Shield spiral machine blade state detection method, device and system

Abstract

The application discloses a shield screw machine blade state detection method, device and system. The application relates to the technical field of shield screw machine blade state detection and aims to solve the problem that the existing shield screw machine blade state detection scheme is difficult to realize high-precision and non-stop real-time monitoring in an engineering site due to harsh sensor installation conditions. Two eddy current sensors are asymmetrically installed on a cylinder body to collect dynamic distance signals of blade rotation; signals at each time point are subjected to Gaussian fitting denoising; a linear equation with a blade wear amount and a central shaft offset amount as unknown quantities is constructed; an over-determined equation group is established based on equation groups at multiple sampling moments; the equation group is solved by using a linear least square method to obtain the wear amount and the offset amount. The application only needs two asymmetrically installed eddy current sensors, reduces installation difficulty and opening risk, realizes non-stop real-time and non-contact detection, and the system has data redundancy and strong robustness.

Description

Technical Field

[0001] This invention relates to the field of shield tunneling auger machine blade condition detection technology, and in particular to a method, device and system for detecting shield tunneling auger machine blade condition. Background Technology

[0002] In current transportation and water conservancy tunnel construction, tunnel boring machines (TBMs) are commonly used for excavation, a common method for underground tunneling. During TBM construction, a screw conveyor is used to transport excavated soil from the soil chamber. During the excavation process, the screw conveyor's blades experience constant friction and collisions with the excavated soil and hard objects within it. This wear and tear, over time, leads to a larger gap between the blades and the tunnel body, significantly increasing the risk of blade misalignment and potentially causing soil stagnation or even overall TBM failure. Therefore, accurately monitoring the blade condition during tunneling is crucial for improving construction efficiency and preventing equipment malfunctions.

[0003] Currently, blade wear monitoring methods mainly involve contact and non-contact measurement methods. For example, CN117722925A describes a screw conveyor blade wear detection device and a tunnel boring machine, which uses the length of the detection rod inserted into the conveying pipe to determine the blade wear. CN223279922U describes a screw conveyor blade wear detection device that uses the displacement generated by a telescopic drive to push the detection plate to measure the wear of the screw shaft blades. CN120948602A describes a non-destructive screw blade wear real-time monitoring system and detection method, which uses a magnetic sensor (sensing unit) installed on the cylinder to convert changes in magnetic induction intensity into a voltage signal output, which is then processed by an algorithm to detect the blade wear. CN105234820B describes a single-sensor eddy current sensor radially mounted on a grinding wheel to measure the change in distance between the sensor and the grinding wheel surface in real time. Based on this change, the grinding wheel radius change is calculated, thus obtaining the roundness error curve and wear amount for one revolution. By comparing the radius changes at different positions before and after grinding, the wear amount at each position on the grinding wheel surface is obtained. Of the four solutions mentioned above, CN117722925A, CN223279922U, and CN120948602A can detect blade wear without large-scale disassembly. However, CN117722925A and CN223279922U require system shutdown for testing, making fully online real-time detection impossible. Furthermore, these two solutions rely on mechanical contact, and severe blade deformation can lead to inaccurate contact, affecting wear calculations. CN120948602A requires sensors to be evenly distributed at equal angles, with at least three sensors required. In practical engineering, sensor installation is often limited, making evenly distributed installation impossible, leading to decreased fitting accuracy. Additionally, drilling holes in the cylinder structure is very cumbersome, and more holes increase the risk. Although CN105234820B is not a method for detecting the wear of screw conveyor blades, it can still be used to detect the wear of blades. Although this scheme can measure the wear using a single sensor, it cannot detect and quantify the magnitude and direction of the offset of the central axis. Therefore, when the blade central axis is offset due to bearing wear or other reasons, the distance change measured by this single sensor is a mixed effect of wear and offset, which cannot be distinguished, resulting in a decrease in accuracy. In addition, if the single sensor fails or is subject to local interference, the entire system will fail or produce significant errors. Summary of the Invention

[0004] The purpose of this invention is to provide a method, device, and system for detecting the condition of shield tunneling auger blades, in order to solve the problem that existing shield tunneling auger blade condition detection schemes are difficult to achieve high-precision, non-stop real-time monitoring on the engineering site due to the harsh sensor installation conditions.

[0005] In a first aspect, the present invention provides a method for detecting the condition of shield tunneling auger machine blades, comprising the following steps:

[0006] Step 1: Acquire the dynamic distance signal of blade rotation collected by two eddy current sensors at multiple sampling time points. The two eddy current sensors are asymmetrically fixed on the shield tunneling auger cylinder.

[0007] Step 2: Perform Gaussian fitting to denoise the multiple blade rotation dynamic distance signals at each sampling time point to obtain the denoised distance value of each eddy current sensor at each sampling time point;

[0008] Step 3: Based on the blade rotational angular velocity, eddy current sensor installation position parameters, original blade radial dimension, and the denoised distance value obtained in Step 2, construct a linear equation with the two-dimensional offset of the blade central axis and the blade radial wear as unknown parameters.

[0009] Step 4: Construct an overdetermined system of equations based on the linear equations corresponding to multiple sampling time points;

[0010] Step 5: Solve the overdetermined equations using the linear least squares method to calculate the two-dimensional offset of the blade central axis and the radial wear of the blade.

[0011] Furthermore, in step one, two eddy current sensors are fixedly installed on the shield tunneling auger cylinder in a non-equidistant angular distribution manner. Each eddy current sensor collects multiple blade rotation dynamic distance signals at each sampling time point. The sampling time points are set at equal intervals according to a preset sampling period, and the sampling period is determined according to the blade rotation angular velocity.

[0012] Further, step two includes: for each sampling time point, acquiring multiple sampled values ​​collected by each eddy current sensor at that sampling time point; based on the assumption that the multiple sampled values ​​follow a Gaussian distribution, calculating the mean of the multiple sampled values ​​as the true distance estimate for that sampling time point, and calculating the variance of the multiple sampled values; based on the calculated variance, using the 3σ criterion to remove outliers from the multiple sampled values; based on the sampled values ​​after removing outliers, recalculating the mean to obtain the denoised distance value.

[0013] Further, step three includes: establishing a fixed planar rectangular coordinate system with the central axis position when there is no shaft offset as the origin; calculating the rotation angle at each sampling time point based on the blade rotational angular velocity, initial rotation angle, and sampling time point; establishing a distance equation from the eddy current sensor to the detection point based on the installation angle of the eddy current sensor, the distance from the eddy current sensor to the original center, the original radial dimension of the blade, the rotation angle, and the distance value after noise reduction; and linearizing the distance equation to obtain a linear equation for the two-dimensional offset of the blade central axis and the radial wear of the blade.

[0014] Furthermore, in step four, for each sampling time point, two linear equations are constructed based on the denoised distance values ​​of the two eddy current sensors to form the equation set corresponding to that sampling time point; the equation sets corresponding to all sampling time points are combined to form an overdetermined equation set containing multiple linear equations. The number of linear equations is greater than the number of unknown parameters in the overdetermined equation set. The unknown parameters include the two components of the two-dimensional offset of the blade center axis and the radial wear of the blade.

[0015] Further, step five includes: constructing the coefficient matrix and constant term vector of the overdetermined system of equations; calculating the product of the transpose of the coefficient matrix and the coefficient matrix, and the product of the transpose of the coefficient matrix and the constant term vector; solving the normal equations to obtain the optimal estimate vector containing the offset of the blade central axis in the x-direction, the offset of the blade central axis in the y-direction, and the radial wear of the blade; extracting the offset of the blade central axis in the x-direction and y-direction from the optimal estimate vector, and calculating the magnitude and direction of the blade central axis offset; and extracting the radial wear of the blade from the optimal estimate vector.

[0016] Furthermore, step five also includes: calculating the total offset magnitude and offset direction of the blade center axis based on the calculated two-dimensional offset of the blade center axis.

[0017] Furthermore, step five also includes: when the radial wear of the blade is less than zero, it is determined that the blade is scaled or deformed, and the calculation results are corrected according to the actual engineering situation.

[0018] Secondly, the present invention provides a shield tunneling auger machine blade condition detection device, comprising:

[0019] The acquisition unit is used to acquire the dynamic distance signal of blade rotation collected by two eddy current sensors at multiple sampling time points. The two eddy current sensors are asymmetrically fixed on the shield tunneling auger cylinder.

[0020] The denoising unit is used to perform Gaussian fitting denoising on the dynamic distance signals of multiple blade rotations at each sampling time point to obtain the denoised distance value of each eddy current sensor at each sampling time point.

[0021] The construction unit is used to construct a linear equation with the two-dimensional offset of the blade central axis and the radial wear of the blade as unknown parameters, based on the blade rotational angular velocity, the installation position parameters of the eddy current sensor, the original radial dimension of the blade, and the denoised distance value obtained in step two.

[0022] The assembly unit is used to assemble an overdetermined system of equations based on the linear equations corresponding to multiple sampling time points.

[0023] The solution unit is used to solve the overdetermined equations using the linear least squares method to calculate the two-dimensional offset of the blade central axis and the radial wear of the blade.

[0024] Thirdly, the present invention provides a shield tunneling auger blade condition detection system, comprising: two eddy current sensors asymmetrically fixedly installed on the shield tunneling auger cylinder and a processor electrically connected to the eddy current sensors;

[0025] The eddy current sensor is used to collect dynamic distance signals of blade rotation at multiple sampling time points;

[0026] The processor is used to acquire blade rotation dynamic distance signals collected by two eddy current sensors at multiple sampling time points; to perform Gaussian fitting denoising on the multiple blade rotation dynamic distance signals at each sampling time point to obtain the denoised distance value of each eddy current sensor at each sampling time point; to construct a linear equation with the two-dimensional offset of the blade central axis and the radial wear of the blade as unknown parameters based on the blade rotation angular velocity, the installation position parameters of the eddy current sensors, the original radial dimension of the blade, and the denoised distance value obtained in step two; to construct an overdetermined equation system based on the linear equations corresponding to multiple sampling time points; and to solve the overdetermined equation system using the linear least squares method to calculate the two-dimensional offset of the blade central axis and the radial wear of the blade.

[0027] The beneficial effects of this invention are as follows: By using two asymmetrically installed eddy current sensors, this invention solves the engineering problem of not being able to achieve uniform installation of multiple sensors at equal angles due to limitations in sensor installation conditions. It also reduces the number of sensors required from three or more in traditional solutions to two, thereby significantly reducing hardware costs, construction drilling risks, and system complexity. By combining angular velocity information, Gaussian fitting denoising algorithms, and linear least squares methods, this solution can simultaneously calculate the radial wear of the auger blades and the two-dimensional offset of the central axis in real time and non-contactly without stopping the tunnel boring machine. This effectively overcomes the shortcomings of existing contact-based measurements requiring machine shutdown and existing non-contact single-sensor solutions that cannot distinguish between the mixed effects of wear and offset and have weak anti-interference capabilities. The dual eddy current sensor design also provides data redundancy, enhancing the system's robustness. Even if a single eddy current sensor fails, the system can still degrade its operation or issue an early warning, thus ensuring the continuity of construction and equipment safety. This achieves more accurate, reliable, and adaptable real-time monitoring of the auger blade status in complex on-site environments. Attached Figure Description

[0028] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0029] Figure 1 This is a structural diagram of an eddy current sensor on the tube wall of a tunnel boring machine used to detect blade wear. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention. The technical solutions provided by various embodiments of this invention will be described in detail below with reference to the accompanying drawings.

[0031] To address the limitations of sensor installation conditions that prevent uniform angular distribution, and the challenge of using two sensors instead of three, this invention proposes an asymmetric eddy current sensor arrangement. This arrangement incorporates a distance-motion model constructed from angular velocity information, along with Gaussian fitting for noise reduction and a real-time detection algorithm using linear least squares. By asymmetrically installing two eddy current sensors, dynamic distance signals during blade rotation are acquired. Noise is removed using Gaussian fitting for greater accuracy, and the distance-motion model is constructed using angular velocity information. The nonlinear distance equations are then linearized, and the overdetermined equations are solved using least squares to calculate blade wear and two-dimensional offset of the central axis in real time. This approach offers high flexibility in eddy current sensor installation, eliminating the need for uniform angular distribution. Furthermore, compared to traditional three-sensor systems, two sensors significantly reduce costs and the risks associated with drilling, making it more adaptable to engineering applications.

[0032] Please see Figure 1 The shield tunneling auger machine blade condition detection method provided in this embodiment of the invention specifically includes the following steps: First, two eddy current sensors are asymmetrically fixedly installed on the shield tunneling auger machine cylinder 100 for detecting the wear of the blades 200. The two eddy current sensors are denoted as S1 and S2, respectively. Then, the algorithm and derivation for calculating the wear amount Δw and the offsets x0 and y0 are performed.

[0033] Coordinate system definition: When the screw conveyor is operating normally and there is no shaft offset, the central axis O is the origin, and a fixed planar rectangular coordinate system O-xy (which does not change with blade rotation) is established, with the horizontal direction set as the x-axis. This coordinate system serves as the unified reference for all position parameters. The core parameters for subsequent algorithm formula derivation are shown in Table 1.

[0034] Table 1 Core parameters for algorithm formula derivation

[0035]

[0036] The position of the eddy current sensor based on the O-xy plane coordinate system: Two eddy current sensors are fixedly installed on the cylinder, and their positions remain unchanged. It is known that the installation angles of the two eddy current sensors S1 and S2 are α1 and α2, and the distance of the eddy current sensors from the origin O(0,0) is r0. Therefore, the polar coordinates of the positions of the eddy current sensors can be obtained as S1(r0, α1) and S2(r0, α2). From the conversion formula between the polar coordinate system and the rectangular coordinate system (1), the expression (2) of the position of the eddy current sensors S1 and S2 in the rectangular coordinate system is obtained:

[0037] (1)

[0038] (2)

[0039] The location of the eddy current sensor detection point: Assuming an eccentricity occurs, the blade rotates around the offset center axis O'(x0,y0). The eddy current sensor can detect a certain point on the blade. The position of this point changes with the angular velocity ω, and the rotation angle is determined by the angular velocity ω: θ(t)=ωt+θ0; After wear, the rotation radius is R-Δw. The location data of this point based on the coordinate system O-xy plane is established as A[X(t),Y(t)], which is obtained from formula (3):

[0040] (3)

[0041] Discrete sampling (t) k = k·T s The distance between the eddy current sensor and the detection point changes from far to near and then back to far as the blade rotates, forming a continuous periodic analog signal that varies with the angular velocity ω in the time domain. This signal is then processed by the eddy current sensor's data processing circuit (AD conversion, microcontroller acquisition and transmission) to obtain the data. However, in actual engineering, the rotation time of one revolution of the propeller blade is quite long, resulting in a large amount of data, making computer processing complex and causing data duplication. Therefore, this invention samples this data again. Thus, a sampling period T can be set. s =1 / f s f s The sampling frequency is t, and the sampling time point is t. k =k·T s k=1,2,…,N, where N is the total number of sampling times. To minimize the impact of noise, M data points are sampled at each sampling time point for easier Gaussian fitting and noise reduction later. Therefore, the data obtained at each sampling time point can be organized into a two-dimensional sequence: d1(t k ,n) (the nth sampling at the kth time point of the eddy current sensor S1), d2(t k(,n) (the nth sampling at the k-th time point of the eddy current sensor S2). Therefore, the sampling time point t can be... k Substituting into formula (3), we can obtain the position of the detection point at each sampling time point as A[X(t)]. k ),Y(t k From formula (4), we can derive:

[0042] (4)

[0043] In the formula Let be the blade rotation angle at the k-th sampling time point.

[0044] Gaussian fitting for denoising: For noisy original sampled data, based on the characteristic that noise follows a Gaussian distribution, Gaussian fitting is used to extract the true distance signal. A single time step t can be set... k M sampled values ​​d i (t k The distribution N(μ) follows a Gaussian distribution. i (t k ),σ²), where μ i (t k σ² represents the true distance (mean), and σ² represents the noise variance. The mean and variance can be solved using maximum likelihood estimation; therefore, μ... i (t k ) and σ² are derived from formula (5).

[0045] (5)

[0046] The obtained data also contained anomalies. To remove these anomalies, this patent application is based on the 3σ criterion (covering 99.73% of normal data) and removes data exceeding […]. i (t k )-3 , i (t k )+3 The outliers can be identified by refitting the fitted data; then, by fitting the fitted data again, t can be obtained. k Real distance after real-time noise reduction i (t k (i=1,2, corresponding to S1, S2).

[0047] Establish the calculation equation: Given the eddy current sensor S i The distance from the detection point is i (t k The coordinates of the detection point are A[X(t)]. k ),Y(tk The coordinates of the eddy current sensor are S. i [x si ,y si According to the formula for the distance between two points in a plane (6), the eddy current sensor S is obtained. i to t k Real distance of dynamic detection points at all times i (t k To satisfy formula (7), for ease of calculation, formula (7) is squared to obtain formula (8):

[0048] (6)

[0049] (7)

[0050] (8)

[0051] Substituting formula (4) into the above equation, we obtain the squared distance equation containing unknown parameters x0, y0, and Δw:

[0052] (9)

[0053] Expanding formula (9) completely yields formula (10):

[0054]

[0055] (10)

[0056] Using the trigonometric identity cos² k +sin² k =1 simplifies formula (10) to formula (11).

[0057]

[0058] (11)

[0059] Will Formula (11) leads to Formula (12)

[0060]

[0061] (12)

[0062] The left side of the equation is the known squared distance value after denoising, and the right side contains three parameters to be solved. Since the power of the unknown is at most 2, this is a nonlinear equation. Solving the nonlinear equation will make the subsequent calculations particularly complicated. In order to facilitate the calculation, the nonlinear equation is transformed into a linear equation through linearization.

[0063] Linearization of nonlinear equations: For the quadratic unknowns Δw², x0², and y0² contained in formula (12), and considering that the risk of equipment failure increases significantly when the wear of the blades Δw ≥ 10 mm in actual engineering, we can set Δw ≤ 5 mm, R ≥ 300 mm, x0 and y0 ≤ 2 mm, and obtain (Δw / R ≤ 1.67%) >> (Δw² / R 2 Since ≤ 0.028% and (x0², y0² ≤ 4mm²) << (R² ≥ 90000mm²), Δw², x0², and y0² can be ignored.

[0064] Therefore, formula (12) can be transformed into formula (13):

[0065] (13)

[0066] Moving all the known terms in equation (13) to the left, we can rearrange it into the final linear equation for x0, y0, and Δw (equation (14)):

[0067] (14)

[0068] The coefficients are all known quantities (including ω, ...). k (e.g., parameters), the specific expressions are shown in Table 2.

[0069] Table 2. Coefficients and their expressions

[0070]

[0071] Constructing an overdetermined system of equations: Given N sampling times, each sampling time point corresponds to two linear equations (one for S1 and one for S2), resulting in a total of 2N linear equations. Since the number of equations 2N is much greater than the number of parameters to be determined (3), the constructed system of equations is an overdetermined system of equations A· =b, where:

[0072] Coefficient matrix A (2N×3): Each row corresponds to the coefficients of one linear equation, arranged in order of sampling time point and eddy current sensor;

[0073] Unknown parameter vector (3×1): The offset and wear amount to be determined;

[0074] The constant term vector b (2N×1): Each row corresponds to one constant term E of a linear equation. i (t k ).

[0075] The following is the specific form of the matrix (Formula (15)).

[0076] (15)

[0077] Least squares method solution: For the overdetermined system of equations A in formula (15) = b has no exact solution, so we can assume the residual vector is r = A - b, therefore the least squares method minimizes the sum of squared residuals J( ) = (Formula (16)), for the residual sum of squares J( Find the partial derivative and set it to 0 (Equation (17)). The normal equation can be derived (Equation (18)). Solving the normal equation yields the optimal estimate. , From formula (19):

[0078] (16)

[0079] (17)

[0080] (18)

[0081] (19)

[0082] In the above formulas: A T Let A be the transpose of the coefficient matrix A. T A is a 3×3 square matrix (reversible, with a rank of 3 guaranteed due to the asymmetric installation of the eddy current sensor).

[0083] Find the offset (x0, y0), wear Δw: from the optimal solution vector Parameters are extracted to obtain the final detection result.

[0084] Two-dimensional offset of the central axis: x0= (1,1), y0= (2,1); Further calculate the offset size. and direction (Required for engineering applications):

[0085] ;

[0086] Blade radial wear: Δw= (3,1) (Δw<0 indicates scaling / deformation, which needs to be corrected based on the actual engineering situation).

[0087] The embodiments of the present invention described above do not constitute a limitation on the scope of protection of the present invention.

Claims

1. A method for detecting the condition of shield tunneling auger blades, characterized in that, Includes the following steps: Step 1: Acquire the dynamic distance signal of blade rotation collected by two eddy current sensors at multiple sampling time points. The two eddy current sensors are asymmetrically fixed on the shield tunneling auger cylinder. Step 2: Perform Gaussian fitting to denoise the multiple blade rotation dynamic distance signals at each sampling time point to obtain the denoised distance value of each eddy current sensor at each sampling time point; Step 3: Based on the blade rotational angular velocity, eddy current sensor installation position parameters, original blade radial dimensions, and the denoised distance value obtained in Step 2, construct a linear equation with the two-dimensional offset of the blade central axis and the blade radial wear as unknown parameters. A fixed Cartesian coordinate system is established, with the central axis position without shaft offset as the origin. The rotation angle at each sampling time point is calculated based on the blade rotational angular velocity, initial rotation angle, and sampling time point. A distance equation from the eddy current sensor to the detection point is established based on the eddy current sensor installation angle, the distance from the eddy current sensor to the original center, the original blade radial dimensions, the rotation angle, and the denoised distance value. The distance equation is then linearized to obtain a linear equation concerning the two-dimensional offset of the blade central axis and the blade radial wear. Step 4: Construct an overdetermined system of equations based on the linear equations corresponding to multiple sampling time points; Step 5: Solve the overdetermined equations using the linear least squares method to calculate the two-dimensional offset of the blade central axis and the radial wear of the blade.

2. The method for detecting the condition of shield tunneling auger machine blades according to claim 1, characterized in that, In step one, two eddy current sensors are fixedly installed on the shield tunneling auger cylinder in a non-equidistant angular distribution manner. Each eddy current sensor collects multiple blade rotation dynamic distance signals at each sampling time point. The sampling time points are set at equal intervals according to a preset sampling period, and the sampling period is determined according to the blade rotation angular velocity.

3. The method for detecting the condition of shield tunneling auger machine blades according to claim 1, characterized in that, Step two includes: For each sampling time point, acquire multiple sampled values ​​collected by each eddy current sensor at that sampling time point; Based on the assumption that the multiple sampled values ​​follow a Gaussian distribution, the mean of the multiple sampled values ​​is calculated as the true distance estimate at the sampling time point, and the variance of the multiple sampled values ​​is calculated. Based on the calculated variance, outliers in the multiple sampled values ​​are removed using the 3σ criterion. Based on the sampled values ​​after removing outliers, the mean is recalculated to obtain the denoised distance value.

4. The method for detecting the condition of shield tunneling auger machine blades according to claim 1, characterized in that, In step four, for each sampling time point, two linear equations are constructed based on the denoised distance values ​​of the two eddy current sensors to form the equation set corresponding to that sampling time point. The equations corresponding to all sampling time points are combined to form an overdetermined equation set containing multiple linear equations. The number of linear equations is greater than the number of unknown parameters in the overdetermined equation set. The unknown parameters include two components of the two-dimensional offset of the blade central axis and the radial wear of the blade.

5. The method for detecting the condition of shield tunneling auger machine blades according to claim 1, characterized in that, Step five includes: Construct the coefficient matrix and constant term vector of the overdetermined system of equations; Calculate the product of the transpose of the coefficient matrix and the coefficient matrix itself, as well as the product of the transpose of the coefficient matrix and the constant term vector. Solving the normal equations yields the optimal estimate vector containing the offset of the blade's central axis in the x-direction, the offset of the blade's central axis in the y-direction, and the radial wear of the blade. Extract the offset of the blade center axis in the x and y directions from the optimal estimate vector, and calculate the magnitude and direction of the blade center axis offset; Extract the radial wear of the blade from the optimal estimate vector.

6. The method for detecting the condition of shield tunneling auger machine blades according to claim 5, characterized in that, Step five further includes: calculating the total offset magnitude and offset direction of the blade center axis based on the calculated two-dimensional offset of the blade center axis.

7. The method for detecting the condition of shield tunneling auger machine blades according to claim 5, characterized in that, Step five also includes: when the radial wear of the blade is less than zero, it is determined that the blade is scaled or deformed, and the calculation results are corrected according to the actual engineering situation.

8. A shield tunneling auger blade condition detection device, characterized in that, include: The acquisition unit is used to acquire the dynamic distance signal of blade rotation collected by two eddy current sensors at multiple sampling time points. The two eddy current sensors are asymmetrically fixed on the shield tunneling auger cylinder. The denoising unit is used to perform Gaussian fitting denoising on the dynamic distance signals of multiple blade rotations at each sampling time point to obtain the denoised distance value of each eddy current sensor at each sampling time point. A construction unit is used to construct a linear equation with the two-dimensional offset of the blade's central axis and the radial wear of the blade as unknown parameters, based on the blade's rotational angular velocity, the installation position parameters of the eddy current sensor, the original radial dimension of the blade, and the distance value after noise reduction. Specifically, a fixed Cartesian coordinate system is established, with the central axis position without shaft offset as the origin. The rotation angle at each sampling time point is calculated based on the blade's rotational angular velocity, the initial rotation angle, and the sampling time point. A distance equation from the eddy current sensor to the detection point is established based on the installation angle of the eddy current sensor, the distance from the eddy current sensor to the original center, the original radial dimension of the blade, the rotation angle, and the distance value after noise reduction. The distance equation is then linearized to obtain a linear equation concerning the two-dimensional offset of the blade's central axis and the radial wear of the blade. The assembly unit is used to assemble an overdetermined system of equations based on the linear equations corresponding to multiple sampling time points. The solution unit is used to solve the overdetermined equations using the linear least squares method to calculate the two-dimensional offset of the blade central axis and the radial wear of the blade.

9. A shield tunneling auger machine blade condition detection system, characterized in that, include: Two eddy current sensors are asymmetrically fixedly installed on the shield tunneling machine cylinder, and a processor is electrically connected to the eddy current sensors; The eddy current sensor is used to collect dynamic distance signals of blade rotation at multiple sampling time points; The processor is used to acquire the blade rotation dynamic distance signals collected by two eddy current sensors at multiple sampling time points; and to perform Gaussian fitting and denoising on the multiple blade rotation dynamic distance signals at each sampling time point to obtain the denoised distance value of each eddy current sensor at each sampling time point. Construct a linear equation with the two-dimensional offset of the blade central axis and the radial wear of the blade as unknown parameters based on the blade rotation angular velocity, the installation position parameters of the eddy current sensor, the original radial dimension of the blade, and the denoised distance value. Among them, establish a fixed plane rectangular coordinate system with the position of the central axis without shaft offset as the coordinate origin. Calculate the rotation angle at each sampling time point according to the blade rotation angular velocity, the initial rotation angle, and the sampling time point. Establish the distance equation from the eddy current sensor to the detection point according to the installation angle of the eddy current sensor, the distance from the eddy current sensor to the original center, the original radial dimension of the blade, the rotation angle, and the denoised distance value. Linearize the distance equation to obtain a linear equation about the two-dimensional offset of the blade central axis and the radial wear of the blade. Based on the linear equations corresponding to multiple sampling time points, form an overdetermined system of equations. Solve the overdetermined system of equations by the linear least squares method to calculate the two-dimensional offset of the blade central axis and the radial wear of the blade.

Images