Crack development prediction device, crack inspection system, and crack development prediction method
By inputting the structural parameters and generating state space models in the fracture development prediction device, and combining with the Kalman filter to update the posterior distribution, the problem of uncertainty in the fracture development amount is solved, quantitative fracture development prediction is achieved, and the structure is improved Accuracy of lifespan management.
Patent Information
- Application Number
- CN202280093333.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2022-03-14
- Publication Date
- 2024-11-15
AI Technical Summary
The prior art cannot effectively reduce the uncertainty of the crack development amount, resulting in the inability to quantitatively represent the accuracy of the crack development amount, which in turn affects the remaining life prediction of the structure.
A crack development prediction device is adopted to input the parameters and uncertainties of the structure through the parameter input part to generate a state space model, combining the ensemble Kalman filter of the crack shape measurement and estimation part, and update it The posterior distribution of fracture shape and parameters quantitatively represents the uncertainty of the amount of fracture development.
It reduces the uncertainty of crack shape measurement, can quantitatively represent the uncertainty of crack development volume with probability distribution, improves the accuracy of crack prediction, and supports more economical maintenance and life management.
Smart Images

Figure CN118974531A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to a crack development prediction device, a crack inspection system and a crack development prediction method. Background Art
[0002] In conventional crack inspections of structures, when predicting the growth amount of a discovered crack, there is uncertainty in the growth characteristics of the crack due to the shape of the measured crack, the shape of the inspection object, the force applied to the inspection object, displacement, temperature, the material used in the inspection object, etc. The growth amount of the crack estimated based on the uncertain information also becomes uncertain, and the uncertainty is reduced as much as possible and the remaining life of the inspection object is estimated based on the predicted growth amount of the crack (for example, see Patent Document 1).
[0003] Prior art literature
[0004] Patent Literature
[0005] Patent Document 1: Japanese Patent Application Publication No. 2009-31124 Summary of the invention
[0006] However, since it is impossible to quantitatively express the uncertainty while reducing the uncertainty of the estimated crack growth amount, there is a problem that the accuracy of the remaining life estimated from the crack growth amount cannot be quantitatively expressed.
[0007] The present application is completed to solve the problems as mentioned above, and its purpose is to provide a crack development prediction device, a crack inspection system and a crack development prediction method that can reduce the uncertainty of the measured crack shape to infer the shape of the crack and can quantitatively represent the uncertainty of the crack development amount with a probability distribution.
[0008] The present application discloses a crack development prediction device for predicting the development of cracks generated in a structure, characterized by comprising:
[0009] A parameter input unit inputs at least the parameters of the shape of the structure, the force applied to the structure, the material properties of the structure, and the initial value of the crack shape, including the uncertainty of each parameter, in a probability distribution;
[0010] A model generation unit generates a state space model for predicting a crack development state composed of a state equation and an observation equation according to input parameters;
[0011] a crack shape measuring unit for measuring a crack shape of a structure; and
[0012] The estimation unit estimates a posterior distribution including the crack shape and parameters based on the probability distribution of the uncertainty caused by the measurement value of the crack shape measured by the crack shape measuring unit and the measurement error and the prior distribution of the crack shape predicted by the state space model.
[0013] The crack development prediction device predicts the development of the crack based on the crack shape and parameters updated by the estimation unit.
[0014] According to the crack growth prediction device disclosed in the present application, the uncertainty of the measured crack shape can be reduced to estimate the crack shape, and the uncertainty of the crack growth amount can be quantitatively expressed by probability distribution. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] Figure 1 This is a functional configuration diagram of the crack development prediction device according to the first embodiment.
[0016] Figure 2 This is a prediction flow chart of the crack development prediction device according to the first embodiment.
[0017] Figure 3 It is a diagram for explaining the parameters of the structure for predicting the development of cracks.
[0018] Figure 4 It is expressed as a probability distribution. Figure 3 Figure 2. Parameters described in Figure 2.
[0019] Figure 5 This is a diagram for explaining an example of a model for predicting the development of cracks according to the first embodiment.
[0020] Figure 6 This is a diagram for explaining the measurement of the shape of a crack in the structure according to the first embodiment.
[0021] Figure 7 This is a flowchart of estimating the shape and parameters of the crack involved in the first embodiment.
[0022] Figure 8 This is a functional configuration diagram of a crack development prediction device according to the second embodiment.
[0023] Fig. 9 This is a prediction flow chart of the crack development prediction device according to the second embodiment.
[0024] Fig.10 This is a flowchart showing the operation of the crack inspection system equipped with the crack development prediction device according to the second embodiment.
[0025] Fig.11 This is a functional configuration diagram of a crack development prediction device according to the third embodiment.
[0026] Fig.12 This is a prediction flow chart of the crack development prediction device according to the third embodiment.
[0027] Fig.13 This is a diagram for explaining an example of prediction by the operating condition prediction unit according to the third embodiment.
[0028] Fig.14 This is a flowchart showing the operation of the crack inspection system equipped with the crack development prediction device according to the third embodiment.
[0029] Fig.15 This is a prediction flow chart of the crack development prediction device according to the fourth embodiment.
[0030] Fig.16 This is a diagram for explaining an example of the hardware configuration of the crack growth prediction device according to the embodiment.
[0031] Fig.17 This is a diagram for explaining another example of the hardware configuration of the crack growth prediction device according to the embodiment. DETAILED DESCRIPTION
[0032] Hereinafter, preferred embodiments of the crack development prediction device, crack inspection system, and crack development prediction method involved in the present application are described with reference to the accompanying drawings. In addition, the same reference numerals are added to the same contents and corresponding parts, and their detailed descriptions are omitted. Similarly, in the subsequent embodiments, repeated descriptions are omitted for the structures added with the same reference numerals. Implementation method 1.
[0033] Figure 1 A functional configuration diagram of a crack development prediction device according to Embodiment 1 of the present application is shown. Figure 2 A prediction flowchart according to the first embodiment of the present application is shown.
[0034] like Figure 1 As shown, the crack development prediction device 1 includes functions of a parameter input unit 2 , a model generation unit 3 , a crack shape measurement unit 4 , and an estimation unit 5 .
[0035] use Figure 2 The prediction flow chart is used to illustrate Figure 1The crack development prediction device 1 shown in the figure predicts the development of cracks. First, the parameters of the structure for predicting the development of cracks are input to the parameter input unit 2 (step S1). In the model generation unit 3, a model for predicting the development of cracks is created (step S2). In the crack shape measurement unit 4, the shape of the crack is measured (step S3), and in the estimation unit 5, the shape of the crack measured by the crack shape measurement unit 4 is updated with the crack shape predicted based on the input parameters and the created model, and the shape and parameters of the crack are estimated (step S4).
[0036] Steps S1 to S4 will be described in more detail.
[0037] (1) Step S1 (Operation of parameter input unit 2)
[0038] Figure 3 (b) shows that Figure 3 An example of parameters for predicting the development of cracks input to the parameter input unit 2 in the structure shown in (a).
[0039] In Figure 3 The following describes an example in which a crack 27 exists in the XY plane, that is, a cross section 28, of a structure 29 shown in (a) and a repeated force 24 is applied. Figure 6 As shown, cracks are detected from surface 25, and the vicinity of cross section 28 where cracks exist is detected. Figure 3 As shown in (b), the parameters 20 of the structure for predicting the development of cracks include the elements of the shape 21 of the structure, the initial crack shape 22, the crack development characteristics 23, and the applied force (including the expressions such as load, load, displacement, etc.) 24. Furthermore, temperature changes may be input as environmental changes occurring in the structure.
[0040] Among the parameters of the structure, the shape 21 of the structure includes the width W and thickness T of the structure. The initial crack shape 22 includes the crack length a and the crack width b. The crack development characteristics 23 are composed of the following formula A. In addition, the applied force 24 is represented by force F. Formula A is represented by the relationship between the crack development speed da / dN and the stress expansion coefficient range ΔK. In the formula, C and n are material constants.
[0041] [Mathematical formula 1]
[0042]
[0043] Figure 3 There are uncertainties in the parameters shown in (b). Figure 4 This shows an example of expressing parameters as probability distributions including uncertainty. That is, we express them as vectors Figure 3 The parameter x shown in the figure represents the probability distribution of the parameter as a normal distribution N(x0, v0 2) is input as the initial value. Figure 4 In , (i) indicates the number of the ensemble member. 0|0 indicates the initial value of "the time of the predicted data|the time of the data used in the prediction".
[0044] (2) Step S2 (Operation of the model generation unit 3)
[0045] Figure 5 An example of a model for predicting the development of cracks is shown. t and variance vector v t A vector 40 of the parameters of the structure that will become the state variable is defined. Next, among the state variables, the observation vector y of the crack length a is t It is defined as vector 41. aM recorded in vector 41 k The vector 41 represents the crack length including the measurement error. The vector 41 is the likelihood distribution of the uncertainty caused by the measured value of the crack shape and the measurement error. Based on the vector 40 of the parameters of the structure defined as the state variable and the vector 41 of the observation data of the crack length a, a state space model 42 for predicting the development state of the crack is created.
[0046] In the state space model 42, the current state vector x t and the state vector x a certain period of time ago t-1 、v t-1 The state equation of the relationship is shown in equation (1). The state vector x is represented by various parameters. t In addition, as shown in equation (2) as the observation equation, the vector y of the observation data t From the state vector x t and the vector w of the probability distribution of the measurement error t The observed data is assumed to be a likelihood distribution consisting of the state vector and the probability distribution of the measurement error. In formula (1), the length of the crack a is used t-1 , crack width b t-1 The function f t There are relationships shown in equations (3), (4), and (5) that represent the crack development characteristics. u Represents a loop, using function g a , g b Indicates the stress expansion coefficient range ΔK a , ΔK b .
[0047] (3) Step S3 (Operation of the Crack Shape Measuring Unit 4)
[0048] Figure 6An example of actually measuring the shape of a crack is shown. The structure 29 is scanned by an ultrasonic flaw detector 50 to measure the length a and the width b of the crack. The ultrasonic flaw detector 50 is controlled by a control unit 51 to perform the measurement. The measured length a and the width b of the crack are used as Figure 5 The observation vector y shown by vector 41 t By using ultrasonic testing in crack inspection, it is possible to inspect cracks in locations that are not visible in a non-destructive state, and the remaining life of the structure can be estimated more accurately through inspection. In addition, the measurement of the crack shape does not need to be performed after the model generation part works, but can also be performed before the model generation work or at the same time as the model generation work.
[0049] (4) Step S4 (Operation of Estimation Unit 5)
[0050] Figure 7 The flow of using the state space model 42 to update the crack shape measured by the crack shape measuring unit 4 with the predicted crack shape and inferring the shape and parameters of the crack is shown. In the inference of the shape and parameters of the crack, an ensemble Kalman filter is used. In the Kalman filter, the inferred state needs to be linearly approximated, but in the ensemble Kalman filter, the nonlinear crack development phenomenon can be directly inferred. As a result, the prediction accuracy of the crack is improved.
[0051] exist Figure 7 First, the initial value is randomly selected within the range of probability distribution of the input crack shape and the initial value of the parameter, and N u The state vector after the loop is predicted to be M (step S10). The mean and variance of the predicted M state vectors are calculated, and the matrix of the prediction error of the state vector is calculated (step S11). The predicted state vector is used as a prior distribution. Based on the prediction error matrix of the state vector, the shape of the crack of the state vector, and the error matrix of the parameters, the Kalman gain is calculated (step S12).
[0052] Next, based on the predicted state vector (prior distribution), observation data (likelihood distribution), and Kalman gain, M filtered inference values of the crack shape and parameters for which the state vector is obtained are obtained (step S13). Based on the M filtered inference values of the crack shape and parameters for which the state vector is obtained, the probability distribution of the filtered inference values is obtained (step S14). Based on the probability distribution of the filtered inference values, the probability distribution of the crack shape and parameters is obtained (step S15) (inference of the posterior distribution).
[0053] Based on the obtained probability distribution of the crack shape and parameters, when the probability distribution is a normal distribution, the mean is set as the estimated result of the crack shape and parameters, and the variance represents the deviation of the estimated result of the crack shape and parameters.
[0054] Thus, in this embodiment, the uncertainty of the measured crack shape can be reduced to estimate the crack shape, and the uncertainty of the crack development amount can be quantitatively expressed by probability distribution. In addition, by obtaining not only the crack but also the estimated value of the parameter related to the development of the crack of the target structure and its deviation, changes other than the crack occurring in the structure can be grasped, and it can be determined whether inspection other than the crack part is required. Thus, appropriate countermeasures can be taken for changes other than the crack, and relatively economical maintenance can be achieved. Implementation method 2.
[0055] Figure 8 A functional configuration diagram of a crack development prediction device 1a according to Embodiment 2 is shown. Fig. 9 The prediction flow chart is shown. Only the part changed from the first embodiment is described, and the rest is the same as the first embodiment, so it is omitted. Figure 8 As shown, in the crack development prediction device 1a, in addition to the structure of the first embodiment, a crack shape prediction unit 6 is additionally included. By adding this structure, as shown in FIG. Fig. 9 As shown, the crack shape at an arbitrary (arbitrary) time can be predicted based on the crack shape and parameters updated by the estimation unit 5 (step S5 ).
[0056] The operation of step S5 is described below. Figure 7 Similarly, in step S10, the predicted state vector is used as the predicted value of the crack shape at an arbitrary time. Specifically, the number of cycles of repeatedly applying load or displacement to the cracked structure until an arbitrary time is set, and the shape and parameters of the crack estimated in step S4 are used as initial values, and M state vectors after the set number of cycles are obtained. By assuming the probability distribution of the M state vectors, the predicted value and deviation are obtained. When the probability distribution is a normal distribution, the predicted value becomes the mean and the deviation becomes the variance. By obtaining the predicted value and deviation of the crack development amount, it is possible to quantitatively determine the repair and replacement period of the inspection object, and to achieve efficient maintenance.
[0057] According to the above, the crack shape prediction unit 6 predicts the probability distribution of the crack shape at any time after the measurement based on the shape of the crack obtained by the estimation unit 5. Thus, the uncertainty of the crack development amount is quantitatively expressed by the probability distribution based on the shape of the crack after reducing the uncertainty of the measured crack shape, so that the repair and replacement period of the inspection object can be quantitatively determined, and efficient maintenance can be achieved.
[0058] Fig.10This is a flowchart showing the operation of the crack inspection system 100 equipped with the crack development prediction device 1a described in Embodiment 2. In the crack inspection system 100, the shape of the structure 29, the force or displacement applied to the structure 29, the crack development characteristics of the material used, and the initial crack shape are input as parameters using the input unit 101. As input means, direct input of data using a keyboard or a mouse, input based on data transmitted using a storage device, reception of data transmitted from other devices in a wireless or wired manner, or reception of data transmitted from the Internet, a cloud, etc. can be considered, but are not limited to these.
[0059] The above-mentioned operation is performed in the crack development prediction device 1a for the input parameters, that is, the probability distribution of each parameter is calculated by the parameter input unit 2, and the state space model 42 for predicting the development state of the crack is prepared in the model generation unit 3. In addition, the shape of the crack is measured by the crack shape measurement unit 4, and the shape of the measured crack is updated by the estimation unit 5 using the crack shape predicted by the state space model 42, and the shape and parameters of the crack are estimated. Based on the shape of the crack obtained by the estimation unit 5, the probability distribution of the crack shape at any time after the measurement is predicted by the crack shape prediction unit 6. Based on the predicted probability distribution, the output unit 102 outputs maintenance-related information such as data related to the next inspection period, repair period, and replacement period, while also considering other factors. Other factors refer to factors other than those related to the structure itself, including the scheduling of inspection personnel, the cost of inspection and repair, the inventory of repair parts, etc. Information related to other factors can also be input from the input unit 101.
[0060] The output unit 102 may also be provided with a determination unit for determining these periods. As a method of using the determination unit to determine the inspection period, repair period or replacement period, the stress expansion coefficient of the crack shape in the next inspection period, repair period or replacement period may be measured, and the fracture toughness value of the rotor parts may be compared to make a determination based on whether it is below a predetermined safety rate. In addition, the output unit 102 may also be provided with a display unit, and the predicted probability distribution, the next inspection period, repair period and replacement period may be displayed on the display unit. In addition, an alarm device such as a buzzer may be provided to notify the replacement period, inspection period, etc. Furthermore, the output unit may also be provided with a communication unit, and the determination results and their data, and predicted data may be sent to other devices by wire or wirelessly. Furthermore, it is also possible to connect to a network to accumulate these data in a cloud, etc., and accumulate a large amount of parts data, so that the accumulated data is not only used for maintenance, but also for manufacturing, sales, etc.
[0061] Next, a specific example of inspecting the parts of the rotor of a generator is described. As parameters of the shape of the structure, there are the diameter or inner diameter of the rotor parts, their dimensional errors, etc. As parameters of the force or displacement applied to the structure, there are temperature changes, centrifugal forces repeated when the rotor is rotating and stopped, centrifugal forces caused by changes in the rotation speed of the rotor, etc. As parameters of the material properties of the rotor, there are the characteristics of crack development used in the rotor parts. Each parameter has uncertainty.
[0062] The shape and uncertainty of the crack measured by the crack shape measuring unit 4 are the measured value and the measurement error assumed for each measurement method. The initial crack shape and uncertainty provided temporarily can be arbitrarily determined or can be set to the crack shape and uncertainty matching the measurement method.
[0063] As described above, the crack inspection system 100 equipped with the crack development prediction device 1a can quantitatively represent the uncertainty of the crack development amount with probability distribution, thereby being able to quantitatively determine the next inspection period of the structure, or the repair or replacement period of the structure, and enabling efficient maintenance. Implementation method 3.
[0064] Fig.11 A functional configuration diagram of a crack development prediction device 1b according to Embodiment 3 is shown. Fig.12 The prediction flow chart is shown. Only the part changed from the first embodiment is described, and the rest is the same as the first embodiment, so it is omitted. Fig.11 As shown, in the crack development prediction device 1b, in addition to the structure of the first embodiment, an operation condition prediction unit 7 is additionally included. In the operation condition prediction unit 7, as shown in FIG. Fig.12 As shown in the flowchart, based on the shape and parameters of the crack updated by the estimation unit 5, the value and reliability of the load that can work until the scheduled maintenance time are calculated according to the probability distribution of the time until the predetermined (predetermined) crack shape is achieved and the probability distribution of the updated load applied to the structure (step S6).
[0065] exist Fig.13 In the following, an example of prediction by the operating condition prediction unit 7 shown in step S6 is described. Fig.13 (a) is a diagram showing the relationship between the force (load) applied to the structure 29 during measurement and the time until a predetermined crack shape is formed. Fig.13 (b) is a diagram showing the relationship between the case where the load is changed based on the prediction result and the time until a predetermined crack shape is achieved. Fig.13The vertical axis of the graphs shown in (a) and (b) represents the load, and the horizontal axis represents the time until the predetermined crack shape is formed. In the graph, contour lines 114 and 117 represent probability distribution. The crack develops due to repeated application of load or displacement, and the time and the number of repetitions can be converted by measuring or assuming the number of repetitions per unit time.
[0066] exist Fig.13 In (a), a contour line 114 is obtained based on the probability distribution 111 of the load estimated by the estimation unit 5 and the probability distribution 112 of the time until the predetermined crack shape is formed. Based on the relationship between the probability distribution 112 of the time until the predetermined crack shape is formed and the repair time point 113, it is possible to grasp whether the crack predicted until the repair will become a crack shape of a predetermined size with a probability including a deviation.
[0067] When the structure 29 to be measured is used until repair, there is a method of changing (reducing) the load. Until the time point 113 of repair, the structure 29 is used with a predetermined reliability, so Fig.13 As shown in (b), the load distribution after the change 115 and the probability distribution of the time after the load change 116 obtained by the probability distribution of the time until the predetermined crack shape is formed, which is estimated based on the crack shape and parameters obtained in step S4, are obtained. The reliability can be calculated based on the relationship between the probability distribution of time 116 and the repair time point 113. In addition, the contour line (new probability distribution) 117 is obtained.
[0068] By outputting the average value based on the obtained new probability distribution 117, for example, the load or displacement that becomes the operating condition for the structure 29 to work until the timing when it can be replaced or repaired can be calculated, and output to the target mechanical structure. Furthermore, a control device for controlling the structure 29 may be provided, and Fig.12 After step S6 of the prediction flowchart, a step of providing a control signal for controlling the structure 29 based on the obtained output is added.
[0069] In this way, by quantitatively expressing the uncertainty of the crack growth amount with probability distribution, the operating conditions under which the structure can operate until replacement or repair is required can be output together with the reliability, and maintenance can be adjusted to an appropriate period.
[0070] Fig.14This is a flowchart showing the operation of the crack inspection system 200 equipped with the crack development prediction device 1b described in Embodiment 3. In the crack inspection system 200, the shape of the structure 29, the force or displacement applied to the structure 29, the crack development characteristics of the material used, and the initial crack shape are input as parameters using the input unit 101. As input means, direct input of data using a keyboard or a mouse, input based on data transmitted using a storage device, reception of data sent from other devices in a wireless or wired manner, or reception of data sent from the Internet, a cloud, etc. can be considered, but are not limited to these.
[0071] The above-mentioned work is performed in the crack development prediction device 1b for the input parameters, that is, the probability distribution of each parameter is calculated by the parameter input unit 2, and the state space model 42 for predicting the development state of the crack is prepared in the model generation unit 3. In addition, the shape of the crack is measured by the crack shape measurement unit 4, and the measured crack shape is updated by the inference unit 5 using the crack shape predicted by the state space model 42, and the shape and parameters of the crack are inferred. The relationship between the probability distribution of the development time until the shape of the crack inferred by the inference unit 5 becomes the predetermined crack shape and the probability distribution of the force applied to the structure in the inferred parameters is obtained by the operating condition prediction unit 7, and the control conditions of the mechanical structure that provide the force that enables the structure to work after the measurement time point until the maintenance timing is calculated are calculated, and the control signal corresponding to the control condition or the probability distribution related to the control condition and the reliability and other maintenance-related information are output from the output unit 201. In addition, the output unit 201 may also have a display unit, or may be as shown in FIG. Fig.13 As shown in (a) and (b), the probability distribution and reliability are displayed. Furthermore, the output unit may be provided with a communication unit, and data based on the control conditions may be sent to other devices by wire or wireless means. Furthermore, it is also possible to connect to a network to accumulate these data in a cloud, etc., and accumulate data related to the control conditions of a large number of parts. Thus, the accumulated data can be used not only for maintenance, but also for the manufacture and sale of manufacturing equipment related to maintenance. In addition, it is also possible to have both the crack shape prediction unit 6 and the operating condition prediction unit 7 described in Implementation Example 2. Regarding the output unit at this time, it is also possible to have the functions of both the output unit 102 and the output unit 201, or it is also possible to have an independent output unit having the functions of the output unit 102 and the output unit 201 in each of the crack shape prediction unit 6 and the operating condition prediction unit 7.
[0072] A specific example of a part of a rotor of a generator is described. As a preferred example of calculating the value and reliability of the load that can be operated until the scheduled maintenance time, there is a temperature change of the rotor accompanying the output change of the generator. There is a possibility that cracks will develop due to thermal stress, wherein the thermal stress is caused by the temperature change of the rotor during the operation and stop of the generator. By controlling the output of the generator to decrease by the above method, the temperature of the rotor is prevented from becoming a certain temperature or above, so that the generator can be operated until the scheduled maintenance time.
[0073] Furthermore, the change in the output of the generator and the temperature change of the rotor parts are different for each generator, and the temperature change of the rotor parts in the control of the output change contains uncertainty. The output change output to the control device is determined by the output unit 201 in consideration of the temperature of the rotor parts output by the operating condition prediction unit 7 and its reliability, and the uncertainty of the temperature change of the rotor parts in the control of the output change. Thus, the rotor parts can be controlled in a manner that becomes the temperature output by the operating condition prediction unit 7. Here, the temperature is used as an example, but the same control can also be performed on a structure whose output changes due to an applied force. The specifications of the generator and information on other generators of the same type can also be obtained from the input unit 101. Implementation method 4.
[0074] Fig.15 The following is a prediction flow chart of the crack development prediction device 1 according to the embodiment of the present application. Only the parts changed from the embodiment 1 are described, and the rest are the same as the embodiment 1 and are therefore omitted. Fig.15 In the process, it is determined whether the next crack inspection needs to be performed (step S7). If it is performed, the parameters estimated in step S4 are used as the parameters input in step S1 (step S8). If it is not performed, the shape and parameters of the crack are output (step S9).
[0075] In this way, by repeatedly measuring the shape of the crack and updating the parameters, the uncertainty can be reduced, and the uncertainty of the estimated crack length can be reduced. Furthermore, based on the estimated result of the crack after reducing the uncertainty, the repair and replacement period of the structure can be postponed as much as possible, and more economical maintenance can be performed.
[0076] Each embodiment includes a processing circuit for executing the functions of the crack development prediction device 1, 1a, 1b. The processing circuit may be dedicated hardware or a CPU (Central Processing Unit, central processing unit, processing device, computing device, microprocessor, microcomputer, processor, DSP (Digital Signal Processor), etc.) that executes a program stored in a memory.
[0077] Fig.16 1 is a diagram for explaining an example of the hardware configuration of the crack development prediction device 1, 1a, 1b. Fig.16 In the embodiment, the processing circuit 601 is connected to the bus 602. When the processing circuit 601 is dedicated hardware, it is equivalent to, for example, a single circuit, a composite circuit, a programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or a combination thereof. Each of the functions of the crack growth prediction device 1, 1a, 1b can be implemented by the processing circuit 601, or the functions can be integrated and implemented by the processing circuit 601.
[0078] Fig.17 FIG. 1 is a diagram for explaining another example of the hardware configuration of the crack development prediction device 1, 1a, 1b. Fig.17 In the embodiment, the processor 603 and the memory 604 are connected to the bus 602. When the processor is a CPU, each of the functions of the crack development prediction device 1, 1a, 1b is implemented by software, firmware, or a combination of software and firmware. The software or firmware is described as a program and stored in the memory 604. The processing circuit implements the functions of each part by reading and executing the program stored in the memory 604. Here, the memory 604 refers to a non-volatile or volatile semiconductor memory such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, EPROM (Erasable Programmable Read Only Memory), EEPROM (Electrically Erasable Programmable Read Only Memory), or a magnetic disk, a floppy disk, an optical disk, a compact disk, a mini disk, a DVD (Digital Versatile Disc), etc.
[0079] In addition, each function of the crack development prediction device 1, 1a, 1b may be partially implemented by dedicated hardware and another part by software or firmware. For example, the model generation unit 3 of each function may be implemented by a processing circuit as dedicated hardware, and the estimation unit 5 may be implemented by the processor 603 reading and executing a program stored in the memory 604.
[0080] In addition, the value described in Implementation Example 1 can also be obtained by numerical analysis. Figure 5 Formulas (4) and (5) in . There are also formulas for the stress expansion coefficient range used to predict the development of cracks through experiments or theories. However, in actual structures, unformulated crack shapes also occur. By using numerical analysis to obtain the stress expansion coefficient range of unformulated crack shapes, there is no need to formulate through experiments or theories, and the time required for preparing the prediction device is shortened.
[0081] The present application describes various exemplary embodiments and examples, but various features, schemes, and functions described in one or more embodiments are not limited to application in specific embodiments and can be applied to the embodiments alone or in various combinations.
[0082] Therefore, within the scope of the technology disclosed in the present specification, numerous modifications that are not illustrated are conceivable, including, for example, a case where at least one component is modified, a case where at least one component is added, or a case where at least one component is omitted, and a case where at least one component is extracted and combined with components of other embodiments.
[0083] (Explanation of symbols)
[0084] 1, 1a, 1b: crack development prediction device; 2: parameter input unit; 3: model generation unit; 4: crack shape measurement unit; 5: estimation unit; 6: crack shape prediction unit; 7: operation condition prediction unit.
Claims
1. A crack development prediction device for predicting the development of cracks generated in a structure, characterized in that: have: A parameter input unit that inputs at least the parameters of the shape of the structure, the force applied to the structure, the material properties of the structure, and the initial value of the crack shape, including the uncertainty of each parameter, in a probability distribution; A model generation unit generates a state space model for predicting a crack development state composed of a state equation and an observation equation based on the input parameters; A crack shape measuring unit, which measures the crack shape of the structure; as well as The estimation unit estimates a posterior distribution including crack shape and parameters based on the likelihood distribution of uncertainty caused by the measurement value and measurement error of the crack shape measured by the crack shape measuring unit and the prior distribution of the crack shape predicted by the state space model.
2. The crack development prediction device according to claim 1, characterized in that: As the parameter, a change in temperature is added.
3. The crack development prediction device according to claim 1 or 2, characterized in that: The material property of the structure is a crack growth property.
4. The crack development prediction device according to any one of claims 1 to 3, characterized in that: A crack shape predicting unit is provided for predicting a probability distribution of a crack shape at an arbitrary time thereafter measured by the crack shape measuring unit, based on the crack shape estimated by the estimating unit.
5. The crack development prediction device according to any one of claims 1 to 3, characterized in that: An operating condition prediction unit is provided, which calculates a probability distribution of a time until a shape of a crack becomes a predetermined shape based on the posterior distribution estimated by the estimation unit.
6. The crack development prediction device according to claim 2, characterized in that: The invention is provided with an operating condition prediction unit, which calculates the relationship between the probability distribution of the time until the shape of the crack becomes a predetermined shape and the force or temperature applied to the structure based on the posterior distribution estimated by the estimation unit, and calculates the control conditions of the force or temperature that enable the structure to work after the measurement time point until the predetermined crack repair time point.
7. The crack development prediction device according to claim 6, characterized in that: A control device is provided, the control device controlling the structure according to the control condition calculated by the operating condition prediction unit.
8. The crack development prediction device according to claim 6, characterized in that: In the operating condition prediction unit, reliability is calculated based on a relationship between a probability distribution of time until a predetermined crack shape is achieved by the force or temperature changed by the control condition and the repair timing.
9. The crack development prediction device according to claim 8, characterized in that: A control device is provided, the control device controlling the structure based on the control condition calculated by the operating condition prediction unit and the reliability.
10. The crack development prediction device according to any one of claims 1 to 9, characterized in that: Every time the shape of the crack is measured by the crack shape measuring unit, the parameter updated by the estimating unit is input to the parameter input unit.
11. The crack development prediction device according to any one of claims 1 to 10, characterized in that: As the estimation unit, a collective Kalman filter is used.
12. The crack development prediction device according to any one of claims 1 to 11, characterized in that: Ultrasonic flaw detection is used in the crack shape measuring section.
13. A crack inspection system comprising: Crack development prediction device; An input unit for inputting parameters into the crack development prediction device; as well as an output unit, which outputs information related to maintenance including at least information on the next inspection time based on the output of the crack development prediction device, The crack development prediction device has the following features: A parameter input unit that inputs at least the parameters of the shape of the structure, the force applied to the structure, the material properties of the structure, and the initial value of the crack shape, including the uncertainty of each parameter, in a probability distribution; A model generation unit generates a state space model for predicting a crack development state composed of a state equation and an observation equation based on the input parameters; A crack shape measuring unit, which measures the crack shape of the structure; an inference unit that infers a posterior distribution including a crack shape and parameters based on a likelihood distribution of uncertainty caused by a measurement value of the crack shape measured by the crack shape measuring unit and a measurement error and a prior distribution of the crack shape predicted by the state space model; and The crack shape prediction unit predicts the probability distribution of the crack shape at an arbitrary time in the future measured by the crack shape measurement unit based on the crack shape estimated by the estimation unit.
14. The crack inspection system according to claim 13, characterized in that: An operating condition prediction unit is provided, which calculates the relationship between the probability distribution of the time until the shape of the crack becomes a predetermined shape and the force applied to the structure based on the posterior distribution estimated by the estimation unit, thereby calculating the force control condition that enables the structure to operate until the predetermined crack repair time point after the measurement time point, The crack inspection system outputs a control signal corresponding to the control condition or information related to maintenance in the output unit or a different output unit based on the output related to the control condition from the operating condition prediction unit.
15. The crack inspection system according to claim 13, characterized in that: An operating condition prediction unit is provided, which adds temperature to the parameter, and outputs the force or temperature at which the structure can operate until the maintenance timing after measurement together with the reliability based on the crack shape of the structure obtained by the estimation unit and the parameter, The crack inspection system inputs the operating conditions output by the operating condition prediction unit, the reliability, and the control signal, and based on the control uncertainty calculated based on the difference between the actual working state of the structure and the indication of the control signal, outputs a signal from the output unit or a different output unit to control the structure in a manner that becomes the force or temperature output by the operating condition prediction unit.
16. A method for predicting the development of cracks, for predicting the development of cracks generated in a structure, comprising: A step of inputting at least the parameters of the shape of the structure, the force applied to the structure, the material properties of the structure, and the initial value of the crack shape, including the uncertainty of each parameter, in the form of probability distribution; A step of generating a state space model for predicting the crack development state composed of a state equation and an observation equation according to the input parameters; The step of measuring the crack shape of the structure; as well as The step of inferring a posterior distribution including crack shape and parameters based on the probability distribution of uncertainty caused by the measurement value of the crack shape measured by the crack shape measuring unit and the measurement error and the prior distribution of the crack shape predicted by the state space model, The crack development prediction method predicts the development of a crack based on the crack shape and parameters updated by inference.
