Estimation device, estimation method, and computer-readable medium.

Description

1 / 41 Estimation device, estimation method, and computer-readable medium. FIELD OF TECHNIQUE

[001] The present disclosure relates to an estimation device, an estimation method, and a computer-readable means for estimating a response from a structure at an unmeasured point. The present disclosure claims priority based on Japanese patent application No. 2020-191656, filed in Japan on November 18, 2020, the contents of which are incorporated herein by reference. BACKGROUND OF THE INVENTION

[002] There has been an increase in the number of industries using floating installations, such as the expansion of offshore oil and gas production using a floating production, storage and offloading system (FPSO) and the use of offshore wind turbines. These floating structures need to be moored and operated in a sea area where they are installed for a long period of time. Therefore, the structures cannot be periodically anchored for inspection and repair, as marine vessels can. Currently, outer circumference inspection has been carried out underwater, and partial tank inspection has been carried out approximately once every five years. These inspections include visual inspection and measurement of the plate thickness at representative points.This is nothing more than an assessment of a state at the point of inspection, meaning that the remaining service life for the long subsequent service period can be quantitatively calculated with the inspection. A structural model can be used to estimate a load acting on it based on statistical data, obtain a structural response, and calculate fatigue damage using the results. Still, this is a simple estimate based on the initial state of the floating structure; that is, an estimate that does not take the current state of the floating structure into account. Therefore... Petition 870260066319, dated 06 / 07 / 2026, page 9 / 108 2 / 41 form, the remaining service life cannot be determined quantitatively while taking actual operating conditions and aging into consideration.

[003] The actual operating state of a floating structure is monitored by monitoring a value measured by sensors attached to the floating structure. However, the sensors can only be attached to an external surface and a dry region of the floating structure, such as an upper platform, which means that an internal structure and a member on an internal surface of the tank, which are typical critical areas, cannot be directly detected. To monitor the state of a critical area, a structural response needs to be calculated using a digital twin, which is a virtual replica of the floating structure. Unfortunately, sufficient accuracy is not obtained with the state of the art.

[004] Patent document 1 discloses a system for monitoring a physical change in an offshore structure using optical sensors.

[005] Patent document 2 discloses a system for assisting ship navigation, with a processing unit that corrects structural load estimates and / or magnitude estimates based on a compensation value. The system has navigation and environmental detection units that provide indications of ship movements and sea waves, respectively. A processing unit calculates a magnitude estimate at a predefined point on the ship's structure during movement. The contributions of the first and second estimates are calculated. A magnitude indication provided by a reference sensor during ship movement is compared with the magnitude estimate to determine a compensation value. The structural load estimates and / or the magnitude estimate are corrected based on the compensation value.

[006] Non-patent document 1 discloses a method for estimating the stress produced in a vessel and a method for estimating fatigue damage to a vessel structure. Non-patent document 2 discloses a method for Petition 870260066319, dated 06 / 07 / 2026, p. 10 / 108 3 / 41 estimate for a wave spectrum based on values ​​measured by sensors attached to the vessel. LIST OF REFERENCES Patent document

[007] Patent Document 1: JP 2016-166001 A

[008] Patent Document 2: WO 2016 / 108183 A1 Non-patent literature

[009] Non-Patent Document 1: Tetsuo Okada, Yasumi Kawamura, Atsushi Kato, Hideyuki Ando, ​​Takeshi Yonezawa, Fumitaka Kimura, Masanobu Toyoda, Akihiko Yamauchi, Toshiro Arima, Masayoshi Oka, Toshiyuki Matsumoto, Hiroaki Kakizaki: 14,000TEU Large Container Ships, Conference Proceedings, Japan Society of Naval Architects and Ocean Engineers, No. 24, 2017. 05, pages 31 to

[010] Non-patent document 2: Yuki Yoshihira, Tetsuo Okada, Yasumi Kawamura, Yuki Terada: A study on the estimation method of an ocean wave spectrum using measured hull responses on 14,000TEU large container ships. 2017. 05, pages 37-42. SUMMARY OF THE INVENTION Technical Problem

[011] There is a demand for an estimation method to estimate a response at a position of a floating structure, such as an FPSO, where a sensor that measures a structural response cannot be attached.

[012] The present disclosure provides an estimation device, an estimation method, and a computer-readable means that can solve the problem described above. Solution to the Problem Petition 870260066319, dated 06 / 07 / 2026, page 11 / 108 4 / 41

[013] A monitoring device, according to the present disclosure, includes: a strain response spectrum capture unit configured to capture a strain response spectrum at an installation position where a strain sensor is installed on a floating structure, wherein the strain response spectrum is calculated based on a measurement value from the strain sensor; a wave spectrum capture unit configured to capture a wave spectrum of a wave at a position where the floating structure is located; a RAO capture unit configured to capture a strain response function (response amplitude operator (RAO)) at the installation position of the strain sensor and a strain response function (RAO) at a non-installation position of the strain sensor, calculated based on a structure model of the floating structure;a unit of calculation for the amount of correction configured to calculate, based on the strain response spectrum and the strain response function (RAO) at the strain sensor installation position, the wave spectrum and a formula that expresses a relationship between the strain response spectrum and the strain response function (RAO) at the strain sensor installation position and the wave spectrum, an amount of correction for a difference between a theoretical value of the strain response spectrum at the strain sensor installation position calculated from the strain response function (RAO) and the wave spectrum and the strain response spectrum at the strain sensor installation position;and a strain response spectrum estimation unit configured to calculate, based on a theoretical value of the strain response spectrum at the non-installation position calculated from the strain response function (RAO) and the wave spectrum at the non-installation position of the strain sensor, and the amount of correction, the spectrum of; Petition 870260066319, dated 06 / 07 / 2026, page 12 / 108 5 / 41 response to deformation in the non-installation position.

[014] An estimation method according to the present disclosure includes: a step of capturing a strain response spectrum at an installation position where a strain sensor is installed on a floating structure, wherein the strain response spectrum is calculated based on a measurement value from the strain sensor; a step of capturing a wave spectrum of a wave at a position where the floating structure is located; a step of capturing a strain response function at the installation position of the strain sensor and a strain response function (SRF) at a non-installation position of the strain sensor, calculated based on a structure model of the floating structure;a calculation step, based on the strain response spectrum and the strain response function (RAO) at the strain sensor installation position, the wave spectrum, and a formula that expresses a relationship between the strain response spectrum and the strain response function (RAO) at the strain sensor installation position and the wave spectrum, a correction amount for a difference between a theoretical value of the strain response spectrum at the strain sensor installation position calculated from the strain response function (RAO) and the wave spectrum and the strain response spectrum at the strain sensor installation position;and a calculation step, based on a theoretical value of the strain response spectrum in the non-installation position calculated from the strain response function (RAO) and the wave spectrum in the non-installation position of the strain sensor, and the amount of correction, the strain response spectrum in the non-installation position.

[015] A computer-readable means according to the present disclosure causes a computer to perform processing including: capturing a strain response spectrum at an installation position in which a sensor Petition 870260066319, dated 06 / 07 / 2026, page 13 / 108 A 6 / 41 strain sensor is installed on a floating structure, and the strain response spectrum is calculated based on a measurement value from the strain sensor; capture a wave spectrum of a wave at a position where the floating structure is located; capture a strain response function (response amplitude operator (RAO)) at the strain sensor installation position and a strain response function (RAO) at a non-installation position of the strain sensor, calculated based on a structure model of the floating structure;Calculate, based on the strain response spectrum and the strain response function (RAO) at the strain sensor installation position, the wave spectrum and a formula that expresses a relationship between the strain response spectrum and the strain response function (RAO) at the strain sensor installation position and the wave spectrum, a correction quantity for a difference between a theoretical value of the strain response spectrum at the strain sensor installation position calculated from the strain response function (RAO) and the wave spectrum and the strain response spectrum at the strain sensor installation position;and calculate, based on a theoretical value of the strain response spectrum in the non-installation position calculated from the strain response function (RAO) and the wave spectrum in the non-installation position of the strain sensor, and the amount of correction, the strain response spectrum in the non-installation position. Advantageous Effects of the Invention

[016] The estimation device, the estimation method and the computer-readable means described above allow the estimation of a response in a position where a sensor that detects a structural response is not installed. BRIEF DESCRIPTION OF THE DRAWINGS

[017] Figure 1 is a block diagram that illustrates an example of a Petition 870260066319, dated 06 / 07 / 2026, p. 14 / 108 7 / 41 monitoring device according to a modality.

[018] Figure 2 is a diagram that illustrates an example of a floating structure according to one modality.

[019] Figure 3 is a diagram that illustrates an example of a floating structure model according to one modality.

[020] Figure 4A is a diagram that illustrates a correction amount according to a modality.

[021] Figure 4B is a diagram that illustrates a correction quantity S according to a modality.

[022] Figure 5 is a flowchart that illustrates an example of estimation processing for a strain response spectrum at an unmeasured position according to a modality.

[023] Figure 6 is a diagram that illustrates an example of a Bayesian network model obtained by total modeling of the strain response spectrum estimation processing at the unmeasured position according to a modality.

[024] Figure 7 is a diagram that illustrates an example of a monitoring device hardware configuration according to a modality. DESCRIPTION OF THE MODALITIES Modalities

[025] A monitoring device according to the present disclosure will be described below with reference to Figure 1 through Figure 7. Settings

[026] Figure 1 is a block diagram that illustrates an example of a monitoring device according to a modality.

[027] A monitoring device 10 monitors the status of an offshore floating structure. Examples of floating structures include an FPSO, an offshore wind turbine, and the like. The monitoring device 10 includes: an apparatus Petition 870260066319, dated 06 / 07 / 2026, page 15 / 108 8 / 41 monitoring device 10A that captures measurement values ​​measured by various sensors attached to the floating structure, calculates a monitoring parameter by performing the necessary calculation or similar on the measurement values, and outputs the monitoring parameter; and a simulator (digital twin) 10B that simulates an operation, an internal state, and similar of the floating structure based on a three-dimensional structure model and similar of the floating structure. The monitoring device 10 includes a data capture unit 11, a strain response spectrum calculation unit 12, a RAO calculation unit 13, a storage unit 14, an output unit 15, an estimation unit 16, and an evaluation condition capture unit 17. For example, the data capture unit 11 and the strain response spectrum calculation unit 12 are functions of the monitoring device 10A.The calculation unit of RAO 13 and the estimation unit 16 are functions of simulator 10B.

[028] The data capture unit 11 captures a strain measured by a strain sensor, an acceleration measured by an accelerometer, a wave spectrum emitted from a wave radar, a measurement value of a plate thickness measured by a plate thickness sensor, and the like. Figure 2 illustrates an example of the floating structure. A wave radar 21, an accelerometer 22, strain sensors 23 to 25, and a plate thickness sensor 26 are attached to the floating structure. These sensors and the monitoring device 10 are communicatively connected to each other. Each of the sensors performs the measurement for a measurement target within a predetermined control interval and transmits a measurement result to the monitoring device 10.The data capture unit 11 captures the wave spectrum emitted from the wave radar 21, as well as the sensor data measured by the accelerometer 22, the strain sensors 23 to 25, and the plate thickness sensor 26, and records and stores each sensor data in the storage unit 14 along with the measurement time. The wave spectrum indicates the... Petition 870260066319, dated 06 / 07 / 2026, page 16 / 108 9 / 41 wave power with each wave direction and frequency. The plate thickness sensor 26 measures the plate thickness of a floating structure (the upper platform of the floating structure, for example), and inputs the measurement result into the monitoring device 10. The data capture unit 11 additionally captures the plate thickness of an internal structure such as an internal tank surface measured during a regular inspection, and records and stores the plate thickness in the storage unit 14 along with the measurement date. When a crack is found during the inspection, the position and length of the crack are recorded and stored in the storage unit 14 along with the measurement date. A position 30 illustrated in Figure 2 is an example of a target evaluation position where the strain sensor cannot be fixed, such as an internal surface structural member of the tank.

[029] The strain response spectrum calculation unit 12 reads time history data on the strain measured by strain sensors 23 to 25 from the storage unit 14 and calculates a strain response spectrum at the positions where strain sensors 23 to 25 are installed. The method for calculating the strain response spectrum is known and therefore not described in this descriptive report. The strain response spectrum indicates the strain (or stress) spectrum for each frequency. The strain response spectrum calculation unit 12 records and stores the calculated strain response spectrum in the storage unit 14. Strain and stress are values ​​that are easily interchangeable with the use of a material constant and therefore can be treated as having the same meaning. In many cases, strain and stress are respectively used in measurement and calculation.In this descriptive report, for example, a description will be given using deformation as an example, that is, using a strain response spectrum and a response function. Petition 870260066319, dated 06 / 07 / 2026, p. 17 / 108 10 / 41 to the deformation that can be replaced respectively by a stress response spectrum and a stress response function.

[030] The RAO 13 calculation unit calculates a strain response function (response amplitude operator (RAO)) through a finite element method (FEM) analysis using a floating structure model or any other similar structure analysis. Figure 3 illustrates an example of a three-dimensional structure model. A 20D structure model is a three-dimensional model that simulates the structural response of a floating structure 20. The RAO 13 calculation unit uses the 20D structure model to perform semi-static FEM analysis for a case where the floating structure receives regular waves, and calculates the strain response function (RAO) at positions 23D to 25D corresponding to the positions where strain sensors 23 to 25 are fixed, and a position 30D which corresponds to a target evaluation position 30.The RAO calculation unit 13 performs FEM analysis using the most recent information on plate thickness stored in storage unit 14. The strain response function (RAO) is a set of strain response magnifications corresponding to a ratio between the vessel length and the wavelength of the waves, for each wave direction. In other words, the strain response function (RAO) is a set of strain response magnifications for each wave frequency in each wave direction. For example, with the strain response function (RAO), a large strain response magnification is obtained with a wave that has a frequency that corresponds to the natural frequency of a structure. The method for calculating the strain response function (RAO) is known and therefore not described in this descriptive report.The RAO 13 calculation unit calculates strain response functions (RAO) for regular waves of various frequencies from various directions. Petition 870260066319, dated 06 / 07 / 2026, p. 18 / 108 11 / 41 and records and stores strain response functions (RAO) in storage unit 14 in advance. Later in this document, the positions where strain sensors 23 to 25 are fixed can be called positions 23 to 25 or measured points 23 and 25. Similarly, the position where plate thickness sensor 26 is fixed can be described as a position 26 or a measured point 26.

[031] When the most recent plate thickness information measured by plate thickness sensor 26 is obtained, a plate thickness estimation unit 167 estimates plate thicknesses at positions 23 to 25 and 30 while taking into account thinning due to corrosion. Then, the RAO calculation unit 13 recalculates the strain response functions (RAO) at positions 23 to 25 and 30 using the estimated plate thickness values ​​at positions 23 to 25 and 30, taking into account the increase in stress due to the reduction in plate thickness.

[032] Storage unit 14 stores the sensor data captured by data capture unit 11, the 20D structure model of the digital twin and the calculation results such as the strain response spectrum and strain response functions (RAO).

[033] Output unit 15 emits the strain response spectrum in a position where the strain sensor is not installed, or similar, to a monitor or an electronic file.

[034] Estimation unit 16 estimates the strain response spectrum at a non-installation position of the strain sensor (position 30), based on the wave spectrum captured by data capture unit 11, the strain response spectrum at the strain sensor installation positions (positions 23 to 25) calculated by strain response spectrum calculation unit 12, and the strain response function (RAO) at the strain sensor installation position and the strain response function (RAO) at the non-installation position calculated by Petition 870260066319, dated 06 / 07 / 2026, page 19 / 108 12 / 41 RAO calculation unit 13. Estimation unit 16 corrects the strain response spectrum described above successively, based on the strain response function (RAO) recalculated by RAO calculation unit 13, using an estimated plate thickness value estimated by plate thickness estimation unit 167, taking into account thinning due to corrosion in the plate thickness sensor at the unmeasured position (positions 23 to 25 and 30). Estimation unit 16 estimates the fatigue level / remaining service life at the non-installation position of the strain sensor. In this way, it is possible to determine the strain response spectrum and the fatigue level / remaining service life at a position where the strain sensor cannot be fixed. The estimation method for the fatigue level / remaining service life based on the strain response spectrum is known and therefore its description will be omitted.

[035] Estimation unit 16 includes a strain response spectrum capture unit 161, a wave spectrum capture unit 162, an ORA capture unit 163, a correction quantity calculation unit 164, a strain response spectrum estimation unit 165, a plate thickness capture unit 166, a plate thickness estimation unit 167 and a degradation estimation unit 168.

[036] The strain response spectrum capture unit 161 reads and thus captures the strain response spectrum at positions 23 to 25 in Figure 2 from the storage unit 14. The strain response spectrum capture unit 161 outputs the strain response spectrum to the correction quantity calculation unit 164.

[037] Wave spectrum capture unit 162 reads and thus captures the wave spectrum from storage unit 14. Wave spectrum capture unit 162 outputs the wave spectrum to correction quantity calculation unit 164 and strain response spectrum estimation unit 165. Petition 870260066319, dated 06 / 07 / 2026, page 20 / 108 13 / 41

[038] The RAO capture unit 163 reads and thus obtains RAO (strain response amplification for each frequency) information corresponding to the same wave direction as the wave direction of the wave spectrum measurement result, from storage unit 14 storing information calculated by the RAO calculation unit 13 for the strain response function (RAO) at positions 23D to 25D and 30D in Figure 3. The RAO capture unit 163 outputs the RAO information to the correction quantity calculation unit 164 and the strain response spectrum estimation unit 165.

[039] The correction quantity calculation unit 164 calculates a correction quantity for the strain response spectrum at position 30D. In general, it is known that the strain response spectrum, the strain response function (RAO) and the wave spectrum satisfy the relationship expressed by the following Formula (1): Strain response spectrum = (strain response function (SRF))2 χ wave spectrum. (1).

[040] The correction quantity calculation unit 164 calculates the correction quantity to correct the difference between the left and right sides in Formula (1), from the strain response spectrum based on the measurement values ​​at positions 23 to 25, the strain response function (RAO) calculated for positions 23D to 25D, the wave spectrum and the relationship expressed by Formula (1). Now reference is made to Figure 4A.

[041] Figure 4A is a diagram that illustrates a correction amount according to the modality.

[042] The vertical geometric axis and the horizontal geometric axis of a graph in Figure 4A represent, respectively, the strain response spectrum and the frequency. A C1 curve indicates the strain response spectrum based on Petition 870260066319, dated 06 / 07 / 2026, p. 21 / 108 14 / 41 in the measurement value at any of the positions 23 to 25. Thus, curve C1 corresponds to the left side in Formula (1) at any of the positions 23 to 25. A curve C2 is a waveform obtained by multiplying the square of the strain response function (RAO) calculated for the position corresponding to curve C1 between positions 23D to 25D, by the wave spectrum. Thus, curve C2 corresponds to the right side in Formula (1). The strain response spectrum, the strain response function (RAO), and the wave spectrum at the measured strain points 23 to 25 are known through measurement and analysis. Thus, a difference obtained by comparing the right side and the left side in Formula (1) is defined as an error in the strain response function (RAO). Thus, the error between the right side and the left side in Formula (1) is considered as an error between the 20D structure model and the actual floating structure 20.With the error thus defined, obtaining the correction amount to correct the error, a difference between the structural response of the floating structure calculated by the 20D structure model established under the design condition and the response of the actual floating structure 20 can be compensated. For a position where the sensor that detects the structure's response is not installed, the response can be estimated with high precision at the position where the sensor is not installed using the response value and the correction amount calculated by the method of the present embodiment. Specifically, the strain response spectrum at the non-installation position of the strain sensor 30 can be estimated.

[043] An example of the definition is described using L, K, and S errors in Figure 4A. The right and left sides of Formula (1) described above may not correspond. Since the comparison is made between distributions, there are three possible differences between them, including an L error in peak position, a K error in peak quantity, and an S error related to width. Petition 870260066319, dated 06 / 07 / 2026, page 22 / 108 15 / 41 distribution between curve C1 and curve C2, as illustrated in the graph in Figure 4A. The error K is a difference between the maximum value of the strain response spectrum of curve C1 and the maximum value of the strain response spectrum of curve C2. The error L is a difference between the position on the horizontal axis where the strain response spectrum of curve C1 has the maximum value, and the position on the horizontal axis where the strain response spectrum of curve C2 has the maximum value. For the error S, first, consideration is given to an integrated value m of a frequency domain of the strain response spectrum f(ω) illustrated in Figure 4B. Specifically, m is defined by the following Formula (2). So, for a predetermined α (α = 5% in Figure 4B), a lower bound and an upper bound of a range of both side sections 100 χ (1 - α)% of m are respectively defined by ωι in Formula (3) and ωυ in Formula (4). [Mathematics 1] [Mathematics 2] [Mathematics 3]

[044] Next, the length of the interval of both side sections Petition 870260066319, dated 06 / 07 / 2026, page 23 / 108 16 / 41 100 χ (1-α) % is defined by a difference S = ωu - ωL between the upper limit value ωu and the lower limit value ωL. Then, using this two-sided interval length S, the error S related to the distribution width is defined as follows. Specifically, the error S in the example illustrated in Figure 4A is the value calculated by S1 / S2, where S1 is a length of a two-sided 95% interval (α = 5%; 95% is an example and, for instance, the value could be 99% (α = 1%) or 90% (α = 10%) of the integrated frequency domain value of the C1 curve strain response spectrum, and S2 is a length of a two-sided 95% interval of the integrated frequency domain value of the C2 curve strain response spectrum. The unit calculation of the correction quantity 164 calculates the error L, the error K, and the error S as correction quantities.In the following description, a correction quantity to correct the error L can be called a correction quantity L, a correction quantity to correct the error K can be called a correction quantity K, and a correction quantity to correct the error S can be called a correction quantity S. Although a method of calculating the correction quantity to estimate the response at the unmeasured position taking into account uncertainties will be described below with reference to Figure 6, as an extremely simple method, the correction quantity calculation unit 164 can calculate the error L, the error K, and the error S using Formula (1) for each type 23 to 25, and obtain an average value of each error type as the correction quantity. Alternatively, based on the positional relationship with the target evaluation position 30, the correction quantity calculation unit 164 can use the correction quantity calculated for the nearest position.For example, when the target valuation position 30 is located between position 24 and position 25, the correction quantity calculation unit 164 obtains the average values ​​of the L errors, K errors, and S errors calculated for positions 24 and 25, as the correction quantity L, correction quantity K, and correction quantity S for the target position. Petition 870260066319, dated 06 / 07 / 2026, page 24 / 108 17 / 41 rating 30.

[045] As described below, the correction quantity calculation unit 164 can calculate the correction quantity L, the correction quantity K, and the correction quantity S for each position and for each wave condition, and can calculate the correction quantity L, the correction quantity K, and the correction quantity S corresponding to the member property in the position, in addition to the position and wave condition.

[046] The strain response spectrum estimation unit 165 calculates a theoretical value of the strain response spectrum at the target evaluation position 30 using the right-hand side in Formula (1), and corrects the theoretical value using the correction amounts (correction amount L, correction amount K, and correction amount S) calculated by the correction amount calculation unit 164, to calculate the estimated value of the strain response spectrum. Specifically, the strain response spectrum estimation unit 165 multiplies the wave spectrum captured by the wave spectrum capture unit 162 by the square of the strain response function (RAO) at position 30D captured by the RAO capture unit 163.Therefore, for the resulting distribution curve (curve C2 in Figure 4A), the position of the peak value is corrected by adding or subtracting the correction amount L for or from the position of the maximum value of the strain response spectrum, the peak value is corrected by adding or subtracting the correction amount K for or from the maximum value, and the distribution width is corrected by multiplying the width of a 95% confidence interval by the correction amount S or similar. The strain response spectrum estimation unit 165 outputs the strain response spectrum estimated in this way for the degradation estimation unit 168.

[047] The plate thickness capture unit 166 reads and thus captures information about the plate thickness of the floating structure 20 measured Petition 870260066319, dated 06 / 07 / 2026, page 25 / 108 18 / 41 by the plate thickness sensor 26 of the storage unit 14. The plate thickness sensor 26 is fixed to a position on the floating structure 20 where the sensor can be easily installed, such as the top platform. The plate thickness sensor 26 constantly measures the plate thickness of the top platform and transmits the latest measurement result to the monitoring device 10.

[048] The plate thickness estimation unit 167 estimates the plate thicknesses at the strain sensor installation position and the non-installation position of the strain sensor, using an estimation formula based on a theoretical solution or actual measurement or similar, based on plate thickness information captured from the plate thickness capture unit 166 and a plate thickness inspection result. For example, the plate thickness estimation unit 167 stores a table (theoretical solution) and a function (empirical formula from actual measurement) indicating the association between the amount of thinning at the installation position (the top platform, for example) of the plate thickness sensor 26 and the amount of thinning at the target plate thickness estimation position.The table and function define, for example, a relationship such as a 1 mm thinning on the upper deck leads to a 0.5 mm thinning at position 25 and a 0.3 mm thinning at position 30. The table and similar instruments are prepared in advance from the results of measurements or experiments on a real vessel, theoretical calculations, or similar. The plate thickness estimation unit 167 estimates the amount of thinning at the target estimation position based on the most recent plate thickness of the upper deck or similar measured by the plate thickness sensor 26 and the table or similar instruments indicating the relationship described above, and estimates the plate thickness of the target estimation position based on the estimated value of the quantity. Petition 870260066319, dated 06 / 07 / 2026, page 26 / 108 19 / 41 thinning. When plate thicknesses at positions 23 to 25 are measured by regular inspection or similar, for example, the error between the actual plate thickness and the estimated plate thickness value estimated using the table or similar described above can be calculated. The plate thickness estimation unit 167 can store the error thus calculated as the correction amount, and add the inspection-based correction amount to the estimated plate thickness value estimated based on the table or similar described above, to estimate the plate thickness at the target estimation position. Alternatively, the plate thickness estimation unit 167 can estimate the plate thickness at the target estimation position by the Bayes method based on the measurement result of the plate thickness sensor 26 and the measurement result of the plate thickness measured by regular inspection or similar.Plate thickness estimation unit 167 outputs estimated plate thickness values ​​from the sensor installation position and the non-installation position to the RAO calculation unit 13 and the degradation estimation unit 168. Based on the most recent plate thickness measurement value, the RAO calculation unit 13 calculates the most recent strain response function (RAO) at the estimated strain sensor installation position (positions 23D to 25D), and the most recent strain response function (RAO) at the non-installation position (position 30D). Correction quantity calculation unit 164 calculates the K, S, and L correction quantities using these most recent strain response functions (RAO). The strain response spectrum estimation unit 165 estimates the strain response spectrum using the most recent strain response function (RAO).

[049] When the plate thickness changes, the strain response spectrum of the structure also changes. In the present embodiment, by updating the response function Petition 870260066319, dated 06 / 07 / 2026, page 27 / 108 20 / 41 to strain (RAO) and the amount of correction based on the estimated value of the most recent plate thickness, the strain response spectrum in the non-installation position of the strain sensor can be estimated with high accuracy.

[050] Degradation estimation unit 168 uses a known method that uses an SN diagram (fatigue diagram) to estimate the fatigue level / remaining service life of the floating structure. Based on the estimated value of the plate thickness at the non-installation position of the strain sensor, or a type of joint of the structure at the non-installation position of the strain sensor, degradation estimation unit 168 selects an SN diagram based on the joint classification including the plate thickness or an SN diagram based on a hotspot stress using the plate thickness as a parameter. Furthermore, the SN diagram is selected while taking into account the corrosion environment, so that service life can be estimated with high accuracy. Degradation estimation unit 168 estimates the current short-term fatigue level of the floating structure, based on the selected SN diagram and the obtained strain response spectrum.The degradation estimation unit 168 estimates the future fatigue level of the floating structure, based on the selected SN diagram and a distribution of the number of times stress occurs over a long period. When a crack is found during a regular inspection, the degradation estimation unit 168 can similarly estimate how the crack will grow in the future using a known method and estimate when the crack will reach a dangerous length.

[051] Evaluation condition capture unit 17 captures the evaluation target data and an evaluation condition for strain response spectrum estimation. Examples of the evaluation condition include the location of the evaluation target position, a wave condition (wave direction and frequency), and the like. Operation Petition 870260066319, dated 06 / 07 / 2026, page 28 / 108 21 / 41

[052] Next, the estimation processing for a strain response spectrum at an unmeasured point will be described.

[053] Figure 5 is a flowchart illustrating an example of the estimation processing for a strain response spectrum at an unmeasured position according to the modality. With processes in steps S1 to S4, the function to output the correction quantity for each position and each wave condition is prepared. With processes in steps S5 to S8, the strain response spectrum and the remaining service life at the non-installation position of the sensor are calculated. Preparing a function to output a correction quantity.

[054] First, data capture unit 11 captures sensor data measured within a predetermined period (step S1). Data capture unit 11 captures a wave spectrum, deformation, acceleration, and plate thickness measured within the predetermined time period.

[055] Next, the strain response spectrum calculation unit 12 calculates the strain response spectrum using a known method based on the measurement values ​​measured by the strain sensor over a certain period of time (step S2).

[056] Next, the RAO calculation unit 13 uses the 20D structure model to calculate the strain response function (RAO) at the measured point that is a sensor installation position and the strain response function (RAO) at the unmeasured point (step S3). The calculation is performed under a regular wave load condition under any wave condition (wave direction and frequency). It is assumed that the strain response function (RAO) is calculated based on the most recent plate thickness information estimated by the plate thickness estimation unit 167.

[057] Next, the correction quantity calculation unit 164 prepares a function to calculate the correction quantity L, the correction quantity K and the Petition 870260066319, dated 06 / 07 / 2026, page 29 / 108 22 / 41 correction quantity S for each position and for each wave condition, based on the concept described with reference to Figure 4A and Figure 4B (step S4). Examples of wave conditions include wave frequency and wave direction. For example, upon receiving the position information and wave condition of position 23, the correction quantity calculation unit 164 prepares the function to output the correction quantity L, the correction quantity K, and the correction quantity S for position 23 that satisfies the wave condition. A method for preparing the function to calculate the correction quantity for each position and for each wave condition will be described below with reference to Figure 6. The correction quantity calculation unit 164 prepares each function to output the correction quantity L, the function to output the correction quantity S, and the function to output the correction quantity K, and records and stores the functions in storage unit 14.The correction quantity calculation unit 164 can prepare a function to calculate a correction quantity for a property of each member such as a general plate structure, a bone structure, and an intersection between bones. For example, when the member at position 23 has property A, the member at position 24 has property B, and the member at position 25 has property A, the correction quantity calculation unit 164 prepares the function corresponding to property A for positions 23 and 25, and prepares the function corresponding to property B for position 24. Estimated response time and remaining lifespan of the sensor in its non-installed position.

[058] Next, evaluation condition capture unit 17 captures the evaluation target data and the evaluation condition (step S5). For example, evaluation condition capture unit 17 captures the strain response function (RAO) of the evaluation target position 30D and a wave spectrum measured within a certain time period, as the evaluation target data. A Petition 870260066319, dated 06 / 07 / 2026, page 30 / 108 23 / 41 Evaluation Condition Capture Unit 17 captures the wave condition (wave direction and frequency) corresponding to the target evaluation position 30 and the wave spectrum of the target evaluation data as the evaluation condition. The target evaluation data may be the data obtained in step S1, or it may be newly captured data. When the wave spectrum is not measured for some reason, the wave spectrum can be estimated using an existing calculation formula using the significant wave height, wave direction, wave period and similar of the target sea area obtained in a different way.

[059] Next, the correction quantity calculation unit 164 calculates the correction quantity corresponding to the evaluation condition (step S6). The correction quantity calculation unit 164 inserts the target evaluation position and the wave condition designated by the evaluation condition into the function for correction quantity K to capture the correction quantity K corresponding to the evaluation condition. Similarly, the correction quantity calculation unit 164 inserts the evaluation condition for the function for correction quantity S to capture the correction quantity S, and inserts the evaluation condition for the function for correction quantity L to capture the correction quantity L.For example, when the member in position 30 has property B, the correction quantity calculation unit 164 inserts the evaluation condition for the respective functions for the correction quantities K, S, and L prepared for property B, to capture the correction quantities K, S, and L corresponding to property B, position, and wave condition.

[060] Next, the strain response spectrum estimation unit 165 estimates the strain response spectrum at the unmeasured point (step S7). The strain response spectrum estimation unit 165 calculates the strain response spectrum at the target evaluation position 30, correcting, using the correction quantities calculated in step S6, a value obtained by multiplication Petition 870260066319, dated 06 / 07 / 2026, page 31 / 108 24 / 41 of the square of the strain response function (RAO) calculated for the 30D evaluation target position from the wave spectrum of the evaluation target data.

[061] Next, the degradation estimation unit 168 estimates the fatigue level / remaining service life at the unmeasured point (step S8). The degradation estimation unit 168 estimates the fatigue level / remaining service life of the structure at the unmeasured point, based on the strain response spectrum estimated by the strain response spectrum estimation unit 165 and the SN diagram based on the joint type, including the most recent plate thickness, or the SN diagram based on the access point using the plate thickness as a parameter. Then, the output unit 15 outputs the strain response spectrum at the target evaluation position 30 and the fatigue level / remaining service life of the structure at the target evaluation position 30 to a screen or an electronic file (step S9). In this way, the strain response spectrum and the fatigue level / remaining service life at the non-installation position can be configured. Details of the correction amount calculation

[062] Next, an example of a method for calculating the amount of correction will be described with reference to Figure 6.

[063] Figure 6 illustrates an example of an estimation model obtained by total modeling, with a Bayesian network, the estimation processing of the strain response spectrum at the unmeasured position according to the modality.

[064] The correction quantity calculation unit 164 creates the estimation model illustrated as an example in Figure 6.

[065] In the estimation model of Figure 6, the correction amount K is expressed by a Gaussian process regression model using the parameters μΐ, σ1, and λ1. These K, μΐ, σ1, and λ1 correspond respectively to variables 312, 303, 304, and 305 in Figure 6. Similarly, the correction amount S is expressed by a Gaussian process regression model using the Petition 870260066319, dated 06 / 07 / 2026, p. 32 / 108 25 / 41 parameters μ2, σ2, and λ2, and the correction quantity L is expressed by a Gaussian process regression model using the parameters μ3, σ3, and λ3. These S, μ2, σ2, and λ2 correspond, respectively, to variables 313, 306, 307, and 308 in Figure 6, and these L, μ3, σ3, and λ3 correspond, respectively, to variables 314, 309, 310, and 311 in Figure 6. The Gaussian process regression models of the correction quantity K, S, L are wave condition and position functions. For example, when the position and wave condition are fed into a function fK expressed by the Gaussian process regression model for the correction quantity K, the Gaussian process regression model calculates the correction quantity K corresponding to the position and wave condition. The same applies to the correction quantities S and L.

[066] In the model illustrated in Figure 6, an arrow indicates that an object pointed to by the arrow depends on an object from which the arrow extends. For example, the amount of correction K depends on the parameters μΐ, σ1 and λ1.Based on the model illustrated in Figure 6, the correction quantity calculation unit 164 performs regression analysis, using a method such as Gaussian process regression, on a relationship between the strain response spectrum based on the measurement value obtained by the strain sensor and a value obtained by correcting the theoretical value of the strain response spectrum using the correction quantity K expressed by μ1, σ1 and λ1, the correction quantity S expressed by μ2, σ2 and λ2, and the correction quantity L expressed by μ3, σ3 and λ3, and adjusts the values ​​of the parameters μ1 to μ3, σ1 to σ3 and λ1 to λ3, to obtain a function fK for the correction quantity K, a function fS for the correction quantity S, and a function fL for the correction quantity L to calculate the correction quantities K, S and L with high precision.When Gaussian process regression is used, with an estimation method taking into account a correlation between input variables (wave condition and position), a trend emerges that a similar input leads to... Petition 870260066319, dated 06 / 07 / 2026, page 33 / 108 26 / 41 a similar output can be evaluated.

[067] A function 301 outputs a strain response function (RAO) corresponding to the wave condition and position, and a function 302 outputs a wave spectrum corresponding to the wave condition. A function 315 is for calculating and outputting the theoretical value of the strain response spectrum (the right-hand side in Formula (1)) at a position ie under a wave condition j, from the strain response function (RAO) output from function 301 and the wave spectrum output from function 302. Similarly, a function 316 is for calculating the theoretical value of the strain response spectrum at a position m under a wave condition n, and a function 317 is for calculating the theoretical value of the strain response spectrum at a position ke under a wave condition l. Position ie and position m are positions where strain sensors are installed, and position l corresponds to a non-installation position.A variable 318 is a value obtained by correcting the theoretical value of the strain response spectrum emitted by function 315 using the correction quantities K, S, and L. Similarly, a variable 319 is the corrected strain response spectrum at position me under wave condition n. A variable 320 is the corrected strain response spectrum at position ke under wave condition l. Conversely, a function 322 is a function that emits a true value of the strain response spectrum. Function 322 emits a variable 323 which is a true value of the strain response spectrum. An observed value 324 is a value actually observed at position ie under wave condition j. An observed value 325 is a value actually observed at position ke under wave condition l.The observed values ​​324 and 325 correspond to values ​​calculated by the strain response spectrum calculation unit 12 from the measurement values ​​obtained by the strain sensors. The observed values ​​324 and 325 include a measurement error. As illustrated in the Figure, the observed value 324 depends on variable 318 and variable 323, and the observed value... Petition 870260066319, dated 06 / 07 / 2026, page 34 / 108 27 / 41 325 depends on variable 319 and variable 323. The remaining service life (variable 326) at position i is estimated based on the strain-corrected response spectrum (variable 318) at position i. The same applies to the remaining service life (variable 327) at position m and the remaining service life (variable 321) at position k. The calculation unit of the correction quantity 164 calculates the parameters μ1 to μ3, σ1 to σ3, and λ1 to λ3, resulting in a small error between variable 318 and the observed value 324 and a small error between variable 319 and the observed value 325, based on the estimation model illustrated as an example in Figure 6.For example, the correction quantity calculation unit 164 captures a large number of strain response spectrum sets, wave spectra and strain response functions (SRF), based on measured values ​​at positions iame under wave conditions jan, and calculates the parameters μ1 to μ3, σ1 to σ3, and λ1 to λ3, resulting in a small difference between the strain response spectrum (observed values ​​324 and 325) based on the measured values ​​and the corrected strain response spectrum (variables 318 and 319), through regression analysis such as Gaussian process regression. When the parameters μ1 to μ3, σ1 to σ3, and λ1 to λ3 are calculated, the function fK for the correction quantity K using the parameters μ1, σ1 and λ1, the function fs for the correction quantity S using the parameters μ2, σ2 and λ2, and the function fL for the correction quantity L using the parameters μ3, σ3 and λ3 are obtained.These functions fK, fs, and fL are functions of the wave condition and position. Thus, the functions for the correction quantities K, S, and L can be calculated with high precision.

[068] In the example illustrated in Figure 6, the correction quantity calculation unit 164 calculates the functions that produce the correction quantities K, S, and L based on the position and wave condition. Alternatively, the correction quantity calculation unit 164 can prepare a Bayesian network model for each member property at each sensor installation position and calculate functions that produce the corresponding correction quantities K, S, and L. Petition 870260066319, dated 06 / 07 / 2026, page 35 / 108 28 / 41 position and wave condition for each member property.

[069] The correction quantities K, S, and L are not limited to Gaussian process regression and may be other regression analysis methods, such as polynomial regression, polynomial polynomial expansion, or neural network regression. By calculating the regression parameters of these regression analysis methods, the strain response spectrum at the non-installation position of the strain sensor can be estimated with high precision.

[070] Through the method of preparing the estimation model illustrated as an example in Figure 6 and the preparation of the functions for the correction quantities K, S and L through regression analysis, the actual measurement data can be accumulated and the correction quantities K, S and L can be updated as appropriate through recalculation. By updating data using monitoring data, the deformation can be estimated with high precision, with the uncertainty further reduced.

[071] The processing of preparing the functions (position and wave condition functions) for the K, S, and L correction quantities using this method will be described with reference to the flowchart in Figure 5. In step S4, the correction quantity calculation unit 164 prepares the respective functions for the K, S, and L correction quantities. In step S6, the position and wave condition defined in step S5 are entered into each of the functions for the K, S, and L correction quantities. As a result, the appropriate K, S, and L correction quantities for the position and wave condition are calculated, and the correction quantities are output to the strain response spectrum estimation unit 165. The other steps are similar to those described above.

[072] Recently, a digital twin has been increasingly used for hull monitoring. In this context, there is room for improvement in terms of organic cooperation between measured values ​​and a simulation result obtained. Petition 870260066319, dated 06 / 07 / 2026, page 36 / 108 29 / 41 by the digital twin. With the present embodiment, a physical quantity (the strain response spectrum and the remaining service life based on the strain response spectrum) can be estimated with high precision for a portion not measured by strain sensors, through data synchronization between the monitoring result and the simulation result. In particular, in a floating structure, such as an FPSO, where inspection and repair are limited, strain is difficult to measure, and it is important to verify the fatigue level, the amount of corrosion, the remaining service life, and similar factors for a member in a tank with a high risk of corrosion. With the present embodiment, the strain response spectrum in such a portion can be estimated with high precision.In this way, the current fatigue level and amount of corrosion can be estimated with high precision, and the remaining future service life can be predicted while taking corrosion into account, to contribute to inspection planning.

[073] Figure 7 is a diagram that illustrates an example of a monitoring device hardware configuration according to the modality.

[074] A computer 900 includes a CPU 901, a main storage device 902, an auxiliary storage device 903, an input / output interface 904 and a communication interface 905, and is connected to a sensor 800.

[075] The monitoring device 10 described above is implemented in computer 900. The functions described above are stored in auxiliary storage device 903 in the form of a program. CPU 901 reads the program from auxiliary storage device 903, deploys the program to main storage device 902, and executes the processing mentioned above according to the program. CPU 901 secures a storage area in main storage device 902 in accordance with the program. CPU 901 secures a storage area to store data under processing in the device. Petition 870260066319, dated 06 / 07 / 2026, page 37 / 108 30 / 41 of auxiliary storage 903 in accordance with the program. The measurement value measured by sensor 800 is entered into computer 900 via input / output interface 904 or communication interface 905, and is processed by CPU 901 to be stored in auxiliary storage device 903.

[076] It should be noted that a program to implement all or part of the functions of the monitoring device 10 may be recorded on a computer-readable recording medium, and a computer system may be required to read and execute the method that is recorded on the recording medium to perform the processing of the respective functional units. The computer system here includes hardware such as an operating system (OS) and peripheral equipment. If a worldwide web (WWW - world wide web) system is used, the computer system also includes a home page provisioning environment (or a display environment). The computer-readable recording medium refers to a portable medium, such as a CD, a DVD or a USB, or a storage device, such as a hard disk built into a computer system.When this program is distributed to computer 900 via a communication line, the receiving computer 900 can deploy the program to the main storage device 902 and perform the processing mentioned above. The program described above can implement some of the functions described above and, in addition, can also implement the functions described above in combination with a program already registered in the computer system.

[077] Previously, certain embodiments of the present disclosure have been described, but all such embodiments are merely illustrative and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the essence of the invention. These embodiments and modifications are included within the scope and essence of the invention and are also included within the scope of the invention described in the claims and their equivalents. Petition 870260066319, dated 06 / 07 / 2026, page 38 / 108 31 / 41

[078] For example, strain sensors can be stress sensors. A stress response spectrum and a stress response function (RAO) can be used instead of the strain response spectrum and strain response function (RAO) of the example described above. In this case, for example, Formula (1) above is replaced by the following Formula (1'): Voltage response spectrum = (voltage response function (RAO))2χ wave spectrum... (1'). Notes

[079] The estimation device, the estimation method and the computer-readable means described in the embodiments may be understood as follows, for example.

[080] An estimating device (monitoring device 10), according to a first aspect, includes: a strain response spectrum capture unit 161 configured to capture a strain response spectrum at an installation position where a strain sensor (sensors 23 to 25) is installed on a floating structure (floating structure 20), wherein the strain response spectrum is calculated based on a measurement value from the strain sensor; a wave spectrum capture unit 162 configured to capture a wave spectrum of a wave at a position where the floating structure is located;a RAO capture unit 163 configured to capture a strain response function (response amplitude operator (RAO)) at the strain sensor installation position and a strain response function (RAO) at a non-strain sensor installation position, calculated based on a floating frame structure model; a correction quantity calculation unit 164 configured to calculate, based on the strain response spectrum and the strain response function (RAO) at the strain sensor installation position; Petition 870260066319, dated 06 / 07 / 2026, page 39 / 108 32 / 41 strain, the wave spectrum, and a formula (Formula (1)) that expresses a relationship between the strain response spectrum and the strain response function (RAO) at the strain sensor installation position and the wave spectrum, a correction quantity (L, K, S) for a difference between a theoretical value of the strain response spectrum at the strain sensor installation position calculated from the strain response function (RAO) and the wave spectrum and the strain response spectrum at the strain sensor installation position; and a strain response spectrum estimation unit 165 configured to calculate, based on a theoretical value of the strain response spectrum at the non-installation position calculated from the strain response function (RAO) and the wave spectrum at the non-installation position of the strain sensor, and the correction quantity, the strain response spectrum at the non-installation position.

[081] With this configuration, the strain response spectrum of a position where the strain sensor is difficult to fix can be estimated. The strain response spectrum at the non-installation position of the strain sensor, calculated using a structure model, is corrected using the amount of correction calculated through data synchronization between a monitoring result and a simulation result. In this way, the strain response spectrum at the non-installation position of the strain sensor can be estimated with high precision.

[082] As described above, since strain and stress are interchangeable, the first aspect described above is synonymous with the following description. That is, an estimating device (monitoring device 10) includes: a strain response spectrum capture unit 161 configured to capture a strain response spectrum at an installation position in which a strain sensor (strain sensors 23 to 25) is installed in a Petition 870260066319, dated 06 / 07 / 2026, page 40 / 108 33 / 41 floating structure (floating structure 20), wherein the strain response spectrum is calculated based on a measurement value from the strain sensor; a wave spectrum capture unit 162 configured to capture a wave spectrum of a wave at a position where the floating structure is located; an RAO capture unit 163 configured to capture a strain response function (amplitude response operator (RAO)) at the strain sensor installation position and a strain response function (RAO) at a non-installation strain sensor position, calculated based on a floating structure model;a correction quantity calculation unit 164 configured to calculate, based on the strain response spectrum and the stress response function (RAO) at the strain sensor installation position, the wave spectrum, and a formula (Formula (1)) that expresses a relationship between the strain response spectrum and the stress response function (RAO) at the strain sensor installation position and the wave spectrum, a correction quantity (L, K, S) for a difference between a theoretical value of the strain response spectrum at the strain sensor installation position calculated from the stress response function (RAO) and the wave spectrum and the strain response spectrum at the strain sensor installation position;and a strain response spectrum estimation unit 165 configured to calculate, based on a theoretical value of the strain response spectrum at the non-installation position calculated from the stress response function (RAO) and the wave spectrum at the non-installation position of the strain sensor, and the amount of correction, the strain response spectrum at the non-installation position.

[083] The same applies to the following second to eleventh aspects below.

[084] (2) An estimating device (monitoring device 10), of Petition 870260066319, dated 06 / 07 / 2026, p. 41 / 108 34 / 41 according to a second aspect, is the estimation device (1), which additionally includes: a plate thickness capture unit 166 configured to capture information about a plate thickness of a structure to which a plate thickness sensor is installed, measured by the plate thickness sensor, the plate thickness sensor being installed in a position that is identical to or different from the strain sensor; a plate thickness estimation unit 167 configured to estimate a plate thickness in the non-installation position of the strain sensor based on the plate thickness measured by the plate thickness sensor and a plate thickness in any position measured during an inspection;and an RAO calculation unit 13 configured to calculate the strain response function (RAO) in the non-installation position of the strain sensor using the plate thickness estimated by the plate thickness estimation unit, wherein the RAO capture unit 163 captures the strain response function (RAO) calculated by the RAO calculation unit 13, and the strain response spectrum estimation unit 165 calculates the strain response spectrum based on the captured strain response function (RAO).

[085] The plate thickness estimation unit 167 estimates the plate thickness while taking into account thinning due to corrosion in the non-installation position of the strain sensor, and the RAO calculation unit 13 recalculates the strain response function (RAO) based on this estimated plate thickness value. In this way, improved accuracy of the strain response spectrum estimated by the strain response spectrum estimation unit 165 can be expected. The most recent plate thickness measured by the plate thickness sensor 26 is used to recalculate the strain response function (RAO) and to recalculate the strain response spectrum. In this way, the current strain response spectrum in the non-installation position of the strain sensor can be Petition 870260066319, dated 06 / 07 / 2026, page 42 / 108 35 / 41 estimated with high precision. Every time the most recent measurement value measured by the plate thickness sensor is obtained, the plate thickness estimation unit 167 estimates the plate thickness at the non-installation position of the strain sensor. Using this result to perform strain response spectrum estimation, it is possible to discover the strain response spectrum transition at the non-installation position of the strain sensor.

[086] (3) An estimating device (monitoring device 10), according to a third aspect, is the estimating device of (2), in which the plate thickness capture unit 166 captures information about a plate thickness measured by the plate thickness sensor installed on an upper platform of the floating structure, and the plate thickness estimating unit 167 estimates the plate thickness in the non-installation position, based on a relationship between an amount of thinning of the upper platform and an amount of thinning of the structure in the non-installation position.

[087] The plate thickness sensor is installed on the upper platform where the plate thickness sensor can be easily installed. Based on a relationship (association) between the amount of thinning of the upper platform measured by the plate thickness sensor and the amount of thinning at a target estimation position (e.g., a position where the strain sensor cannot be installed), the plate thickness at the target estimation position is estimated, so that information about the plate thickness of the upper platform can be captured from the plate thickness sensor. In this way, the plate thickness at the target estimation position, where the plate thickness sensor cannot be fixed, can be estimated.

[088] (4) An estimating device (monitoring device 10), according to a fourth aspect, is the estimating device of (1) to (3), in which the correction quantity calculation unit calculates a correction quantity Petition 870260066319, dated 06 / 07 / 2026, page 43 / 108 36 / 41 to correct a difference in a peak position (L), a correction amount to correct a difference in a peak value (K), and a correction amount to correct a difference in a distribution width (S) between a waveform expressing a frequency distribution of a strain response spectrum indicated by the theoretical value of the strain response spectrum and a waveform expressing a frequency distribution of a strain response spectrum indicated by the strain response spectrum at the strain sensor installation position.

[089] With this configuration, the amount of correction for the strain response spectrum that indicates the stress frequency distribution can be calculated.

[090] (5) An estimating device (monitoring device 10), according to a fifth aspect, is the estimating device of (1) to (4), in which the correction quantity calculation unit calculates the correction quantity corresponding to each of different positions of the floating structure.

[091] For example, a difference between the theoretical value and the measured value of the strain response spectrum differs between the front, rear, and middle parts of the vessel. By calculating the amount of correction for each position, the strain response spectrum can be estimated with high precision.

[092] (6) An estimating device (monitoring device 10), according to a sixth aspect, is the estimating device of (1) to (5), in which the correction quantity calculation unit calculates the correction quantity for each wave condition indicated by the wave spectrum at a position where the floating structure is situated.

[093] For example, when a condition (wave direction, frequency) of the wave received by the floating structure changes, the difference between the theoretical value and the measurement value of the strain response spectrum changes even at the same position. Petition 870260066319, dated 06 / 07 / 2026, page 44 / 108 37 / 41 Calculate the amount of correction for each condition (wave spectrum) of the wave, and the strain response spectrum can be estimated with high precision.

[094] (7) An estimating device (monitoring device 10), according to a seventh aspect, is the estimating device of (1) to (6), in which the correction quantity calculation unit calculates the correction quantity for each property of a floating structure member.

[095] The difference between the theoretical value and the measured value of the strain response spectrum differs when the properties of the members (a plate structure, a bone structure, an intersection between bones and the like) differ even in close positions. By calculating the amount of correction according to the member property, the strain response spectrum can be estimated with high precision.

[096] (8) An estimation device (monitoring device 10), according to an eighth aspect, is the estimation device of (4), in which the correction quantity calculation unit uses a Bayesian network to model a relationship between the correction quantity L to correct the difference in peak position, the correction quantity K to correct the difference in peak position, the correction quantity S to correct the difference in distribution width, the strain response function (RAO) (301), the wave spectrum (302), a function (315 to 317) that outputs the theoretical value of the strain response spectrum calculated from the strain response function (RAO) and the wave spectrum, the strain response spectrum (318 to 320) after correction obtained by correcting the theoretical value of the strain response spectrum using the correction quantity L, the correction quantity K and the correction quantity S,and a measured value (324 to 325) of the strain response spectrum calculated based on a value measured by the strain sensor, and calculates the correction amount L, the correction amount K and the correction amount S through back-analysis based on, Petition 870260066319, dated 06 / 07 / 2026, page 45 / 108 38 / 41 measured value of the strain response spectrum.

[097] A large number of measured values ​​of the strain response spectrum are prepared, and the correction quantity L, the correction quantity K, and the correction quantity S are calculated, with the correction quantities being able to be calculated with high precision.

[098] (9) An estimating device (monitoring device 10), according to a ninth aspect, is the estimating device of (1) to (8), additionally including a degradation estimating unit 168 configured to estimate the remaining service life of the structure in the non-installation position of the strain sensor based on the strain response spectrum in the non-installation position estimated by the strain response spectrum estimating unit, and an SN diagram.

[099] By using the strain response spectrum at the non-installation position of the strain sensor estimated by the strain response spectrum estimation unit 165, the remaining service life of the structure at the non-installation position of the strain sensor can be estimated with high accuracy.

[0100] (10) A method of estimation according to a tenth aspect includes: capturing a strain response spectrum at an installation position where a strain sensor is installed on a floating structure, wherein the strain response spectrum is calculated based on a measurement value from the strain sensor; capturing a wave spectrum of a wave at a position where the floating structure is located; capturing a strain response function (response amplitude operator (RAO)) at the installation position of the strain sensor and a strain response function (RAO) at a non-installation position of the strain sensor, calculated based on a structure model of the floating structure; calculating, based on the strain response spectrum and the strain response function (RAO) at the position of Petition 870260066319, dated 06 / 07 / 2026, page 46 / 108 39 / 41 installation of the strain sensor, the wave spectrum and a formula that expresses a relationship between the strain response spectrum and the strain response function (RAO) at the strain sensor installation position and the wave spectrum, a correction amount for a difference between a theoretical value of the strain response spectrum at the strain sensor installation position calculated from the strain response function (RAO) and the wave spectrum and the strain response spectrum at the strain sensor installation position; and calculate, based on a theoretical value of the strain response spectrum at the non-installation position calculated from the strain response function (RAO) and the wave spectrum at the non-installation position of the strain sensor, and the correction amount, the strain response spectrum at the non-installation position.

[0101] (11) A computer-readable means according to an eleventh aspect causes a computer 900 to perform processing including: capturing a strain response spectrum at an installation position where a strain sensor is installed on a floating structure, wherein the strain response spectrum is calculated based on a measurement value from the strain sensor; capturing a wave spectrum of a wave at a position where the floating structure is located; capturing a strain response function (response amplitude operator (RAO)) at the installation position of the strain sensor and a strain response function (RAO) at a non-installation position of the strain sensor, calculated based on a structure model of the floating structure;Based on the strain response spectrum and the strain response function (SRF) at the strain sensor's installation position, calculate the wave spectrum and a formula that expresses a relationship between the strain response spectrum and the strain response function (SRF) at the strain sensor's installation position. Petition 870260066319, dated 06 / 07 / 2026, p. 47 / 108 40 / 41 strain and wave spectrum, a correction amount for a difference between a theoretical value of the strain response spectrum at the strain sensor installation position calculated from the strain response function (RAO) and wave spectrum and the strain response spectrum at the strain sensor installation position; and calculate, based on a theoretical value of the strain response spectrum at the non-installation position calculated from the strain response function (RAO) and wave spectrum at the non-installation position of the strain sensor, and the correction amount, the strain response spectrum at the non-installation position. Industrial Applicability

[0102] The estimation device, the estimation method, and the computer-readable means described above allow for the estimation of a response in a position where a sensor that detects a structural response is not installed. List of Reference Numbers Monitoring device Data capture unit Unit of calculation of the strain response spectrum RAO calculation unit Storage unit Output unit Unit of estimate 161 Strain Response Spectrum Capture Unit 162 Waveform Spectrum Capture Unit 163 RAO Capture Unit 164 Unit of calculation for the amount of correction 165 Unit for estimating the strain response spectrum 166 Plate thickness capture unit Petition 870260066319, dated 06 / 07 / 2026, page 48 / 108 41 / 41 167 Plate thickness estimation unit 168 Degradation estimation unit 900 Computer 901 CPU 902 Main storage device 903 Auxiliary storage device 904 Input / Output Interface 905 Communication Interface Petition 870260066319, dated 06 / 07 / 2026, page 49 / 108

Claims

1 / 8 CLAIMS 1. Estimating device (10) comprising: a strain response spectrum capture unit (161) configured to capture a strain response spectrum at an installation position where a strain sensor (23, 24, 25) is installed on a floating structure, wherein the strain response spectrum is calculated based on a measurement value from the strain sensor (23, 24, 25); a wave spectrum capture unit (162) configured to capture a wave spectrum of a wave at a position where the floating structure is located;an RAO capture unit (163) configured to capture a strain response function (response amplitude operator (RAO)) at the strain sensor installation position (23, 24, 25) and a strain response function (RAO) at a non-installation strain sensor position (23, 24, 25), calculated based on a floating structure model;a correction quantity calculation unit (164) configured to calculate, based on the strain response spectrum and the strain response function (RAO) at the strain sensor installation position (23, 24, 25), the wave spectrum and a formula that expresses a relationship between the strain response spectrum and the strain response function (RAO) at the strain sensor installation position (23, 24, 25) and the wave spectrum, a correction quantity for a difference between a theoretical value of the strain response spectrum at the strain sensor installation position (23, 24, 25) calculated from the strain response function (RAO) and the wave spectrum and the strain response spectrum at the strain sensor installation position (23, 24, 25);and Petition 870260066319, dated 06 / 07 / 2026, page 50 / 108 2 / 8 a strain response spectrum estimation unit (165) configured to estimate, based on a theoretical value of the strain response spectrum at the non-installation position calculated from the strain response function (RAO) and the wave spectrum at the non-installation position of the strain sensor (23, 24, 25), and the amount of correction, the strain response spectrum at the non-installation position, CHARACTERIZED in that it comprises: a plate thickness capture unit (166) configured to capture information about a plate thickness of a structure to which a plate thickness sensor (26) is installed, measured by the plate thickness sensor (26), wherein the plate thickness sensor (26) is installed in a position that is identical to or different from the strain sensor (23, 24, 25);a plate thickness estimation unit (167) configured to estimate a plate thickness at the non-installation position of the strain sensor (23, 24, 25) based on the plate thickness measured by the plate thickness sensor (26) and a plate thickness at any position measured during an inspection; and a RAO calculation unit (13) configured to calculate the strain response function (RAO) at the non-installation position of the strain sensor (23, 24, 25) using the plate thickness estimated by the plate thickness estimation unit (167), wherein the RAO capture unit (163) configured to capture the strain response function (RAO) calculated by the RAO calculation unit (13), and the strain response spectrum estimation unit (165) configured to calculate the strain response spectrum based on the captured strain response function (RAO).

2. Estimating device (10), according to claim 1, Petition 870260066319, dated 06 / 07 / 2026, page 51 / 108 3 / 8 CHARACTERIZED in that: the plate thickness capture unit (166) is configured to capture information about a plate thickness measured by the plate thickness sensor (26) installed on an upper platform of the floating structure, and the plate thickness estimation unit (167) is configured to estimate the plate thickness in the non-installation position, based on a ratio between a plate thickness thinning amount of the upper platform and a structure thinning amount in the non-installation position.

3. Estimating device (10), according to claim 1 or 2, CHARACTERIZED in that the correction quantity calculation unit (164) is configured to calculate a correction quantity to correct a difference in a peak position, a correction quantity to correct a difference in a peak value, and a correction quantity to correct a difference in a distribution width between a waveform expressing a frequency distribution of a strain response spectrum indicated by the theoretical value of the strain response spectrum and a waveform expressing a frequency distribution of a strain response spectrum indicated by the strain response spectrum at the strain sensor installation position (23, 24, 25).

4. Estimating device (10), according to any one of claims 1 to 3, CHARACTERIZED in that the correction quantity calculation unit (164) is configured to calculate the correction quantity corresponding to each of different positions of the floating structure.

5. Estimating device (10), according to any one of claims 1 to 4, CHARACTERIZED in that the correction quantity calculation unit (164) is configured to calculate the correction quantity Petition 870260066319, dated 06 / 07 / 2026, page 52 / 108 4 / 8 for each wave condition indicated by the wave spectrum at a position where the floating structure is situated.

6. Estimating device (10), according to any one of claims 1 to 5, CHARACTERIZED in that the correction quantity calculation unit (164) is configured to calculate the correction quantity for each property of a floating structure member.

7. Estimation device (10), according to claim 3, CHARACTERIZED in that the correction quantity calculation unit (164) is configured to use a Bayesian network to model a relationship between the correction quantity L to correct the difference in peak position, the correction quantity K to correct the difference in peak value, the correction quantity S to correct the difference in distribution width, the strain response function (RAO), the wave spectrum, a function that outputs the theoretical value of the strain response spectrum calculated from the strain response function (RAO) and the wave spectrum, the strain response spectrum after correction obtained by correcting the theoretical value of the strain response spectrum using the correction quantity L, the correction quantity K and the correction quantity S,and a measured value of the response spectrum calculated based on a value measured by the strain sensor (23, 24, 25), and is configured to calculate the correction amount L, the correction amount K and the correction amount S through back-analysis based on the measured value of the strain response spectrum.

8. Estimating device (10), according to any one of claims 1 to 7, CHARACTERIZED in that it further comprises a degradation estimation unit (168) configured to estimate the remaining service life of the structure in the non-installation position of the strain sensor (23, 24, 25) based on the strain response spectrum in the non-installation position Petition 870260066319, dated 06 / 07 / 2026, page 53 / 108 5 / 8 estimated by the strain response spectrum estimation unit (165) and an SN diagram.

9. Estimation method of the estimation device, as defined in claim 1, the method CHARACTERIZED in that it comprises: capturing a strain response spectrum at an installation position where a strain sensor (23, 24, 25) is installed on a floating structure, wherein the strain response spectrum is calculated based on a measurement value from the strain sensor (23, 24, 25); capturing a wave spectrum of a wave at a position where the floating structure is located; capturing a strain response function (response amplitude operator (RAO)) at the installation position of the strain sensor (23, 24, 25) and a strain response function (RAO) at a non-installation position of the strain sensor (23, 24, 25), calculated based on a structure model of the floating structure;Calculate, based on the strain response spectrum and the strain response function (RAO) at the installation position of the strain sensor (23, 24, 25), the wave spectrum and a formula that expresses a relationship between the strain response spectrum and the strain response function (RAO) at the installation position of the strain sensor (23, 24, 25) and the wave spectrum, a correction quantity for a difference between a theoretical value of the strain response spectrum at the installation position of the strain sensor (23, 24, 25) calculated from the strain response function (RAO) and the wave spectrum and the strain response spectrum at the installation position of the strain sensor (23, 24, 25);calculate, based on a theoretical value of the strain response spectrum in the non-installation position calculated from the strain response function Petition 870260066319, dated 06 / 07 / 2026, page 54 / 108 6 / 8 (RAO) and the wave spectrum in the non-installation position of the strain sensor (23, 24, 25), and the amount of correction, the strain response spectrum in the non-installation position; acquire information on the sheet thickness of a structure in which a sheet thickness sensor (26) is installed, measured by the sheet thickness sensor (26), with the sheet thickness sensor (26) installed in an identical or different position from the strain sensor (23, 24, 25); estimate the thickness of a sheet in the non-installation position of the strain gauge (23, 24, 25) based on the sheet thickness measured by the sheet thickness gauge (26) and a sheet thickness at any position measured during an inspection;Calculate the strain response function (RAO) at the non-installation position of the strain sensor (23, 24, 25) using the estimated plate thickness; acquire the calculated strain response function (RAO); and calculate the strain response spectrum based on the acquired strain response function (RAO).

10. Computer-readable means CHARACTERIZED in that it comprises instructions which, when read by a computer, execute the steps of the method as defined in claim 9, comprising: capturing a strain response spectrum at an installation position where a strain sensor (23, 24, 25) is installed on a floating structure, wherein the strain response spectrum is calculated based on a measurement value from the strain sensor (23, 24, 25); capturing a wave spectrum of a wave at a position where the floating structure is located; capture a strain response function (amplitude response operator (RAO)) at the strain sensor installation position (23, 24, 25) and a strain response function (RAO) at a non-installation position of the strain sensor (23, 24, 25), calculated based on a floating structure model;Calculate, based on the strain response spectrum and the strain response function (RAO) at the installation position of the strain sensor (23, 24, 25), the wave spectrum and a formula that expresses a relationship between the strain response spectrum and the strain response function (RAO) at the installation position of the strain sensor (23, 24, 25) and the wave spectrum, a correction quantity for a difference between a theoretical value of the strain response spectrum at the installation position of the strain sensor (23, 24, 25) calculated from the strain response function (RAO) and the wave spectrum and the strain response spectrum at the installation position of the strain sensor (23, 24, 25);Calculate, based on a theoretical value of the strain response spectrum in the non-installation position calculated from the strain response function (RAO) and the wave spectrum in the non-installation position of the strain sensor (23, 24, 25), and the amount of correction, the strain response spectrum in the non-installation position; acquire information on the sheet thickness of a structure in which a sheet thickness sensor (26) is installed, measured by the sheet thickness sensor (26), with the sheet thickness sensor (26) installed in an identical or different position from the strain sensor (23, 24, 25); estimate the thickness of a sheet in the non-installation position of the strain sensor (23, 24, 25) based on the sheet thickness measured by the sheet thickness sensor (26) and a sheet thickness in any position measured during an inspection;Petition 870260066319, dated 06 / 07 / 2026, p. 56 / 108 8 / 8 calculate the strain response function (SRF) at the non-installation position of the strain sensor (23, 24, 25) using the estimated plate thickness; acquire the calculated strain response function (SRF); and calculate the strain response spectrum based on the acquired strain response function (SRF). Petition 870260066319, dated 06 / 07 / 2026, p. 57 / 108;

Images