Floating structure overall structure response prediction method based on few sensors

Abstract

The invention provides a floating structure overall structure response prediction method based on a small number of sensors, and relates to the technical field of ship and ocean engineering, and the method mainly comprises the steps: building a finite element simulation model of a floating structure whole structure field, according to the finite element simulation model, determining a first simulation response data set of the whole structure field of the floating structure under various wave working conditions; determining a monitoring point according to the correlation between the response data set of the candidate monitoring point in the first simulation response data set and the response data set of the prediction point; selecting a plurality of wave working conditions as a characteristic wave load working condition combination through a preset optimization algorithm according to a response data set of a prediction point in the first simulation response data set under a plurality of wave working conditions; the response data sets of the monitoring points under the multiple wave working conditions and the feature wave load weighting coefficients corresponding to the wave working conditions serve as training data, and a proxy model is trained based on a preset algorithm; and finally, outputting the proxy model, and simulating any actual wave working condition into a combination of a plurality of characteristic wave load working conditions and weighting coefficients thereof through the proxy model obtained by training, thereby realizing prediction of the overall structure response of the floating structure. According to the structure response prediction method, real-time prediction of a whole structure field can be achieved, the prediction precision is high, and then fatigue damage calculation of structure fatigue hot spots and prediction of the residual life of the structure can be achieved.

Description

Technical Field

[0001] The invention relates to the field of marine engineering technology, and in particular to a method for predicting the overall structural response of a floating structure based on a small number of sensors. Background Art

[0002] All kinds of offshore transport or engineering ships, floating devices for oil, gas and mineral mining and storage, as well as floating wind power, photovoltaic, hydrogen and ammonia production platforms, can be collectively referred to as floating structures. As floating structures will continue to bear the ever-changing wind, wave and current loads at sea, their structures will suffer yield failure of structural materials or fatigue damage at structural hot spots under the continuous action of alternating loads, thus affecting the safety and functional operation of floating structures. Therefore, full life cycle and real-time structural response monitoring of floating structures is of great significance to the operation and maintenance and asset integrity management of floating structures.

[0003] At present, the traditional structural stress monitoring scheme based on sensor measurement data is limited by the number of sensors and space constraints, and it is difficult to reflect the entire structural field, especially the structural response of key positions; the scheme that uses the measured wave data at sea and the finite element simulation model for simultaneous real-time calculation is affected by the accuracy of the wave data, the simulation calculation error is large, and the calculation amount of the structural response data is large, and real-time monitoring cannot be achieved; there are also some monitoring technologies that collect data through sensors for load inversion and then obtain the structural response data through simulation calculation. The anti-interference ability of the monitoring results is poor, and it is also difficult to achieve real-time monitoring due to the large amount of calculation.

[0004] Therefore, there is a need for a monitoring technology solution that can monitor the structural response data of the structural points of the floating structure in real time with high accuracy by collecting structural response data from a small number of sensors. Summary of the invention

[0005] One purpose of the present application is to provide a method for predicting the overall structural response of a floating structure based on a small number of sensors, so as to solve the problem that it is difficult to perform high-precision real-time monitoring of the entire structural field of a floating structure under the existing technology.

[0006] To achieve the above objectives, some embodiments of the present application provide a method for predicting the overall structural response of a floating structure based on a small number of sensors, the method comprising:

[0007] Establish a finite element simulation model of the entire structural field of the floating structure;

[0008] Determine a first simulation response data set of the entire structural field of the floating structure under various wave conditions according to the finite element simulation model, wherein the parameters of the wave conditions at least include a wave direction angle, a period, and a phase angle;

[0009] Determining the monitoring point according to the correlation between the response data set of the candidate monitoring point and the response data set of the prediction point in the first simulation response data set;

[0010] According to the response data set of the prediction point in the first simulation response data set under multiple wave conditions, a plurality of wave conditions are selected as a characteristic wave load condition combination through a preset optimization algorithm, and the characteristic wave load condition combination includes a plurality of characteristic wave load conditions;

[0011] The response data sets of the monitoring points under various wave conditions and the characteristic wave load weighting coefficients of the corresponding wave conditions are used as training data, and the proxy model is trained based on the preset algorithm. The characteristic wave load weighting coefficient is determined according to the response data sets of the monitoring points under the wave conditions and the response data sets of the monitoring points under the characteristic wave load condition combination;

[0012] The proxy model is output, and any actual wave condition is simulated as a combination of several characteristic wave load conditions and their weighted coefficients through the proxy model to predict the overall structural response of the floating structure.

[0013] Furthermore, according to the finite element simulation model, a first simulation response data set of the entire structural field of the floating structure under various wave conditions is determined, including:

[0014] Generate a first wave condition set for simulation according to the step length and range of each parameter of the initial wave condition;

[0015] According to the first wave operating condition set and the finite element simulation model, a first simulation response data set of the entire structural field of the floating structure under the first wave operating condition set is determined.

[0016] Furthermore, the response data set includes a plurality of response data, and the response data includes structural response data or motion response data.

[0017] Structural response data includes one or more combinations of the following: stress data, strain data, deformation data,

[0018] The motion response data includes one or more combinations of the following: roll data, pitch data, heave data, and acceleration data.

[0019] Further, determining the monitoring point according to the correlation between the response data set of the candidate monitoring point and the response data set of the prediction point in the first simulation response data set includes:

[0020] The candidate monitoring points in the first simulation response data set with high correlation with the response data set of the prediction point and with low correlation between the response data sets are determined as the final monitoring points.

[0021] Furthermore, according to the response data set of the prediction point in the first simulation response data set under multiple wave conditions, a plurality of wave conditions are selected as characteristic wave load condition combinations through a preset optimization algorithm, including:

[0022] According to the response data sets of the prediction point under multiple wave conditions in the first simulation response data set, the multiple wave conditions are determined as the preliminary selected characteristic wave load condition combination through a preset optimization target and a corresponding optimization algorithm, wherein the optimization target includes at least one of the following: the sum of the dot products of the vectors composed of the response data of the prediction point under each characteristic wave load condition is minimized; or the characteristic wave load weighting coefficient is obtained by solving the characteristic wave load condition combination, and then the average error of the prediction point response data set obtained according to the characteristic wave load weighting coefficient in a large number of random verification samples is minimized.

[0023] Furthermore, the response data sets of the monitoring points under various wave conditions and the characteristic wave load weighting coefficients of the corresponding wave conditions are used as training data, and the agent model is trained based on the preset method, including:

[0024] The response data sets of the monitoring points under multiple wave conditions and the characteristic wave load weighting coefficients of the corresponding wave conditions are used as training data, and are trained by multiple proxy model training methods to obtain multiple proxy models. The proxy models are used to output the characteristic wave load weighting coefficients corresponding to the wave conditions according to the response data sets of the monitoring points under the wave conditions;

[0025] According to the response data set of the monitoring point under the verification sample working condition set, the characteristic wave load weighting coefficients of the corresponding verification sample working conditions output by multiple proxy models are obtained, and according to the characteristic wave load weighting coefficients of the corresponding verification sample working conditions and the simulation response data set of the prediction point under the characteristic wave load working condition, the prediction response data set of the prediction point under the verification sample working condition is determined;

[0026] The optimal proxy model is determined based on the average error between the predicted response data set of the prediction point under the verification sample condition and the simulated or measured response data set of the prediction point under the verification sample condition.

[0027] Furthermore, surrogate model training methods include radial basis functions, kriging, and deep neural networks.

[0028] Furthermore, after outputting the proxy model, it also includes:

[0029] Acquire measured response data collected by multiple sensors set at monitoring points under actual wave conditions;

[0030] The measured response data are input into the proxy model to obtain the characteristic wave load weighting coefficient corresponding to the actual wave condition;

[0031] Determine the predicted response data of the whole structure field of the floating structure under the actual wave condition according to the characteristic wave load weighting coefficient corresponding to the actual wave condition and the simulation response data of the whole structure field of the floating structure under multiple characteristic wave load conditions;

[0032] Output the predicted response data of the entire structural field of the floating structure under actual wave conditions.

[0033] Some embodiments of the present application further provide a computer-readable medium having computer-readable instructions stored thereon, wherein the computer-readable instructions can be executed by a processor to implement the aforementioned method for predicting the overall structural response of a floating structure based on a small number of sensors.

[0034] Some embodiments of the present application also provide an electronic device, which includes a memory for storing computer program instructions and a processor for executing the computer program instructions, wherein when the computer program instructions are executed by the processor, the electronic device executes the aforementioned method for predicting the overall structural response of a floating structure based on a small number of sensors.

[0035] Compared with the prior art, the solution provided by the present application establishes a finite element simulation model of the entire structural field of the floating structure, and determines the first simulation response data set of the entire structural field of the floating structure under various wave conditions according to the finite element simulation model, determines the monitoring point according to the correlation between the response data set of the candidate monitoring point and the response data set of the prediction point in the first simulation response data set, and selects several wave conditions as characteristic wave load condition combinations according to the response data set of the prediction point under various wave conditions in the first simulation response data set through a preset optimization algorithm, and then uses the response data set of the monitoring point under various wave conditions and the characteristic wave load weighting coefficients of the corresponding wave conditions as training data, trains the proxy model based on the preset algorithm, and finally outputs the proxy model, and the proxy model obtained by training can simulate any actual wave condition as a combination of several characteristic wave load conditions and their weighting coefficients, thereby realizing the prediction of the overall structural response of the floating structure. The structural response prediction method of the present application can realize real-time prediction of the entire structural field with high prediction accuracy, and can also realize the calculation of fatigue damage of structural fatigue hotspots and the prediction of the remaining life of the structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0037] Figure 1 A flowchart of a method for predicting the overall structural response of a floating structure based on a small number of sensors is provided for some embodiments of the present application.

[0038] Figure 2A flowchart of obtaining a proxy model by training according to a response data set of a monitoring point and a characteristic wave load weighting coefficient is provided for some embodiments of the present application.

[0039] Figure 3 A flowchart for using a proxy model to predict response data of a prediction point under actual wave conditions is provided for some embodiments of the present application. DETAILED DESCRIPTION

[0040] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0041] Here, the method for predicting the overall structural response of a floating structure based on a small number of sensors in the embodiment of the present application is suitable for the scenario of real-time prediction of the structural response of the entire structural field of a floating structure.

[0042] In this scenario, floating structures are continuously subjected to changing wind, wave and current loads at sea. Under the continuous action of alternating loads, the structures themselves may suffer yield failure of structural materials or fatigue damage at structural hot spots, thus affecting the functional realization and safe production of the floating structures. Therefore, it is necessary to conduct full life cycle and real-time structural response monitoring of floating structures to improve the safety of floating structures and ensure the realization of functionality. However, many key structural points in floating structures, especially fatigue hot spots, cannot be monitored in real time by setting sensors. In addition, the existing monitoring schemes for obtaining structural stress through simulation technology are greatly affected by the accuracy of wave data, resulting in large errors. At the same time, the acquisition of structural responses is affected by computing power and it is difficult to achieve real-time monitoring.

[0043] In actual sea conditions, all sea waves are irregular waves. In engineering applications, in order to simplify the problem, the irregular wave conditions can be considered to be the superposition of many regular wave conditions with different wave directions, different periods, different amplitudes, and random phases. The formula for the rise of the irregular wave surface can be expressed as follows:

[0044]

[0045] Among them, ξ Ai represents the amplitude of the ith component wave, x, t represent the position and time of the recording point respectively, ε i represents the initial phase of the i-th long-peaked irregular wavelet, which can take any value between 0 and 2π. i The wave number of the i-th component wave is equal to 2π times the number of waves appearing per unit distance. i Indicates the wave circle frequency.

[0046] Therefore, under certain assumptions, the structural response or motion response of floating structures caused by actual sea conditions can also be considered to be the superposition of structural responses or motion responses caused by many regular wave conditions. At any time point in actual sea conditions, the structural response or motion response formula of floating structures can be expressed as follows:

[0047] R=R 1 +R 2 +…+R i

[0048] Where R is the structural response or motion response of the floating structure caused by the actual sea conditions at that time point, and R 1 To R i is the structural response or motion response caused by i regular wave conditions decomposed from the irregular wave condition at that time point.

[0049] Furthermore, it is assumed that the structural response or motion response of a floating structure caused by a regular wave condition with arbitrary wave direction, period, phase and amplitude can also be characterized by the linear weighting of the structural response or motion response caused by several characteristic wave load conditions of regular wave conditions, where the characteristic wave load condition refers to the wave condition of unit amplitude under several specific wave directions, periods and phases. The linear weighting formula of the structural response or motion response can be expressed as follows:

[0050]

[0051] Among them, R1 to R n are the structural responses or motion responses caused by n characteristic wave load conditions, to is the characteristic wave load weighting coefficient of n characteristic wave load conditions for the regular wave condition.

[0052] Therefore, at any time point in actual sea conditions, the structural response or motion response formula of the floating structure caused by irregular wave conditions can be expressed as follows:

[0053]

[0054] That is, R = a1R1 + a2R2 + ... + a n R n

[0055] Among them, R1 to R n are the structural responses or motion responses caused by n characteristic wave load conditions, a1 to a n is the weighting coefficient of the characteristic wave load at that time point in the actual sea conditions for n characteristic wave load conditions.

[0056] By installing sensors at multiple designated locations on the floating structure, the sensors may be, for example, strain sensors, attitude sensors, acceleration sensors, etc., so that m response data can be obtained at any time point, the response data including structural response data and motion response data, m is the number of sensors installed, m is greater than or equal to n, that is, the number of sensors installed must be greater than or equal to the number of characteristic wave load conditions.

[0057] Under n characteristic wave load conditions, the response data collected by these m sensors can be expressed by the following matrix M:

[0058]

[0059] Among them, r 11 To r mn They are the response data collected by m sensors under n characteristic wave load conditions.

[0060] Therefore, for any time point in the actual sea conditions, the structural response or motion response of the floating structure caused by irregular wave conditions can be expressed in the following matrix form:

[0061]

[0062] The above matrix form can be simplified as follows:

[0063] M2A=R ′

[0064] in, is the characteristic wave load weighting coefficient vector of n characteristic wave loads, is the response data vector collected by m sensors.

[0065] Therefore, the characteristic wave load weighting coefficient vector A can be expressed as follows:

[0066] A=M -1 ·R ′

[0067] For the pseudo-inverse matrix M -1 Solve, R ′ The characteristic wave load weighting coefficient vector A can be calculated by using the response data collected by the sensor, thereby obtaining the characteristic wave load weighting coefficients corresponding to the n characteristic wave loads at any time point.

[0068] After obtaining the characteristic wave load weighting coefficients corresponding to n characteristic wave loads at any time point, we can use the formula R = a1R1+a2R2+…+a n R n The structural response and motion response of the floating structure under actual sea conditions at any time point are calculated.

[0069] The present application provides a method for predicting the overall structural response of a floating structure based on a small number of sensors, by establishing a finite element simulation model of the overall structural field of the floating structure, and determining the first simulation response data set of the overall structural field of the floating structure under various wave conditions according to the finite element simulation model; determining the monitoring point according to the correlation between the response data set of the candidate monitoring point and the response data set of the prediction point in the first simulation response data set; and selecting several wave conditions as a characteristic wave load condition combination according to the response data set of the prediction point under various wave conditions in the first simulation response data set through a preset optimization algorithm; then using the response data set of the monitoring point under various wave conditions and the characteristic wave load weighting coefficient of the corresponding wave condition as training data, training the proxy model based on the preset algorithm; finally outputting the proxy model, and being able to simulate any actual wave condition as a combination of several characteristic wave load conditions and their weighting coefficients through the proxy model obtained by training, thereby realizing the prediction of the overall structural response of the floating structure. The structural response prediction method of the present invention can realize real-time prediction of the overall structural field, and has high prediction accuracy, and can also realize the calculation of fatigue damage of structural fatigue hot spots and the prediction of the remaining life of the structure.

[0070] In some embodiments of the present application, the electronic device executes the process of the method for predicting the overall structural response of a floating structure based on a small number of sensors, such as Figure 1 As shown, the method may include the following steps:

[0071] Step S101, establishing a finite element simulation model of the entire structural field of the floating structure.

[0072] It will be appreciated that electronic devices may include, but are not limited to, laptop computers, desktop computers, tablet computers, mobile phones, wearable devices, head-mounted displays, servers, mobile email devices, portable game consoles, portable music players, reader devices, televisions having one or more processors embedded or coupled thereto, or other electronic devices capable of accessing a network, etc.

[0073] Here, floating structures may include but are not limited to: various types of offshore transport ships, engineering ships, oil, gas, and mineral mining and storage floating devices, floating wind power, photovoltaic, hydrogen and ammonia production platforms, etc.

[0074] The full structural field refers to the overall structure of a floating structure. For example, a ship is a full ship structure, and an offshore engineering platform is a full platform structure. Specifically, it may include the materials of the outer plate, bulkhead, frame, steel grade, plate thickness specifications, etc., as well as the position of each structural component in space.

[0075] Finite Element Analysis (FEA) is a method that uses mathematical approximation to simulate real physical systems (geometry and load conditions). Finite element analysis uses simple and interacting elements (i.e. units) to approximate real systems with infinite unknowns with a finite number of unknowns. Finite element analysis replaces complex problems with simpler ones and then solves them. It regards the solution domain as consisting of many small interconnected subdomains called finite elements. It assumes a simpler approximate solution for each unit, and then derives the conditions that satisfy the domain, such as the equilibrium conditions of the structure, to obtain the solution to the problem. Because the actual problem is replaced by a simpler problem, this solution is not an exact solution, but an approximate solution. Since most actual problems are difficult to obtain exact solutions, finite elements not only have high calculation accuracy, but can also adapt to various complex shapes, making them an effective engineering analysis tool.

[0076] The finite element simulation model is a model established when using the finite element analysis method. It is a combination of units that are connected only at nodes, transmit forces only at nodes, and are constrained only at nodes. It is the result of the discretization of the mechanical model and is a digital model for numerical calculations.

[0077] Here, the finite element simulation model of the entire structural field for the floating structure can be established by using third-party finite element simulation software, such as the SESAM software package of Det Norske Veritas (DNV), FEmap, patran / nastran, Ansys, etc.

[0078] Step S102: determining a first simulation response data set of the entire structural field of the floating structure under various wave conditions according to the finite element simulation model.

[0079] After obtaining the finite element simulation model, the finite element simulation model is used to obtain a structural response data set of the entire structural field of the floating structure under a typical loading and different wave conditions through simulation. This data set is the first simulation response data set, which contains the simulation response data of all structural points in the floating structure.

[0080] Here, the response data set includes multiple response data. In some embodiments of the present application, the response data may include structural response data and motion response data. The structural response data includes, but is not limited to, the following data: stress data, strain data, deformation data, etc. For example, the yield stress data of a certain structural point is 150MPa, and the deformation data is a vertical deformation of 5mm. The motion response data may include, but is not limited to, the following data: roll data, pitch data, heave data, acceleration data, etc. For example, the roll data of the floating structure is 15°, the pitch data is 5°, the heave data is -2m, and the acceleration of a certain structural point is 1m / s in the x-direction. 2 , y direction 2m / s 2 , z direction 0.5m / s 2 .

[0081] It can be understood that the parameters used to describe the wave working conditions include at least the wave direction angle, period and phase angle, and may also include other more parameters, and the parameters used to describe the wave working conditions are not limited here.

[0082] Here, multiple wave conditions can be generated by setting a certain step size of wave parameter changes within the range of sea conditions that the floating structure may actually encounter. In some embodiments of the present application, a first wave condition set for simulation can be generated based on the step size and range of each parameter of the initial wave condition, and the first wave condition set includes multiple wave conditions. Then, based on the first wave condition set and the finite element simulation model, a first simulation response data set of the entire structural field of the floating structure under the first wave condition set is determined. For example, the wave parameters of the initial wave condition are, for example, a period of 3s, a wave direction angle of 0°, and a phase angle of 0°. The step sizes of the parameters, for example, set the period step size to 1s, the wave direction angle step size to 15°, and the phase angle step size to 30°. Wave condition A is obtained according to the initial wave condition and the step sizes of the parameters. A simulation response data set of the full-structure field under wave condition A is obtained according to wave condition A and the finite element simulation model. The simulation response data set of the full-structure field under wave condition A is added to the first simulation response data set to obtain a new first simulation response data set. Wave condition B is then obtained according to wave condition A and the step sizes of the parameters. A simulation response data set of the full-structure field under wave condition B is obtained and added to the first simulation response data set. This is repeated until the parameter range of the wave condition is exceeded.

[0083] In some embodiments of the present application, a full-ship finite element simulation model of a floating production storage and offloading system (FPSO) is established by using the SESAM software package of Det Norske Veritas (DNV), and the number of units of the simulation model, i.e., the number of structural points, is about 700,000. Then, a first simulation response data set is established by the following process: defining wave parameter variables, performing hydrodynamic analysis by SESAM-HydroD, performing structural response analysis by SESAM-Sestra, and extracting simulation data by SESAM-Xtract. The range of the wave parameter variables includes: period 3s to 30s, wave angle 0° to 360°, and phase angle 0° to 360°.

[0084] In some embodiments of the present application, sensors can be set at multiple key structural positions of a floating structure that has been built and operated to collect a measured response data set of the floating structure under actual sea conditions, and the union of the measured response data set and the response data set obtained by simulation is used as the first simulation response data set.

[0085] Step S103, determining the monitoring point according to the correlation between the response data set of the candidate monitoring point and the response data set of the prediction point in the first simulation response data set.

[0086] Here, the multiple candidate monitoring points and multiple prediction points of the floating structure are predetermined key structural points. Specifically, according to the monitoring requirements of the floating structure and actual engineering experience, some key structural points can be selected in advance in the whole structural field of the floating structure as candidate monitoring points and prediction points. The candidate monitoring points can be structural points that are easy to characterize the overall stress state of the floating structure. These structural points can collect corresponding response data by setting sensors at these points. The prediction points can be key structural strength points or fatigue hot spots of the structure that are prone to fatigue in the floating structure. Response data cannot often be collected by setting sensors. For example, 81 points can be selected as candidate monitoring points from about 700,000 structural points of the floating structure, and 136 key strength points and fatigue hot spots can be selected as prediction points.

[0087] It can be understood that the number of candidate monitoring points is greater than the number of monitoring points, and the monitoring points finally obtained are only a small part of the candidate monitoring points. The finally obtained monitoring points are used to obtain the corresponding measured response data set by setting sensors. For example, 19 candidate monitoring points can be selected from 81 candidate monitoring points as the final monitoring points, and strain sensors can be set at the 19 monitoring points to obtain measured strain response data.

[0088] Here, the final monitoring point is selected from the multiple candidate monitoring points based on the correlation between the response data sets of the candidate monitoring points and the prediction points, and the correlation between the response data sets of the candidate monitoring points.

[0089] In some embodiments of the present application, candidate monitoring points having high correlation between the first simulation response data set and the response data set of the prediction point and low correlation between the response data sets are determined as final monitoring points.

[0090] The correlation analysis between response data sets can adopt a variety of analysis methods, such as Pearson, Spearman and Kendall, etc. Specifically, the correlation analysis can be completed by third-party software that implements the correlation analysis algorithm or self-developed software. For example, the correlation analysis function built into the third-party software Optimus can be used to analyze the correlation between the response data sets of candidate monitoring points and prediction points and the correlation between the response data sets of candidate monitoring points.

[0091] After the correlation analysis results are obtained, the candidate monitoring points that simultaneously satisfy the high correlation with the response data set of the prediction point and the low correlation with the response data sets of other candidate monitoring points are determined as the final monitoring points.

[0092] Step S104, according to the response data set of the prediction point in the first simulation response data set under multiple wave conditions, a plurality of wave conditions are selected as a characteristic wave load condition combination through a preset optimization algorithm, wherein the characteristic wave load condition combination includes a plurality of characteristic wave load conditions.

[0093] After all prediction points are determined, corresponding response data sets of all prediction points under various wave conditions are obtained from the first simulation response data set, and characteristic wave load condition combinations are determined through the response data sets.

[0094] It can be understood that the pre-set optimization algorithm can be a different optimization analysis algorithm, such as a differential evolution algorithm, a sequential quadratic programming algorithm, a generalized parsimonious gradient algorithm, an adaptive evolutionary algorithm, and a single-objective particle swarm algorithm, etc. The embodiment of the present application does not impose specific restrictions on the optimization algorithm used.

[0095] In some embodiments of the present application, the pre-set optimization algorithm is a differential evolution algorithm (DifferentialEvolution, DE). Differential evolution algorithm is a kind of population-based adaptive global optimization algorithm, which belongs to a kind of evolutionary algorithm. It has the characteristics of simple structure, easy implementation, fast convergence and strong robustness. It is widely used in data mining, pattern recognition, digital filter design, artificial neural network, electromagnetism and other fields. Similar to genetic algorithm, differential evolution algorithm is also an optimization algorithm based on modern intelligence theory, which guides the direction of optimization search through the group intelligence generated by mutual cooperation and competition between individuals in the group. The basic idea of ​​the algorithm is to start from a randomly generated initial population, generate new individuals by summing the vector difference of any two individuals in the population with the third individual, and then compare the new individuals with the corresponding individuals in the contemporary population. If the fitness of the new individual is better than the fitness of the current individual, the new individual is used to replace the old individual in the next generation, otherwise the old individual is still preserved. Through continuous evolution, excellent individuals are retained, inferior individuals are eliminated, and the search is guided to approach the optimal solution.

[0096] In some embodiments of the present application, based on the response data set of the prediction point in the first simulation response data set under multiple wave conditions, the multiple wave conditions can be determined as a characteristic wave load condition combination through a preset optimization target and a corresponding optimization algorithm, and the characteristic wave load condition combination includes multiple characteristic wave load conditions.

[0097] Here, the optimization objectives include at least one of the following: the sum of the dot products of the response data sets of the prediction points under the characteristic wave load condition combination is minimized, or the characteristic wave load weighting coefficient is obtained by solving the characteristic wave load condition combination, and then the average error of the prediction point response data set obtained according to the characteristic wave load weighting coefficient in a large number of random verification samples is minimized.

[0098] Under each wave condition, the response data of all prediction points are taken as a set of vectors. The smaller the dot product of the response data vectors of all prediction points under two wave conditions, the better the orthogonality of the response data sets under the two wave conditions, that is, the smaller the correlation. Therefore, by selecting multiple wave conditions with small correlations from multiple wave conditions as characteristic wave load condition combinations, irregular wave conditions can be described with a small number of characteristic wave load conditions.

[0099] In some embodiments of the present application, the sum of the dot products of the response data sets of the prediction points under the characteristic wave load conditions is less than the sum of the dot products of the response data sets of the prediction points under the non-characteristic wave load conditions. Here, the sum of the dot products of the response data of all the prediction points under different wave condition combinations is calculated, and the corresponding wave characteristic load combination is determined according to the sum of the dot products. For example, the sum of the dot products is sorted from small to large, and the wave condition combinations in the front row are determined as the characteristic wave load condition combinations.

[0100] Here, a variety of possible wave condition combinations can be calculated using a preset optimization algorithm. For example, the optimization algorithm can be set to calculate 20,000 times, and the best wave condition combinations can be found from the results. One of the best wave condition combinations is selected as the characteristic wave load condition combination.

[0101] For example, the differential evolution algorithm can be used to find k wave conditions within the parameter range of wave conditions that the floating structure may encounter in practice, with a relatively small sum of dot products, and k may be 5-20. These k wave conditions are used as the characteristic load condition combination of the floating structure. In addition, the number of characteristic load conditions in the characteristic load condition combination can be determined by comprehensively considering the number of input signals of the monitoring point sensors.

[0102] In some embodiments of the present application, the average error of the response data set of all prediction points obtained by solving the characteristic wave load weighting coefficient through the characteristic wave load condition combination in a large number of random verification samples is smaller than the average error of the response data set of all prediction points obtained by solving the characteristic wave load weighting coefficient through the non-characteristic wave load condition combination in a large number of random verification samples.

[0103] Specifically, multiple selections can be made from a variety of wave conditions, and several wave conditions are selected each time to obtain multiple wave condition combinations. Then, the corresponding characteristic wave load weighting coefficient is determined for each wave condition combination through the proxy model, and the response data sets of all prediction points are calculated according to the obtained characteristic wave load weighting coefficient. Then, the average error of the response data sets of all prediction points in a large number of random verification samples is calculated, and the wave condition combination with the smallest average error is taken as the characteristic wave load condition combination, so that the average error of the response data sets of all prediction points obtained by solving the characteristic wave load weighting coefficient through the characteristic wave load condition combination in a large number of random verification samples is smaller than the average error of the response data sets of all prediction points obtained by solving the characteristic wave load weighting coefficient through the non-characteristic wave load condition combination in a large number of random verification samples.

[0104] Step S105, using the response data sets of the monitoring points under various wave conditions and the characteristic wave load weighting coefficients of the corresponding wave conditions as training data, and training the proxy model based on a preset algorithm.

[0105] It can be understood that the response data set of all monitoring points under various wave conditions can be a simulated response data set obtained by simulating or artificially generating various wave conditions through a finite element simulation model, causing corresponding responses of the floating structure at all monitoring points; it can also be a measured response data set collected by setting sensors at monitoring points in a built and operated floating structure under various actual wave conditions; it can also be a combined response data set of a simulated response data set and a measured response data set. The embodiment of the present application does not specifically limit the source of the response data set of the monitoring points under wave conditions.

[0106] Here, the characteristic wave load weighting coefficient corresponding to a wave condition is determined based on the response data vector of all monitoring points under the wave condition and the response data set of all monitoring points under the characteristic wave load condition combination. Specifically, it can be determined according to the above formula A=M -1 ·R ′ Calculate and get, here M -1 is the pseudo-inverse matrix of the response data matrix of all monitoring points under multiple characteristic wave load conditions, R ′ is the response data vector of all monitoring points under this wave condition.

[0107] The proxy model training method is a pre-selected method, which may include but is not limited to: Radial Basis Function (RBF), Kriging and Deep Neural Networks (DNN), etc. It can be understood that the proxy model training method can output the corresponding characteristic wave load weighting coefficient vector according to the input response data set, and the embodiment of the present application does not specifically limit the proxy model training method used.

[0108] Figure 2 The process of training the proxy model according to the response data set of the monitoring point and the characteristic wave load weighting coefficient in some embodiments of the present application is shown, such as Figure 2 As shown, the response data sets of all monitoring points under various wave conditions and the characteristic wave load weighting coefficients of the corresponding wave conditions are used as training data, and training is performed according to a preset proxy model training method to obtain a proxy model, which may include the following steps:

[0109] Step S201, using the response data of the monitoring point under multiple wave conditions and the characteristic wave load weighting coefficients of the corresponding wave conditions as training data, respectively training by multiple proxy model training methods to obtain multiple proxy models.

[0110] The response data under multiple wave conditions and the corresponding characteristic wave load weighting coefficients are used as training samples, and an appropriate proportion of sample points are reserved as the validation set and do not participate in the proxy model training. The remaining sample points are used as the training set for proxy model training. Among them, the response data of the monitoring point under each sample condition in the training set is used as the input data of the proxy model, and the characteristic wave load weighting coefficient corresponding to the sample condition is used as the output data of the proxy model. The proxy model training is carried out to finally obtain the proxy model.

[0111] Similarly, the response data of all monitoring points under each wave condition in the training sample and the corresponding characteristic wave load weighting coefficient are used as training data for training through other proxy model training methods, and finally a plurality of trained proxy models are obtained.

[0112] Each trained proxy model can be used to output the characteristic wave load weighting coefficient corresponding to a certain wave condition based on the response data of all monitoring points under the wave condition.

[0113] Step S202, based on the response data of the monitoring point under the verification sample condition set, obtain the characteristic wave load weighting coefficients of the corresponding verification sample conditions output by multiple proxy models, and determine the predicted response data of the prediction point under the verification sample conditions based on the characteristic wave load weighting coefficients of the corresponding verification sample conditions and the simulation response data set of the prediction point under the characteristic wave load conditions.

[0114] It can be understood that the response data of all monitoring points under the verification sample conditions can be the simulation response data obtained according to the finite element simulation model, or the response data under the measured wave conditions, or a combination of the simulation response data and the measured response data. The embodiments of the present application do not impose specific restrictions on this.

[0115] In addition, the prediction point simulation response data under the final characteristic wave load condition combination is generated according to the finite element simulation model of the floating structure. The specific method can refer to the generation process of the first simulation response data set mentioned above, which will not be repeated here.

[0116] Here, the proxy model outputs the characteristic wave load weighting coefficient corresponding to the verification sample condition based on the input response data of all monitoring points under the verification sample condition, and then matrix-multiplies the obtained characteristic wave load weighting coefficient with the simulation response data of all prediction points under the characteristic wave load condition, so as to obtain the response data of all prediction points under the verification sample condition. The response data is the predicted response data calculated based on the characteristic wave load weighting coefficient.

[0117] Multiple proxy models output their own corresponding characteristic wave load weighting coefficients according to the input response data of all monitoring points under the verification sample conditions. The obtained characteristic wave load weighting coefficients are then matrix-multiplied with the simulation response data of all prediction points under the characteristic wave load conditions to obtain the prediction point response data corresponding to different proxy models.

[0118] Step S203, determining the optimal proxy model according to the average error between the predicted response data set of the predicted point under the validation sample working condition and the response data set of the predicted point under the validation sample working condition.

[0119] Here, the simulated or measured response data sets of all prediction points under the verification sample conditions are response data obtained from the finite element simulation model or response data obtained from actual ship sensors, which are used as evaluation criteria for judging the performance of the proxy model.

[0120] All predicted response data obtained according to different proxy models are compared with the simulated or measured response data of all prediction points under the verification sample conditions, and the corresponding average errors are calculated. The proxy model corresponding to the predicted response data with the smallest average error is determined as the optimal proxy model.

[0121] In some embodiments of the present application, the optimal proxy model may be further optimized by adding additional wave conditions and corresponding response data sets of monitoring points under the additional wave conditions.

[0122] Step S106, outputting the proxy model, simulating any actual wave condition as a combination of several characteristic wave load conditions and their weighted coefficients through the proxy model, so as to predict the overall structural response of the floating structure.

[0123] After selecting the proxy model with the best performance from multiple proxy models, the optimal proxy model is output for users to use to predict the response data of all prediction points under actual wave conditions.

[0124] Figure 3 The process of using the proxy model to predict the response data of the prediction point under the actual wave condition in some embodiments of the present application is shown, such as Figure 3 As shown, the following steps may be specifically included:

[0125] Step S301, obtaining a measured response data set collected by multiple sensors set at monitoring points under actual wave conditions.

[0126] Here, the actual wave condition is the irregular wave condition experienced by the floating structure in the actual offshore environment at a certain point in time, and multiple monitoring points in the floating structure are respectively provided with sensors for collecting response data, and the multiple sensors collect measured response data sets at this time point.

[0127] Step S302: input the measured response data into the proxy model to obtain the characteristic wave load weighting coefficient corresponding to the actual wave condition.

[0128] The measured response data under actual wave conditions collected at all monitoring points are input into the optimal proxy model as input data, and the optimal proxy model outputs the characteristic wave load weighting coefficient corresponding to the actual wave conditions.

[0129] Step S303, determining predicted response data of the predicted point under the actual wave condition according to the characteristic wave load weighting coefficient corresponding to the actual wave condition and the simulation response data of the predicted point under multiple characteristic wave load conditions.

[0130] After obtaining the characteristic wave load weighting coefficient corresponding to the actual wave condition, it is matrix-multiplied with the simulation response data set of all prediction points under multiple characteristic wave load conditions to obtain the predicted response data of all prediction points under the actual wave condition.

[0131] Step S304, outputting the predicted response data of the predicted point under the actual wave condition.

[0132] Similarly, the measured response data of the monitoring point's time history is input into the optimal proxy model, and the optimal proxy model outputs the corresponding time history characteristic wave load weighting coefficient, thereby outputting the time history predicted response data corresponding to all prediction points, thereby calculating the fatigue damage of the structural hotspots and predicting the remaining life of the structure. The output time history predicted response data, fatigue damage calculation results and structural remaining life prediction results can also be displayed in a graphical and visual way, which is convenient for users to have an intuitive understanding of the structural fatigue conditions of all prediction points in the floating structure.

[0133] In summary, the solution provided by the present application establishes a finite element simulation model of the entire structural field of the floating structure, and determines the first simulation response data set of the entire structural field of the floating structure under various wave conditions according to the finite element simulation model, determines the monitoring point according to the correlation between the response data set of the candidate monitoring point and the response data set of the prediction point in the first simulation response data set, and selects several wave conditions as a characteristic wave load condition combination according to the response data set of the prediction point under various wave conditions in the first simulation response data set through a preset optimization algorithm, and then uses the response data set of the monitoring point under various wave conditions and the characteristic wave load weighting coefficient of the corresponding wave condition as training data, trains the proxy model based on the preset algorithm, and finally outputs the proxy model, and the proxy model obtained by training can simulate any actual wave condition as a combination of several characteristic wave load conditions and their weighting coefficients, thereby realizing real-time prediction of the overall structural response of the floating structure. The structural response prediction method of the present application can realize real-time prediction of the entire structural field with high prediction accuracy, and can also realize calculation of fatigue damage of structural fatigue hotspots and prediction of the remaining life of the structure.

[0134] It is obvious to those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and that the present application can be implemented in other specific forms without departing from the spirit or basic features of the present application. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-restrictive, and the scope of the present application is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present application. Any figure mark in the claims should not be regarded as limiting the claims involved. In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The words first, second, etc. are used to indicate names, and do not indicate any particular order.

Claims

1. A method for predicting the overall structural response of a floating structure based on a small number of sensors, characterized in that: The method includes: Establish a finite element simulation model of the entire structural field of the floating structure; Determining a first simulation response data set of a full structural field of the floating structure under a plurality of wave conditions according to the finite element simulation model, wherein the parameters of the wave conditions at least include a wave direction angle, a period, and a phase angle; Determining a monitoring point according to a correlation between a response data set of a candidate monitoring point and a response data set of a prediction point in the first simulation response data set; According to the response data sets of the prediction points in the first simulation response data set under multiple wave conditions, a plurality of wave conditions are selected as characteristic wave load condition combinations through a preset optimization algorithm, wherein the characteristic wave load condition combination includes a plurality of characteristic wave load conditions; The response data set of the monitoring point under the plurality of wave conditions and the characteristic wave load weighting coefficient corresponding to the wave conditions are used as training data, and the proxy model is trained based on a preset algorithm, wherein the characteristic wave load weighting coefficient is determined according to the response data set of the monitoring point under the wave condition and the response data set of the monitoring point under the combination of the characteristic wave load conditions; The proxy model is output, and any actual wave condition is simulated as a combination of several characteristic wave load conditions and their weighted coefficients through the proxy model, so as to realize the prediction of the overall structural response of the floating structure.

2. The prediction method according to claim 1, characterized in that: Determining a first simulation response data set of the entire structural field of the floating structure under various wave conditions according to the finite element simulation model includes: Generate a first wave condition set for simulation according to the step length and range of each parameter of the initial wave condition; A first simulation response data set of the entire structural field of the floating structure under the first wave operating condition set is determined according to the first wave operating condition set and the finite element simulation model.

3. The prediction method according to claim 1, characterized in that: The response data set includes a plurality of response data, wherein the response data includes structural response data or motion response data. The structural response data includes one or more combinations of the following: stress data, strain data, deformation data, The motion response data includes one or more combinations of the following: roll data, pitch data, heave data, and acceleration data.

4. The prediction method according to claim 1, characterized in that: Determining the monitoring point according to the correlation between the response data set of the candidate monitoring point and the response data set of the prediction point in the first simulation response data set includes: The candidate monitoring points in the first simulation response data set, which have a high correlation with the response data set of the prediction point and a low correlation between the response data sets, are determined as the final monitoring points.

5. The prediction method according to claim 1, characterized in that: According to the response data set of the prediction point in the first simulation response data set under multiple wave conditions, a plurality of wave conditions are selected as characteristic wave load condition combinations through a preset optimization algorithm, including: According to the response data sets of the prediction point under multiple wave conditions in the first simulation response data set, the multiple wave conditions are determined as a characteristic wave load condition combination through a preset optimization target and a corresponding optimization algorithm, wherein the optimization target includes at least one of the following: the sum of the dot products of the vectors composed of the response data of the prediction point under the characteristic wave load conditions is minimized; or the characteristic wave load weighting coefficient is obtained by solving the characteristic wave load condition combination, and then the average error of the prediction point response data set obtained according to the characteristic wave load weighting coefficient in a large number of random verification samples is minimized.

6. The prediction method according to claim 1, characterized in that: The response data sets of the monitoring points under the various wave conditions and the characteristic wave load weighting coefficients corresponding to the wave conditions are used as training data, and the agent model is trained based on a preset method, including: The response data sets of the monitoring point under multiple wave conditions and the characteristic wave load weighting coefficients of the corresponding wave conditions are used as training data, and are trained by multiple proxy model training methods to obtain multiple proxy models, wherein the proxy models are used to output the characteristic wave load weighting coefficients corresponding to the wave conditions according to the response data sets of the monitoring point under the wave conditions; According to the response data set of the monitoring point under the verification sample working condition set, characteristic wave load weighting coefficients corresponding to the verification sample working condition output by the plurality of the proxy models are obtained, and according to the characteristic wave load weighting coefficients corresponding to the verification sample working condition and the simulation response data set of the prediction point under the characteristic wave load working condition, a prediction response data set of the prediction point under the verification sample working condition is determined; The optimal proxy model is determined according to an average error between a predicted response data set of the predicted point under the validation sample condition and a simulated or measured response data set of the predicted point under the validation sample condition.

7. The prediction method according to claim 1, characterized in that: The agent model training method includes radial basis function, Kriging and deep neural network.

8. The prediction method according to claim 1, characterized in that: After outputting the proxy model, the method further includes: Acquiring measured response data collected by a plurality of sensors disposed at the monitoring point under actual wave conditions; Inputting the measured response data into the proxy model to obtain a characteristic wave load weighting coefficient corresponding to the actual wave condition; Determining predicted response data of the whole structure field of the floating structure under the actual wave condition according to the characteristic wave load weighting coefficient corresponding to the actual wave condition and the simulation response data of the whole structure field of the floating structure under a plurality of the characteristic wave load conditions; Output the predicted response data of the entire structural field of the floating structure under the actual wave condition.

Images