TBM tunneling control method and system

Abstract

The invention relates to the technical field of tunnel boring machines, in particular to a TBM tunneling control method and system, and the method comprises the steps: obtaining TBM vibration acceleration data and a rock slag image to obtain a rock mass integrity coefficient, rock compressive strength, a TBM main beam vibration power spectrum density characteristic and a TBM main beam vibration root-mean-square amplitude, thereby setting an optimal TBM tunneling parameter; the vibration amplitude and the resonance phenomenon of the TBM are reduced by acquiring the vibration condition of the TBM main beam and setting the current optimal tunneling parameters according to the factors influencing the vibration of the TBM main beam, such as the rock compressive strength and the rock mass integrity coefficient.

Description

Technical Field

[0001] This invention relates to the field of tunnel boring machine technology, and more specifically, to a TBM tunneling control method and system. Background Technology

[0002] A tunnel boring machine (TBM) is a large machine that uses a rotating cutterhead to break rocks and propel itself forward, providing tunnel support and muck removal during the process. TBM tunneling replaces traditional explosive blasting, offering a more controllable and safer process. However, TBM tunneling also involves many uncontrollable factors, such as the geological parameters of the underground rock formations. Especially when encountering rock masses with high integrity and strong compressive strength, the thrust of the TBM's cutterhead during cutting and shearing of hard rock intensifies. When the rock acts with a large reaction force on the cutterhead's rollers, it is instantly broken into rock fragments by another roller on the cutterhead, causing the reaction force to disappear momentarily. This process generates vibration in the cutterhead. This vibration is transmitted to other TBM components such as the saddle, main beam, and support shoes, and severe vibration can cause abnormal wear on these mechanical parts. To reduce vibration, in addition to conducting thorough geological surveys and adjusting tunneling parameters to suit the current rock strata before construction, it is also necessary to continuously monitor construction data during the construction process and adjust TBM tunneling parameters in a timely manner to avoid TBM jamming and abnormal wear and tear on the mechanical structure. Besides the impact of vibration on the wear and tear of the TBM's components, the construction process of urban subway tunnels is constrained by the special environment, requiring minimal environmental impact. The construction vibration caused by TBM mechanical tunneling in the confined space of urban subway tunnels can also affect the stability and safety of existing buildings around the tunnel. Summary of the Invention

[0003] (1) Technical problems to be solved

[0004] To address the shortcomings of existing technologies, the present invention aims to provide a TBM tunneling control method and system that, by acquiring current vibration parameters and combining them with rock properties, sets the optimal tunneling parameters to reduce TBM vibration and ensure that the vibration impact caused by the TBM is within a controllable range.

[0005] (2) Technical solution

[0006] To achieve the above objectives, the embodiments of this application provide the following technical solutions:

[0007] On one hand, embodiments of this application provide a TBM tunneling control method, the method comprising:

[0008] Initializing TBM tunneling parameters includes setting the cutterhead rotation speed, cutterhead thrust, and penetration depth. Once the TBM cutterhead contacts the tunnel face, the tunneling parameters are read in real time. When fluctuations in the penetration depth are detected and the average value of these fluctuations exceeds a set penetration depth threshold, acceleration data from at least one acceleration sensor positioned on the TBM main beam in the steady-state section is acquired. This steady-state acceleration data is obtained using an acceleration sensor positioned to measure the vertical acceleration of the TBM main beam. The root mean square amplitude of the TBM main beam vibration is calculated based on the acceleration data.

[0009] The system acquires a first image of the rock slag conveyed by the cutterhead belt conveyor using a camera. It then performs a logarithmic transformation on the RGB values ​​of each pixel in the first rock slag image to obtain a first logarithmic image. Next, it performs a Gaussian blur on the rock slag image according to a set pixel radius and performs a logarithmic transformation on the RGB values ​​of each pixel to obtain a second logarithmic image. A difference image is obtained by subtracting the RGB values ​​of the second image from the RGB values ​​of the first logarithmic image. The inverse difference image is obtained by calculating the inverse logarithm of the RGB values ​​of each pixel in the difference image. Finally, a second rock slag image is obtained by normalizing the R channel to the range of 0–255 and scaling the G and B channels proportionally to the R channel scaling ratio.

[0010] The second rock debris image is converted into a grayscale image using the grayscale maximum value algorithm and its edges are detected using the Canny algorithm. The number of closed spaces enclosed by the edges is counted as the first quantity, and the number of closed spaces whose elongation is within a set elongation threshold is counted as the second quantity. The ratio of the second quantity to the first quantity is recorded as the first ratio, and the first ratio has a mapping relationship with the rock mass integrity coefficient.

[0011] Based on the second rock debris image, edges are detected using the Canny algorithm. Image regions with closed spaces enclosed by the edges and whose closed space threshold is greater than a set area threshold are saved as a third rock debris image set. The texture features of the third rock debris images in the third rock debris image set are extracted using the LBP algorithm, and the color features are extracted using the color moment method. The texture and color features are then used to determine the rock type of the third rock debris using a rock recognition model. The percentage of each rock type within the third rock debris image set is calculated. The compressive strength of each rock type within the third rock debris image set is multiplied by its corresponding percentage and summed to obtain a second ratio. This second ratio has a mapping relationship with the rock compressive strength. The rock recognition model establishes a mapping relationship between rock color features, rock texture features, and rock type using a neural network model.

[0012] During TBM tunneling, when the root mean square amplitude of the TBM main beam vibration exceeds a set root mean square amplitude threshold, the first ratio, the second ratio, and the root mean square amplitude of the TBM main beam vibration are input into the first tunneling parameter model to obtain the first compensated cutterhead speed and the first compensated cutterhead thrust. The cutterhead speed is then superimposed with the first compensated cutterhead speed to obtain the first corrected cutterhead speed. The first tunneling parameter model establishes a mapping relationship between the first ratio, the second ratio, the set root mean square amplitude threshold, the cutterhead thrust, and the cutterhead speed using a multiple linear regression algorithm. When the cutterhead speed obtained through the first tunneling parameter model satisfies the condition that the root mean square amplitude of the TBM main beam vibration does not exceed the set root mean square amplitude threshold, and the cutterhead thrust is the minimum value that satisfies the condition that the root mean square amplitude of the TBM main beam vibration does not exceed the set root mean square amplitude threshold, the tunneling parameter model is determined to be accurate.

[0013] Furthermore, the control method further includes:

[0014] The steady-state acceleration data is filtered out of its DC component by a DC high-pass filter and converted into frequency domain signal data by wavelet transform. The power spectral density characteristics of the TBM main beam vibration are calculated based on the frequency domain signal data.

[0015] During TBM tunneling, when the power spectral density characteristics of the TBM main beam vibration are not within the set power spectral density interval threshold, the first ratio, the second ratio, and the power spectral density characteristics of the TBM main beam vibration are input into the second tunneling parameter model to obtain the second compensated cutterhead thrust. The cutterhead thrust is then superimposed on the second compensated cutterhead thrust to obtain the second corrected cutterhead thrust. The first ratio, the second ratio, the root mean square amplitude of the TBM main beam vibration, and the second corrected cutterhead thrust are input into the first tunneling parameter model to obtain the second compensated cutterhead rotation speed. The cutterhead rotation speed is then superimposed on the second compensated cutterhead rotation speed to obtain the second corrected cutterhead rotation speed. The second tunneling parameter model establishes a mapping relationship between the first ratio, the second ratio, the set power spectral density interval threshold, and the cutterhead thrust through a neural network. When the cutterhead thrust obtained through the second tunneling parameter model satisfies the requirement that the power spectral density characteristics of the TBM main beam vibration are within the set power spectral density interval threshold, the tunneling parameter model is considered to be operating within the specified range.

[0016] Furthermore, the control method further includes:

[0017] Before the TBM cutterhead contacts the tunnel face, acceleration data is acquired. The DC component of the acceleration data is filtered out by a DC high-pass filter and converted into wavelet transform into the spectrum data of the tunneling section. The power spectral density characteristics of the TBM main beam vibration in the tunneling section are calculated based on the spectrum data of the tunneling section and recorded as the power spectral characteristics of the tunneling section. Based on the power spectral characteristics of the tunneling section, peak values ​​with power spectral density greater than a set threshold are detected by a peak detection algorithm. A narrow frequency band centered on the peak value is set, and the frequency range of the narrow frequency band is set as the parameters of the band-stop filter.

[0018] Before performing wavelet transform on the steady-state acceleration data, the steady-state acceleration data is filtered to remove its DC component by passing a DC high-pass filter and to remove its vibration frequency component in the push-out section by passing a band-stop filter to obtain filtered acceleration data. The filtered acceleration data is then converted into frequency domain signal data by wavelet transform, and the power spectral density characteristics of the TBM main beam vibration are calculated based on the frequency domain signal data.

[0019] Furthermore, the control method further includes:

[0020] When the TBM tunneling parameters are run at the second corrected cutterhead speed exceeding the set number of times, if the difference between the root mean square amplitude of the TBM main beam vibration and the set root mean square amplitude threshold is greater than the set first difference threshold, then the first tunneling parameter model is regenerated.

[0021] Furthermore, the control method further includes:

[0022] When the TBM tunneling parameters are operated with the second corrected cutterhead thrust exceeding the set number of times, and the difference between the power spectrum characteristic density of the TBM main beam vibration and the set power spectrum density interval threshold is greater than the set second difference threshold, the second tunneling parameter model is regenerated.

[0023] Secondly, based on the same inventive concept, this embodiment provides a TBM tunneling control system, the system comprising:

[0024] The vibration amplitude acquisition module is used to initialize TBM tunneling parameters, including cutterhead rotation speed, cutterhead thrust, and penetration depth. When the TBM cutterhead contacts the working face, the TBM tunneling parameters are read in real time. When fluctuations in the penetration depth are detected and the average value of these fluctuations exceeds a set penetration depth threshold, acceleration data from at least one acceleration sensor located on the TBM main beam in the steady-state section is acquired. This steady-state acceleration data is obtained from an acceleration sensor positioned to measure the vertical acceleration of the TBM main beam. The root mean square amplitude of the TBM main beam vibration is calculated based on the acceleration data.

[0025] The image enhancement module is used to acquire the current first rock slag image conveyed by the cutterhead belt conveyor via a camera, perform a logarithmic transformation on the RGB values ​​of each pixel of the first rock slag image to obtain a first logarithmic image, and then perform a Gaussian blur on the rock slag image according to a set pixel radius and perform a logarithmic transformation on the RGB values ​​of each pixel to obtain a second logarithmic image; subtract the RGB values ​​of the second pixel image from the RGB values ​​of the first logarithmic image to obtain a difference image, calculate the inverse logarithm of the RGB values ​​of each pixel of the difference image to obtain an inverse difference image, and obtain a second rock slag image by normalizing its R channel to the range of 0 to 255 and scaling the G and B channels proportionally to the R channel scaling ratio to the corresponding values;

[0026] The first ratio acquisition module is used to convert the second rock debris image into a grayscale image using a grayscale maximum value algorithm and detect its edges using a Canny algorithm; count the number of closed spaces enclosed by the edges as a first quantity, count the number of closed spaces whose elongation is within a set elongation threshold as a second quantity; and record the ratio of the second quantity to the first quantity as a first ratio, which has a mapping relationship with the rock mass integrity coefficient.

[0027] The second ratio acquisition module is used to detect the edges of the second rock debris image using the Canny algorithm, and save the image regions enclosed by the edges with a closed space threshold greater than a set area threshold as a third rock debris image set; sequentially, the texture features of the third rock debris images in the third rock debris image set are extracted using the LBP algorithm and the color features are extracted using the color moment method; the texture features and color features are used to obtain the rock type of the third rock debris through a rock recognition model, and the percentage of each rock type in the third rock debris image set is calculated; the compressive strength of each rock type in the third rock debris image set is multiplied by its corresponding percentage and accumulated to obtain the second ratio, which has a mapping relationship with the rock compressive strength; the rock recognition model establishes the mapping relationship between rock color features, rock texture features and rock type through a neural network model.

[0028] The first tunneling parameter adjustment module is used to, during TBM tunneling, when the root mean square amplitude of the TBM main beam vibration exceeds a set root mean square amplitude threshold, input a first ratio, a second ratio, and the root mean square amplitude of the TBM main beam vibration into a first tunneling parameter model to obtain a first compensated cutterhead speed and a first compensated cutterhead thrust, and then superimpose the cutterhead speed with the first compensated cutterhead speed to obtain a first corrected cutterhead speed; the first tunneling parameter model establishes a mapping relationship between the first ratio, the second ratio, the set root mean square amplitude threshold, the cutterhead thrust, and the cutterhead speed through a multiple linear regression algorithm; when the cutterhead speed obtained through the first tunneling parameter model satisfies the condition that the root mean square amplitude of the TBM main beam vibration does not exceed the set root mean square amplitude threshold, and the cutterhead thrust is the minimum value that satisfies the condition that the root mean square amplitude of the TBM main beam vibration does not exceed the set root mean square amplitude threshold.

[0029] Furthermore, the control system also includes:

[0030] The vibration spectrum acquisition module is used to filter out the DC component of the steady-state acceleration data through a DC high-pass filter and convert it into frequency domain signal data through wavelet transform. Based on the frequency domain signal data, the power spectral density characteristics of the TBM main beam vibration are calculated.

[0031] The second tunneling parameter adjustment module is used to, during TBM tunneling, when the power spectral density characteristics of the TBM main beam vibration are not within a set power spectral density range threshold, input the first ratio, the second ratio, and the power spectral density characteristics of the TBM main beam vibration into the second tunneling parameter model to obtain the second compensated cutterhead thrust, and then superimpose the cutterhead thrust with the second compensated cutterhead thrust to obtain the second corrected cutterhead thrust; input the first ratio, the second ratio, the root mean square amplitude of the TBM main beam vibration, and the second corrected cutterhead thrust into the first tunneling parameter model to obtain the second compensated cutterhead rotation speed, and then superimpose the cutterhead rotation speed with the second compensated cutterhead rotation speed to obtain the second corrected cutterhead rotation speed; the second tunneling parameter model establishes a mapping relationship between the first ratio, the second ratio, the set power spectral density range threshold, and the cutterhead thrust through a neural network; when the cutterhead thrust obtained through the second tunneling parameter model satisfies the requirement that the power spectral density characteristics of the TBM main beam vibration are within the set power spectral density range threshold.

[0032] Furthermore, the control system also includes:

[0033] The vibration acquisition module for the tunnel boring machine (TBM) is used to acquire acceleration data before the cutterhead contacts the tunnel face. Based on the acceleration data, the DC component is filtered out by a DC high-pass filter and converted into spectral data for the tunnel boring machine using wavelet transform. The power spectral density characteristics of the TBM main beam vibration in the tunnel boring section are calculated based on the spectral data and recorded as the power spectral characteristics of the tunnel boring machine. Based on the power spectral characteristics of the tunnel boring machine, a peak detection algorithm is used to detect peaks whose power spectral density is greater than a set threshold. A narrow frequency band centered on the peak is set, and the frequency range of the narrow frequency band is set as the parameters of the band-stop filter.

[0034] The steady-state section vibration filtering module is used to filter out the DC component of the steady-state section acceleration data by passing it through a DC high-pass filter and the vibration frequency component of the air thrust section by passing it through a band-stop filter before performing wavelet transform on the steady-state section acceleration data. The filtered acceleration data is then converted into frequency domain signal data by wavelet transform, and the power spectral density characteristics of the TBM main beam vibration are calculated based on the frequency domain signal data.

[0035] Furthermore, the control system also includes:

[0036] The first parameter model adjustment module is used to regenerate the first tunneling parameter model when the difference between the root mean square amplitude of the TBM main beam vibration and the set root mean square amplitude threshold is greater than the set first difference threshold after the TBM tunneling parameters have been running at the second corrected cutterhead speed for more than a set number of times.

[0037] Furthermore, the control system also includes:

[0038] The second parameter model adjustment module is used to regenerate the second tunneling parameter model when the TBM tunneling parameters are operated with the second corrected cutterhead thrust more than a set number of times, and the difference between the power spectrum characteristic density feature of the TBM main beam vibration and the set power spectrum density interval threshold is greater than the set second difference threshold.

[0039] (3) Beneficial effects

[0040] The beneficial effects of this invention are as follows:

[0041] By acquiring the vibration data of the TBM main beam and determining the optimal tunneling parameters based on factors affecting the TBM main beam vibration, such as rock compressive strength and rock mass integrity coefficient, the vibration amplitude and resonance phenomenon of the TBM can be reduced. Attached Figure Description

[0042] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a flowchart of the method according to Embodiment 1 of the present invention;

[0044] Figure 2 This is a system module block diagram of Embodiment 2 of the present invention. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0046] It should be noted that similar reference numerals or letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0047] Before providing examples, the application scenarios involved in this invention need further explanation. During TBM tunneling, the cutterhead rotates continuously and advances forward, with the cutters cutting out fracture pits and cracks on the rock surface. When the cutters penetrate the rock cracks formed by the previous cutter cutters, the cracks peel off, forming rock fragments, which are then transported outside the tunnel by a belt conveyor. An unavoidable problem is that the cutting and squeezing of the cutterhead and rock during TBM tunneling inevitably causes cutterhead vibration. This severe vibration is transmitted to the main beam, and then from the main beam to the saddle and support shoes. The severe vibration of the main beam leads to wear on the saddle slide rails, and simultaneously increases the wear and tear on the cutterhead, cutters, and mechanical structural components on the equipment bridge. Therefore, the purpose of this invention is to monitor the vibration of the TBM to adjust the TBM tunneling parameters in real time, aiming to prevent the vibration during TBM tunneling from exceeding a certain threshold and causing abnormal wear on structural components such as the saddle slide rails, cutterhead, and cutterhead support. Therefore, it is necessary to monitor the vibration of the TBM during tunneling. However, the working conditions at the cutterhead are extremely harsh, making it difficult to place sensors for detection. A better solution is to place a triaxial accelerometer on the main beam of the TBM. In addition to being affected by tunneling parameters, the vibration of the TBM is also affected by rock properties, such as rock mass integrity and rock compressive strength. Therefore, this invention sets the optimal tunneling parameters by detecting the vibration of the TBM and the properties of the rock.

[0048] Example 1

[0049] like Figure 1 As shown in the figure, this embodiment provides a TBM tunneling control method, the method including:

[0050] Initializing TBM tunneling parameters includes controlling cutterhead rotation speed, cutterhead thrust, and penetration depth. Once the TBM cutterhead contacts the working face, these parameters are read in real-time. When fluctuations in the penetration depth are detected, and the average value of these fluctuations exceeds a set penetration depth threshold, acceleration data from at least one location on the TBM main beam in the steady-state section is acquired. This steady-state acceleration data is obtained using acceleration sensors positioned to measure the vertical acceleration of the TBM main beam. The root mean square amplitude of the TBM main beam vibration is calculated based on the acceleration data. Penetration depth is correlated with cutterhead thrust and rock properties; therefore, significant fluctuations in penetration depth indicate intense operation between the cutterhead and the working face (the plane formed by the cutterhead cutting through the rock). A set fluctuation threshold in penetration depth data signifies that the penetration depth is no longer stable but fluctuates. This fluctuation threshold refers to the average value of the fluctuating penetration depth data exceeding a set threshold. A complete TBM tunneling process is divided into a shutdown phase and a tunneling phase. The tunneling phase is further divided into an unsupported thrust phase, an ascent phase, and a steady-state phase. The steady-state phase is the most valuable for monitoring TBM main beam vibration. Therefore, the shutdown phase is for maintenance or cutter replacement. The unsupported thrust phase occurs before the cutterhead contacts the tunnel face, and the TBM main beam vibration is very small at this time, approximately one-tenth of the vibration data in the steady-state phase. The ascent phase is the process of the TBM main beam vibration gradually increasing to the steady-state phase, and its time is relatively short, about 2-4 minutes. TBM main beam vibration occurs in the axial, lateral, and vertical directions. Extensive experiments have shown that during the same tunneling process, the vertical vibration amplitude of the TBM main beam is the largest, followed by the axial and lateral directions. Therefore, measuring vertical vibration data is optimal in this embodiment. However, using lateral or axial vibration to replace vertical vibration in this area can also achieve vibration data acquisition. Vibration data is acquired using accelerometers. These accelerometers are typically mounted on or near the surface of the object being measured. When the object vibrates, the accelerometer senses the acceleration and converts it into a corresponding electrical signal. The root mean square amplitude represents the average amplitude of the vibration signal, reflecting the vibration energy.

[0051] The system acquires a first image of the rock debris being transported by the cutterhead conveyor belt using a camera. A first logarithmic image is obtained by performing a logarithmic transformation on the RGB values ​​of each pixel in the first rock debris image. Then, the rock debris image is Gaussian blurred according to a set pixel radius, and a second logarithmic image is obtained by performing a logarithmic transformation on the RGB values ​​of each pixel in the second image. A difference image is obtained by subtracting the RGB values ​​of the second image from the RGB values ​​of the first image. An inverse difference image is obtained by calculating the inverse logarithm of the RGB values ​​of each pixel in the difference image. Based on the inverse difference image, the R channel is normalized to a value between 0 and 255, and the G and B channels are scaled proportionally to their corresponding values ​​according to the scaling ratio of the R channel to obtain the second rock debris image. Since the camera is positioned above the conveyor belt, and the lighting conditions are poor in the tunnel environment, the acquired first rock debris image requires enhancement processing. In this embodiment, the camera acquires the current first rock debris image from the belt conveyor section near the cutterhead. Some belt conveyors have a long transport distance, and some sections of the belt conveyor are even closer to the tunnel entrance with better lighting. However, in this embodiment, the camera is still positioned very close to the belt conveyor section near the cutterhead because the first rock debris image acquired at this time is the closest to the cutterhead. Controlling the TBM tunneling parameters based on the first rock debris image and the vibration of the TBM main beam is more timely; otherwise, there would be a significant time lag in this process.

[0052] The second rock debris image is converted into a grayscale image using a grayscale maximum value algorithm, and its edges are detected using the Canny algorithm. The number of closed spaces enclosed by these edges is recorded as the first quantity, and the number of closed spaces whose extension length falls within a set extension length threshold is recorded as the second quantity. The ratio of the second quantity to the first quantity is recorded as the first ratio, which has a mapping relationship with the rock mass integrity coefficient. Ideally, the rock mass integrity coefficient should be measured at the face of the cutterhead. However, during construction, the cutterhead is in close contact with the face of the cutterhead, making it impossible to stop the machine and take pictures for identification. Therefore, the first rock debris image must be analyzed as a secondary option. Since the first rock debris has already been broken by the cutterhead, the only way to determine its rock mass integrity coefficient is to perform statistical analysis of the shape of the broken rock after various rock mass integrity coefficients. Firstly, the first rock debris contains a wealth of stratigraphic rock information. Taking a fault fracture zone from the inventor's practical operation as an example, the results show that the rock debris excavated by the TBM in this section contains approximately 90% rock blocks and approximately 5% rock flakes, consistent with a rock mass integrity coefficient of Class V. In another section of rock mass integrity coefficient III, the rock debris analysis results show approximately 20% rock blocks and approximately 80% rock flakes. The second rock debris image edge detection statistically analyzes various closed space shapes. Closed spaces with an elongation length within a set threshold are considered rock blocks; otherwise, they are considered rock flakes. The ratio of the second quantity to the first quantity represents the percentage of rock blocks, and this percentage has a corresponding mapping relationship with the integrity coefficient. The closed spaces, i.e., the shape of the rock debris, are irregular, thus requiring determination of whether they are blocky or flaky. Blocky shapes are closer to square, while flaky shapes are more elongated. The elongation of an irregular shape is obtained from its elongation from the side. When the irregular shape is a perfect square, the elongation is 1. When the irregular shape is elongated, the more elongated it is, the closer the elongation is to 0.

[0053] Based on the second rock debris image, edges are detected using the Canny algorithm. Image regions with closed spaces enclosed by the edges and whose closed space threshold is greater than a set area threshold are saved as a third rock debris image set. The texture features of the third rock debris images in the third rock debris image set are extracted using the LBP algorithm, and the color features are extracted using the color moment method. The texture and color features are then used to determine the rock type of the third rock debris using a rock recognition model. The percentage of each rock type within the third rock debris image set is calculated. The compressive strength of each rock type within the third rock debris image set is multiplied by its corresponding percentage and summed to obtain a second ratio, which is correlated with the rock compressive strength. The rock recognition model establishes a mapping relationship between rock color features, rock texture features, and rock type using a neural network model. The second rock debris image is an enhanced version of the first rock debris image under low light conditions. Edges of the second rock debris image are detected, and larger rock areas are identified. For some rock debris that has been broken into fine fragments, it is difficult to obtain its surface characteristics, including color and texture. Therefore, the rock that occupies a large range is identified and its rock classification is calculated. In actual field, the types of rock debris are mixed, so the rock compressive strength of rock debris is affected by the content of various types of rock. The method adopted in this embodiment is to multiply the rock compressive strength of each type of rock by its corresponding percentage to obtain the comprehensive rock compressive strength.

[0054] During TBM tunneling, when the root mean square amplitude of the TBM main beam vibration exceeds a set root mean square amplitude threshold, the first ratio, the second ratio, and the root mean square amplitude of the TBM main beam vibration are input into the first tunneling parameter model to obtain the first compensated cutterhead speed and the first compensated cutterhead thrust. The cutterhead speed is then superimposed with the first compensated cutterhead speed to obtain the first corrected cutterhead speed. The first tunneling parameter model establishes a mapping relationship between the first ratio, the second ratio, the set root mean square amplitude threshold, the cutterhead thrust, and the cutterhead speed using a multiple linear regression algorithm. When the cutterhead speed obtained through the first tunneling parameter model satisfies the condition that the root mean square amplitude of the TBM main beam vibration does not exceed the set root mean square amplitude threshold, and the cutterhead thrust is the minimum value that satisfies the condition that the root mean square amplitude of the TBM main beam vibration does not exceed the set root mean square amplitude threshold, the tunneling parameter model is determined to be accurate. The root mean square (RMS) amplitude of TBM main beam vibration is positively correlated with the cutterhead rotation speed and cutterhead thrust. The faster the cutterhead rotation speed, the more times the cutterhead contacts and collides with the tunnel face per unit time, resulting in more severe TBM main beam vibration and a larger RMS amplitude. Conversely, the greater the cutterhead thrust, the greater the penetration per unit time (i.e., the amount of rock cut by the cutterhead), and the stronger the reaction force on the cutterhead, leading to more severe TBM main beam vibration and a larger RMS amplitude. The first compensation cutterhead rotation speed and thrust can be positive or negative, primarily correcting the original cutterhead rotation speed and thrust. Reducing the RMS amplitude of TBM main beam vibration mainly reduces the intensity of the vibration, thereby mitigating a series of negative impacts, such as reducing abnormal friction between structural components. However, this vibration reduction is not unlimited; it is controlled within a manageable range, meaning the RMS amplitude of TBM main beam vibration does not exceed a set RMS amplitude threshold. The cutterhead thrust and cutterhead rotation speed output by the first tunneling parameter model both affect the root mean square amplitude of the TBM main beam vibration. Therefore, the cutterhead thrust and rotation speed should be selected such that the root mean square amplitude of the TBM main beam vibration is within the set root mean square amplitude threshold, i.e., close to the root mean square amplitude threshold. While multiple sets of parameters can satisfy this condition for cutterhead thrust and rotation speed, this embodiment selects the minimum cutterhead thrust because, in practice, cutter wear is most sensitive to cutterhead thrust. Generally, to reduce wear, the cutterhead thrust is controlled between 9000 and 11000 kN during construction.

[0055] Furthermore, the control method further includes:

[0056] The steady-state acceleration data is filtered out of its DC component by a DC high-pass filter and converted into frequency domain signal data by wavelet transform. The power spectral density characteristics of the TBM main beam vibration are calculated based on the frequency domain signal data.

[0057] During TBM tunneling, when the power spectral density characteristics of the TBM main beam vibration are outside the set power spectral density interval threshold, the first ratio, the second ratio, and the power spectral density characteristics of the TBM main beam vibration are input into the second tunneling parameter model to obtain the second compensated cutterhead thrust. The cutterhead thrust is then superimposed on the second compensated cutterhead thrust to obtain the second corrected cutterhead thrust. Similarly, the first ratio, the second ratio, the root mean square amplitude of the TBM main beam vibration, and the second corrected cutterhead thrust are input into the first tunneling parameter model to obtain the second compensated cutterhead rotation speed. The cutterhead rotation speed is then superimposed on the second compensated cutterhead rotation speed to obtain the second corrected cutterhead rotation speed. The second tunneling parameter model establishes a mapping relationship between the first ratio, the second ratio, the set power spectral density interval threshold, and the cutterhead thrust through a neural network. The cutterhead thrust obtained through the second tunneling parameter model satisfies the condition that the power spectral density characteristics of the TBM main beam vibration are within the set power spectral density interval threshold. When the measured object vibrates, the power spectral density represents the magnitude of the vibration energy contained in each frequency unit; the larger the value, the larger the amplitude value in the corresponding frequency band. It is important to note that the power spectral density characteristic of TBM main beam vibration refers to the frequency distribution characteristic of TBM main beam vibration. The inventors' research shows that when the cutterhead rotation speed changes, regardless of whether it is low, medium, or high speed, the frequency distribution of TBM main beam vibration remains almost unchanged. In the 0–150 Hz range, 85% of the vibration power energy is still distributed in the 0–100 Hz range, while approximately 15% is distributed in the 100–150 Hz range. However, when the cutterhead thrust increases, under a specific rock property, as the cutterhead penetration increases (e.g., from 3 mm / revolution to 9 mm / revolution), the proportion of power energy in the high-frequency portion of the 0–150 Hz range increases from 5% to 55%. Therefore, there is a positive correlation between penetration and the power spectral density characteristic of TBM main beam vibration, but no significant correlation with the power spectral density characteristic of TBM main beam vibration and cutterhead rotation speed. Controlling the power density distribution of vibration involves staggering the natural frequencies of the various structural components of the TBM, minimizing the concentration of higher-frequency vibration energy at the natural frequencies of critical structural components to prevent resonance and thus avoid further damage to the connections between structural components. In other words, the power spectral density characteristics of the TBM main beam vibration are kept within a set power spectral density range threshold. It is important to note that in this embodiment, the power spectral density characteristics of the TBM main beam vibration are first considered, and the cutterhead thrust is limited using a second tunneling parameter model. Then, the cutterhead rotation speed is limited using a first tunneling parameter model. Here, the cutterhead thrust calculated using the first tunneling parameter model is set to a fixed value.

[0058] Furthermore, the control method further includes:

[0059] Before the TBM cutterhead contacts the tunnel face, acceleration data is acquired. The DC component of the acceleration data is filtered out by a DC high-pass filter and converted into wavelet transform into the spectrum data of the tunneling section. The power spectral density characteristics of the TBM main beam vibration in the tunneling section are calculated based on the spectrum data of the tunneling section and recorded as the power spectral characteristics of the tunneling section. Based on the power spectral characteristics of the tunneling section, peak values ​​with power spectral density greater than a set threshold are detected by a peak detection algorithm. A narrow frequency band centered on the peak value is set, and the frequency range of the narrow frequency band is set as the parameters of the band-stop filter.

[0060] Before performing wavelet transform on the steady-state acceleration data, the steady-state acceleration data is filtered to remove its DC component using a DC high-pass filter and to remove its vibration frequency component during the no-load thrust phase using a band-stop filter, resulting in filtered acceleration data. This filtered acceleration data is then converted into frequency domain signal data using wavelet transform, and the power spectral density characteristics of the TBM main beam vibration are calculated based on the frequency domain signal data. The TBM main beam also experiences some vibration during the no-load thrust phase. This vibration originates from errors in TBM machining accuracy, such as the center error of the TBM propulsion spindle. Some vibration also arises from abnormal wear of other structural components, such as the saddle slide, cutter head, and tool holder, leading to inaccurate fit between components and causing vibration. In this embodiment, the aim is to remove vibrations caused by the TBM's own structure and filter out vibrations caused by the interaction between the TBM cutter head and the rock during steady-state operation. The narrow frequency band here is defined as one that surrounds frequencies with smaller peaks to prevent filtering out valuable frequency ranges.

[0061] Furthermore, the control method further includes:

[0062] When the TBM tunneling parameters are run at the second corrected cutterhead speed exceeding the set number of times, if the difference between the root mean square amplitude of the TBM main beam vibration and the set root mean square amplitude threshold is greater than the set first difference threshold, the first tunneling parameter model is regenerated. Even after the TBM has been running for a considerable period, the root mean square amplitude of the TBM main beam vibration remains proportional to the rock mass integrity coefficient, rock compressive strength, cutterhead speed, and cutterhead thrust. However, due to other conditions such as zero-point drift of the accelerometer or the presence of external stable interference sources, the parameter model may become distorted, requiring model regeneration to maintain the accuracy of the first tunneling parameter model.

[0063] Furthermore, the control method further includes:

[0064] When the TBM tunneling parameters are operated at the second corrected cutterhead thrust more than the set number of times, and the difference between the power spectrum characteristic density of the TBM main beam vibration and the set power spectrum density interval threshold is greater than the set second difference threshold, the second tunneling parameter model is regenerated. Even after the TBM has been running for a long time, the power spectrum characteristic density of the TBM main beam vibration remains related to the rock mass integrity coefficient, rock compressive strength, and cutterhead thrust. However, due to other conditions such as zero-point drift of the accelerometer or the presence of external stable interference sources, the parameter model may be distorted, requiring model regeneration to maintain the accuracy of the second tunneling parameter model.

[0065] Example 2

[0066] Based on the same inventive concept, such as Figure 2 As shown, this embodiment provides a TBM tunneling control system, the system comprising:

[0067] The vibration amplitude acquisition module is used to initialize TBM tunneling parameters, including cutterhead rotation speed, cutterhead thrust, and penetration depth. When the TBM cutterhead contacts the working face, the TBM tunneling parameters are read in real time. When fluctuations in the penetration depth are detected and the average value of these fluctuations exceeds a set penetration depth threshold, acceleration data from at least one acceleration sensor located on the TBM main beam in the steady-state section is acquired. This steady-state acceleration data is obtained from an acceleration sensor positioned to measure the vertical acceleration of the TBM main beam. The root mean square amplitude of the TBM main beam vibration is calculated based on the acceleration data.

[0068] The image enhancement module is used to acquire the current first rock slag image conveyed by the cutterhead belt conveyor via a camera, perform a logarithmic transformation on the RGB values ​​of each pixel of the first rock slag image to obtain a first logarithmic image, and then perform a Gaussian blur on the rock slag image according to a set pixel radius and perform a logarithmic transformation on the RGB values ​​of each pixel to obtain a second logarithmic image; subtract the RGB values ​​of the second pixel image from the RGB values ​​of the first logarithmic image to obtain a difference image, calculate the inverse logarithm of the RGB values ​​of each pixel of the difference image to obtain an inverse difference image, and obtain a second rock slag image by normalizing its R channel to the range of 0 to 255 and scaling the G and B channels proportionally to the R channel scaling ratio to the corresponding values;

[0069] The first ratio acquisition module is used to convert the second rock debris image into a grayscale image using a grayscale maximum value algorithm and detect its edges using a Canny algorithm; count the number of closed spaces enclosed by the edges as a first quantity, count the number of closed spaces whose elongation is within a set elongation threshold as a second quantity; and record the ratio of the second quantity to the first quantity as a first ratio, which has a mapping relationship with the rock mass integrity coefficient.

[0070] The second ratio acquisition module is used to detect the edges of the second rock debris image using the Canny algorithm, and save the image regions enclosed by the edges with a closed space threshold greater than a set area threshold as a third rock debris image set; sequentially, the texture features of the third rock debris images in the third rock debris image set are extracted using the LBP algorithm and the color features are extracted using the color moment method; the texture features and color features are used to obtain the rock type of the third rock debris through a rock recognition model, and the percentage of each rock type in the third rock debris image set is calculated; the compressive strength of each rock type in the third rock debris image set is multiplied by its corresponding percentage and accumulated to obtain the second ratio, which has a mapping relationship with the rock compressive strength; the rock recognition model establishes the mapping relationship between rock color features, rock texture features and rock type through a neural network model.

[0071] The first tunneling parameter adjustment module is used to, during TBM tunneling, when the root mean square amplitude of the TBM main beam vibration exceeds a set root mean square amplitude threshold, input a first ratio, a second ratio, and the root mean square amplitude of the TBM main beam vibration into a first tunneling parameter model to obtain a first compensated cutterhead speed and a first compensated cutterhead thrust, and then superimpose the cutterhead speed with the first compensated cutterhead speed to obtain a first corrected cutterhead speed; the first tunneling parameter model establishes a mapping relationship between the first ratio, the second ratio, the set root mean square amplitude threshold, the cutterhead thrust, and the cutterhead speed through a multiple linear regression algorithm; when the cutterhead speed obtained through the first tunneling parameter model satisfies the condition that the root mean square amplitude of the TBM main beam vibration does not exceed the set root mean square amplitude threshold, and the cutterhead thrust is the minimum value that satisfies the condition that the root mean square amplitude of the TBM main beam vibration does not exceed the set root mean square amplitude threshold.

[0072] Furthermore, the control system also includes:

[0073] The vibration spectrum acquisition module is used to filter out the DC component of the steady-state acceleration data through a DC high-pass filter and convert it into frequency domain signal data through wavelet transform. Based on the frequency domain signal data, the power spectral density characteristics of the TBM main beam vibration are calculated.

[0074] The second tunneling parameter adjustment module is used to, during TBM tunneling, when the power spectral density characteristics of the TBM main beam vibration are not within a set power spectral density range threshold, input the first ratio, the second ratio, and the power spectral density characteristics of the TBM main beam vibration into the second tunneling parameter model to obtain the second compensated cutterhead thrust, and then superimpose the cutterhead thrust with the second compensated cutterhead thrust to obtain the second corrected cutterhead thrust; input the first ratio, the second ratio, the root mean square amplitude of the TBM main beam vibration, and the second corrected cutterhead thrust into the first tunneling parameter model to obtain the second compensated cutterhead rotation speed, and then superimpose the cutterhead rotation speed with the second compensated cutterhead rotation speed to obtain the second corrected cutterhead rotation speed; the second tunneling parameter model establishes a mapping relationship between the first ratio, the second ratio, the set power spectral density range threshold, and the cutterhead thrust through a neural network; when the cutterhead thrust obtained through the second tunneling parameter model satisfies the requirement that the power spectral density characteristics of the TBM main beam vibration are within the set power spectral density range threshold.

[0075] Furthermore, the control system also includes:

[0076] The vibration acquisition module for the tunnel boring machine (TBM) is used to acquire acceleration data before the cutterhead contacts the tunnel face. Based on the acceleration data, the DC component is filtered out by a DC high-pass filter and converted into spectral data for the tunnel boring machine using wavelet transform. The power spectral density characteristics of the TBM main beam vibration in the tunnel boring section are calculated based on the spectral data and recorded as the power spectral characteristics of the tunnel boring machine. Based on the power spectral characteristics of the tunnel boring machine, a peak detection algorithm is used to detect peaks whose power spectral density is greater than a set threshold. A narrow frequency band centered on the peak is set, and the frequency range of the narrow frequency band is set as the parameters of the band-stop filter.

[0077] The steady-state section vibration filtering module is used to filter out the DC component of the steady-state section acceleration data by passing it through a DC high-pass filter and the vibration frequency component of the air thrust section by passing it through a band-stop filter before performing wavelet transform on the steady-state section acceleration data. The filtered acceleration data is then converted into frequency domain signal data by wavelet transform, and the power spectral density characteristics of the TBM main beam vibration are calculated based on the frequency domain signal data.

[0078] Furthermore, the control system also includes:

[0079] The first parameter model adjustment module is used to regenerate the first tunneling parameter model when the difference between the root mean square amplitude of the TBM main beam vibration and the set root mean square amplitude threshold is greater than the set first difference threshold after the TBM tunneling parameters have been running at the second corrected cutterhead speed for more than a set number of times.

[0080] Furthermore, the control system also includes:

[0081] The second parameter model adjustment module is used to regenerate the second tunneling parameter model when the TBM tunneling parameters are operated with the second corrected cutterhead thrust more than a set number of times, and the difference between the power spectrum characteristic density feature of the TBM main beam vibration and the set power spectrum density interval threshold is greater than the set second difference threshold.

[0082] It should be noted that the specific methods by which each module performs operations in the system described in the above embodiments have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0083] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A TBM excavation control method, characterized by, The method comprises: The initialization of the TBM tunneling parameters comprises reading the TBM tunneling parameters in real time through the speed of the cutter head, the thrust of the cutter head and the penetration degree when the cutter head of the TBM contacts the tunnel face; when the fluctuation data of the penetration degree value is detected and the mean value of the fluctuation data is greater than the set penetration degree threshold, the acceleration data of at least one of the steady-state sections arranged on the main beam of the TBM is obtained, the steady-state section acceleration data is obtained through the acceleration sensor arranged to measure the vertical acceleration of the main beam of the TBM; and the root mean square amplitude of the vibration of the main beam of the TBM is calculated according to the acceleration data; A first rock residue image currently conveyed by the cutter head belt conveyor is obtained through the camera, the RGB value of each pixel of the first rock residue image is logarithmically transformed to obtain a first logarithmic image, and the rock residue image is Gaussian blurred according to a set pixel radius and the RGB value of each pixel thereof is logarithmically transformed to obtain a second logarithmic image; a difference image is obtained by subtracting the RGB value of the second pixel image from the RGB value of the first logarithmic image, and an inverse difference image is obtained by solving the inverse logarithm of the RGB value of each pixel of the difference image; a second rock residue image is obtained by normalizing the R channel of the inverse difference image to the interval value of 0-255 and proportionally scaling the G channel and the B channel to the corresponding values according to the scaling ratio of the R channel; The second rock residue image is converted into a gray image through a gray maximum value algorithm, and the edges thereof are detected through a Canny algorithm; the number of closed spaces surrounded by the edges is recorded as a first number, and the number of closed spaces with a length greater than a set length threshold is recorded as a second number; a first ratio is recorded according to the ratio of the second number to the first number, and the first ratio and a rock mass integrity coefficient have a mapping relationship; The edges of the second rock residue image are detected through a Canny algorithm, and the image regions of the closed spaces surrounded by the edges and having a closed space threshold greater than a set area threshold are saved as a third rock residue image set; the texture features of the third rock residue images in the third rock residue image set are extracted through an LBP algorithm, and the color features thereof are extracted through a color moment method; the rock types of the third rock residue are obtained through a rock identification model, and the percentages of various rock types in the third rock residue image set are counted; the rock compressive strengths of various rock types in the third rock residue image set are multiplied by the corresponding percentages, and the second ratio is obtained by accumulating the products; the second ratio and the rock compressive strength have a mapping relationship; and the rock identification model is a neural network model for establishing the mapping relationship among the rock color features, the rock texture features and the rock types. During the TBM tunneling, when the root mean square amplitude of the TBM girder vibration exceeds the set root mean square amplitude threshold, the first ratio, the second ratio, and the root mean square amplitude of the TBM girder vibration are input into the first tunneling parameter model to obtain the first compensated cutterhead rotating speed and the first compensated cutterhead thrust, and the cutterhead rotating speed is superimposed with the first compensated cutterhead rotating speed to obtain the first corrected cutterhead rotating speed; the first tunneling parameter model is established by a multiple linear regression algorithm to map the first ratio, the second ratio, the set root mean square amplitude threshold, and the cutterhead thrust and the cutterhead rotating speed; when the cutterhead rotating speed obtained by the first tunneling parameter model satisfies that the root mean square amplitude of the TBM girder vibration does not exceed the set root mean square amplitude threshold, and the cutterhead thrust at this time is the minimum value that satisfies the condition that the root mean square amplitude of the TBM girder vibration does not exceed the set root mean square amplitude threshold.

2. A TBM excavation control method as claimed in claim 1, wherein, The control method further comprises: The steady-state segment acceleration data is filtered of its direct current component by a direct current high-pass filter and is converted into frequency domain signal data by wavelet transform, and the power spectral density characteristics of the TBM girder vibration are calculated according to the frequency domain signal data; During the TBM tunneling, when the power spectral density characteristics of the TBM girder vibration are not within the set power spectral density interval threshold, the first ratio, the second ratio, and the power spectral density characteristics of the TBM girder vibration are input into the second tunneling parameter model to obtain the second compensated cutterhead thrust, and the cutterhead thrust is superimposed with the second compensated cutterhead thrust to obtain the second corrected cutterhead thrust; the first ratio, the second ratio, the root mean square amplitude of the TBM girder vibration, and the second corrected cutterhead thrust are input into the first tunneling parameter model to obtain the second compensated cutterhead rotating speed, and the cutterhead rotating speed is superimposed with the second compensated cutterhead rotating speed to obtain the second corrected cutterhead rotating speed; the second tunneling parameter model is established by a neural network to map the first ratio, the second ratio, the set power spectral density interval threshold, and the cutterhead thrust; when the cutterhead thrust obtained by the second tunneling parameter model satisfies that the power spectral density characteristics of the TBM girder vibration are within the set power spectral density interval threshold.

3. A TBM excavation control method as claimed in claim 2, wherein, The control method further comprises: Before the cutterhead of the TBM contacts the tunnel face, acceleration data is obtained, the acceleration data is filtered of its direct current component by a direct current high-pass filter and is converted into empty thrust segment spectrum data by wavelet transform, the power spectral density characteristics of the TBM girder vibration in the empty thrust segment are calculated according to the empty thrust segment spectrum data and are recorded as empty thrust segment power spectral characteristics; a peak value is detected by a peak value detection algorithm according to the empty thrust segment power spectral characteristics, the peak value is centered to set a narrow frequency band, and the frequency interval of the narrow frequency band is set as the parameter of a band-stop filter; Before the steady-state segment acceleration data is subjected to wavelet transform, the steady-state segment acceleration data is filtered of its direct current component by a direct current high-pass filter and is filtered of its empty thrust segment vibration frequency component by a band-stop filter to obtain filtered acceleration data, the filtered acceleration data is converted into frequency domain signal data by wavelet transform, and the power spectral density characteristics of the TBM girder vibration are calculated according to the frequency domain signal data.

4. A TBM excavation control method as claimed in claim 2, wherein, The control method further comprises: When the TBM tunneling parameters are operated at the second corrected cutterhead speed exceeding the set number of times, the difference between the root mean square amplitude of the TBM main beam vibration and the set root mean square amplitude threshold is greater than the set first difference threshold, and the first tunneling parameter model is regenerated.

5. A TBM excavation control method as claimed in claim 2, wherein, The control method further comprises: When the TBM tunneling parameters are operated at the second corrected cutterhead thrust exceeding the set number of times, the difference between the power spectrum feature density feature of the TBM main beam vibration and the set power spectrum density interval threshold is greater than the set second difference threshold, and the second tunneling parameter model is regenerated.

6. A TBM excavation control system, characterized by, The system comprises: The vibration amplitude acquisition module is configured to initialize the TBM tunneling parameters including the cutterhead speed, the cutterhead thrust and the penetration, read the TBM tunneling parameters in real time after the cutterhead of the TBM contacts the tunnel face, acquire at least one acceleration data of the TBM main beam arranged in the stable state segment when the fluctuation data of the penetration value is detected and the mean value of the fluctuation data is greater than the set penetration threshold, the stable state segment acceleration data is acquired by the acceleration sensor arranged to measure the vertical acceleration of the TBM main beam, and calculate the root mean square amplitude of the TBM main beam vibration according to the acceleration data; The image enhancement module is configured to acquire the current first rock residue image conveyed by the cutterhead belt conveyor through the camera, perform logarithmic transformation on the RGB value of each pixel of the first rock residue image to obtain a first logarithmic image, perform Gaussian blur on the rock residue image according to the set pixel radius, perform logarithmic transformation on the RGB value of each pixel of the rock residue image to obtain a second logarithmic image, subtract the RGB value of the second pixel image from the RGB value of the first logarithmic image to obtain a difference image, solve the inverse logarithm of the RGB value of each pixel of the difference image to obtain an inverse difference image, normalize the R channel of the inverse difference image to the interval value of 0-255, and scale the G channel and the B channel to the corresponding values according to the scaling ratio of the R channel to obtain a second rock residue image; The first ratio acquisition module is configured to convert the second rock residue image into a gray image by the gray maximum value algorithm and detect the edge thereof by the Canny algorithm, count the number of closed spaces surrounded by the edge as a first number, count the number of closed spaces with a length within a set length threshold as a second number, and record the ratio of the second number to the first number as a first ratio, wherein the first ratio and the rock mass integrity coefficient have a mapping relationship. The first ratio acquisition module is configured to convert the second rock residue image into a gray image by the gray maximum value algorithm and detect the edge thereof by the Canny algorithm, count the number of closed spaces surrounded by the edge as a first number, count the number of closed spaces with a length within a set length threshold as a second number, and record the ratio of the second number to the first number as a first ratio, wherein the first ratio and the rock mass integrity coefficient have a mapping relationship. The second ratio acquisition module is configured to detect edges of the second rock slag image by a Canny algorithm, save an image region of a closed space surrounded by the edges and having a closed space threshold greater than a set area threshold as a third rock slag image set, extract texture features of the third rock slag image of the third rock slag image set by an LBP algorithm and color features by a color moment method, obtain a rock type of the third rock slag by the rock identification model, and count percentages of various rock types in the third rock slag image set; multiply rock compressive strengths of the various rock types in the third rock slag image set by the corresponding percentages, and obtain a second ratio by accumulating the products; the second ratio has a mapping relationship with the rock compressive strength; and the rock identification model is established by a neural network model to map relationships among rock color features, rock texture features and rock types. The first tunneling parameter adjustment module is configured to, during TBM tunneling, input the first ratio, the second ratio and a root mean square amplitude of TBM girder vibration into a first tunneling parameter model when the root mean square amplitude of the TBM girder vibration exceeds a set root mean square amplitude threshold, to obtain a first compensated cutter head rotating speed and a first compensated cutter head thrust, and to obtain a first corrected cutter head rotating speed by superimposing the cutter head rotating speed and the first compensated cutter head rotating speed; the first tunneling parameter model is established by a multiple linear regression algorithm to map relationships among the first ratio, the second ratio, the set root mean square amplitude threshold and the cutter head thrust and the cutter head rotating speed; and the cutter head rotating speed obtained by the first tunneling parameter model satisfies that the root mean square amplitude of the TBM girder vibration does not exceed the set root mean square amplitude threshold, and the cutter head thrust at this time is the minimum value that satisfies the condition that the root mean square amplitude of the TBM girder vibration does not exceed the set root mean square amplitude threshold.

7. A TBM excavation control system as claimed in claim 6, wherein, The control system further comprises: The vibration frequency spectrum acquisition module is configured to filter out a direct current component of the steady-state segment acceleration data by a direct current high-pass filter and convert the steady-state segment acceleration data into frequency domain signal data by wavelet transform, and calculate a power spectral density characteristic of the TBM girder vibration according to the frequency domain signal data. The second tunneling parameter adjustment module is configured to, during TBM tunneling, input the first ratio, the second ratio and the power spectral density characteristic of the TBM girder vibration into a second tunneling parameter model when the power spectral density characteristic of the TBM girder vibration is not within a set power spectral density interval threshold, to obtain a second compensated cutter head thrust, and to obtain a second corrected cutter head thrust by superimposing the cutter head thrust and the second compensated cutter head thrust; input the first ratio, the second ratio, the root mean square amplitude of the TBM girder vibration and the second corrected cutter head thrust into the first tunneling parameter model to obtain a second compensated cutter head rotating speed, and to obtain a second corrected cutter head rotating speed by superimposing the cutter head rotating speed and the second compensated cutter head rotating speed; the second tunneling parameter model is established by a neural network to map relationships among the first ratio, the second ratio, the set power spectral density interval threshold and the cutter head thrust; and the cutter head thrust obtained by the second tunneling parameter model satisfies that the power spectral density characteristic of the TBM girder vibration is within the set power spectral density interval threshold.

8. A TBM excavation control system as claimed in claim 7, wherein, The control system further comprises: The empty pushing section vibration acquisition module is configured to acquire acceleration data before the cutter head of the TBM contacts the tunnel face, filter the direct current component of the acceleration data through a direct current high-pass filter, and convert the acceleration data into empty pushing section spectrum data through wavelet transform, calculate the power spectrum density characteristic of the TBM main beam vibration in the tunneling empty pushing section as an empty pushing section power spectrum characteristic according to the empty pushing section spectrum data, detect a peak value with a power spectrum density greater than a set threshold value through a peak value detection algorithm according to the empty pushing section power spectrum characteristic, set a narrow frequency band with the peak value as a center, and set a frequency interval of the narrow frequency band as a band-stop filter parameter. The steady state section vibration filtering module is configured to filter the direct current component of the steady state section acceleration data through a direct current high-pass filter and filter the vibration frequency component of the empty pushing section of the steady state section acceleration data through a band-stop filter to obtain filtered acceleration data before wavelet transform of the steady state section acceleration data, convert the filtered acceleration data into frequency domain signal data through wavelet transform, and calculate the power spectrum density characteristic of the TBM main beam vibration according to the frequency domain signal data.

9. A TBM excavation control system as claimed in claim 8, wherein, The control system further comprises: The first parameter model adjustment module is configured to regenerate the first tunneling parameter model when the difference between the root mean square amplitude of the TBM main beam vibration and the set root mean square amplitude threshold value is greater than the set first difference threshold value after the TBM tunneling parameter runs at the second corrected cutter head rotating speed more than a set number of times.

10. A TBM excavation control system as claimed in claim 8, wherein, The control system further comprises: The second parameter model adjustment module is configured to regenerate the second tunneling parameter model when the difference between the power spectrum density characteristic of the TBM main beam vibration and the set power spectrum density interval threshold value is greater than the set second difference threshold value after the TBM tunneling parameter runs at the second corrected cutter head thrust more than a set number of times.

Images