Digital-twin-based method for evaluating a tunnel structure reinforced by a fiber composite arch

Abstract

The application discloses a kind of digital twin's fiber composite arch reinforced tunnel structure evaluation method, it is related to tunnel health monitoring technical field, establishes bonding slip model and total stress model, is combined with coupling temperature variation equation, coupling humidity variation equation, constructs damage rate model;Real-time data are collected, total stress is calculated after Gaussian process interpolation processing, calculate residual life and first assess whether to replace;When no replacement is needed in first assessment, future state is simulated using time series reasoning network and finite element method and life is assessed again to decide whether to replace lattice reinforced fiber composite arch, realize the dynamic evaluation and life prediction of lattice reinforced fiber composite arch reinforced tunnel structure, improve the evaluation accuracy and forwardness.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of tunnel health monitoring, and particularly relates to a fiber composite arch reinforced tunnel structure evaluation method based on digital twinning. BACKGROUND

[0002] The subway shield tunnel long-term bears the surrounding rock pressure, train vibration, temperature and humidity change synergy, and the lining is prone to cracking, peeling, water leakage and other diseases, which needs to be protected by reinforcement technology to ensure the operation safety. In the traditional reinforcement scheme, the self-weight of the concrete reinforcement is large, the space occupied is large, and the construction risk is high. The steel ring reinforcement has high cost, is prone to corrosion, and is difficult to maintain in the later period. The durability of the pasted fiber composite sheet depends on the performance of the bonding glue, which is easy to fail in a humid environment. Moreover, none of them is designed for the tunnel arc structure, and the adhesion to the segment is poor, the stress transmission is uneven, and it is difficult to meet the requirements of "reinforcement + disease repair".

[0003] The lattice reinforced fiber composite arch is widely used in shield tunnel structure reinforcement engineering due to its light weight, high strength, corrosion resistance and other characteristics. The lattice reinforced fiber composite arch is wrapped with lattice reinforced fiber composite arch fiber cloth outside the hard filling core material, and glass fiber cloth is laid on the bottom layer and the top layer respectively. The whole member is integrated by resin bonding and curing. The lattice reinforced fiber composite arch is connected by multiple sections, and the connection is fixed by external reinforcing plates to ensure the integrity of the structure. It has the advantages of large specific stiffness, large specific strength and corrosion resistance. The lattice reinforced fiber composite arch is bonded with the inner arc surface of the tunnel segment for structural reinforcement. The lattice reinforced fiber composite arch and the segment are bonded with structural adhesive with a thickness of not less than 5 mm and fixed with chemical anchors to form a double connection guarantee, realize the cooperative stress of the member and the segment, and can simultaneously repair the segment cracks and surface peeling. It has the functions of reinforcement and disease repair. The density of the lattice reinforced fiber composite arch is much lower than that of the traditional steel material, which can greatly reduce the additional load of the tunnel structure, and is resistant to moisture and corrosion, which can adapt to the complex tunnel operation environment for a long time and reduce the need for later maintenance.

[0004] However, during the operation of the shield tunnel, the lattice reinforced fiber composite arch long-term bears the external effects of soil pressure, vehicle dynamic load, etc., and is affected by environmental temperature and humidity changes. The glue between the lattice reinforced fiber composite arch and the segment is prone to adhesive slip, which leads to the attenuation of the reinforcement effect. Therefore, it is necessary to evaluate the reinforcement effect of the lattice reinforced fiber composite arch in real time to ensure the safety of the tunnel.

[0005] Existing assessment methods for tunnels reinforced with lattice-reinforced fiber composite arches still have problems and are difficult to match the actual needs of lattice-reinforced fiber composite arch reinforcement: First, they rely on static monitoring data and do not fully consider the bond-slip effect between the lattice-reinforced fiber composite arch and the tunnel lining segments, the random characteristics of dynamic loads, and the impact of temperature and humidity coupling on the performance of the lattice-reinforced fiber composite arch, resulting in a large deviation between the stress analysis and the actual reinforcement situation of the lattice-reinforced fiber composite arch; Second, they do not combine the collaborative stress mechanism of the lattice-reinforced fiber composite arch, nor do they incorporate the strength evolution law of the lattice-reinforced fiber composite arch, making it impossible to accurately quantify the damage state of the lattice-reinforced fiber composite arch reinforcement system; Third, they lack the ability to predict the future state of the lattice-reinforced fiber composite arch, making it difficult to provide early warning of the performance degradation trend of the reinforcement system, and making it impossible to scientifically formulate replacement decisions for the lattice-reinforced fiber composite arch, which may easily lead to hidden dangers in tunnel operation safety. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention proposes a digital twin-based evaluation method for fiber-reinforced composite arch-based tunnel structures. By establishing bond slip, total stress, coupled temperature and humidity changes, and damage rate models, and combining Gaussian process interpolation, temporal inference networks, and the finite element method, real-time data is collected and future states are simulated. The remaining lifespan of the reinforcement system based on the lattice-reinforced fiber-reinforced composite arch is dynamically calculated to determine whether to replace the lattice-reinforced fiber-reinforced composite arch.

[0007] The technical solution adopted to achieve the purpose of this invention is as follows:

[0008] A digital twin-based method for evaluating fiber composite arch-reinforced tunnel structures includes the following steps:

[0009] A bond-slip model is established by combining the arch dimensions and bilinear constitutive relations. A total stress model is constructed considering earth pressure, dynamic load effects, and the influence of bond-slip. This model is based on the laws of heat conduction, diffusion, and temperature. ,humidity The coupling relationship is used to establish coupled temperature change equations and coupled humidity change equations, and the Hachson failure criterion is combined with temperature. ,humidity The influence of the overall stress model on the construction of the damage rate model;

[0010] Collection of the first Earth pressure at any moment ,strain sparse glide vector Temperature matrix Humidity matrix Gaussian interpolation is used to interpolate sparse slip vectors. Expanded into dense slip vector And the slip function is obtained by fitting. The calculation of the first [unclear] based on the bond-slip model and the total force model Total force at time ;

[0011] Calculate the first Average temperature at time and average humidity and the Total force at time Substitute into the damage rate model, combined with the first Remaining lifespan at any given moment Calculate the first Remaining lifespan at any given moment and the lower limit of lifespan Comparison of decisions regarding replacing lattice-reinforced fiber composite arches or the first Total force at time Input temporal reasoning network generates the first Simulated total force at time The coupled temperature variation equation and coupled humidity variation equation are spatially discretized and approximated by time derivatives using the finite element method and the implicit Euler backward method. Based on the first... Temperature matrix at time Humidity matrix The reasoning leads to the first Simulated temperature matrix at time 1 and simulated humidity matrix ;

[0012] Calculate the first Simulated average temperature at time and simulated average humidity and the Simulated total force at time Substitute into the damage rate model, combined with the first Remaining lifespan at any given moment Calculate the first Simulated remaining lifespan at any time and the lower limit of lifespan The comparison will be repeated to decide whether to replace the lattice-reinforced fiber composite arch or wait for the next... The time has come.

[0013] Specifically, the bond-slip model is based on the first Time, major axis coordinate slip function Calculation of the relationship between lattice-reinforced fiber composite arches and segments in the first stage Time, major axis coordinate Adhesion force Adhesion With the slip function Satisfying the bilinear constitutive relation, when the glide function Less than the critical slip At that time, adhesive force Equal to the slip function With bond stiffness The product of the slip function Greater than or equal to the critical slip amount At that time, adhesive force Equal to maximum adhesive force , to bond strength Integrate along the major axis of the lattice-reinforced fiber composite arch and multiply by the minor axis length , obtained the Adhesive force at all times .

[0014] Furthermore, the overall force model is determined based on force analysis, the first... Total force at time With component area The product equals the first Earth pressure at any moment Add dynamic load force Subtract adhesive force , among which, the Dynamic load at any moment static load The neighborhood follows a Gaussian distribution, and according to Hooke's law, the th static load at time equal to the Response at all times With elastic modulus Component area The product of.

[0015] Specifically, based on the law of thermal conductivity, temperature Over time rate of change With air density Specific heat capacity of air The product of and equals the thermal conductivity. With temperature Hamiltonian operator The product of temperature Hamiltonian operator equal to temperature Regarding the second-order gradient of spatial coordinates, based on the latent heat of water phase change. By superimposing the coupled correlation caused by humidity changes, we obtain the coupled temperature change equation, i.e., temperature. Over time rate of change With air density Specific heat capacity of air The product of and equals the thermal conductivity. With temperature Hamiltonian operator The product of water phase change and latent heat With humidity Over time rate of change The coupled temperature change equation is obtained.

[0016] Specifically, temperature is introduced into the diffusion law. Humidity Basic diffusion coefficient The influence of air density is used to construct a coupled wet-variance equation, namely air density. With humidity Over time rate of change The product of and equals humidity Coupling diffusion coefficient With humidity Hamiltonian operator The product of, where humidity Coupling diffusion coefficient Equal to the basic diffusion coefficient Compared with the natural constant as the base and empirical coefficients The opposite number of temperature The ratio is the product of powers of the exponents.

[0017] Furthermore, based on the hash factor failure criterion, the first Failure index at time equal to the Total force at time With the Average temperature at time and average humidity The ultimate total force The square of the ratio minus 1, where the first... Average temperature at time and average humidity The ultimate total force Compared with standard temperature and standard humidity The ultimate total force Proportional, the proportion equals 1 minus the missing proportion, the missing proportion equals the first... Average temperature at time Compared with standard temperature The difference multiplied by the total stress temperature compensation coefficient Gadi Average humidity at any time Compared with standard humidity The difference multiplied by the total force wet compensation coefficient When the first Failure index at time When greater than or equal to 0, the first Damage rate at time Equal to damage evolution coefficient , No. Failure index at time With total force The product of.

[0018] Furthermore, Gaussian process interpolation includes the following steps:

[0019] The physical law showing a positive correlation between the axial coordinate spacing and the slip amount was characterized using the Matern kernel;

[0020] Calculate the first Sparse glide vector at time The Middle Major axis coordinates of a displacement sensor With the Major axis coordinates of a displacement sensor The Martern core between and superimposed the first Standard deviation of acquisition noise of each displacement sensor and The product of the Kronecker function values ​​yields the first Major axis coordinates of a displacement sensor With the Major axis coordinates of a displacement sensor sparse covariance between , construct the first Sparse covariance matrix at time step The Kronecker function takes the value 1 when the two input values ​​are the same, and takes the value 0 when they are different.

[0021] Determine the dense slip vector The interpolated major axis coordinate vector Calculate the interpolated major axis coordinate vector The Middle Interpolation major axis coordinates of each interpolation point With the Sparse glide vector at time The Middle Major axis coordinates of a displacement sensor The Martern core between And as the first Interpolation major axis coordinates of each interpolation point With the Major axis coordinates of a displacement sensor Interactive covariance , construct the first Interaction covariance matrix at time step ;

[0022] Calculate the first Interpolated major axis coordinate vector at time t The Middle Interpolation major axis coordinates of each interpolation point With the Interpolation major axis coordinates of each interpolation point The Martern core between and superimposed the first Standard deviation of interpolation noise at each interpolation point and The product of the Kronecker function values ​​yields the first Interpolation major axis coordinates of each interpolation point With the Interpolation major axis coordinates of each interpolation point Dense covariance between , construct the first Dense covariance matrix at time step ;

[0023] The first Interaction covariance matrix at time step transpose, the first Sparse covariance matrix at time step With the Sparse glide vector at time Multiply to get the first The vector of the distribution mean at time t. , will the Dense covariance matrix at time step Subtract the first Interaction covariance matrix at time step transpose, the first Sparse covariance matrix at time step The inverse matrix and the first Interaction covariance matrix at time step The product of , we get the first vector of standard deviation of distribution at time Construct a Gaussian distribution as the first Dense slip vector at time The interpolated posterior distribution;

[0024] The nuclear parameters of the Matrn nucleus, The standard deviation of the acquisition noise of each displacement sensor and The standard deviation of interpolation noise at each interpolation point is used as a variable. To maximize the marginal log-likelihood function To achieve the objective, an optimization algorithm is used for iterative optimization to obtain the optimal variables. Recalculate the first The optimal distribution mean vector at time 1 And as the first Dense slip vector at time Among them, the marginal log-likelihood function equal to one-negative half, the first Sparse glide vector at time transpose, the first Sparse covariance matrix at time step The inverse matrix and the first Sparse glide vector at time The product of the product plus one-negative half and the first Sparse covariance matrix at time step The product of the natural logarithm of the absolute value plus one-negative half, and the total number of displacement sensors. and The product of the natural logarithms.

[0025] Specifically, time 1 is defined as the initial use time and the remaining lifetime at time 1. Lower limit of lifespan , No. Remaining lifespan at any given moment equal to the Remaining lifespan at any given moment Subtract the first Damage rate at time Acquisition interval The product of , if the first Remaining lifespan at any given moment Less than or equal to the lower limit of lifespan Replace the lattice-reinforced fiber composite arch, if the first Remaining lifespan at any given moment Greater than the lower limit of lifespan Then the simulation calculates the first Simulated total force at time Simulated average temperature and simulated average humidity Substitute into the damage rate model to determine the first Simulated damage rate at time step And calculate the first Simulated remaining lifespan at any time If the first Simulated remaining lifespan at any time Greater than the lower limit of lifespan Waiting for the first When the time arrives, if the first Simulated remaining lifespan at any time Less than or equal to the lower limit of lifespan Replace the lattice-reinforced fiber composite arch.

[0026] Specifically, call the first Before the moment The historical forces at each moment construct the first Pre-time total force sequence And input into the temporal inference network, respectively through Different base learners capture the first Pre-time total force sequence The time-series change trend, generating the first Latent temporal representation sequence at time point And input into the meta-learner, using a linear layer to... Latent temporal representation sequence at time point middle The first output of the seed base learner The latent temporal representations at time t are mapped to 1-dimensional values ​​and weighted summed using LR linear regression to generate the th... Simulated total force at time .

[0027] Furthermore, the finite element method discretizes the lattice-reinforced fiber composite arch into tetrahedral elements. Each tetrahedral element includes four vertices, and each vertex corresponds to a deployed temperature sensor and a humidity sensor. Based on the small-scale approximation principle, element temperature and humidity expressions are constructed. The point temperature or humidity at any point within an element can be obtained by linear interpolation of the vertex temperature or humidity at the four vertices within the element combined with the corresponding volume functions. The volume function of a single vertex is defined as the ratio of the volume of the tetrahedron formed by any point within the element and the three vertices other than the single vertex to the volume of the tetrahedral element containing the single vertex. The volume function of each vertex in the element temperature expression is used as a weight function, and the element temperature expression is integrated within the element to obtain the element temperature matrix equation. The element temperature matrix equation includes the element heat conduction matrix. Unit heat capacity matrix Coupled with unit temperature and humidity matrix Unit heat conduction matrix The element's thermal capacity matrix describes the effect of heat conduction at each vertex on the temperature changes at the other three vertices. The temperature-humidity coupling matrix of the cell describes the impact of temperature changes at each vertex on the heat storage of the other three vertices. This describes the coupling effect of humidity changes at each vertex within a cell on the temperature changes at the other three vertices. The volume function of each vertex in the cell humidity expression is used as a weight function, and the expression is integrated within the cell using volume. This simplifies to obtain the cell humidity matrix equation, which includes the cell humidity coupling matrix. and unit humidity quality matrix Unit humidity coupling matrix The humidity-quality matrix describes the coupled effect of temperature changes at each vertex within a cell on humidity changes at the other three vertices. The effect of humidity changes at each vertex within a unit on the moisture content of the other three vertices is described. The unit temperature matrix equations and unit humidity equations of all units are combined to form a set of temperature matrix equations and a set of humidity matrix equations for a lattice-reinforced fiber composite arch.

[0028] Furthermore, the implicit Euler backward method is used to transform the temperature matrix. Regarding time Discretize the differential terms over time to obtain the temperature matrix. From the Time to the Partial differential term at time Approximately the first Temperature matrix at time Subtract the first Temperature matrix at time The obtained temperature change divided by the sampling interval Humidity matrix From the Time to the Partial differential term at time Approximately the first Humidity matrix at time Subtract the first Humidity matrix at time The obtained humidity change divided by the sampling interval Replace the approximate equations with the temperature matrix equations and humidity matrix equations and substitute them into the first... Temperature matrix at time Humidity matrix , obtained the Simulated temperature matrix at time 1 and simulated humidity matrix .

[0029] Compared with existing technologies, this invention accurately characterizes the bonding effect and dynamic load effect between the lattice-reinforced fiber composite arch and the tunnel lining segments by establishing a bond-slip model and a total stress model. By combining coupled temperature and humidity change equations with the Hachson failure criterion, it quantifies the impact of environmental factors on the reinforcement performance of the lattice-reinforced fiber composite arch. By introducing Gaussian process interpolation, time-series inference networks, and the finite element method, it achieves the expansion of slip data and prediction of future stress, temperature, and humidity conditions. The remaining life is dynamically calculated through a damage rate model to assist in replacement decisions. This solves the problems of inaccurate stress analysis and lack of predictive and early warning capabilities in traditional assessments, significantly improving the accuracy, dynamism, and foresight of the assessment of lattice-reinforced fiber composite arch-reinforced tunnel structures, and providing reliable technical support for the safe operation of tunnels. Attached Figure Description

[0030] Figure 1 A flowchart for an evaluation method of fiber composite arch-reinforced tunnel structures based on digital twins;

[0031] Figure 2 Here is a flowchart for calculating adhesive strength;

[0032] Figure 3 Here is a flowchart of the Gaussian process interpolation;

[0033] Figure 4 Here is a flowchart of the temporal reasoning network;

[0034] Figure 5 This is a front view of a lattice-reinforced fiber composite arch.

[0035] Figure 6 This is a cross-sectional view of the interior of a lattice-reinforced fiber composite arch.

[0036] Figure 7 This is a side view of a lattice-reinforced fiber composite arch. Detailed Implementation

[0037] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0038] Example 1

[0039] like Figure 1 As shown in the figure, a specific embodiment of the present invention discloses an evaluation method for fiber composite arch-reinforced tunnel structures based on digital twins, comprising the following steps:

[0040] Arch dimensions are extracted from the tunnel BIM model, and a bond-slip model is established by combining bilinear constitutive relations. A total stress model is constructed by synergistically considering earth pressure, dynamic load effects, and bond-slip influence. Temperature is then considered synergistically based on the laws of heat conduction and diffusion. With humidity Coupled temperature change equations and coupled humidity change equations were established to correlate the coupling changes. The influence of temperature and humidity on the performance of lattice-reinforced fiber composite arches was introduced by the Hashiin failure criterion. The damage rate model was constructed by combining the total stress model with the overall stress model.

[0041] Collection of the first Earth pressure at any moment ,strain sparse glide vector Temperature matrix Humidity matrix Gaussian interpolation is used to interpolate sparse slip vectors. Expanded into dense slip vector The slip function is obtained through a fitting algorithm. The calculation of the first [unclear] based on the bond-slip model and the total force model Total force at time Earth pressure ,strain Data were collected using earth pressure gauges and strain gauges, respectively. Earth pressure gauges were deployed at key stress sections on the outer arc surface of the tunnel segment, while strain gauges were deployed at the center of the inner arc surface of the tunnel segment. Sparse slip vectors were also collected. Temperature data was collected using displacement sensors deployed along the long axis of the lattice-reinforced fiber composite arch. Humidity matrix Temperature and humidity data are collected by distributed temperature and humidity sensors. The fitting algorithms include, but are not limited to, cubic spline fitting, polynomial regression, Fourier series fitting, B-spline fitting, and locally weighted regression.

[0042] Calculate the first Average temperature at time and average humidity and the Total force at time Substitute into the damage rate model to determine the first Damage rate at time And combined with the first Remaining lifespan at any given moment Calculate the first Remaining lifespan at any given moment With lower limit of life Compare the decision to replace the lattice-reinforced fiber composite arch with a new one or to reinforce it. Total force at time Input temporal reasoning network generates the first Simulated total force at time The coupled temperature-varying equation and coupled humidity-varying equation are spatially discretized using the finite element method. The implicit Euler backward method is used to approximate the time derivative, based on the first... Temperature matrix at time and humidity matrix the simulated temperature matrix at the time instant and the simulated humidity matrix wherein the average temperature at the time instant and the average humidity are the mean values of the temperature matrix and the humidity matrix respectively;

[0043] the simulated average temperature at the time instant and the simulated average humidity are calculated and substituted into the damage rate model with the simulated total force at the time instant to determine the simulated damage rate at the time instant and combined with the residual life at the time instant to calculate the simulated residual life at the time instant , which is compared again with the lower limit life to decide whether to reinforce the lattice reinforced fiber composite arch immediately or to wait until the time instantis reached, wherein the simulated average temperature at the time instant and the simulated average humidity are the mean values of the simulated temperature matrix and the simulated humidity matrix respectively.

[0044] As shown in Figure 2 , specifically, the interface slip is continuously distributed along the length direction of the arch of the lattice reinforced fiber composite arch, and the bond slip model is based on the slip function at the time instantand the long axis coordinate to calculate the bond force between the lattice reinforced fiber composite arch and the segment at the time instant and the long axis coordinate , wherein the long axis direction of the lattice reinforced fiber composite arch is the length direction of the arch, and the bond force satisfies the bilinear constitutive relationship with the slip function , specifically as follows:

[0045]

[0046] wherein is the bond stiffness, which is determined through the test of the lattice reinforced fiber composite arch, The critical slip amount is determined by an interface shear test, The maximum bond force is confirmed by consulting the manual of lattice reinforced fiber composite arch, the arch size is extracted from the tunnel BIM model, the arch size includes the short axis length And the long axis length The bond force Is integrated along the long axis direction of the lattice reinforced fiber composite arch and multiplied by the short axis length To obtain the bond force At the first Time, under the premise of the long axis coordinate At the first Time, the bond force Of any short axis coordinate Is the same, equal to the bond force .

[0047] Further, the total force model considers the soil pressure At the first Time, the dynamic load force And the bond force Real-time influence on the total force At the first Time, the total force model is determined based on force analysis, as follows:

[0048] ,

[0049] Where, The component area, the dynamic load force At the first Time is subject to Gaussian distribution near the static load force , and according to Hooke's law, the static load force At the first Time is equal to the product of the strain At the first Time and the elastic modulus , the component area , as follows:

[0050] ,

[0051] Where, Indicates Gaussian distribution conforming to zero mean and strain variance The strain variance Is calculated by statistical calculation through multiple measurements, and the elastic modulus Is determined by pouring the tunnel concrete mixture into a standard prism test piece and testing.

[0052] Specifically, based on the heat conduction law, the temperature Changes with time The temperature change equation is as follows:

[0053] ,

[0054] wherein, , and are air density, air specific heat capacity and thermal conductivity respectively, is the Hamiltonian of temperature , which is equal to the second order gradient of temperature with respect to spatial coordinates, reflecting the spatial conduction change of temperature , is the change rate of temperature with respect to time , the humidity is determined by the water content, and the change of the water content will cause the change of the temperature, and the change rate of the humidity with respect to time is superimposed on the latent heat of water phase change , the coupling temperature change equation is obtained, and the specific equation is as follows:

[0055] .

[0056] Specifically, based on the diffusion law, the influence of temperature on the basic diffusion coefficient of humidity is introduced, and the coupling humidity change equation describing the change of humidity with respect to time is as follows:

[0057] ,

[0058] wherein, is the Hamiltonian of humidity , which is equal to the second order gradient of humidity with respect to spatial coordinates, reflecting the spatial diffusion change of humidity , is the coupling diffusion coefficient of humidity , reflecting the influence of temperature on the spatial diffusion of humidity , and the specific equation is as follows:

[0059] ,

[0060] wherein, is an empirical coefficient.

[0061] Further, based on the Hashin failure criterion, the failure index at the first time is constructed to quantify the failure of the lattice reinforced fiber composite arch at the first Average temperature at time Average humidity Total Force The relative correlation between the failure critical state and the first case Failure index at time Specifically as follows:

[0062] ,

[0063] in, For collaborative considerations, the first Average temperature at time and average humidity The ultimate total load that the lattice-reinforced fiber composite arch can withstand is as follows:

[0064] ,

[0065] in, To be at standard temperature and standard humidity The ultimate total stress that the lattice-reinforced fiber composite arch can withstand, as determined below. and These are the temperature compensation coefficient and humidity compensation coefficient for total stress, respectively, determined by repeatedly measuring and fitting the ultimate total stress under different temperature and humidity conditions, with a standard temperature... 25 degrees Celsius, standard humidity The relative humidity is 0.4. Relative humidity is the ratio of the water vapor content in the air to the saturated water vapor content under the same conditions. Failure index at time When it is greater than or equal to 0, that is, when the lattice-reinforced fiber composite arch is in the first... The time has reached or exceeded the failure critical condition, the first Damage rate at time Equal to damage evolution coefficient , No. Failure index at time With total force The product of, i.e. Damage evolution coefficient The durability of the fiber-reinforced composite arch was determined through a lattice-reinforced structure arch durability test.

[0066] like Figure 3 As shown, Gaussian process interpolation further includes the following steps:

[0067] Based on the Sparse glide vector at time ,in, For the first The displacement sensor in the first The amount of slip collected at any time. For the first Major axis coordinates of a displacement sensor This represents the total number of displacement sensors.

[0068] The Matern kernel is used to characterize the physical law that the relationship between the axial spacing and the slip is positively correlated, for any major axis coordinate. With major axis coordinates The Martern core between Specifically as follows:

[0069] ,

[0070] in, It is an exponential function. and These are the kernel standard deviation controlling the amplitude of slip fluctuation and the kernel scale controlling the smoothness of the curve, respectively.

[0071] Calculate the first Sparse glide vector at time sparse covariance matrix , among which, the Major axis coordinates of a displacement sensor With the Major axis coordinates of a displacement sensor sparse covariance between Specifically as follows:

[0072] ,

[0073] in, For the first Major axis coordinates of a displacement sensor With the Major axis coordinates of a displacement sensor The Martern core between, For the first Standard deviation of the noise acquired by each displacement sensor For input and The Kronecker function value is given when the two inputs are equal; it takes the value 1 when the two inputs are unequal and 0 when the two inputs are not equal. ;

[0074] Determine the dense slip vector The interpolated major axis coordinate vector ,in, For the first Time of the first The slippage of each interpolation point For the first Interpolated major axis coordinates of each interpolation point Let be the total number of interpolation points, and Much larger ;

[0075] Calculate the first Sparse glide vector at time With dense slip vector Interaction covariance matrix , among which, the Interpolation major axis coordinates of each interpolation point With the Major axis coordinates of a displacement sensor Interactive covariance equal to the Interpolation major axis coordinates of each interpolation point With the Major axis coordinates of a displacement sensor The Martern core between ;

[0076] Calculate the first Dense slip vector at time Dense covariance matrix , among which, the Interpolation major axis coordinates of each interpolation point With the Interpolation major axis coordinates of each interpolation point Dense covariance between Specifically as follows:

[0077] ,

[0078] in, For the first Interpolation major axis coordinates of each interpolation point With the Interpolation major axis coordinates of each interpolation point The Martern core between, For the first Standard deviation of interpolation noise at each interpolation point;

[0079] Calculate the first Dense slip vector at time The interpolated posterior distribution follows the principle of the first... The vector of the distribution mean at time t. and the vector of standard deviation of distribution at time The constructed Gaussian distribution is as follows:

[0080] ,

[0081] ,

[0082] in, and Let these represent the transpose and inverse of a matrix, respectively. The vector of the distribution mean at time t. Used to reflect the Dense slip vector at time middle The average slip at the interpolation point, the first vector of standard deviation of distribution at time Then it reflects the first Dense slip vector at time middle Uncertainty at each interpolation point;

[0083] nuclear standard deviation nuclear scale , The standard deviation of the acquisition noise of each displacement sensor and The standard deviation of interpolation noise at each interpolation point is used as a variable. Construct the marginal log-likelihood function The details are as follows:

[0084] ,

[0085] To maximize the marginal log-likelihood function To achieve the objective, an optimization algorithm is used for iterative optimization to reduce uncertainty and obtain the optimal variables. And recalculate the first The optimal distribution mean vector at time 1 , will the The optimal distribution mean vector at time 1 Treat it directly as the first Dense slip vector at time Among them, the optimization algorithms include, but are not limited to, L-BFGS algorithm, genetic algorithm, particle swarm optimization algorithm, gray wolf optimization algorithm, mouse swarm optimization algorithm and whale foraging algorithm.

[0086] Specifically, the first moment is defined as the moment when the lattice-reinforced fiber composite arch is first put into use after installation, and the remaining life at the first moment is... Lower limit of lifespan When the first Average temperature at time Average humidity Total Force At that time, the damage rate model was used to calculate the first... Damage rate at time And multiply by the collection interval get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point get the remaining life of the first time point get the damage amount from the first time point to the second time point

[0087] As shown in FIG. 8, specifically, the historical total stress of the previous n time points before the first time point is called to construct the pre-total stress sequence of the first time point, and the total stress of the first time point is calculated to get the total stress of the first time point Figure 4 ​​​​​​The pre-total force sequence of the time point The input time sequence reasoning network adopts an ensemble learning framework, and respectively captures the time sequence change trend of the pre-total force sequence of the time point t by different base learners The pre-total force sequence of the time point The time sequence change trend of the pre-total force sequence of the time point t The latent time sequence representation sequence of the time point t , wherein The latent time sequence representation of the time point t output by the base learner The latent time sequence representation of the time point t output by the base learner The base learner includes but is not limited to LSTM, GRU, CNN, RNN, CNN-LSTM, Transformer, BP, GBDT, XGBoost and LightGBM The latent time sequence representation sequence of the time point t Input the meta-learner, because different base learners have different advantages and disadvantages, the meta-learner is used to make up for the shortcomings of the base learners The calculation complexity of the time sequence reasoning network, because The base learner has respectively learned the time sequence change trend of the pre-total force sequence of the time point t from different angles The base learner has respectively learned the time sequence change trend of the pre-total force sequence of the time point t from different angles The base learner has respectively learned the time sequence change trend of the pre-total force sequence of the time point t from different angles The latent time sequence representation sequence of the time point t The latent time sequence representation of the time point t output by the base learner The latent time sequence representation of the time point t output by the base learner The latent time sequence representation of the time point t output by the base learner The preliminary simulation total force of the time point t , wherein The preliminary simulation total force of the time point t output by the base learner The preliminary simulation total force of the time point t output by the base learner The preliminary simulation total force of the time point t output by the base learner The simulation total force of the time point t .

[0088] ​Further, the lattice reinforced fiber composite arch is discretized into a finite number of tetrahedral elements by the finite element method, each tetrahedral element includes 4 vertices, and each vertex corresponds to a deployed temperature sensor and humidity sensor, that is, when the temperature sensor and humidity sensor are distributed, the temperature sensor and humidity sensor are deployed on the vertices divided by the finite element method, the element temperature expression and the element humidity expression are constructed based on the small-scale approximation principle, the point temperature of any point in the element can be obtained by linear interpolation of the vertex temperature of the 4 vertices in the element combined with the corresponding volume function, and the point humidity of any point in the element can be obtained by linear interpolation of the vertex humidity of the 4 vertices in the element combined with the corresponding volume function, wherein the volume function of a single vertex is defined as the volume ratio of the tetrahedron formed by any point in the element and the three vertices other than the single vertex to the volume of the tetrahedral element where the single vertex is located, the volume function of each vertex in the element temperature expression is taken as the weight function, and the volume integration of the element temperature expression in the element is performed to offset the residual of the coupled temperature change equation solution, the volume integration is calculated and arranged to obtain the element temperature matrix equation, as follows:

[0089] ,

[0090] wherein, is the element heat conduction matrix, which is used to describe the influence of the heat conduction of each vertex in the element on the temperature change of the remaining 3 vertices, and are the element temperature vector and the element humidity vector respectively, which include the temperature and humidity of the 4 vertices in the element, is the element heat capacity matrix, which is used to describe the influence of the temperature change of each vertex in the element on the heat storage of the remaining 3 vertices, is the element temperature and humidity coupling matrix, which describes the coupling influence of the humidity change of each vertex in the element on the temperature change of the remaining 3 vertices, the volume function of each vertex in the element humidity expression is taken as the weight function, and the volume integration of the element humidity expression in the element is performed to offset the residual of the coupled humidity change equation solution, the volume integration is calculated and arranged to obtain the element humidity matrix equation, as follows:

[0091] ,

[0092] wherein, is the element humidity coupling matrix, which describes the coupling influence of the temperature change of each vertex in the element on the humidity change of the remaining 3 vertices, is the element humidity mass matrix, which is used to describe the influence of the humidity change of each vertex in the element on the water content of the remaining 3 vertices, the element temperature matrix equation and the element humidity equation of all elements are combined respectively to form the temperature matrix equation set and the humidity matrix equation set of the lattice reinforced fiber composite arch, as follows:

[0093] ,

[0094] ,

[0095] wherein, is a heat conduction matrix, and are a temperature matrix and a humidity matrix, respectively, is a heat capacity matrix, is a temperature-humidity coupling matrix, is a humidity coupling matrix, is a humidity mass matrix, are obtained by splicing a unit temperature matrix equation and a unit humidity matrix equation, and the finite element method discretely calculates the Hamiltonian operator in the coupled temperature variation equation and the coupled humidity variation equation at the unit vertex. Since the volume integral calculation process is lengthy and conventional, it is not described in detail here.

[0096] Furthermore, since the temperature matrix and the humidity matrix vary with time , the implicit Euler backward method discretizes the differential term of the temperature matrix with respect to time . The implicit Euler backward method shows that the partial differential term of the temperature matrix from the first moment to the second moment can be approximated as the temperature change amount obtained by subtracting the temperature matrix at the first moment from the temperature matrix at the second moment divided by the collection interval . According to the same method, the partial differential term of the humidity matrix from the first moment to the second moment can be approximated as the humidity change amount obtained by subtracting the humidity matrix at the first moment from the humidity matrix at the second moment divided by the collection interval . Substituting the approximate equations into the temperature matrix equation and the humidity matrix equation and substituting the temperature matrix and the humidity matrix at the first moment, the simulated temperature matrix and the simulated humidity matrix at the second moment can be obtained.

[0097] Example 2

[0098] As shown in Figures 5-7 A lattice reinforced fiber composite arch, comprising a fiber composite shell, a hard filling core material, a flange, a lattice rib and a reinforcing plate;

[0099] The fiber composite shell serves as the outer protective and basic load-bearing component of the arch structure. It uses fiber-reinforced composite materials and one-step forming process to accurately shape the curved surface of the arch, providing basic mechanical support and environmental protection for the arch as a whole, and can adapt to complex curved surface modeling requirements.

[0100] The hard filling core material is filled inside the fiber composite shell and cooperates with the fiber composite shell to bear stress. It enhances the overall stiffness and stability of the arch structure by its own characteristics. The core material is filled in the internal space of the fiber composite shell, and the appropriate material and density of the core material can be selected according to the structural performance requirements to optimize the mechanical response of the arch structure.

[0101] The flange is arranged at a specific position of the arch to expand the stress surface, optimize the stress transfer path, and enhance the mechanical properties of the connection between the arch and other structures. The material is FRP material consistent with the fiber composite shell, which is integrated into the arch structure through one-step forming or subsequent connection process.

[0102] The lattice rib is an optional structure arranged inside the fiber composite shell in a lattice form to play a stiffening role, improve the local and overall carrying capacity of the fiber composite shell, and disperse the load. It forms a composite load-bearing system together with the fiber composite shell and the hard filling core material. The FRP designability is used to optimize the lattice form and spacing parameters according to the stress characteristics of the arch.

[0103] The reinforcing plate is an optional structure that is added at the stress concentration or strengthening parts on the inner arc surface side and side surface of the arch according to the actual stress calculation results. The reinforcing plate is made of FRP material and connected with the main fiber composite shell through bonding process to improve the local carrying capacity and ensure the stress safety of the arch structure.

[0104] The application discloses a kind of digital twin's fiber composite arch reinforcement tunnel structure evaluation method, bond slip model is established in combination with arch size and bilinear constitutive relation, the change law of the adhesion of lattice reinforced fiber composite arch and segment with slip amount is described;Considering the influence of soil pressure, dynamic load and bond slip, a total stress model is constructed to comprehensively reflect the superposition effect of various loads on the structure;Based on the heat conduction law, diffusion law and temperature and humidity coupling relationship, the coupling temperature change and humidity change equation is established, which can accurately capture the synergistic effect of environmental factors on the structure;Hashin failure criterion is used to construct a damage rate model based on temperature, humidity and total stress, which provides a scientific method for quantifying structural damage evolution;After collecting real-time soil pressure, strain, sparse slip vector, temperature matrix and humidity matrix, the sparse slip vector is expanded to dense slip vector by Gaussian process interpolation, which makes up for the data sparsity problem caused by limited displacement sensor deployment, and a continuous slip function is obtained;Based on the bond slip model and total stress model, the total stress at this time is calculated to accurately quantify the current stress state of the structure;Calculate the average temperature and humidity at this time, and input them into the damage rate model together with the total stress, and calculate the current remaining life based on the remaining life at the last time, and compare it with the lower limit life to preliminarily decide whether to replace the lattice reinforced fiber composite arch, to realize real-time evaluation of the current health status of the structure;If the current remaining life is acceptable, input the current total stress into the time series reasoning network, capture the time series trend through multiple base learners and fuse through element learners to generate the next time simulation total stress, which improves the prediction accuracy of future stress;The lattice reinforced fiber composite arch is discretized into tetrahedral elements using the finite element method, and the coupled temperature change and humidity change equation is discretized and time approximated using the implicit Euler backward method, and the next time simulation temperature and humidity matrix is inferred based on the current temperature and humidity matrix to accurately deduce the future state of environmental factors;Calculate the simulation average temperature and humidity and simulation total stress, and input them into the damage rate model to calculate the next time simulation remaining life, and compare it with the lower limit life again to make a decision, which realizes early warning of potential risks of the structure by predicting future state in advance, and balances safety and economy.

[0105] The above only describes the preferred embodiments of the present application, and the protection scope of the present application is not limited to the above-mentioned embodiments. Any technical solutions falling within the scope of the present application should be considered within the protection scope of the present application. It should be noted that, for ordinary skilled persons in the art, some improvements and refinements without departing from the principles of the present application should also be considered within the protection scope of the present application.

Claims

1. A method for assessing a tunnel structure reinforced with a digital-twinning fiber composite arch, characterized by, The method comprises the following steps: A bond-slip model is established by combining arch size and a bilinear constitutive relationship, a total stress model is constructed by considering the effects of soil pressure, dynamic load and bond-slip, a coupled temperature change equation and a coupled humidity change equation are established based on the heat conduction law, the diffusion law and the temperature-humidity coupling relationship, and a damage rate model is constructed by combining the Hashin failure criterion with the effects of temperature and humidity and the total stress model; The soil pressure, strain, sparse slip vector, temperature matrix and humidity matrix at the current time are collected, the total stress at the current time is calculated after Gaussian process interpolation processing, the remaining life at the current time is calculated and compared with the lower limit life, and it is initially evaluated whether the lattice reinforced fiber composite arch needs to be replaced; If replacement is not needed, the simulated total stress at the next time is generated by using a time series reasoning network, the coupled temperature change equation and the coupled humidity change equation are processed by using the finite element method and the implicit Euler backward method, the simulated temperature matrix and the simulated humidity matrix at the next time are obtained based on the temperature matrix and the humidity matrix at the current time, and the life is re-evaluated based on the simulated total stress, the simulated temperature matrix and the simulated humidity matrix at the next time to decide whether to replace; The bond-slip model is based on a slip function of the current time and a long-axis coordinate, and the bond force between the lattice reinforced fiber composite arch and the segment at the current time and at any long-axis coordinate is calculated, the bond force and the slip function satisfy a bilinear constitutive relationship, the bond force is integrated along the long-axis direction of the lattice reinforced fiber composite arch and multiplied by the length of the short axis to obtain the bond force at the current time; In the total stress model, the product of the total stress at the current time and the area of the component is equal to the soil pressure at the current time plus the dynamic load minus the bond force at the current time, the dynamic load follows a Gaussian distribution around the static load, and the static load is equal to the product of the strain at the current time, the elastic modulus and the area of the component according to Hooke's law; The Gaussian process interpolation is used to expand the sparse slip vector into a dense slip vector, including: based on the sparse slip vector at the current time, the relevance between the axial coordinate interval and the slip is described by using a Matern kernel; a sparse covariance matrix, an interaction covariance matrix and a dense covariance matrix at the current time are constructed; a distribution mean vector and a distribution standard deviation vector at the current time are calculated, and the dense slip vector at the current time is obtained by optimizing the variables to maximize the marginal log-likelihood function, and the dense slip vector is used to fit the slip function; The damage rate model is constructed based on the Hashin failure criterion, and the failure index at the current time is obtained by subtracting 1 from the square of the limit total stress at the average temperature and the average humidity minus the total stress at the current time, when the failure index at the current time is greater than or equal to 0, the damage rate at the current time is equal to the product of the damage evolution coefficient, the failure index at the current time and the total stress at the current time.

2. A method of assessing a tunnel structure reinforced with a digital-twin fiber composite arch as claimed in claim 1, wherein, The remaining life at the current time is equal to the remaining life at the last time minus the product of the damage rate at the current time and the collection interval, in the initial evaluation, if the remaining life at the current time is less than or equal to the lower limit life, the lattice reinforced fiber composite arch is replaced, otherwise, the life is re-evaluated, and the lower limit life is a set value.

3. A method of assessing a tunnel structure reinforced with a digital-twin fiber composite arch as claimed in claim 1, wherein, The timing inference network calls the historical total force before the present moment, constructs the pre-total force sequence of the present moment and inputs the network, captures the timing trend through multiple base learners to generate the potential timing sequence, and the meta-learner generates the simulated total force of the next moment through linear mapping and weighted summation of LR linear regression.

4. A method of assessing a tunnel structure reinforced with a digital-twin fiber composite arch as claimed in claim 1, wherein, The finite element method discretizes the lattice reinforced fiber composite arch into tetrahedral elements, each vertex corresponding to a temperature sensor and a humidity sensor, linearly interpolates the temperature and humidity in the element through volume functions, obtains the element temperature matrix equation and the element humidity matrix equation through volume integration, combines them into the temperature matrix equation set and the humidity matrix equation set, approximates the time derivative using the implicit Euler backward method, and obtains the simulated temperature matrix and the simulated humidity matrix of the next moment based on the temperature matrix and the humidity matrix of the present moment.

5. A method of assessing a tunnel structure reinforced with a digitally twinned fiber composite arch as claimed in claim 1, wherein, The coupled temperature change equation is established based on the heat conduction law, and the product of the temperature time-varying term and the air density and the air specific heat capacity is equal to the product of the thermal conductivity coefficient and the Hamiltonian operator of the temperature plus the latent heat of water phase change and the rate of change of the humidity with time, which is used to describe the temperature change from the present moment to the next moment.

6. A method of assessing a tunnel structure reinforced with a digitally twinned fiber composite arch as claimed in claim 1, wherein, The coupled humidity change equation is constructed based on the diffusion law, and the product of the air density and the rate of change of the humidity with time is equal to the product of the coupled diffusion coefficient of the humidity and the Hamiltonian operator of the humidity; the coupled diffusion coefficient is the product of the base diffusion coefficient and the power with the natural constant as the base and the ratio of the inverse of the empirical coefficient and the temperature as the exponent, which is used to describe the humidity change from the present moment to the next moment.

Images