery, the deformation monitoring results of the deep foundation pit of the pipeline gallery are obtained.

2. The intelligent detection method for deformation of deep foundation pit of pipe gallery according to claim 1 is characterized in that: The method of analyzing the mutation signal point according to the change trend of each IMF component signal of each vibration signal in each short-time window, and obtaining the vibration propagation sequence of each vibration signal in each short-time window and the first-stage signal corresponding to each IMF specifically includes: For any short-time window of any vibration signal; The second-order difference value of each IMF component signal is obtained respectively, and the moment when the second-order difference value is greater than the preset threshold is recorded as the mutation point, and the signal corresponding to the first mutation point in each IMF component signal is recorded as the initial mutation signal; The initial mutation signals of all IMF component signals in the short-time window constitute the vibration propagation sequence of the vibration signal in the short-time window; on each IMF component signal in the short-time window, the signal segment corresponding to the time period corresponding to the initial mutation signal in the vibration propagation sequence is obtained as the first stage signal corresponding to each IMF in the short-time window.

3. The intelligent detection method for deformation of deep foundation pit of pipe gallery according to claim 1 is characterized in that: The first-stage correction coefficient is obtained according to the similarity between the IMF component signal and the corresponding vibration signal of each vibration signal in each short-time window, combined with the similarity between the first-stage signals corresponding to adjacent IMFs in the short-time window, specifically including: For any short-time window of any vibration signal, any IMF component signal is recorded as the target component signal; Determining a first characteristic factor based on the similarity between the first stage signal corresponding to the target component signal and the original vibration signal; Obtaining the number of repetitions between all frequency components in the target component signal and all frequency components in the next adjacent IMF component signal as the second characteristic factor; A first-stage correction coefficient of the target component signal is determined based on a negative correlation coefficient of a ratio between the first characteristic factor and the second characteristic factor.

4. The intelligent detection method for deformation of deep foundation pit of pipe gallery according to claim 2 is characterized in that: The step of obtaining the second-stage signal corresponding to each IMF according to the vibration propagation sequence and the IMF component signal of each vibration signal in each short-time window specifically includes: The time period after the last initial mutation point of all IMF component signals in the short-time window is recorded as the second stage; The signal corresponding to each IMF component signal in the short-time window in the second stage is used as the second stage signal corresponding to each IMF in the short-time window.

5. The intelligent detection method for deformation of deep foundation pit of pipe gallery according to claim 2 is characterized in that: The obtaining of the second-stage correction coefficient according to the signal correlation relationship between the first-stage signal and the second-stage signal specifically includes: For any short-time window of any vibration signal; Obtain the first-stage signal function and the second-stage signal function corresponding to each IMF; calculate the product between each first-stage signal function and the second-stage signal function in the short-time window, and the integral value at the corresponding time as the cross-correlation characteristic value between each first-stage signal and each second-stage signal; Record any IMF component signal as a selected component signal, obtain the maximum value of the cross-correlation eigenvalue between the selected component signal and each first-stage signal and record it as the maximum eigenvalue of the selected component signal; The time length between the first stage signal corresponding to the maximum eigenvalue of the selected component signal and the second stage signal corresponding to the selected component signal is obtained, and the ratio of the time length to the maximum eigenvalue is normalized to obtain the second stage correction coefficient of the selected component signal in the short time window.

6. The intelligent detection method for deformation of deep foundation pit of pipe gallery according to claim 5 is characterized in that: The correcting the signal in the vibration propagation sequence by using the first-stage correction coefficient and the second-stage correction coefficient specifically includes: For any short-time window of any vibration signal; Obtain the time interval between the initial mutation point of each IMF component signal corresponding to the vibration propagation sequence and the initial mutation point of the next adjacent IMF component signal to obtain the characteristic time length of each IMF component signal; Calculate the product of the mean value between the first-stage correction coefficient and the second-stage correction coefficient of each IMF component signal and the characteristic time length to obtain the moving time length of each IMF component signal; The initial mutation point of each IMF component signal corresponding to the vibration propagation sequence is shifted to the right by the moving time length to obtain a moving signal which is recorded as a modified propagation signal of each IMF component signal; A sequence composed of the corrected propagation signals of all IMF component signals within the short-time window is used as a corrected propagation sequence, and the corrected propagation sequence is a correction result.

7. The intelligent detection method for deformation of deep foundation pit of pipe gallery according to claim 6 is characterized in that: The method of obtaining the vibration source signal segment corresponding to each monitoring point according to the correction result and the similarity between the IMF component signals of each vibration signal in each short-time window specifically includes: The average phases corresponding to all IMF component signals in the short-time window are arranged in ascending order of the average phases to form a propagation phase sequence; According to the time length between the vibration propagation signals of two adjacent IMF component signals in each short-time window, combined with the difference between two adjacent average phases in the propagation phase sequence and the time series length of the vibration propagation sequence, the similarity index between each short-time window and the modified propagation sequence is obtained; According to the change of the similarity index between each short-time window of each vibration signal and the modified propagation sequence, the vibration source signal segment of each vibration signal corresponding to the monitoring point is obtained.

8. The intelligent detection method for deformation of deep foundation pit of pipe gallery according to claim 7 is characterized in that: The calculation formula of the similarity index between each short-term window and the modified propagation sequence is: in, It represents the similarity index between the u-th short-time window of any vibration signal and the modified propagation sequence, It represents the time interval between the initial mutation point of the ith IMF component signal and the initial mutation point of the next adjacent IMF component signal in the modified propagation sequence within the uth short-time window. Represents the total number of IMF component signals contained in the short-time window, represents the time length of the modified propagation sequence, The absolute value of the average phase difference between the ith IMF component signal and the next adjacent IMF component signal in the propagation phase sequence of the u-th short-time window of the vibration signal, that is, the phase delay, Represents the maximum value of all phase delays in the u-th short-time window of the vibration signal, represents an exponential function with the natural constant e as the base, is the normalization function.

9. The intelligent detection method for deformation of deep foundation pit of pipe gallery according to claim 7 is characterized in that: The step of obtaining the vibration source signal segment of the monitoring point corresponding to each vibration signal according to the change of the similarity index between each short-time window of each vibration signal and the modified propagation sequence specifically includes: For any vibration signal, calculate the first-order difference value of the similarity index corresponding to each short-time window in the vibration signal, and take the previous short-time window adjacent to the short-time window corresponding to the minimum value of all first-order difference values ​​as the first vibration source signal segment; According to the same method as the first vibration source signal segment, the mutation point is reacquired within the short time window after the vibration source signal segment to determine the second vibration source signal segment, and so on, to obtain all vibration source signal segments corresponding to the monitoring points of the vibration signal.

10. An intelligent detection system for deformation of a deep foundation pit of a pipe gallery, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the computer program is executed by the processor, the steps of the intelligent detection method for deformation of a deep foundation pit of a pipe gallery as described in any one of claims 1 to 9 are implemented.

Images