A dam structural performance evaluation method and platform based on energy analysis

Abstract

This invention belongs to the field of dam structural performance evaluation technology, specifically involving a method and platform for dam structural performance evaluation based on energy analysis. The method includes: constructing a finite element model of the dam; using monitoring data to invert the mechanical parameters of the dam materials; calculating the actual stress and displacement fields of the dam; calculating the average density and distribution entropy of the dam's elastic energy and dissipated energy based on the stress and displacement fields; constructing a comprehensive energy characterization index set for the dam; extracting the trend components of the dam's energy using multivariate variational mode decomposition; determining whether the trend components of the dam's energy have undergone a trend change; if a change occurs, proceeding to the next step; fitting the trend components of the dam's energy using a polynomial model; and using a cusp catastrophe model to detect abrupt changes in the trend components of the dam's energy; if a catastrophe occurs, determining that the dam's structural performance is unstable; otherwise, determining that the dam's structural performance is normal. This invention can accurately reflect the service performance of the dam structure under real-world conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of dam structure performance evaluation, and in particular to a dam structure performance evaluation method and platform based on energy analysis. Background Art

[0002] The traditional evaluation method of dam structural deformation is based on measured data, but the number of monitoring points arranged on the dam is limited, and there are a large number of unmonitored parts of the dam. Due to the weak integrity of the dam, the observed structural response cannot well reflect the structural properties of the dam's monitoring blind area. The structural numerical analysis method can fully grasp the structural properties of the dam. At present, this method mainly analyzes and evaluates the structural properties of the dam from the perspective of stress, strain and deformation, but stress, strain and deformation mainly reflect the local structural properties of the dam. The dam is an open nonlinear system, and the evolution of its structural properties is the result of energy exchange between the dam and the external environment. The structural properties of the dam are closely related to its energy state. Therefore, the traditional evaluation methods of dam structural properties cannot accurately reflect the service properties of the dam structure under real conditions. Summary of the invention

[0003] Purpose of the invention: The purpose of the present invention is to address the deficiencies in the prior art and to provide a dam structural performance evaluation method and platform based on energy analysis, which can accurately reflect the service performance of the dam structure under real conditions, realize the digital and intelligent management of dam safety diagnosis, and has outstanding advantages in improving diagnostic accuracy, enhancing engineering practicality and reducing costs.

[0004] Technical solution: The dam structure performance evaluation method based on energy analysis of the present invention comprises the following steps:

[0005] A1. Construct a finite element model of the dam according to the actual situation of the dam, use the monitoring data to invert the mechanical parameters of the dam material, input the mechanical parameters of the dam material into the finite element model of the dam, and calculate the real stress field and displacement field of the dam;

[0006] A2. Based on the stress field and displacement field, the average density and distribution entropy of the dam's elastic energy and dissipated energy are calculated, thereby constructing a comprehensive characterization index set of the dam's energy;

[0007] A3. Use multivariate variational mode decomposition to extract the low-frequency components of the comprehensive characterization index set of dam energy and obtain the trend component of dam energy;

[0008] A4. Determine whether the trend component of the dam energy has a trend change. If so, proceed to the next step. Otherwise, determine that the dam structure is normal.

[0009] A5. Use polynomials to fit the trend component of the dam energy, and use the cusp mutation model to identify the mutation of the trend component of the dam energy. If a mutation occurs, the dam structure is judged to be unstable, otherwise the dam structure is judged to be normal.

[0010] To further improve the above technical solution, A1 includes:

[0011] The actual situation of the dam includes the dam structure design, construction, geological survey data, and the construction of a finite element model of the dam under the action of seepage-stress coupling;

[0012] The monitoring data include environmental monitoring data of the dam site area, the monitoring data are divided into a training sample set and a test sample set according to the time span, and the training sample set is input into the finite element model of the dam;

[0013] Determine the parameters to be inverted, use the orthogonal design table to formulate N different parameter combinations, input them into the dam finite element model, and obtain the finite element simulation values ​​of the dam displacement under N different parameter combinations;

[0014] The training samples were normalized and then input into the neural network model for training. After learning and training, the accuracy of the neural network model reached the expected level, and the mean square error after iteration dropped below 0.05. Thus, the nonlinear mapping relationship between the finite element simulation value of the dam displacement and the inversion parameters was established.

[0015] Taking the minimum error between the finite element simulation value of displacement and the dam displacement monitoring data as the objective function, the optimal combination of inversion parameters is obtained. The objective function expression is as follows:

[0016] f=min|HH * |,

[0017] Where: f is the objective function, H is the displacement finite element simulation value, and H* is the measured value of displacement;

[0018] The best combination of the test sample set and the inversion parameters is input into the dam finite element model to obtain the displacement finite element simulation value, which is compared with the actual value in the dam displacement monitoring data to calculate the relative error value. If the relative error meets the requirements, the inversion parameters are output, otherwise the neural network model is returned to re-learning and training;

[0019] According to the monitoring data of reservoir water level, rainfall and temperature at different times in history, the stress and strain of the dam are calculated based on the finite element model of the dam and the parameter inversion values.

[0020] Further, the A2 comprises:

[0021] Dam elastic energy U e and dissipated energy U dThe change in the dam structure state evolution process satisfies: W = U e +U d ,

[0022] Where: W is the work done by the external load on the dam, U e is the elastic strain energy stored in the dam, U d is the dissipated energy during the deformation of the dam;

[0023] The finite element model of the dam has N units, where the energy q(k) and volume V(k) of the kth unit are:

[0024]

[0025] Where: Ω k is the volume domain of the kth unit, υ is the energy density, and the elastic energy density of the dam υ e and dissipated energy density υ d Calculate as follows:

[0026]

[0027] Where: ij , and Respectively stress tensor, elastic strain tensor, plastic strain tensor, rheological strain tensor and wetting strain tensor;

[0028] The total energy U and total volume V of the dam are obtained by adding up the energy q(k) and volume V(k) of all units:

[0029]

[0030] Divide the total energy U of the dam by the total volume V to get the average energy density of the dam

[0031] Define the energy distribution entropy S:

[0032] in:

[0033]

[0034] Where: λ(k) is the proportion of the energy of the kth unit of the dam in the total energy U, satisfying:

[0035]

[0036] Calculate the dam elastic energy U e and dissipated energy U d The average density And distribution entropy S, thereby constructing a comprehensive characterization index set U of dam energy I :

[0037]

[0038] Where: and S e The dam elastic energy U e The average density and distribution entropy of and S d are the dam dissipated energy U d The average density and distribution entropy of .

[0039] Furthermore, the A3 adopts multivariate variational mode decomposition to extract the trend component of dam energy including:

[0040] The 4-element energy signal U corresponding to the dam energy characterization index set I (t) is uniformly denoted as x(t) = [x 1 (t) x 2 (t)x 3 (t) x 4 (t)],

[0041] Assume that the 4-element signal x(t) is decomposed into n 4-element intrinsic mode function IMF components u k (t)(k=1,2,…n):

[0042]

[0043] Where: u k (t)=[u k,1 (t)u k,2 (t)u k,3 (t)u k,4 (t)],u k,c (t) = a k,c (t)cos(φ k,c (t)), c = 1, 2, 3, 4,

[0044] Where: u k,c (t) is an AM / FM signal; a k,c (t) and φ k,c (t) are the signals u k,c (t) Amplitude and phase functions;

[0045] For each IMF component u k (t)(k=1,2,…n), using Hilbert transform Obtaining an analytical signal with a single-sided spectrum

[0046]

[0047] in:

[0048]

[0049] Where: j is the imaginary unit;

[0050] Will With exponential term Multiply to adjust its center frequency ω k , so that the spectrum of each IMF component signal is modulated to the corresponding baseband and the harmonic mixing signal The gradient of L 2 The squared norm estimates the bandwidth of each IMF component,

[0051] With the goal of minimizing the sum of the IMF component bandwidths of all signals, the 4-element signal x(t) is decomposed into the following constrained optimization problem, namely:

[0052]

[0053] Where: represents the partial derivative with respect to time t; Indicates L 2 The square of the norm;

[0054] To solve the constrained optimization problem, construct its augmented Lagrangian function:

[0055]

[0056] Where: α is the penalty factor; λ c (t) is the Lagrangian multiplier; <, > represents the inner product;

[0057] The alternating direction multiplier method is used to transform the optimization problem into a series of sub-optimization problems, and each sub-problem is iteratively updated in the frequency domain, where the IMF component u k,c , center frequency ω k and the Lagrangian multiplier λ c The update formulas are as follows:

[0058]

[0059]

[0060]

[0061] Where: and u k,c (t), x c (t) and λ c(t) is the Fourier transform; the superscript i is the number of iterations; τ is the update coefficient;

[0062] Press u k,c ,ω k and λ c The update formula is iterated repeatedly until the condition is met The iteration ends when , and finally the optimal IMF component u of the original 4-element signal x(t) is obtained k (t) and its center frequency ω k (k=1,2,…n), where the 4-element IMF component corresponding to the lowest center frequency is the trend component of the dam energy

[0063]

[0064] Where: is the trend component of the mean density, is the trend component of the distribution entropy.

[0065] Furthermore, it is necessary to determine whether the dam energy has undergone a trend change, including the average dam energy density. The trend increases, or the energy distribution entropy The trend is decreasing.

[0066] Furthermore, the trend component of dam energy is fitted using a 4th-order polynomial On this basis, the cusp catastrophe model is used to determine the trend component of dam energy. Whether a mutation occurs. If so, the dam structure is considered unstable. Otherwise, the dam is considered stable, including:

[0067] Trend component of dam energy Expressed as a continuous function of time t, Taylor expansion is performed and the first 5 terms are obtained:

[0068]

[0069] Where: U is the trend component of dam energy; a i (i=0,1,2,3,4) are polynomial coefficients;

[0070] Using Tschirnhaus transformation, let t = xL, where Then the trend component U of the dam energy becomes:

[0071] U=b 4 x 4 +b 2 x 2 +b 1 x+b 0

[0072] in:

[0073]

[0074] Divide both sides of the trend component U of the dam energy by 4b 4 , and order Then the standard potential function V of the cusp mutation can be obtained as:

[0075]

[0076] Where: x is the state variable of the potential function; u and v are the control variables of the potential function;

[0077] Derivative the expression of the standard potential function V with respect to the state variable x and set its derivative to zero, and the equilibrium surface of the potential function V is obtained as follows:

[0078] x 3 +ux+v=0,

[0079] The set of mutation points of the state formed by the points of the vertical tangent on the equilibrium surface is:

[0080]

[0081] The projection of the mutation point set on the control plane constitutes a bifurcation set, which is the set of all points that cause the state variable x to mutate. By connecting the potential function V equilibrium surface equation and the mutation point set equation to eliminate x, the bifurcation set equation is:

[0082] Δ=4u 3 +27v 2 =0,

[0083] When the control variable (u, v) is outside the bifurcation set, that is, Δ>0, the trend component of the dam energy does not change suddenly, and the dam structure is in a stable state; when the control variable (u, v) is inside the bifurcation set, that is, Δ<0, the trend component of the dam energy changes suddenly, and the dam structure becomes unstable.

[0084] The platform for implementing the above-mentioned dam structural performance evaluation method based on energy analysis includes a dam multi-source information collection and management module, a parameter inversion module and a dam structural performance evaluation module.

[0085] The dam multi-source information collection and management module is used to collect the actual situation of the dam and the dam safety monitoring data. The actual situation of the dam includes the dam structure design, construction, and geological survey data. The dam safety monitoring data includes the reservoir water level and rainfall environmental monitoring data measured by the water level gauge and rain gauge, and the seepage pressure, seepage volume, and displacement dam safety monitoring data measured by the osmometer, water measuring weir, and displacement meter.

[0086] The parameter inversion module includes an orthogonal design module, an artificial neural network and a dam finite element model. The orthogonal design module uses orthogonal design to obtain an inversion parameter combination based on the actual dam situation and dam safety monitoring data provided by the dam multi-source information collection and management module, inputs the combination into the dam finite element model to calculate the dam displacement, and establishes a nonlinear mapping relationship between the dam displacement finite element simulation value and the inversion parameter through the artificial neural network, and finally obtains the best combination of inversion parameters according to the objective function;

[0087] The dam structure performance evaluation module is used to determine whether the dam structure is in a stable state, including constructing a comprehensive energy characterization index set, extracting the trend component of the dam energy and the stability judgment of the cusp mutation model. On the basis of the two types of energy, the elastic energy and the dissipated energy of the dam, two energy characterization indicators, namely the average energy density and the energy distribution entropy, are proposed to construct a comprehensive energy characterization index set for the dam structure performance; multivariate variational modal decomposition is used to extract the trend component of the dam energy from the comprehensive energy characterization index set, and a cusp mutation model is established to determine whether the dam structure performance is stable.

[0088] Beneficial effects: Compared with the prior art, the advantages of the present invention are: 1. The present invention is provided with a dam structural state evaluation module. On the basis of the two types of energy, namely, elastic energy and dissipated energy of the dam, two energy characterization indicators, namely, energy average density and energy distribution entropy, are proposed to construct a comprehensive energy characterization indicator set for the dam structural state; multivariate variational modal decomposition is adopted to extract the trend component of the dam energy from the comprehensive energy characterization indicator set, and a mutation model is established to determine whether the dam structural state is stable; it can accurately reflect the overall structural state of the dam, which is more sensitive and has a lower dimension than the traditional stress-strain method to reflect the structural state of the dam, and has automatic pre- and post-processing functions, which can automatically acquire data, start analysis, and generate results, thereby avoiding the influence of human operation and reducing the use requirements of the module.

[0089] 2. The present invention is provided with a parameter inversion module, which uses a proxy model to invert the constitutive model parameters used by the main materials of the dam according to the working environment of the dam, so as to facilitate users to accurately grasp the mechanical characteristics of the dam.

[0090] 3. The present invention is provided with a dam multi-source information collection and management module, which can realize automatic data collection based on the automated monitoring system, simplify the operation, improve the collection efficiency, save manpower, and facilitate data retrieval by users. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] Figure 1 It is a flow chart of energy analysis and evaluation of dam structural properties based on energy analysis;

[0092] Figure 2It is a diagram of energy changes during the typical evolution of the dam's structural properties;

[0093] Figure 3 It is a schematic diagram of the cusp catastrophe model;

[0094] Figure 4 It is a schematic diagram of the structure of the dam structural performance evaluation platform based on energy analysis. DETAILED DESCRIPTION

[0095] The technical solution of the present invention is described in detail below with reference to the accompanying drawings, but the protection scope of the present invention is not limited to the embodiments.

[0096] Embodiment 1: First, the scenario of this embodiment is described. The dam can be an earth-rock dam or a concrete dam. In order to ensure the safety of the dam, it is necessary to analyze whether the structural state of the dam is stable. The present invention includes a dam multi-source information collection and management module, a parameter inversion module and a dam structural state evaluation module, wherein the dam multi-source information collection and management module is used to collect data such as dam structural design instructions, construction drawings, address survey reports, and can automatically collect environmental monitoring data such as reservoir water level and rainfall measured by water level gauges and rain gauges, and dam safety monitoring data such as seepage pressure, seepage volume, and displacement measured by piezometers, water measuring weirs, and displacement meters. The parameter inversion module can be an agent model, including orthogonal design, artificial neural network and dam finite element model. The dam structural state evaluation module includes constructing an energy comprehensive characterization index set, extracting the trend component of dam energy, and cusp mutation model stability judgment.

[0097] See also Figure 1 , is a flow chart of dam structure performance energy analysis and evaluation provided by an embodiment of the present invention, and the dam structure performance evaluation is realized through the following steps, including A1-A5:

[0098] A1. According to the actual filling, water storage and operation conditions of the dam, a finite element model of the dam is constructed, and the mechanical parameters of the dam material are inverted using the monitoring data. The parameters are input into the finite element model of the dam to calculate the real stress field and displacement field of the dam.

[0099] In this embodiment, A1 (according to the actual filling, water storage and operation of the dam, a finite element model of the dam is constructed, and the mechanical parameters of the dam material are inverted using the monitoring data, which are input into the finite element model of the dam to calculate the real stress field and displacement field of the dam) includes A11-A17:

[0100] A11, retrieve the dam structure design, construction, geological survey and other data in the dam multi-source information collection and management module, and call the dam finite element model in the parameter inversion module to construct the dam finite element model under the action of seepage-stress coupling.

[0101] A12, retrieve the environmental monitoring data of the dam site area in the dam multi-source information collection and management module, divide it into a training sample set and a test sample set according to the time span, and first input the training sample set into the dam finite element model.

[0102] A13, the dam is a homogeneous earth dam, and the dam body fill is clay. The Duncan EB model is used to determine the parameters to be inverted: elastic modulus coefficient K, elastic modulus index n, bulk modulus coefficient Kb, and bulk modulus index m. The remaining parameters are measured using indoor test values ​​or on-site test values. The orthogonal design table in the parameter inversion module is used to formulate N different parameter combinations, which are input into the dam finite element model in A11 to obtain the finite element simulation values ​​of the dam displacement under N different parameter combinations.

[0103] A14, MATLAB was used to normalize the training samples and then input them into the neural network model in the parameter inversion module for training. After learning and training, the accuracy of the neural network model reached the expected level, and the mean square error after iteration dropped below 0.05, thereby establishing a nonlinear mapping relationship between the finite element simulation value of the dam displacement and the inversion parameters.

[0104] A15, taking the minimum error between the finite element simulation value of displacement and the dam displacement monitoring data as the objective function, the parameters K, n, K b , the best combination of m, the objective function expression is as follows:

[0105] f=min|HH * |,

[0106] Where: f is the objective function; H is the displacement finite element simulation value; H* is the measured value of displacement.

[0107] A16, the test sample set and parameters K, n, K b The best combination of K, n, and K is input into the finite element model of the dam to obtain the finite element simulation value, which is compared with the measured value to calculate the relative error value. If the relative error is less than 0.05%, the output parameters K, n, and K b , m, otherwise return to A14.

[0108] A17, the monitoring data of reservoir water level, rainfall and temperature at different times in history provided by the dam multi-source information collection and management module and the parameter inversion values ​​provided by the parameter inversion module are input into the dam finite element model to perform finite element analysis and calculate the dam stress and strain.

[0109] A2, the dam structural performance evaluation module constructs a comprehensive characterization index set of dam energy: based on the stress-strain results obtained by finite element analysis of the dam, the dam elastic energy U is calculated. e and dissipated energy U d The average density and distribution entropy of dam energy are used to construct a comprehensive characterization index set of dam energy.

[0110] like Figure 2 As shown, the elastic energy of the dam U e and dissipated energy U d Changes in the dam structure state evolution process meet the following requirements:

[0111] W=U e +U d ,

[0112] Where: W is the work done by the external load on the dam; U e is the elastic strain energy stored in the dam; U d It is the dissipated energy during the deformation process of the dam, which can be quantified by the irreversible deformation work such as the dam's plastic deformation, viscous rheology and wetting deformation.

[0113] The stress states and stress paths at different parts of the dam are different, resulting in different energy densities at different parts of the dam. The total elastic energy and total dissipated energy of the dam alone cannot characterize the spatial differences in dam energy. e and dissipated energy U d Based on the two types of energy, two energy characterization indicators, namely energy average density and energy distribution entropy, are proposed. Based on these indicators, a comprehensive energy characterization indicator set for dam structural properties is constructed, including A21-A25:

[0114] A21, the finite element model of the dam has N units, where the energy q(k) (elastic energy U e Or dissipated energy U d ) and volume V(k) are:

[0115]

[0116] Where: Ω k is the volume domain of the kth unit; υ is the energy density, and the elastic energy density of the dam υ e and dissipated energy density υ d It can be calculated as follows:

[0117]

[0118] Where: ij , and They are stress tensor, elastic strain tensor, plastic strain tensor, rheological strain tensor and wetting strain tensor respectively.

[0119] A22, the energy q(k) and volume V(k) of all units are accumulated to obtain the total energy U and total volume V of the dam, that is:

[0120]

[0121] A23, divide the total energy U of the dam by its total volume V to get the average energy density of the dam Right now:

[0122]

[0123] A24, in order to comprehensively reflect the spatial distribution of dam energy, according to the information entropy principle, this embodiment defines energy distribution entropy S, that is:

[0124]

[0125] in:

[0126]

[0127] Where: λ(k) is the proportion of the energy of the kth unit of the dam in the total energy U, which satisfies:

[0128]

[0129] The energy distribution entropy S defined by the above formula satisfies the completeness and non-negativity of information entropy. When the dam energy is evenly distributed, the energy distribution entropy S is the largest, and the dam system is in a disordered state. When the dam has energy concentration areas such as stress concentration or strain localization, the energy distribution entropy S decreases, and the dam system tends to an ordered state.

[0130] It should be pointed out that dam filling is generally simulated using the birth-death unit method. In order to ensure the consistency of the energy calculation area during the dam filling period and the operation period, this embodiment calculates the average energy density of all units of the dam. And the distribution entropy S, the energy of the dam unit that has not been filled and activated is zero.

[0131] A25, calculate the dam elastic energy U from A21-A24 respectively e and dissipated energy U d The average density And distribution entropy S, thereby constructing a comprehensive characterization index set U of dam energy I ,Right now:

[0132]

[0133] Where: and S e are the average density and distribution entropy of the dam's elastic energy, respectively; and S d are the average density and distribution entropy of the dam dissipated energy, respectively.

[0134] It can be understood that the energy characterization index set U IIt can reflect the structural properties of the dam as a whole. Compared with stress-strain, it is more sensitive and has a lower dimension in reflecting the structural properties of the dam.

[0135] A3, dam structural performance evaluation module extracts the trend component of dam energy: multivariate variational mode decomposition is used to extract the comprehensive characterization index set U of dam energy I The low-frequency component of (t), i.e., the trend component of the dam energy

[0136] The 4-element energy signal U corresponding to the dam energy characterization index set I (t) is uniformly denoted as x(t) = [x 1 (t)x 2 (t)x 3 (t)x 4 (t)], multivariate variational mode decomposition (MVMD) is used to extract the trend components of dam energy, including A31-A36:

[0137] A31, assuming that the 4-element signal x(t) is decomposed into n 4-element intrinsic mode function IMF components u k (t)(k=1,2,…n), that is:

[0138]

[0139] in:

[0140] u k (t)=[u k,1 (t)u k,2 (t)u k,3 (t)u k,4 (t)]

[0141] u k,c (t) = a k,c (t)cos(φ k,c (t)), c = 1, 2, 3, 4

[0142] Where: u k,c (t) is an AM / FM signal; a k,c (t) and φ k,c (t) are the signal u k,c (t) is the amplitude and phase function of the

[0143] A32, for each IMF component u k (t) (k=1,2,…n), using Hilbert transform Obtaining an analytical signal with a single-sided spectrum Right now:

[0144]

[0145] in:

[0146]

[0147]

[0148] Where: j is the imaginary unit.

[0149] Will With exponential term Multiply to adjust its center frequency ω k , so that the spectrum of each IMF component signal is modulated to its corresponding baseband and mixed with the harmonic signal The gradient of L 2 The squared norm estimates the bandwidth of each IMF component.

[0150] A33, with the goal of minimizing the sum of the IMF component bandwidths of all signals, decomposes the 4-element signal x(t) into the following constrained optimization problem, namely:

[0151]

[0152] Where: represents the partial derivative with respect to time t; Indicates L 2 The square of the norm.

[0153] A34, in order to solve the constrained optimization problem in step A33, construct its augmented Lagrangian function, namely:

[0154]

[0155] Where: α is the penalty factor; λ c (t) is the Lagrangian multiplier; represents the inner product.

[0156] A35, using the Alternate Direction Method of MultiplierA (ADMM) to transform the optimization problem in step A34 into a series of sub-optimization problems, and iteratively update each sub-problem in the frequency domain, where the IMF component u k,c , center frequency ω k and the Lagrangian multiplier λ c The update formulas are as follows:

[0157]

[0158]

[0159]

[0160] Where: and u k,c (t), x c (t) and λ c (t); the superscript i is the number of iterations; τ is the update coefficient.

[0161] A36, iterate according to the update formula in step A35 until the condition is met The iteration ends when , and finally the optimal IMF component u of the original 4-element signal x(t) is obtained k (t) and its center frequency ω k (k=1,2,…n). The 4-element IMF component corresponding to the lowest center frequency is the trend component of the dam energy. Right now:

[0162]

[0163] Where: are the average density and trend components of the distribution entropy of the dam’s elastic energy, and They are respectively the average density of dissipated energy and the trend component of distribution entropy.

[0164] A4, determine whether the dam energy has a trend change. If the average dam energy density The trend increases, or the energy distribution entropy If the trend decreases, the next step A5 is entered; otherwise, the dam structure is judged to be normal;

[0165] It is understandable that a dam is an open nonlinear system. The evolution of its structural state is the result of energy exchange between the dam and the external environment. The structural state of the dam is closely related to its energy state. Under the influence of complex internal and external factors, when the dam has structural abnormalities such as cracking, excessive deformation or instability, the dam energy usually shows a trend change, which is specifically manifested as: the total energy of the dam increases, accompanied by energy concentration, and the average energy density increases. The trend increases, and the energy distribution entropy The trend decreases. When this trend change suddenly occurs, it indicates that the dam structure is unstable. Therefore, the structural state of the dam can be evaluated by analyzing the trend component of the dam energy.

[0166] A5, dam structural performance evaluation module cusp catastrophe model stability judgment: using a 4th-order polynomial to fit the trend component of dam energy On this basis, the cusp catastrophe model is used to determine the trend component of dam energy. Whether a mutation occurs. If so, the dam structure is judged to be unstable; otherwise, the dam is judged to be in a stable state.

[0167] This embodiment uses a cusp mutation model to identify the trend change of dam energy, including A51-A55:

[0168] A51, Trend component of dam energy It can be expressed as a continuous function of time t. Taylor expansion of the function is performed and the first 5 terms are obtained:

[0169]

[0170] Where: U is the trend component of dam energy; a i (i=0,1,2,3,4) are the polynomial coefficients.

[0171] A52, using Tschirnhaus transformation, let t = xL, where Then the trend component U of the dam energy becomes:

[0172] U=b 4 x 4 +b 2 x 2 +b 1 x+b 0

[0173] in:

[0174]

[0175] A53, divide both sides of the trend component U of the dam energy in step A52 by 4b 4 , and order Then the standard potential function V of the cusp mutation can be obtained as:

[0176]

[0177] Where x is the state variable of the potential function; u and v are the control variables of the potential function; the equilibrium surface and bifurcation set of the potential function V are as follows: Figure 3 shown.

[0178] A54, take the derivative of the expression of the standard potential function V in step A53 with respect to the state variable x, and set its derivative to zero, and the equilibrium surface of the potential function V is obtained as follows:

[0179] x 3 +ux+v=0.

[0180] like Figure 3As shown, the equilibrium surface consists of three lobes: upper, middle and lower. The upper and lower lobes are stable, while the middle lobe is unstable. No matter which path (u, v) changes along, the phase point (x, u, v) changes in a balanced manner only in the upper lobe (or lower lobe). When it reaches the edge of the lobe, it jumps over the middle lobe. Therefore, the points on the equilibrium surface composed of the vertical tangents constitute the mutation point set (i.e., singular point set) of the state, and its equation is:

[0181]

[0182] A55, the projection of the mutation point set on the control plane constitutes a bifurcation set, which is the set of all points that cause the state variable x to mutate. By combining the potential function V equilibrium surface equation and the mutation point set equation in step A54 and eliminating x, the bifurcation set equation is:

[0183] Δ=4u 3 +27v 2 =0

[0184] It can be understood that when the control variable (u, v) is outside the bifurcation set, that is, Δ>0, the trend component of the dam energy does not mutate, and the dam structure is in a stable state; when the control variable (u, v) is inside the bifurcation set, that is, Δ<0, the trend component of the dam energy mutates, and the dam structure becomes unstable.

[0185] Example 2: See Figure 4 , is a schematic diagram of a dam structural behavior evaluation platform based on energy analysis provided by an embodiment of the present invention, including a dam multi-source information collection and management module, a parameter inversion module and a dam structural behavior evaluation module:

[0186] The dam multi-source information collection and management module is used to collect information on dam structure design, construction, geological survey, etc., and collect environmental monitoring data such as reservoir water level and rainfall measured by water level meters and rain gauges, and dam safety monitoring data such as seepage pressure, seepage volume, displacement, etc. measured by piezometers, water measuring weirs, and displacement meters, and provide corresponding data retrieval services based on query instructions.

[0187] The parameter inversion module is used to invert the constitutive model parameters used by the dam material. The parameter inversion module is an agent model, including orthogonal design, artificial neural network and dam finite element model. According to the dam structure design, construction, geological survey and other data and safety monitoring data provided by the dam multi-source information collection and management module, the inversion parameter combination is obtained by orthogonal design, which is input into the dam finite element model to calculate the dam displacement, and the nonlinear mapping relationship between the dam displacement finite element simulation value and the inversion parameter is established through the artificial neural network. The optimal combination of inversion parameters is finally obtained according to the objective function.

[0188] The dam structure performance evaluation module is used to determine whether the dam structure is in a stable state, including the construction of a comprehensive energy characterization index set, the extraction of the trend component of the dam energy, and the stability judgment of the cusp mutation model. Based on the two types of energy, the dam elastic energy and the dissipated energy, two energy characterization indicators, the average energy density and the energy distribution entropy, are proposed to construct a comprehensive energy characterization index set for the dam structure performance; multivariate variational mode decomposition is used to extract the trend component of the dam energy from the comprehensive energy characterization index set, and a mutation model is established to determine whether the dam structure performance is stable.

[0189] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as limiting the present invention itself. Various changes in form and details may be made without departing from the spirit and scope of the present invention as defined in the appended claims.

Claims

1. A dam structural performance evaluation method based on energy analysis, characterized in that: The steps include: A1. Construct a finite element model of the dam, use monitoring data to invert the mechanical parameters of the dam material, input the mechanical parameters of the dam material into the finite element model of the dam, and calculate the real stress field and displacement field of the dam; A2. Based on the stress field and displacement field, the average density and distribution entropy of the dam's elastic energy and the average density and distribution entropy of the dissipated energy are calculated, and a comprehensive characterization index set of the dam's energy is constructed accordingly; A3. Use multivariate variational mode decomposition to extract the low-frequency components of the comprehensive characterization index set of dam energy and obtain the trend component of dam energy; A4. Determine whether the trend component of the dam energy has a trend change. If so, proceed to the next step. Otherwise, determine that the dam structure is normal. A5. Use polynomial fitting to measure the trend component of dam energy, and use cusp mutation model to identify the mutation of the trend component of dam energy. If a mutation occurs, the dam structure is judged to be unstable, otherwise the dam structure is judged to be normal.

2. The dam structural performance evaluation method based on energy analysis according to claim 1 is characterized in that: The A1 includes: The actual situation of the dam includes the dam structure design, construction, geological survey data, and the construction of a finite element model of the dam under the action of seepage-stress coupling; The monitoring data include environmental monitoring data of the dam site area, the monitoring data are divided into a training sample set and a test sample set according to the time span, and the training sample set is input into the finite element model of the dam; Determine the parameters to be inverted, use the orthogonal design table to formulate N different parameter combinations, input them into the dam finite element model, and obtain the finite element simulation values ​​of the dam displacement under N different parameter combinations; The training samples were normalized and then input into the neural network model for training. After learning and training, the accuracy of the neural network model reached the expected level, and the mean square error after iteration dropped below 0.

05. Thus, the nonlinear mapping relationship between the finite element simulation value of the dam displacement and the inversion parameters was established. Taking the minimum error between the finite element simulation value of displacement and the dam displacement monitoring data as the objective function, the optimal combination of inversion parameters is obtained. The objective function expression is as follows: f=min|H-H * |, Where: f is the objective function, H is the displacement finite element simulation value, and H* is the measured value of displacement; The best combination of the test sample set and the inversion parameters is input into the dam finite element model to obtain the displacement finite element simulation value, which is compared with the actual value in the dam displacement monitoring data to calculate the relative error value. If the relative error meets the requirements, the inversion parameters are output, otherwise the neural network model is returned to re-learning and training; According to the monitoring data of reservoir water level, rainfall and temperature at different times in history, the stress and strain of the dam are calculated based on the finite element model of the dam and the parameter inversion values.

3. The dam structure performance evaluation method based on energy analysis according to claim 2 is characterized in that: The A2 includes: Dam elastic energy U e and dissipated energy U d The change in the dam structure state evolution process satisfies: W = U e +U d , Where: W is the work done by the external load on the dam, U e is the elastic strain energy stored in the dam, U d is the dissipated energy during the deformation of the dam; The finite element model of the dam has N units, where the energy q(k) and volume V(k) of the kth unit are: Where: Ω k is the volume domain of the kth unit, υ is the energy density, and the elastic energy density of the dam υ e and dissipated energy density υ d Calculate as follows: Where: ij , and Respectively stress tensor, elastic strain tensor, plastic strain tensor, rheological strain tensor and wetting strain tensor; The total energy U and total volume V of the dam are obtained by adding up the energy q(k) and volume V(k) of all units: Divide the total energy U of the dam by the total volume V to get the average energy density of the dam Define the energy distribution entropy S: in: Where: λ(k) is the proportion of the energy of the kth unit of the dam in the total energy U, satisfying: Calculate the dam elastic energy U e and dissipated energy U d The average density And distribution entropy S, thereby constructing a comprehensive characterization index set U of dam energy I : Where: and S e The dam elastic energy U e The average density and distribution entropy of and S d are the dam dissipated energy U d The average density and distribution entropy of .

4. The dam structure performance evaluation method based on energy analysis according to claim 3 is characterized in that: The A3 adopts multivariate variational mode decomposition to extract the trend components of dam energy, including: The 4-element energy signal U corresponding to the dam energy characterization index set I (t) is uniformly denoted as x(t) = [x1(t)x2(t)x3(t)x4(t)], Assume that the 4-element signal x(t) is decomposed into n 4-element intrinsic mode function IMF components u k (t)(k=1,2,…n), then: where: u k (t) = [u k,1 (t)u k,2 (t)u k,3 (t)u k,4 (t)], u k,c (t)=a k,c (t)cos(φ k,c (t)),c=1,2,3,4, Where: u k,c (t) is an AM / FM signal; a k,c (t) and φ k,c (t) are the signals u k,c (t) Amplitude and phase functions; For each IMF component u k (t)(k=1,2,…n), using Hilbert transform Obtaining an analytical signal with a single-sided spectrum in: Where: j is the imaginary unit; Will With exponential term Multiply to adjust the center frequency ω k , so that the spectrum of each IMF component signal is modulated to the corresponding baseband and the harmonic mixing signal The L2 norm square of the gradient estimates the bandwidth of each IMF component, With the goal of minimizing the sum of the IMF component bandwidths of all signals, the 4-element signal x(t) is decomposed into the following constrained optimization problem: Where: represents the partial derivative with respect to time t; represents the square of L2 norm; To solve the constrained optimization problem, construct its augmented Lagrangian function: Where: α is the penalty factor; λ c (t) is the Lagrangian multiplier; <, > represents the inner product; The alternating direction multiplier method is used to transform the optimization problem into a series of sub-optimization problems, and each sub-problem is iteratively updated in the frequency domain, where the IMF component u k,c , center frequency ω k and the Lagrangian multiplier λ c The update formulas are as follows: Where: and u k,c (t), x c (t) and λ c (t) is the Fourier transform; the superscript i is the number of iterations; τ is the update coefficient; Press u k,c ,ω k and λ c The update formula is iterated repeatedly until the condition is met The iteration ends when , and finally the optimal IMF component u of the original 4-element signal x(t) is obtained k (t) and its center frequency ω k (k=1,2,…n), where the 4-element IMF component corresponding to the lowest center frequency is the trend component of the dam energy Where: is the trend component of the mean density, is the trend component of the distribution entropy.

5. The dam structure performance evaluation method based on energy analysis according to claim 1 is characterized in that: Determine whether the dam energy has changed in a trend, including the average dam energy density The trend increases, or the energy distribution entropy The trend is decreasing.

6. The dam structure performance evaluation method based on energy analysis according to claim 5 is characterized in that: A fourth-order polynomial is used to fit the trend component of the dam energy. The cusp catastrophe model is used to determine the trend component of dam energy Whether a mutation occurs. If so, the dam structure is considered unstable. Otherwise, the dam is considered stable, including: Trend component of dam energy Expressed as a continuous function of time t, Taylor expansion is performed and the first 5 terms are obtained: Where: U is the trend component of dam energy, a i (i=0,1,2,3,4) are polynomial coefficients; Using Tschirnhaus transformation, let t = xL, where Then the trend component U of the dam energy becomes: U=b4x 4 +b2x 2 +b1x+b0 in: Divide both sides of the trend component U of the dam energy by 4b4 and let Then the standard potential function V of the cusp mutation is expressed as: Where: x is the state variable of the potential function; u and v are the control variables of the potential function; Derivative the expression of the standard potential function V with respect to the state variable x and set its derivative to zero, and the equilibrium surface of the standard potential function V is obtained as follows: x 3 +ux+v=0, The set of mutation points of the state formed by the points of the vertical tangent on the equilibrium surface is: The projection of the mutation point set on the control plane constitutes a bifurcation set, which is the set of all points that cause the state variable x to mutate. Using the standard potential function V to balance the surface equation and the mutation point set equation to eliminate x, the bifurcation set equation is: Δ=4u 3 +27v 2 =0, When Δ>0, the control variable (u, v) is outside the bifurcation set, and the trend component of the dam energy does not change suddenly, so the dam structure is judged to be normal; when Δ<0, the control variable (u, v) is inside the bifurcation set, the trend component of the dam energy changes suddenly, so the dam structure is judged to be unstable.

7. A platform for implementing the dam structure performance evaluation method based on energy analysis as described in claim 1, characterized in that: It includes dam multi-source information collection and management module, parameter inversion module and dam structural performance evaluation module. The dam multi-source information collection and management module is used to collect the actual situation of the dam and the dam safety monitoring data. The actual situation of the dam includes the dam structure design, construction, and geological survey related data. The dam safety monitoring data includes the reservoir water level and rainfall environmental monitoring data measured by the water level gauge and rain gauge, and the seepage pressure, seepage volume, and displacement dam safety monitoring data measured by the osmometer, water measuring weir, and displacement meter. The parameter inversion module includes an orthogonal design module, an artificial neural network and a dam finite element model. The orthogonal design module uses orthogonal design to obtain an inversion parameter combination based on the actual dam situation and dam safety monitoring data provided by the dam multi-source information collection and management module, inputs the combination into the dam finite element model to calculate the dam displacement, and establishes a nonlinear mapping relationship between the dam displacement finite element simulation value and the inversion parameter through the artificial neural network, and finally obtains the best combination of inversion parameters according to the objective function; The dam structure performance evaluation module is used to determine whether the dam structure is in a stable state, including constructing an energy comprehensive characterization index set, extracting the trend component of the dam energy and the stability judgment of the cusp mutation model. On the basis of two types of energy, namely, the elastic energy and the dissipated energy of the dam, an energy comprehensive characterization index set of the dam structure performance is constructed through two energy characterization indicators, namely, the average energy density and the energy distribution entropy. Multivariate variational modal decomposition is used to extract the trend component of the dam energy from the energy comprehensive characterization index set, and a cusp mutation model is established to determine whether the dam structure performance is stable.

Images