Deformation estimation device, deformation estimation system and deformation estimation method

Abstract

A deformation estimating device (60) comprises a control unit (50) that estimates a deformation occurring in a structural body (1) based on a measurement representing a deformation on a measurement plane of a surface of the structural body (1). The control unit (50) predicts deformation information, which—by means of a probability distribution—represents a deformation occurring on the measurement plane of the structural body (1) under a plurality of boundary conditions applied to the structural body (1).The control unit (50) specifies a first observation equation which represents a correlation between a first observation variable, which represents the measured value, an error of the measured value by a probability distribution, and the deformation prediction information, which serves as a first state variable, and estimates a deformation that occurs in the structural body, based on the first observation equation.

Description

Technical field

[0001] The present invention relates to a deformation estimation device, a deformation estimation system and a deformation estimation method. State of the art

[0002] For example, accurately estimating and detecting deformations occurring in various structural bodies, such as drive equipment, steel towers, and steel bridges, is important for lifetime evaluation, maintenance, and repair work. However, prediction errors, measurement errors, and the like occur in the process of predicting and measuring deformation in such structural bodies. To reduce such errors and accurately measure deformation occurring in a structural body, a material tester is known as a deformation estimation device with the following configuration.

[0003] In a measurement method for measuring a displacement field and a stress field using the material tester as a conventional deformation estimation device, a fixed perturbation pattern in an image of a test specimen, captured in grayscale by a video camera, is stochastically modeled on the basis of darkness, and the uncertainty of a measured value of the centroid of the points brought onto a surface of the test specimen when a load is applied is evaluated using the stochastic model of the fixed perturbation pattern.

[0004] From the measurement result for the center of gravity of the points, a point displacement value, a displacement field, and a stress field are then stochastically calculated based on an uncertainty propagation rule. By stochastically performing a measurement based on a point-center-of-gravity tracking method, the uncertainty propagates to each physical quantity derived by calculation from an observed value, and the reliability of the measured value is presented to a user (see, for example, patent document 1). Bibliography Patent document

[0005] Patent document 1: WO2015 / 049757 A1 (paras.

[0028] to

[0032] ) Summary of the invention Problem to be solved with the invention

[0006] However, in the measurement procedure for measuring a displacement field and a stress field using the material tester as a conventional deformation estimation device, as described above, an error that occurs in a process for measuring a structural body cannot be sufficiently reduced, so that it may be impossible to accurately estimate a deformation that occurs in the structural body.

[0007] The present invention was designed to solve the above problem, and the object of the present invention is to provide a deformation estimation device, a deformation estimation system and a deformation estimation method that can accurately estimate a deformation occurring in a structural body. Ways to solve the problem

[0008] A deformation estimation device according to the present invention comprises a control unit that estimates a deformation occurring in a structural body based on a measured value representing a deformation on a measurement plane of a surface of the structural body. The control unit predicts deformation information, which—by means of a probability distribution—represents a deformation occurring on the measurement plane of the structural body under a plurality of boundary conditions applied to the structural body.The control unit specifies a first observation equation, which represents a correlation between a first observation variable, which represents the measured value, an error of the measured value through a probability distribution, and the deformation prediction information, which serves as a first state variable, and estimates a deformation that occurs in the structural body, based on the first observation equation.

[0009] A deformation estimation system according to the present invention comprises: a deformation estimation device configured as described above; and an output unit that displays an output from the control unit. The control unit specifies—among the longitudinal modulus of elasticity, density, and Poisson's ratio of the structural body—at least the longitudinal modulus of elasticity and the Poisson's ratio as material property values ​​of the structural body when predicting the deformation. The control unit performs a solidity determination to ascertain whether the probability distribution of an estimated deformation occurring in the structural body and its error exceeds a first threshold value, which serves as an index value representing the solidity. The control unit displays the result of the determination on the output unit.

[0010] A deformation estimation method according to the present invention is a deformation estimation method that uses the deformation estimation device described above, wherein the method comprises the following steps: the control unit predicts deformation prediction information, which—by means of a probability distribution—represents a deformation occurring on the measurement plane of the structural body under a plurality of boundary conditions applied to the structural body; and the control unit specifies a first observation equation, which represents a correlation between a first observation variable (u*), representing the measured value, and an error of the measured value by means of a probability distribution, and the deformation prediction information, which serves as a first state variable, and estimates a deformation occurring in the structural body based on the first observation equation. Effect of the invention

[0011] With the deformation estimation device and the deformation estimation method according to the present invention, it becomes possible to specify a deformation estimation device and a deformation estimation method that can accurately estimate a deformation that occurs in a structural body.

[0012] The deformation estimation system according to the present invention accurately estimates deformation occurring in a structural body and determines whether the deformation is equal to or less than a tolerable value. Consequently, a manager can appropriately perform service life evaluations, maintenance, repairs, and the like for a structural body. Brief description of the drawings Fig. Figure 1 is a block diagram showing a schematic configuration of a deformation estimation device according to embodiment 1 and a deformation estimation system comprising the deformation estimation device. Fig. Figure 2 is a flowchart showing the sequence of a control process of a variable output unit included in the deformation estimating device according to embodiment 1. Fig. Figure 3 is a flowchart showing the sequence of a control process of a model generation unit included in the deformation estimation device according to embodiment 1. Fig. Figure 4 is a flowchart showing the sequence of a control process of an estimating unit contained in the deformation estimating device according to embodiment 1. Fig. Figure 5 is a perspective sectional view showing a partial section of a structural body that is a verification target in embodiment 1. Fig. Figure 6A illustrates a process for specifying a boundary condition in a cross-section of the structural body in embodiment 1. Fig. Figure 6B illustrates a process for specifying a boundary condition in a different cross-section of the structural body in embodiment 1. Fig. Figure 7 is a flowchart showing the sequence of a control process of the estimating unit contained in the deformation estimating device according to embodiment 1. Fig. Figure 8 is a flowchart showing the sequence of a control process of the estimating unit contained in the deformation estimating device according to embodiment 1. Fig. Figure 9 is a block diagram showing a schematic configuration of a deformation estimation device according to embodiment 2 and a deformation estimation system comprising the deformation estimation device. Fig. Figure 10 is a block diagram showing a schematic configuration of a measuring unit included in the deformation estimation device according to embodiment 2. Fig. Figure 11 is a flowchart showing the sequence of a control process of a variable output unit contained in a measuring unit according to embodiment 2. Fig. Figure 12 is a flowchart showing the sequence of a control process of a model generation unit contained in a measuring unit according to embodiment 2. Fig. Figure 13 is a flowchart showing the sequence of a control process of the measuring unit according to embodiment 2. Fig. Figure 14 conceptually illustrates a control process of the measuring unit for a structural body, which is a verification target in embodiment 2. Fig. Figure 15 is a block diagram showing a schematic configuration of a deformation estimation device according to embodiment 3 and a deformation estimation system comprising the deformation estimation device. Fig. Figure 16 is a flowchart showing the sequence of a control process of an estimating unit and a crack estimating unit included in the deformation estimating device according to embodiment 3. Fig. Figure 17 is a block diagram showing a schematic configuration of a deformation estimation system according to embodiment 4. Fig. Figure 18 shows an example of a hardware configuration of a control unit. Fig. Figure 19 shows an example of a hardware configuration of the control unit. Description of embodiments Embodiment 1

[0013] Preferred embodiments of a deformation estimation device, a deformation estimation system, and a deformation estimation method according to the present embodiment are described below with reference to the drawings. Identical or corresponding elements and parts are designated with the same reference numerals, and their detailed descriptions are omitted. Similarly, in the other embodiments, components designated with the same reference numerals are not described repeatedly.

[0014] Fig. Figure 1 is a block diagram showing a schematic configuration of a deformation estimation device 60 according to embodiment 1 and a deformation estimation system 100 comprising the deformation estimation device 60.

[0015] Fig. Figure 2 is a flowchart showing the sequence of a control process for a variable output unit 20, which is contained in the deformation estimating device 60, which is in Fig. 1 is shown.

[0016] Fig. Figure 3 is a flowchart showing the sequence of a control process in a model generation unit 30, which is contained in the deformation estimation device 60, which is in Fig. 1 is shown.

[0017] Fig. Figure 4 is a flowchart showing the sequence of a control process for an estimating unit 40 contained in the deformation estimating device 60, which is located in Fig. 1 is shown.

[0018] Fig. Figure 5 is a perspective section view showing a partial section of a structural body 1, which is a verification target.

[0019] Fig. Figure 6A illustrates a process for specifying a boundary condition in a cross-section 1A of the structural body, which is located in Fig. 5 is shown.

[0020] Fig. Figure 6B illustrates a process for specifying a boundary condition in a cross-section 1B of the structural body, which is located in Fig. 5 is shown.

[0021] The deformation estimating device 60 of the present embodiment, which is in Fig. Figure 1 shows an estimate of a deformation that occurs in the structural body 1, which is the verification target.

[0022] The deformation estimation system 100 of the present embodiment comprises the deformation estimation device 60 and an output unit 5 which outputs an estimation result from the deformation estimation device 60, so that it reports the estimated deformation to a manager who performs maintenance and repair work via the output unit 5.

[0023] Output unit 5, which notifies a manager, can have a display function, like a monitor, or it can have an audio output function, like a loudspeaker. The configuration of output unit 5 is not restricted, as long as output unit 5 has a function to report an estimation result from the deformation estimation device 60 to a manager.

[0024] Below the planes 1S, 1A, 1B, 1C 1D of the structural body 1, which is shown in a partially cutaway perspective view of the structural body 1, which is in Fig. Figure 5 shows a plane that can be directly observed, surface 1S, and planes that cannot be directly observed are the cross-sections 1A, 1B, 1C, and 1D of the structural body 1 along the X-axis and Y-axis directions. The deformation estimating device 60 measures a measurement plane 1M, which is a predefined area on surface 1S that can be observed, as described below.

[0025] A configuration of the deformation estimation device 60 is described below.

[0026] As in Fig. As shown in Figure 1, the deformation estimating device 60 has a measuring unit 10 that measures a deformation on the measuring plane 1M of the surface 1S of the structural body 1 as a measured value, and a control unit 50 that initiates a deformation occurring in the structural body 1 on the basis of the measured value.

[0027] Although the detailed configuration of the measuring unit 10 is not limited, for example the position of a point on the measuring plane 1M can be recorded before a deformation occurs, the position of that point after the deformation has occurred can be recorded, and then the deformation can be obtained from a displacement of the position.

[0028] The control unit 50 comprises the variable output unit 20, the model generation unit 30, and the estimation unit 40. The control processes in these units are described below.

[0029] First, the control process of the variable output unit 20 is described with reference to the flowchart of the control process of the variable output unit 20, which is shown in Fig. 2 is shown.

[0030] Variable output unit 20 generates a shape model, obtained by modeling the structure and shape of structural body 1, and specifies a plurality of boundary conditions for the shape model. Here, a load condition, specifying the application states of an external force, a forced displacement, a temperature change, and the like, exerted on structural body 1, and a constraint condition for restrictive displacements in directions of structural body 1 and its direction of rotation are specified together as a single boundary condition f (step S21). In the processing in step S21, variable output unit 20 specifies a plurality of boundary conditions fi (i = 1, 2, 3, ..., n) that assume various stresses, loads, and constraints exerted on structural body 1.

[0031] Next, the variable output unit 20 selects a boundary condition fi from the given plurality of boundary conditions fi (step S22).

[0032] Next, the variable output unit 20 performs a numerical analysis to analyze the structure of the structural body 1 based on the selected boundary condition fi (step S23).

[0033] In the numerical analysis, the variable output unit 20 specifies the cross-section 1A and the cross-section 1B, which are in Fig. 5 are shown and are planes on which the boundary condition fi is applied to the shape model.

[0034] The variable output unit 20 performs a numerical analysis, for example using the following finite element method.

[0035] The variable output unit 20 divides the cross-section 1A of the structural body 1 into a plurality of square unit planes 2 in a grid shape, which are formed from n unit planes in the X-axis direction and m unit planes in the Y-axis direction, as shown in Fig. 6A shown. The variable output unit 20 divides the cross-section 1B into a plurality of square unit planes 2 in a grid shape, which are formed from c unit planes in the X-axis direction and d unit planes in the Y-axis direction, as shown in Fig. 6B shown.

[0036] The unit planes 2 are shown with square shapes, but without restriction, the unit planes 2 can also have trapezoidal shapes, for example.

[0037] Then, points where parallel lines intersecting cross-sections 1A and 1B in the majority of unit planes 2, i.e., vertices of each unit plane 2, are defined as nodes 2P. In cross-section 1A, nodes 2P are designated by numbers (0, 0) to (n, m), and in cross-section 1B, nodes 2P are designated by numbers (0, 0) to (c, d).

[0038] Then the variable output unit 20 applies the selected boundary condition fi to each node 2P in the cross-sections 1A, 1B and performs a numerical analysis to predict a deformation vector ui, which serves as a first predictive deformation indicating a deformation that occurs on the measurement plane 1M under the boundary condition fi.

[0039] In the numerical analysis, the variable output unit 20 specifies and stores a boundary condition vector fIi, as shown in the following formula (1), where the boundary conditions of node 2P in cross sections 1A, 1B are represented together as one vector. Mathematical expression 1 {fIi}=(fIi,A(0,0)⋮fIi,A(n,m)fIi,B(0,0)⋮fIi,B(n,m))

[0040] When specifying the boundary condition vector fIi in the above formula (1), for example in cross-section 1A, the boundary condition imposed on node 2P(0, 0) is specified as a boundary condition fIi,A(0, 0), and the boundary condition imposed on node 2P(n, m) is specified as a boundary condition fIi,A(n, m). Furthermore, in cross-section 1B, the boundary condition imposed on node 2P(0, 0) is a boundary condition fIi,B(0, 0), and the boundary condition imposed on node 2P(n, m) is a boundary condition fIi,B(c, d).

[0041] Next, the variable output unit 20 stores the deformation vector ui on the measurement plane 1M, which is predicted by the numerical analysis (step S24).

[0042] The deformation vector ui can be a vector of displacement change values ​​representing the difference between before and after a deformation at each node occurring on the measurement plane 1M, or the like.

[0043] Next, the variable output unit 20 determines whether or not—for all boundary conditions fi specified in step S21—the numerical analysis shown in steps S22 to S24 has been performed and the deformation vectors ui have been predicted (step S25). If there is a boundary condition fi that has not yet been subjected to a numerical analysis (step S25: NO), then the number i of the boundary condition fi is incremented by 1 in step S25A, and the process from step S22 to step S25 is repeated.

[0044] After the deformation vectors ui (i = 1, 2, ..., n) have been calculated for all of the given boundary conditions fi (i = 1, 2, ..., n) (Step S25: YES), a matrix F[fI1, ..., fIi, ..., fIn] and a matrix U[u1, ..., ui, ..., un] are generated from the boundary condition vectors fli and the deformation vectors ui, and a matrix A is derived, which is shown by the following formula (2), which gives the correlation between the matrix F and the matrix U (Step S26). In formula (2), the matrix F is a pseudoinverse matrix. Mathematical expression 2 A=UF†

[0045] Next, the variable output unit 20 derives a basis vector Ur of a linear space of the deformation field of the measurement plane 1M based on the matrix A. The basis vector Ur can be used with reduced rank, allowing the deformation field to be expressed with the minimum necessary information by singular value decomposition (step S27).

[0046] Next, the variable output unit 20 derives the probability distribution of a coefficient vector fvr as an element forming a linear space of the deformation field by superimposing the basis vector Ur and an uncertainty including a prediction error and the like of an initial value of the coefficient vector fvr (step S28).

[0047] The coefficient vector fvr is derived as a coefficient for superimposing the basis vector Ur, but a scalar can also be used instead of a vector.

[0048] Parallel to step S28 above, the variable output unit 20 acquires a displacement vector u* as a measured value, which is input via a network from the measuring unit 10 and represents the actual deformation on the measuring plane 1M of the structural body 1. The variable output unit 20 then derives a probability distribution of the displacement vector u* and an uncertainty ε including a measurement error and the like for the displacement vector u* (step S29).

[0049] Next, the variable output unit 20 outputs the derived basis vector Ur, the derived probability distribution including the uncertainty of the coefficient vector fvr, and the probability distribution including the uncertainty of the displacement vector u* as variables (step S30).

[0050] The output variables are entered into the model generation unit 30 at the next stage.

[0051] Next, the control process of the model generation unit 30 is described with reference to the flowchart of the control process of the model generation unit 30, as shown in Fig. 3 shown.

[0052] First, the model generation unit 30 captures the basis vector Ur, the probability distribution including the uncertainty of the coefficient vector fvr and the probability distribution including the uncertainty ε of the displacement vector u*, which are the variables that were entered by the variable output unit 20 in the previous stage (step S31).

[0053] Next, the model generation unit 30 specifies the coefficient vector fvr as a first state variable and computes a first state equation representing a time transition of the coefficient vector fvr, using the coefficient vector fvr and the uncertainty of the initial value (step S32).

[0054] As an example, an equation of state in a case where the deformation of the structural body does not change over time is shown by the following formula (3).

[0055] Here, the uncertainty of the initial value of the coefficient vector fvr is contained in {fvr}. Mathematical expression 3 {fvr}t={fvr}t−1

[0056] Next, the model generation unit 30 specifies the displacement vector u* as a first observation variable and computes a first observation equation that represents the correlation between the displacement vector u* and the coefficient vector fvr (step S33). The first observation equation is shown in the following formula (4). Mathematical expression 4 {u*}t=Ur{fvr}t+ε

[0057] As shown in formula (4) above, the displacement vector u*, which serves as the first observation variable in the first observation equation, is expressed by multiplying the coefficient vector fvr, which serves as the first state variable in the first state equation, with the basis vector Ur, which forms the deformation field.

[0058] Next, the model generation unit 30 generates a first state space model to estimate a deformation occurring in the structural body 1, using the first state equation and the first observation equation (step S34).

[0059] The generated first state space model is entered into the estimation unit 40 at the subsequent stage.

[0060] Next, the control process of estimating unit 40 will be described with reference to the flowchart of the control process of estimating unit 40, which is shown in Fig. 4 is shown.

[0061] First, the estimation unit 40 captures the first state space model, entered by the model generation unit 30 at the previous stage (step S41).

[0062] Next, the estimation unit 40 calculates a first prior distribution based on the first state-space model (step S42). That is, it calculates the probability distribution that includes the uncertainty of the coefficient vector fvr, which serves as the first state variable in the first state equation.

[0063] Next, the estimation unit 40 updates the first prior distribution by a first likelihood function and derives a first posterior distribution (step S43).

[0064] Regarding the first likelihood function, the actual deformation on the measurement plane 1M is measured as a measurement by the measuring unit 10 (step S45), and from the measurement the first likelihood function is derived as a function that represents a likelihood distribution of the displacement vector u* and the uncertainty ε, which is its error (step S46).

[0065] Next, the estimation unit 40 determines whether or not a measurement on the measurement plane 1M has been carried out a predetermined first number of times in step S45 (step S44).

[0066] If the number of measurements has not reached the specified first number of times (step S44: NO), then the A posteriori distribution derived in step S43 is specified as a new A priori distribution (step S47).

[0067] If a measurement on the measurement plane 1M has been performed the specified first number of times and consequently an update of the prior distribution by the first likelihood function has been performed the specified first number of times (step S44: YES), then the estimation unit 40 estimates the deformation occurring in the structural body 1 on the basis of the updated posterior distribution (step S48).

[0068] The deformation in structural body 1 can be derived by multiplying the coefficient vector fvr and the basis vector Ur, derived on the basis of the a posteriori distribution.

[0069] Examples of methods for calculating the a posteriori distribution include: recursive Bayesian estimation, a particle filter, and MCMC (Markov chain Monte Carlo). In a case where the deformation to be predicted is an elastic deformation and a measurement error and a system error, assumed in a state-space model, can be assumed to behave according to a Gaussian distribution, a Kalman filter allows an estimation to be performed with reduced computational costs.

[0070] Next, the estimation unit 40 outputs the estimated deformation (step S49).

[0071] The deformation, estimated by the estimation unit 40, is fed via the network into the output unit 5, which is in Fig. 1 is shown, and the output unit 5 displays the estimated deformation on a display or the like and thereby reports it to a manager.

[0072] Consequently, the manager can appropriately carry out maintenance and repair work on structural body 1 based on the estimated deformation.

[0073] The following describes a control process that differs from the control process of the estimation unit 40, which is in Fig. 4 is shown.

[0074] Fig. Figure 7 is a flowchart showing another example of the control process of the estimating unit 40 of the present embodiment.

[0075] Compared to the control process of the estimating unit 40, which is shown above in Fig. As shown in Figure 4, steps S48-A1, S48-A2, and S48-A3 have been added. The control operations in steps S41 to S48 are the same as those in Figure 4. Fig. 4, and therefore their description is omitted.

[0076] After the deformation in structural body 1 has been estimated by multiplying the coefficient vector fvr and the basis vector Ur in step S48, the estimation unit 40 calculates an error that is the difference between the estimated deformation and the displacement vector u* measured by the measuring unit 10 (step S48-A1). The calculated error is stored in the estimation unit 40.

[0077] Next, the estimation unit 40 determines whether or not the calculated error is the smallest among the past stored errors (step S48-A2).

[0078] If the calculated error is not the smallest (step S48-A2: NO), the estimation unit 40 changes the value of a parameter that is given as a constant and assigns the displacement vector u* and the coefficient vector fvr to each other in the first observation equation in the above formula (4) (step S48-A3).

[0079] The control process from step S41 to step S48-A1 is repeated, using the first state space model configured with the new parameter value, and the estimation unit 40 changes the parameter value until the calculated error is minimized.

[0080] When the error has become the smallest (step S48-A2: YES), then the estimation unit 40 outputs the deformation in structural body 1, which is estimated in step S48 when the error has become the smallest (step S49).

[0081] As described above, the deformation estimation is performed on the basis of the first state space model, where an optimal value for an unknown parameter in the first state space model is specified so that the deformation can be estimated accurately with reduced error.

[0082] When determining the smallest value for the error in step S48-A2, a past error is not stored as a comparison target at the time of the initial error calculation. Therefore, without performing the smallest value determination, an error can be calculated each time the parameter is changed a predetermined number of times. After the errors for the respective parameter values ​​have been stored, the smallest value determination can then be performed in step S48-A2.

[0083] If the displacement vector u* measured in step S45 is stored, it is not necessary to repeat the measurement when determining the smallest error value in step S48-A2. The determination of the smallest error value in step S48-A2 can be performed using the stored displacement vector u*.

[0084] The above description performs a deformation estimation based on the first state-space model, which is configured using the first observation equation and the first state equation. However, the deformation estimation can also be performed using only the first observation equation. In this case, steps S32 for computing the first state equation and S32 for generating the first state-space model are omitted.

[0085] In the above description, the displacement vector u* is specified as the first observation variable, and the first observation equation, which gives the correlation between the displacement vector u* and the coefficient vector fvr, which serves as the first state variable, is used, but the present invention is not limited thereto. The first state variable, which is different from the coefficient vector fvr, can also be the deformation vector ui or another element that forms the linear space of the deformation field of the measurement plane 1M, e.g., the basis vector Ur.This means that, as deformation prediction information, which specifies the prediction of a deformation occurring on the measurement plane of the structural body under a plurality of boundary conditions imposed on the structural body, at least one of the coefficient vector fvr, the deformation vector u*i, or the basis vector Ur can be used, and this can be used as the first state variable. The first observation equation can specify the correlation between the first observation variable, which is the displacement vector u*, and the above deformation prediction information, which serves as the first state variable.

[0086] In a case where the basis vector Ur is used as the first state variable, the basis vector Ur can specify a probability distribution.

[0087] The following describes a control process that differs from the control process of the estimation unit 40, which is in Fig. 4 is shown.

[0088] Fig. Figure 8 is a flowchart showing another example of the control process of the estimating unit 40 of the present embodiment.

[0089] Compared to the control process of the estimating unit 40, which is shown above in Fig. As shown in Figure 4, step S48-B1 has been added. A measuring unit 10 is an imaging device capable of capturing an image of the measuring plane 1M of the structural body 1.

[0090] Here, in a stage that precedes the control process of the estimation unit 40, the boundary conditions are specified in the variable output unit 20, which are in Fig. Figure 2 shows that at least the longitudinal modulus of elasticity and the Poisson's ratio are specified as material property values ​​of the structural body 1. The estimation unit 40 performs a deformation estimation described below, based on the first state-space model and using the variables that reflect the material property values.

[0091] In the processing in step S42, the estimation unit 40 derives a probability distribution of at least one displacement, tilt or inclination or stress, which is a deformation occurring in the structural body 1, as well as its error, as the first a priori distribution, based on the probability distribution of the first state variable and its error, estimated using the first state space model.

[0092] In the processing step S46, the estimation unit 40 derives a probability distribution of at least one of the displacement, tilt or inclination or stress of the structural body 1 and their errors as the likelihood distribution of the first likelihood function, based on an image of the measurement plane 1M, captured by the measurement unit 10.

[0093] Regarding the measurement of stress, for example, images before and after deformation on the measurement plane 1M can be subjected to an analytical calculation, and thus the displacement on the measurement plane 1M can be calculated so that the stress can be visualized.

[0094] In step S43, the estimation unit 40 updates the first a priori distribution, which represents the probability distribution of at least one of the displacement, tilt or inclination, or tension, by the first likelihood function, which represents the probability distribution of at least one of the displacement, tilt or inclination, or tension, derived from images, so that the first a posteriori distribution is derived.

[0095] Then the estimation unit 40 determines the solidity of the structural body 1 on the basis of the deformation occurring in the structural body, estimated on the basis of the first a posteriori distribution (S48-B 1).

[0096] In the solidity determination, the estimation unit 40 determines whether the estimated probability distribution of the deformation and its errors exceeds a predefined first threshold. The first threshold is defined as an index value representing whether the structural body 1 is in a state where solidity can be guaranteed, and a tolerable value is specified for at least one of the deformations estimated as a stress, tilt, or displacement.

[0097] In a case where the stress is specified as the first threshold value, the estimation unit 40 outputs the probability distribution representing the stress distribution as the deformation in the structural body, estimated based on the first a posteriori distribution. The stress distribution is a differential value of the displacement and can therefore represent a very small displacement. Consequently, the deformation distribution on the measurement plane 1M can be determined in greater detail, and the solidity of the structural body 1 can be accurately determined.

[0098] For at least one of the stresses, inclinations, or displacements, estimated as deformations, the solidity is determined based on the probability distribution. Consequently, the reliability can be calculated, taking into account uncertainties such as an external force exerted on the structural body 1, and the actual solidity of the structural body 1 can be quantitatively determined from the solidity calculated above.

[0099] The estimation unit 40 outputs the deformation estimated in step S48 and the solidity of the structural body 1 determined in step S48-B1 (step S49).

[0100] Consequently, a manager can appropriately perform a service life evaluation, maintenance, upkeep work and the like for structural body 1 based on the issued solidity.

[0101] The deformation estimation device of the present embodiment, configured as described above, comprises a control unit that estimates a deformation occurring in a structural body based on a measurement representing a deformation on a measurement plane of a surface of the structural body. The control unit predicts deformation information, which—through a probability distribution—represents a deformation occurring on the measurement plane of the structural body under a plurality of boundary conditions applied to the structural body.The control unit specifies a first observation equation, which represents a correlation between a first observation variable, representing the measured value, an error of the measured value through a probability distribution, and the deformation prediction information, and estimates a deformation that occurs in the structural body, based on the first observation equation.

[0102] As described above, the control unit predicts a deformation occurring on the measurement plane of the structural body under a plurality of boundary conditions, including an external force applied to the structural body, as deformation prediction information. That is, the control unit specifies a plurality of boundary conditions and performs a numerical analysis so that the deformation prediction information is generated in advance for each of the specified boundary conditions.

[0103] The control unit then specifies the first observation equation, which represents the correlation between the deformation prediction information obtained for each boundary condition and the first observation variable, which represents a deformation derived from a measurement. The control unit then estimates a deformation occurring in the structural body based on this first observation equation.

[0104] This means that the deformation prediction information, which is used as the first state variable in the first observation equation, is derived based on various and multiple boundary conditions assumed to impose on the structural body. Consequently, it becomes possible to accurately estimate a deformation of the structural body.

[0105] Furthermore, the first state variable is represented by a probability distribution including its errors, and therefore it takes on a value for which an uncertainty, such as an external force acting on the structural body, is also taken into account.

[0106] In the deformation estimation device of the present embodiment, which is configured as described above, the following applies: By using the deformation prediction information as the first state variable, the control unit specifies a first state equation representing the time transition of the first state variable, specifies the first observation equation representing the correlation between the first observation variable and the first state variable, and generates a first state space model for estimating a deformation of the structural body based on the first state equation and the first observation equation.The control unit provides a probability distribution of the first state variable and its errors, estimated from the first state space model, as a first a priori distribution, derives a first likelihood function with respect to a probability distribution of a displacement of the structural body and its errors obtained from the measured value, derives a first a posteriori distribution based on the first a priori distribution and the first likelihood function, and estimates a deformation occurring in the structural body based on the first a posteriori distribution.

[0107] As described above, the variable specified as the first state variable in the first state equation is derived based on various and multiple boundary conditions, which are assumed to impose themselves on the structural body. Consequently, it becomes possible to accurately estimate a deformation of the structural body.

[0108] Furthermore, the first state variable is represented by the probability distribution including its errors, and therefore it takes on a value for which an uncertainty, such as an external force acting on the structural body, is also taken into account.

[0109] Furthermore, the first likelihood function is derived with respect to the probability distribution of a deformation of the structural body and its errors, obtained from the measured value of the measuring unit. The first a posteriori distribution is derived based on the first a priori distribution and the first likelihood function, and a deformation occurring in the structural body is estimated based on the first a posteriori distribution. Consequently, the first a priori distribution is updated using the first likelihood function, which represents the probability distribution of the deformation that most closely matches the information contained in the first a priori distribution, thus enabling a more accurate deformation estimation.

[0110] In the deformation estimation device of the present embodiment, which is configured as described above, the control unit - as the first state variable - derives a probability distribution which represents a first predicted deformation, which represents a difference between before and after a deformation occurring on the measurement plane under a plurality of boundary conditions imposed on the structural body, or a correlation between the boundary condition and the first predicted deformation.

[0111] As described above, the deformation vector ui can be used as the first predicted deformation obtained through numerical analysis. This vector represents the deformation occurring at the measurement plane of the structural body. Alternatively, it can be used as the first state variable, representing the correlation between the deformation vector ui and the boundary condition fi. By using the deformation vector ui, derived from the measurement or from another relevant element, as the first state variable, it becomes possible to perform manifold deformation estimations based on the state of the structural body, the boundary condition, and other factors, thus enabling more accurate deformation estimations.

[0112] In the deformation estimation device of the present embodiment, which is configured as described above, the control unit derives a probability distribution of an element forming a deformation field of the measurement plane with respect to the boundary condition and an error of the element, as a probability distribution that represents the correlation between the boundary condition and the first predicted deformation.

[0113] This means that an element representing the correlation between the boundary condition and the deformation vector ui, which is the first predicted deformation and specified as the first state variable, is used to form the deformation field of the structural body with respect to a plurality of different and multiple boundary conditions. Consequently, it becomes possible to estimate the deformation of the structural body more accurately. Furthermore, the element forming the deformation field is represented by the probability distribution, including its errors, and therefore becomes a value for which an uncertainty, such as an external force exerted on the structural body, is also taken into account.

[0114] In the deformation estimation device of the present embodiment, which is configured as described above, the control unit derives a coefficient of each basis vector forming a linear space in the deformation field, namely those elements that form the deformation field.

[0115] As described above, the probability distribution of the coefficient of each basis vector forming the linear space of the deformation field and the error of the coefficient are derived as an element that forms the deformation field of the measurement plane and serves as the first state variable in the first state equation.

[0116] Since the probability distribution of the coefficient of each basis vector forming the linear space of the deformation field is given as the first state variable, it becomes possible to accurately represent the deformation field that occurs in the structural body and therefore to derive an accurate deformation.

[0117] In the deformation estimation device of the present embodiment, which is configured as described above, the control unit derives the probability distribution of the first observation variable in the first observation equation by multiplying the first state variable of the first state equation with the basis vector.

[0118] As described above, the actual displacement of the structural body, which serves as the first observation variable, is expressed by multiplying the coefficient vector, which is the first state variable, by the basis vector. Consequently, the information contained in the prior distribution is minimized, and the first observation variable is expressed by multiplying the elements that form the deformation field. This reduces the computation time required to specify the variables to be used in the first state-space model and improves computational accuracy.

[0119] Furthermore, a Kalman filter is used in the estimation of the deformation of the structural body to derive the first a posteriori distribution; a measurement is performed the predetermined first number of times for a deformation of the same structural body, and the deformation to be estimated is updated sequentially according to the measurement, so that the estimation accuracy can be improved. Design 2

[0120] The following describes embodiment 2 of the present invention with reference to the drawings, focusing primarily on the differences from embodiment 1 above. Components identical to those in embodiment 1 above are designated with the same reference numerals, and their descriptions are omitted.

[0121] Fig. Figure 9 is a block diagram showing a schematic configuration of a deformation estimation device 260 according to embodiment 2 and a deformation estimation system 200 comprising the deformation estimation device 260.

[0122] Fig. Figure 10 is a block diagram showing a schematic configuration of a measuring unit 210 contained in the deformation estimating device 260, which is in Fig. 9 is shown.

[0123] Fig. Figure 11 is a flowchart showing the sequence of a control process for a variable output unit 270, which is contained in the measuring unit 210, which is in Fig. 10 is shown.

[0124] Fig. Figure 12 is a flowchart showing the sequence of a control process of a model generation unit 280, which is contained in the measuring unit 210, which is in Fig. 10 is shown.

[0125] Fig. 13 is a flowchart showing the sequence of a control process for measuring unit 210, which is located in Fig. 10 is shown.

[0126] Fig. Figure 14 conceptually illustrates a control process of the measuring unit 210, which is located in Fig. Figure 10 shows the structural body 1, which is a verification target.

[0127] The deformation estimation system 200 of the present embodiment, which is in Fig. Figure 9 shows the deformation estimation device 260, as well as the output unit 5, which outputs an estimation result from the deformation estimation device 260.

[0128] The deformation estimating device 260 of the present embodiment differs from the deformation estimating device 60, which is described in [reference], in the configuration of the measuring unit 210. Fig. 1 in embodiment 1. The measuring unit 10 according to embodiment 1 performs a measurement on the measuring plane 1M of the structural body 1 several times from a fixed position. The measuring unit 210 of the present embodiment 2 performs a measurement on the measuring plane 1M of the structural body 1 several times while moving, as shown in Fig. Figure 14 shows that the measurement error for measurement plane 1M is consequently reduced. It is assumed that the structural body 1 does not move.

[0129] In the present embodiment 2, the coordinate system of the measuring unit 10 is shown two-dimensionally.

[0130] A configuration of the measuring unit 210 is described.

[0131] As in Fig. As shown in Figure 10, the measuring unit 210 has a measuring device 211 that measures the position of a feature point i on the measuring plane 1M of the structural body 1 as a measured value, and a control unit 212 that accurately estimates the position of the feature point i on the measuring plane 1M of the structural body 1 on the basis of the measured value.

[0132] Consequently, the deformation estimating device 260 of the present embodiment comprises as control units the control unit 50, which is shown in embodiment 1, and the control unit 212, which is arranged in the measuring unit 210.

[0133] The control unit 212 comprises the variable output unit 270, the model generation unit 280, an estimation unit 290, and an output unit 6. The control processes in these units are described below.

[0134] First, the control process of the variable output unit 270 is described with reference to the flowchart of the control process of the variable output unit 270, which is shown in Fig. 11 is shown.

[0135] The variable output unit 270 specifies the variables to be used in the model generation unit 280 at the subsequent stage.

[0136] In step S71, the variable output unit 270 specifies – as a variable – initial coordinates (Xw,i, Yw,i, Zw,i) that represent the position of feature point i on the measurement plane 1M of the structural body 1 in a three-dimensional coordinate system. Here, the initial coordinates are defined in a world coordinate system.

[0137] The variable output unit 270 provides - as a variable - vectors that represent conversion information between the first coordinates (Xw,i, Yw,i, Zw,i) of the feature point i in the world coordinate system and second coordinates (ui, vi) of the feature point i in the two-dimensional coordinate system on the measuring device 211, here: a quaternion q that specifies the rotation in a three-dimensional coordinate system, and a three-axis vector s that describes the translation, as shown by the following formula (5) as an example.

[0138] As in Fig. Figure 14 shows a function for calculating a parameter for rotating coordinates from the quaternion q, denoted by E. {g}=(q1q2q3q4){s}=(s1s2s3)

[0139] The variable output unit 270 specifies - as a variable - a motion vector representing the motion velocity of the measuring device 211 and its uncertainty, and is given by the following formula (6), using the above conversion information vectors. {q˙}=d{q} / dt,hg{s˙}=d{s} / dt,hs

[0140] Here, hg and hs are errors that represent uncertainties in rotation and velocity.

[0141] Next, the variable output unit 270 - as a variable - provides a probability distribution of the position (ui, vi) of the feature point i in the two-dimensional coordinate system on the measuring device 211 and its uncertainty (εui, εvi) (step S72).

[0142] Next, the variable output unit 270 outputs the derived variables described above (step S73).

[0143] The output variables are entered into the model generation unit 280 at the next stage.

[0144] Next, the control process of the model generation unit 280 is described with reference to the flowchart of the control process of the model generation unit 280, as shown in Fig. 12 shown.

[0145] First, the model generation unit 280 acquires the first coordinates of the feature point i in the three-dimensional coordinate system, the quaternion q and the vector s, which are conversion information vectors, (d{q} / dt, d{s} / dt), which is a motion vector, which is the motion velocity of the measuring device 211, and its uncertainty (ηq, ηs), as well as the probability distribution (ui, vi) of the position of the feature point i in the two-dimensional coordinate system, measured by the measuring device 211, and its uncertainty, which are the variables that are input by the variable output unit 270 at the previous stage (step S81).

[0146] Next, the model generation unit 280 specifies the first coordinates of the feature point i in the three-dimensional coordinate system, the quaternion q, the vector s and the velocity of motion (d{q} / dt, d{s} / dt) of the measuring device 211 as second state variables, and it computed second state equations, which are given by the following formulas, which represent the time transition of the second state variables (step S82). (Xwi,t,Yw,i,t,Zw,i,t)=(Xz,i,t−1,Yw,i,t−1,Zw,i,t−1){q}T={q}T−1+{q ˙}T−1×ΔT{s}T={s}T−1+{s˙}T−1×ΔT{q˙}t={q˙}t−1+ηq,{s˙}t={s˙}t−1+ηs,

[0147] In the second set of equations of state, the position (first coordinates) of the feature point in the world coordinate system does not change, even if the time of the measurement changes. On the other hand, the vector (q, s) of rotation and translation and the rate of change (d{q} / dt, d{s} / dt) of rotation and translation, which represent the motion of the measuring device 211, change over time.

[0148] Next, the model generation unit 280 outputs the probability distribution (ui, vi) including the position of the feature point i in the two-dimensional coordinate system on the measuring device 211 and the uncertainty, which is its error, as a second observation variable.

[0149] Then the model generation unit 280 calculates second observation equations that represent the correlation between the position (ui, vi) and the first coordinates of the feature point i in the three-dimensional coordinate system, the quaternion q, the vector s and the velocity of motion (d{q} / dt, d{s} / dt) of the measuring device 211 and its uncertainty (ηq, ηs), which are the second state variables (step S83).

[0150] The second set of observation equations is given below. ui=hu({x}i)+εuivi=huv({x}i)+εui {x}i=(Xw,i,tYw,i,tZw,i,t{q}T{s}T{q˙}T{s˙}T)

[0151] Here are hu and hv functions for converting a point in the world coordinate system into a point in the two-dimensional coordinate system on the measuring device 211.

[0152] Next, the model generation unit 280 generates a second state-space model to estimate the first coordinates, which are the position of feature point i of the structural body 1 in the three-dimensional coordinate system, using the second state equation and the second observation equation (step S84).

[0153] The generated second state space model is entered into the estimation unit 290 at the subsequent stage.

[0154] Next, the control process of estimating unit 290 will be described with reference to the flowchart of the control process of estimating unit 290, which is shown in Fig. 13 is shown.

[0155] First, the estimation unit 290 acquires the second state space model, entered by the model generation unit 280 at the previous stage (step S91).

[0156] Next, the estimation unit 290 calculates the second a priori distribution based on the second state-space model (step S92). That is, the first coordinates (Xw,i, Yw,i, Zw,i) of the feature point i in the three-dimensional coordinate system, the quaternion q, the vector s, and a probability distribution containing the velocity (d{q} / dt, d{s} / dt) of the measuring device 211 and its uncertainty (ηq, ηs), which are the second state variables in the second state equation, are calculated.

[0157] Next, the estimation unit 290 updates the second prior distribution using a second likelihood function and thus derives the second posterior distribution (step S93).

[0158] Regarding the second likelihood function, the feature point i is measured by the measuring device 211 (step S95), and on the basis of the measured value, the second likelihood function is derived as a function that represents the probability distribution (ui, vi) of the position of the feature point i in the two-dimensional coordinate system on the measuring device 211 and its uncertainty (step S96).

[0159] Next, the estimation unit 290 determines whether or not a measurement on the measurement plane 1M in step S95 has been carried out a predetermined second number of times (step S94).

[0160] If the number of measurements has not reached the specified second number of times (step S94: NO), the second a posteriori distribution derived in step S93 is updated as a new second a priori distribution (step S97), and it is used as an a priori distribution in step S92.

[0161] Examples of methods for calculating the second a posteriori distribution include recursive Bayesian estimation and a particle filter. In a case where a measurement error and a system error, assumed in a state-space model, can be assumed to be Gaussian, the use of an ensemble Kalman filter allows for estimation with reduced computational cost.

[0162] If the measurement on the measurement plane 1M has been carried out the specified second number of times and consequently an update of the prior distribution using the second likelihood function has been carried out the specified second number of times (step S94: YES), then the estimation unit 290 estimates the position of the feature point i of the structural body 1 in the world coordinate system on the basis of the updated second posterior distribution and outputs the estimated position (steps S98 and S99).

[0163] The first coordinates (Xw,i, Yw,i, Zw,i) of the feature point i of the structural body 1, defined in the world coordinate system, have values ​​that differ between the second coordinates Co1 and Co2 in the two-dimensional coordinate system on the measuring device 211, and consequently, with the second coordinates Co1 and Co2 on the measuring device 211, it becomes possible to perform an accurate estimation of the first coordinates in the world coordinate system.

[0164] In this way, the measuring unit 210 accurately derives the position of the feature point i on the structural body 1, thereby reducing measurement errors and the like. Therefore, the deformation estimation device 260 of the present embodiment can accurately estimate a deformation in the structural body 1 based on the position information of the feature point i of the structural body 1, which is measured accurately.

[0165] The measuring device 211, which measures the feature point, can be a strain gauge, or it can be a device such as a laser displacement meter or an optical interferometer.

[0166] The measuring device 211 can measure the feature point i from an image. Measuring the position of feature point i from an image enables contactless measurement over a wide area. Since multiple points can be measured contactlessly at once, the measurement time is reduced.

[0167] One specific method for measuring the position of feature point i from an image involves digital image correlation. Using digital image correlation allows the measuring device 211 to perform a measurement using only a digital camera, thus reducing the size of the measuring device 211.

[0168] When an ensemble Kalman filter is used in the estimation unit 290, the measurement for a deformation in the same structural body is performed several times while the measuring device is moved, and the position of the feature point to be estimated is updated sequentially according to the measurement, thus improving the estimation accuracy.

[0169] In the deformation estimation device of the present embodiment, which is configured as described above and which uses as a second state variable the first coordinates of a feature point on the measurement plane of the structural body in a three-dimensional coordinate system, a conversion information vector for converting the first coordinates into second coordinates in a two-dimensional coordinate system that the measuring unit has, and a probability distribution of a motion vector that uses the conversion information vector, from the measuring unit to measure the structural body while it moves in the three-dimensional coordinate system, and its errors, the control unit specifies a second state equation that represents the time transition of the second state variable.

[0170] Using the second coordinates of the structural body in the two-dimensional coordinate system of the measuring unit as the second observation variable, the control unit specifies a second observation equation that represents the correlation between the second observation variable and the second state variable. The control unit then generates a second state-space model to estimate the first coordinates in the three-dimensional coordinate system of the feature point of the structural body, based on the second state equation and the second observation equation.The control unit provides a probability distribution of the first coordinates in the three-dimensional coordinate system of the feature point and its errors, estimated using the second state-space model, as a second a priori distribution, derives a second likelihood function with respect to a probability distribution of the second coordinates in the two-dimensional coordinate system of the structural body, measured by the measuring unit, and its errors, derives a second a posteriori distribution on the basis of the second a priori distribution and the second likelihood function, and estimates the first coordinates of the feature point in the three-dimensional coordinate system of the structural body on the basis of the second a posteriori distribution.

[0171] If a movement of the measuring unit is assumed when measuring the feature point, the movement is consequently included in the second a priori distribution in the control unit of the measuring unit, so that a measurement error of the position of the feature point due to the positional displacement of the measuring unit can be reduced.

[0172] For example, when a deformation is measured of an image captured with a camera held in a human hand or attached to a robot, an error in the measurement positions occurs as a result of the three-dimensional displacement between the measured images. However, the measuring unit 210 of the present embodiment can reduce such an error.

[0173] Furthermore, when capturing images multiple times, it is unnecessary to mount a camera or use multiple cameras to reduce measurement errors. Mounting a camera increases the overall size and setup complexity due to the preparation of a measuring gauge and the like. Using multiple cameras also presents the problem of increased setup size. With the measuring unit of the present embodiment, even in a case where images of the structural body are captured with a single camera while a manager moves around without the camera being mounted, the error between images can be reduced. This is achieved by using the correlation information between the images, which exhibits the three-dimensional displacement between them. Thus, a deformation of the structural body 1, which is the imaging target, can be measured with high accuracy.

[0174] Consequently, the deformation estimation device of the present embodiment can accurately detect the displacement of the structural body, so that an accurate estimation of the deformation occurring in the structural body can be carried out. embodiment 3

[0175] The following describes embodiment 3 of the present invention with reference to the drawings, focusing primarily on the differences from embodiment 1 above. Parts identical to those in embodiment 1 above are designated with the same reference numerals, and their descriptions are omitted.

[0176] Fig. Figure 15 is a block diagram showing a schematic configuration of a deformation estimation device 360 ​​according to embodiment 3 and a deformation estimation system 300 comprising the deformation estimation device 360.

[0177] Fig. Figure 16 is a flowchart showing the sequence of a control process for the deformation estimation device 360, which is located in Fig. 15 is shown.

[0178] As in Fig. As shown in Figure 15, the deformation estimating device 360 ​​of the present embodiment differs from the deformation estimating device 60 described in embodiment 1 in that a crack estimating unit 341 is provided.

[0179] In the control process of the variable output unit 20, which differs in part from the control process in embodiment 1, a crack to be estimated within the structural body 1 is specified when the load condition, which specifies the application states of an external force and the like acting on the structural body 1, and the constraint condition for restrictive displacements in directions of the structural body 1 and their direction of rotation are specified as the boundary condition f. More precisely: In a case where, for example, a crack has occurred in cross-section 1A, which in Fig. As shown in Figure 5, the variable output unit 20 specifies the cross-section 1A, which is a surface containing a part where the occurrence of the crack is predicted, as a crack candidate plane.

[0180] A constraint is then applied to node 2P, where a crack is presumed to be present, among the majority of nodes 2P in cross-section 1A such that it has no constraints for displacement. The variable output unit 20 then performs a numerical analysis based on the constraint and calculates a deformation vector ui on the measurement plane 1M. The variable output unit 20 then sequentially changes the position of node 2P, where a crack is presumed to be present, and performs a numerical analysis. Consequently, the deformation vector ui is calculated and stored for each node 2P.

[0181] One result is information about the crack in cross-section 1A contained in the coefficient vector fvr and in the basis vector Ur, which serve as elements forming the linear space of the deformation field of the measurement plane 1M, derived from the variable output unit 20.

[0182] Since information about the crack is contained in the coefficient vector fvr and the basis vector Ur, which are variables captured by the model generation unit 30, this information is reflected in a first spatial model. This model is the same as the first spatial model in embodiment 1, except that information about the crack is contained in the basis vector Ur and the coefficient vector fvr.

[0183] As in embodiment 1, the estimating unit 40 estimates a deformation on the measurement plane 1M and outputs the estimated deformation (step S49).

[0184] From the estimated result for the deformation on the measurement plane 1M, estimated in the estimation unit 40, the crack estimation unit 341 selects the constraint that causes a deformation closest to the estimated deformation. That is: Since the constraint in the variable output unit 20 contains information about a crack occurring at each node 2P in the cross-section 1A, the shape of the crack is obtained from the selected constraint.

[0185] The crack estimation unit 341 outputs information about the crack detected (step S349A-2).

[0186] The information provided about the crack specifies its shape and position. This information can be represented by numerical values ​​or displayed using an image.

[0187] Consequently, the cross-section 1A, which is a surface containing a part where the occurrence of a crack within the structural body 1 is predicted, and data of a deformation on the measurement plane 1M, which occurs when a crack is present at a node 2P in the cross-section 1A, are prepared in advance for each node by numerical analysis, so that training data are prepared that represent the relation between the cross-section 1A and the measurement plane 1M.

[0188] In the deformation estimation device of the present embodiment, which is configured as described above, the control unit specifies a constraint that forms the boundary condition of the numerical analysis, such that a surface containing a part where the occurrence of a crack within the structural body is predicted is specified as a crack candidate plane, the crack candidate plane is divided into a plurality of unit planes, and a node that is specified to have a crack is specified under a plurality of nodes that form the unit planes such that it has no constraints for displacement.

[0189] The control unit predicts the first predicted deformation at the measurement plane through numerical analysis based on the boundary condition for which the constraint condition has been specified. The control unit estimates the presence of a crack at the crack candidate plane based on the constraint condition according to the first predicted deformation and the deformation occurring in the structural body, estimated based on the first a posteriori distribution.

[0190] Since a crack occurring within the structural body is estimated from a deformation on the measuring plane, it is therefore possible to carry out service life evaluation, maintenance, preservation work and the like in a suitable manner even for a structural body that needs to be carefully managed. Design 4

[0191] The following describes embodiment 4 of the present invention with reference to the drawings, focusing primarily on the differences from embodiment 1 above. Components identical to those in embodiment 1 above are designated with the same reference numerals, and their descriptions are omitted.

[0192] Fig. Figure 17 is a block diagram showing a schematic configuration of a deformation estimation system 400 according to embodiment 4.

[0193] The deformation estimation system 400 of the present embodiment is a verification system for drive equipment that uses the deformation estimation device 60 from embodiment 1. A drive unit connected to a power supply line is specified as the structural body 1, which is a verification target.

[0194] Compared with the deformation estimation system 100 described in embodiment 1, the deformation estimation system 400 differs in that the control unit 50 has a diagnostic unit 407 and a structural body control unit 408.

[0195] Here, in a stage that precedes the control of the estimating unit 40, when specifying the boundary conditions, the load condition, which defines the application states of an external force and the like, which is exerted on the drive equipment as structural body 1, and the constraint condition for restrictive displacements in directions of the drive equipment and its direction of rotation are jointly specified as one boundary condition f.

[0196] Furthermore, the longitudinal modulus of elasticity, the density and the Poisson number of the drive equipment are specified as material property values.

[0197] The estimation unit 40 performs a deformation estimation based on the first state space model using the variables that reflect the material property values.

[0198] The diagnostic unit 407 determines the solidity and reliability of the drive equipment based on the deformation that occurs in the drive equipment, which is estimated on the basis of the first A-posteriori distribution.

[0199] During the solidity assessment, the diagnostic unit 407 determines whether the probability distribution of deformation and its errors, estimated based on the first a posteriori distribution, exceeds a predefined first threshold. This first threshold is defined as an index value representing whether the drive equipment is in a state where solidity can be guaranteed, and may, for example, specify a tolerable value for stress, tilt or inclination, displacement, or similar parameters.

[0200] During reliability testing, the diagnostic unit 407 determines whether the probability distribution of deformation occurring in the structural body and its associated failures, estimated based on the first a posteriori distribution, exceeds a second threshold value. This second threshold serves as an index value representing reliability. For example, the second threshold might be a specified output power, operating temperature, operating period, or operating frequency of the drive equipment, defining the operating conditions under which the drive equipment can be used while remaining undamaged.

[0201] Next, the structural body control unit 408 controls the drive equipment based on a result from the diagnostic unit 407. For example, in a case where the output power of the drive equipment needs to be reduced, a control to reduce the output power is carried out.

[0202] In the above description, the solidity of the drive equipment is determined based on a deformation on the measurement plane 1M of the drive equipment. However, the estimation that a crack occurs within the drive equipment can also be performed using a configuration that includes the crack estimation unit 341, which is described in embodiment 3.

[0203] For example, at least one shift to restrict the drive equipment in a case where no crack has occurred in the drive equipment, or a shift to restrict the drive equipment in a case where a crack has occurred in the drive equipment, can be specified for the boundary condition to determine whether or not there is a crack.

[0204] In the description above, the drive equipment connected to the power line is shown as structural body 1. However, structural body 1 could also be, for example, an external structure such as a steel bridge or a steel tower.

[0205] In this case, a displacement to constrain the steel bridge or steel tower, as well as the shape of the steel bridge or steel tower, are entered as a boundary condition, and a longitudinal modulus of elasticity, a density, and a Poisson number are entered as material property values ​​of the steel bridge or steel tower.

[0206] In a case where the presence / absence of a crack is inspected, at least one constraint can be specified as a boundary condition in a case where there is no crack within the steel bridge or steel tower, or in a case where there is a crack in the steel bridge or steel tower.

[0207] The diagnostic unit 407 can determine a usage area, usage period, usage frequency, or repair area of ​​the steel bridge or steel tower after estimation based on the estimated deformation, and it can display the usage area, usage period, or repair area of ​​the steel bridge or steel tower, issue an alarm in a case where the usage frequency is exceeded, and provide a time to carry out a repair in the repair area, so that it consequently reports to a manager.

[0208] Then, for example, a drone can be specified as a structural body control unit 408, the repair area can be assigned to the drone, and a repair can be carried out by the drone.

[0209] Based on the estimated information, the system determines the condition in which the steel bridge or tower can be used while remaining structurally sound, the repair area, the usage range, the usage period, the repair area, and the repair time. An alarm regarding the repair frequency is also generated and reported to a manager. Consequently, the steel bridge or tower, which requires careful management, is controlled and monitored to ensure its long-term operation while maintaining structural integrity.

[0210] Furthermore, since the solidity can be diagnosed using the probability distribution, a control output can be determined for the structural body that exhibits the uncertainty of the control, thus improving the accuracy of the control.

[0211] In a case where this deformation estimation device is applied to a pressure vessel or pressure line used for the propulsion equipment, an external force to be specified as the boundary condition may be the internal pressure, and a nominal or design pressure or a pressure recorded during a past inspection may be specified.

[0212] In a case where the structural unit is drive equipment, the structural unit's control unit can perform power control to adjust the output power or similar parameters of the drive equipment, which is a monitoring objective based on the results of reliability testing. In a case where the structural unit is drive equipment, such as a power plant, the temperature of the water or coolant (e.g., hydrogen) supplied to the power plant can be controlled as an environmental factor to adjust the power plant's operating environment. The ambient temperature where the drive equipment is located or a temperature control device within the drive equipment can also be used as an environmental factor to adjust the operating environment of the drive equipment.

[0213] The deformation estimation system of the present embodiment, configured as described above, comprises the following: the deformation estimation device, configured as described above; and an output unit that displays output from the control unit. The control unit outputs—given a longitudinal modulus of elasticity, a density, and a Poisson's ratio of the structural body—at least the longitudinal modulus of elasticity and the Poisson's ratio as material property values ​​of the structural body in the numerical analysis and performs the numerical analysis.The control unit performs a solidity determination to ascertain whether or not the probability distribution of an estimated deformation occurring in the structural body and its error exceeds a first threshold value, which serves as an index value representing the solidity, based on the first a posteriori distribution, and it displays a result of the determination to the output unit.

[0214] Since a manager can confirm the solidity of the structural body, which is measured on the basis of an accurately estimated deformation occurring in the structural body, it is consequently possible to appropriately carry out a service life evaluation, maintenance, upkeep work and the like, even for a structural body that needs to be carefully managed so that operation can be carried out for a long period of time.

[0215] In the deformation estimation system of the present embodiment, configured as described above, the measuring unit acquires an image by capturing the measurement plane of the structural body. The control unit derives a probability distribution of at least one displacement, tilt, or inclination occurring in the structural body and its errors, as the first a priori distribution, based on the probability distribution of the first state variable and its errors, estimated by the first state-space model. It then derives a probability distribution of at least one displacement, tilt, or inclination of the structural body and its errors based on the image, as a likelihood distribution of the first likelihood function.The control unit derives the first a posteriori distribution based on the first a priori distribution and the first likelihood function in order to perform a solidity determination based on the first a posteriori distribution.

[0216] In the deformation estimation system of the present embodiment, configured as described above, the control unit specifies an external structure or drive equipment connected to a power supply line as the structural body. The control unit specifies—given a longitudinal modulus of elasticity, a density, and a Poisson's ratio of the drive equipment or external structure—at least the longitudinal modulus of elasticity and the Poisson's ratio as the material property values ​​for the numerical analysis in order to perform the numerical analysis.The control unit performs a reliability determination to ascertain whether or not the estimated probability distribution of the deformation occurring in the structural body and its error exceeds a second threshold, which serves as an index value representing reliability, based on the first a posteriori distribution, and it displays the result of the determination on the output unit.

[0217] Since the structural integrity and reliability are determined based on deformation estimated by the state-space model using numerical analysis that reflects material property values, a manager can verify the structural integrity calculated based on an accurately estimated deformation occurring within the structure. This makes it possible to appropriately perform lifetime evaluations, maintenance, repairs, and the like, even for a structure requiring careful management. By controlling and monitoring the structural structure based on the results of the integrity or reliability assessment, its operation can be carried out over a long period.

[0218] In the deformation estimation system of the present embodiment, which is configured as described above, in a configuration where the drive equipment is specified as a structural body, the control unit controls at least one parameter of the power for the drive equipment or an environmental factor for setting the operating environment of the drive equipment, based on the reliability determination.

[0219] Since the power control of the drive equipment or an environmental factor for adjusting the operating environment of the drive equipment is set based on the result of the reliability determination, excessive use of the drive equipment can be suppressed, for example, or a suitable temperature environment can be ensured so that operation can be carried out for a long period of time.

[0220] The hardware configuration of the control unit is described below.

[0221] In each embodiment, a processing circuit for performing the processing is arranged in the control unit. The processing circuit can be dedicated hardware or a CPU (a central processing unit, also referred to as a processing device, a computing device, a microprocessor, a microcomputer, a processor, or a DSP) that executes a program stored in memory.

[0222] Fig. Figure 18 illustrates an example of a hardware configuration of the control unit.

[0223] In Fig. In Figure 18, a processing circuit 52 is connected to a bus B1. In a case where the processing circuit 52 is dedicated hardware, the processing circuit 52 can be, for example, a single circuit, a complex circuit, a programmed processor, an ASIC, an FPGA, or a combination thereof. The functions of the units, such as the control unit of the deformation estimation device and the deformation estimation system, described in each embodiment, can each be implemented by the processing circuit 52, or the functions of the units can be implemented jointly by the processing circuit 52.

[0224] Fig. Figure 19 illustrates an example of a hardware configuration of the control unit.

[0225] In Fig.In the 19th case, a processor 54 and a memory 53, which is a storage device, are connected by a bus B2. In a case where the processing circuit is a CPU, the functions of the units of the estimating device are implemented by software, firmware, or a combination of software and firmware. The software or firmware is described as a program and stored in memory 53. The processing circuit reads the program stored in memory 53 and executes it, thus implementing the function of each unit.

[0226] This means that the estimating device has memory 53 for storing such a program that processing steps are executed when the program is run by the processing circuit. The program can be viewed as causing a computer to execute a procedure or process. Here, memory 404 can be a non-volatile or volatile semiconductor memory, such as RAM, ROM, flash memory, EPROM, or EEPROM, or it can be a magnetic disk, a flexible disk, an optical disk, a compact disc, a minidisc, a DVD, or the like.

[0227] Among the functions of the deformation estimation unit, some can be implemented by dedicated hardware, and others by software or firmware. For example, the model generation unit can be implemented as dedicated hardware by the processing circuit. The diagnostic unit, which diagnoses the crack condition, can be implemented by having the processing circuit read and execute the program stored in memory 53.

[0228] As described above, the processing circuit can implement the above functions through hardware, software, firmware, or a combination of these.

[0229] Although the invention is described above in the form of various exemplary embodiments and implementations, it is understood that the various features, aspects and functionalities described in one or more of the individual embodiments are not limited in their applicability to the particular embodiment in which they are described, but can instead also be applied - alone or in various combinations - with one or more of the embodiments of the invention.

[0230] It is therefore understood that numerous modifications can be devised that are not described by way of example without deviating from the scope of the present invention. For example, at least one of the components can be modified, added, or omitted. At least one of the components mentioned in at least one of the preferred embodiments can be selected and combined with the components mentioned in another preferred embodiment. Reference symbol list 1 structural body 10,210 units of measurement 50,212 Control unit 60, 260, 360 Deformation estimation device 1M measuring plane 100, 200, 300, 400 Deformation Estimation System QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] WO 2015 / 049757 A1

[0005]

Claims

[1] Deformation estimating device comprising a control unit which estimates a deformation occurring in a structural body, based on a measurement representing a deformation on a measurement plane of a surface of the structural body, wherein the control unit a deformation prediction information that – through a probability distribution – represents a deformation occurring on the measurement plane of the structural body, under a plurality of boundary conditions applied to the structural body, and It specifies a first observation equation that represents a correlation between a first observation variable, which represents the measured value, an error of the measured value through a probability distribution, and the deformation prediction information, which serves as a first state variable, and estimates a deformation that occurs in the structural body based on the first observation equation. [2] Deformation estimation device according to claim 1, wherein the control unit using the deformation prediction information as the first state variable, specifies a first state equation representing the time transition of the first state variable, specifies the first observation equation representing the correlation between the first observation variable and the first state variable, and generates a first state-space model for estimating a deformation of the structural body based on the first state equation and the first observation equation, and a probability distribution of the first state variable and its errors, estimated by the first state space model, as a first a priori distribution, derives a first likelihood function with respect to a probability distribution of a displacement of the structural body and its errors obtained from the measured value, derives a first a posteriori distribution based on the first a priori distribution and the first likelihood function, and estimates a deformation occurring in the structural body based on the first a posteriori distribution. [3] Deformation estimation device according to claim 2, wherein the control unit - as the first state variable - derives a probability distribution which represents a first predicted deformation, which represents a difference between before and after a deformation occurring on the measurement plane, under a plurality of boundary conditions imposed on the structural body, or a correlation between the boundary condition and the first predicted deformation. [4] Deformation estimation device according to claim 3, the control unit derives a probability distribution of an element forming a deformation field of the measurement plane with respect to the boundary condition and an error of the element, namely as the probability distribution representing the correlation between the boundary condition and the first predicted deformation. [5] Deformation estimation device according to claim 4, wherein the control unit derives a coefficient of each basis vector forming a linear space in the deformation field, namely as those elements that form the deformation field. [6] Deformation estimation device according to claim 5, wherein the control unit derives the probability distribution of the first observation variable in the first observation equation by multiplying the first state variable of the first state equation with the basis vector. [7] Deformation estimation device according to one of claims 2 to 6, wherein the control unit the first observation equation including a parameter that assigns the first observation variable and the first state variable to each other, and estimates a deformation occurring in the structural body by changing a value of the parameter such that the difference between a deformation of the structural body estimated from the first a posteriori distribution and a deformation of the structural body obtained from the measured value is minimized. [8] Deformation estimation device according to one of claims 2 to 7, wherein the measurement for the measured value on the measurement plane is carried out a predetermined first number of times, and the control unit derives the first likelihood function with respect to the probability distribution of the displacement of the structural body and its errors on the basis of the measured value each time the measurement is carried out, and performs a derivation of the first a posteriori distribution on the basis of the first a priori distribution and the first likelihood function the first number of times. [9] Deformation estimation device according to one of claims 2 to 8, wherein the control unit The second state variable consists of the first coordinates of a feature point on the measurement plane of the structural body in a three-dimensional coordinate system, a conversion information vector for converting the first coordinates into second coordinates in a two-dimensional coordinate system, which has a measuring unit for measuring the measured value, and uses a probability distribution of a motion vector, using the conversion information vector of the measuring unit for measuring the structural body while it moves in the three-dimensional coordinate system, and its error, a second state equation that specifies the time transition of the second state variable. Using the second coordinates of the structural body in the two-dimensional coordinate system of the measuring unit as the second observation variable, a second observation equation is specified that represents the correlation between the second observation variable and the second state variable. a second state-space model for estimating the first coordinates in the three-dimensional coordinate system of the feature point of the structural body is generated on the basis of the second state equation and the second observation equation, and a probability distribution of the first coordinates in the three-dimensional coordinate system of the feature point and its errors, estimated using the second state-space model, is specified as a second a priori distribution, a second likelihood function is derived with respect to a probability distribution of the second coordinates in the two-dimensional coordinate system of the structural body, measured by the measuring unit, and its errors, a second a posteriori distribution is derived on the basis of the second a priori distribution and the second likelihood function, and the first coordinates of the feature point in the three-dimensional coordinate system of the structural body are estimated on the basis of the second a posteriori distribution. [10] Deformation estimation device according to claim 9, wherein the measuring unit captures an image that is taken by capturing the measuring plane of the structural body, and The control unit derives the second coordinates in the two-dimensional coordinate system of the feature point from the image. [11] Deformation estimation device according to claim 9 or 10, wherein the measuring unit performs a measurement for the feature point a predetermined second number of times while moving relative to the structural body, and The control unit derives the second likelihood function with respect to the probability distribution of the second coordinates in the two-dimensional coordinate system and its error based on the measured value each time the measurement is performed by the measuring unit, and performs a derivation of the second a posteriori distribution based on the second a priori distribution and the second likelihood function the second number of times. [12] Deformation estimation device according to one of claims 3 to 6, wherein the control unit a constraint that forms the boundary condition, such that a surface containing a part where the occurrence of a crack within the structural body is predicted is specified as a crack candidate plane, the crack candidate plane is divided into a plurality of unit planes, and a node specified to have a crack is specified under a plurality of nodes forming the unit planes such that it has no constraints for displacement. the first predictive deformation on the measurement plane is based on the boundary condition for which the constraint condition has been specified, and a crack that is present at the crack candidate plane, is estimated on the basis of the constraint condition according to the first predicted deformation and the deformation that occurs in the structural body, estimated on the basis of the first a posteriori distribution. [13] Deformation estimation system which includes the following: the deformation estimating device according to any one of claims 1 to 12; and an output unit that displays an output from the control unit, wherein the control unit under a longitudinal elastic modulus, a density and a Poisson number of the structural body, at least the longitudinal elastic modulus and the Poisson number are specified as material property values ​​of the structural body when predicting the deformation prediction information, and performs a solidity determination to determine whether or not the probability distribution of an estimated deformation occurring in the structural body and its error exceeds a first threshold value, which serves as an index value representing the solidity and indicates a result of the determination of the output unit. [14] Deformation estimation system according to claim 13, wherein an image taken by capturing the measurement plane of the structural body is captured as the measurement value, and the control unit using the deformation prediction information as the first state variable, specifies a first state equation representing the time transition of the first state variable, specifies the first observation equation representing the correlation between the first observation variable and the first state variable, and generates a first state-space model for estimating a deformation of the structural body based on the first state equation and the first observation equation, and a probability distribution of the first state variable and its errors, estimated by the first state-space model, as a first a priori distribution; derives a first likelihood function with respect to a probability distribution of a displacement of the structural body and its errors, obtained from the measured value; derives a first a posteriori distribution based on the first a priori distribution and the first likelihood function; and estimates a deformation occurring in the structural body based on the first a posteriori distribution. derives a probability distribution of at least one displacement, tilt or inclination, or stress occurring in the structural body, and its errors, namely as the first a priori distribution, based on the probability distribution of the first state variable and its errors, estimated by the first state-space model; derives a probability distribution of at least one displacement, tilt or inclination, or stress of the structural body and its errors based on the image, namely as a likelihood distribution of the first likelihood function; and derives the first a posteriori distribution based on the first a priori distribution and the first likelihood function, in order to perform a solidity determination based on the first a posteriori distribution. [15] Deformation estimation system according to claim 13 or 14, wherein the control unit specifies as a structural body an external structure or drive equipment connected to a power supply line, under a longitudinal modulus of elasticity, a density and a Poisson's ratio of the drive equipment or the external structure, at least the longitudinal modulus of elasticity and the Poisson's ratio are specified as the material property values, and a reliability determination is performed to determine whether or not the estimated probability distribution of the deformation occurring in the structural body and its error exceeds a second threshold, which serves as an index value representing the reliability, based on the first a posteriori distribution, and displays the result of the determination on the output unit. [16] Deformation estimation system according to claim 15, wherein the control unit in a configuration where the drive equipment is specified as the structural body, at least one of the power for the drive equipment or an environmental factor for setting the operating environment of the drive equipment is controlled, based on the reliability determination. [17] Deformation estimation method which uses a deformation estimation device including a control unit which estimates a deformation occurring in a structural body on the basis of a measurement which represents a deformation on a measurement plane of a surface of the structural body, wherein the method comprises the following steps: Predictions, by the control unit, of a deformation prediction information which - through a probability distribution - represents a deformation that occurs on the measurement plane of the structural body, under a plurality of boundary conditions that are applied to the structural body; and Specifying, by the control unit, a first observation equation, which represents a correlation between a first observation variable, which represents the measured value, and an error of the measured value through a probability distribution, and the deformation prediction information, which serves as a first state variable, and estimating a deformation that occurs in the structural body, based on the first observation equation.

Images