A soft steel arm single-point mooring system structure safety prediction and early warning method

Abstract

The application discloses a kind of soft steel arm single-point mooring system structure safety prediction early warning method, comprising: S1, establishes the coupling model of FPSO and soft steel arm single-point mooring system;S2, the single-point structure finite element model of single-point structure in ANSYS is established by soft steel arm single-point mooring system;S3, the time domain simulation of the coupling model of FPSO and soft steel arm single-point mooring system is carried out and the finite element calculation of the single-point structure system finite element model is carried out;S4, based on the result of time domain simulation and finite element calculation, establish the sample set of ocean environment and ship body feature and single-point structure stress;S5, obtain single-point mooring system structure safety prediction proxy model;S6, obtain single-point structure stress prediction value;S7, compare single-point structure stress prediction value and threshold value to realize prediction early warning;If prediction value reaches or exceeds threshold value;The single-point mooring system structure safety prediction proxy model carries out early warning;The application predicts fast, stability is strong, by prediction early warning, to provide reference for field decision.

Description

Technical Field

[0001] This invention belongs to the field of marine engineering, specifically relating to a method for predicting and warning the structural safety of a soft steel arm single-point mooring system. Background Technology

[0002] FPSOs are widely used in oil and gas development and have become the mainstream marine engineering equipment in the world today. Because single-point mooring systems exhibit a "weathervane effect" under the influence of wind, waves, and currents, placing the moored object in the position of least stress, this has become one of the main mooring methods for FPSOs. Compared with the anchor chain single-point mooring system commonly used in deep seas, the flexible steel arm single-point mooring system is commonly used in shallow seas, offering advantages such as fixed single-point position, minimal hull displacement, and fixed manifold. Given the relatively mild nature of the shallow sea environment, research on extending the lifespan of FPSO flexible steel arm single-point mooring systems has become a major research focus. Installing monitoring devices to track the operational status of the flexible steel arm single-point mooring system is the most direct method for obtaining the stress on the flexible steel arm structure.

[0003] However, strain gauges and other monitoring devices cannot be installed inside the mooring leg bearings, making it difficult to obtain the stress at the mooring leg hinge point. The stress can only be calculated by setting monitoring points nearby, and the accuracy needs improvement. The same applies to the connection between the YOKE arm and the YOKE ballast tank. Furthermore, factors such as corrosion, vibration, and limited lifespan reduce the stability of the monitoring equipment. Summary of the Invention

[0004] To address the problems existing in current technologies, this invention proposes a method for predicting and warning the structural safety of a single-point mooring system with a flexible steel boom. This method uses numerical simulation to obtain hull draft and environmental parameters, as well as a corresponding set of single-point structural stress samples. A neural network algorithm is then used to obtain the mapping relationship between the input and output; by inputting the hull draft and environmental parameters, the single-point structural stress is obtained, enabling prediction and warning.

[0005] To address the technical problems existing in the prior art, the present invention adopts the following technical solution:

[0006] A method for predicting and warning the structural safety of a single-point mooring system with a flexible steel boom includes the following steps:

[0007] S1. A coupled model of FPSO and soft steel arm single-point mooring system is established by using the hydrodynamic model of FPSO floating production storage and offloading vessel.

[0008] S2, a single-point structure finite element model of the single-point mooring system with a soft steel arm is established in ANSYS.

[0009] S3. Based on the following correlation formula, determine the correlation parameters, perform time-domain simulation of the FPSO and soft steel boom single-point mooring system coupling model and finite element calculation of the single-point structure finite element model;

[0010]

[0011] Where: X is the environmental parameter or hull parameter, Y is the single-point structural stress, Cov(X,Y) is the covariance, Var[X] is the variance, and r(X,Y) is the linear correlation coefficient;

[0012] S4, a sample set of marine environment and hull characteristics and single-point structural stress is established based on the results of time-domain simulation and finite element calculation;

[0013] S5, using the sample set to train and test the neural network model to obtain a single-point mooring system structural safety prediction proxy model;

[0014] S6, the predicted value of the input feature of the safety prediction agent model is used to obtain the predicted value of the single-point structural stress;

[0015] Among them: the features include FPSO bow draft, FPSO stern draft, significant wave height, spectral peak period, wind speed, relative wind direction, current speed and relative current direction, and single-point structural stress includes the maximum stress at the mooring leg hinge point and the maximum stress at the YOKE critical location.

[0016] S7, compare the predicted value of single-point structural stress with the threshold to realize forecasting and early warning; if the predicted value reaches or exceeds the threshold, the safety prediction agent model will issue an early warning.

[0017] Furthermore, in steps S1 and S2, the process of establishing the finite element model of the single-point mooring system is as follows:

[0018] Based on the FPSO type value table and typical loading conditions, a hydrodynamic model of FPSO covering different drafts was established in AQWA for hydrodynamic analysis.

[0019] The hydrodynamic analysis results were imported into Orcaflex to establish a coupled model of the FPSO and soft steel boom single-point mooring system.

[0020] Based on the single-point structure drawings, a single-point structure finite element model is established in ANSYS; wherein: the single-point mooring system finite element model includes left and right mooring legs and a YOKE, and the connection position between the YOKE head and the jacket is simplified to a fixed end constraint; the loading method of the single-point structure finite element model is to apply displacement loads to the upper ends of the left and right mooring legs.

[0021] Furthermore, in step S3, trial calculations are performed on the coupled model and the finite element model to analyze the correlation between different hull parameters, marine environment parameters and single-point structural stress.

[0022] The features are: FPSO bow draft, FPSO stern draft, significant wave height, spectral peak period, wind speed, relative wind direction, current velocity, and relative current direction. Based on the value range of the features and the mutual influence between significant wave height, wind speed, and current velocity, the features are arranged and combined into different calculation conditions to perform time-domain simulation of the coupled model.

[0023] For a single working condition, extract the time-domain data of the displacement load at the upper end of the left and right mooring legs from the coupled model;

[0024] In the single-point structural finite element model, the displacement load time-domain data is applied as displacement load to the upper end of the left and right mooring legs for finite element calculation.

[0025] Furthermore, in step 5, the process of using a sample set to train and test the neural network model to obtain a single-point mooring system structural safety prediction proxy model is as follows:

[0026] A preliminary neural network model is established based on the features and the stress variables of the single-point structure to be predicted.

[0027] The sample set is standardized according to the following formula;

[0028]

[0029] Where: X is the original value, x is the standardized value, μ is the mean score of the sample space, and σ is the standard deviation of the sample space;

[0030] The sample set is divided into a training set and a test set;

[0031] The neural network is trained by adjusting the batch size, the number of neurons in the hidden layer, and the number of hidden layers.

[0032] The mean square error of the test set is calculated using the following formula. When the mean square error reaches its minimum, the surrogate model for predicting the structural safety of a single-point mooring system is obtained:

[0033]

[0034] Among them: O p T is the model output value. p is the expected output value, and n is the number of input values.

[0035] Beneficial effects

[0036] Compared with the prior art, the advantages of the present invention are as follows:

[0037] Compared to monitoring equipment which is limited by its installation location, this invention can obtain stress at any location on the soft steel arm structure, such as inside the bearing, at the connection between the YOKE arm and the ballast tank, etc.

[0038] Compared to monitoring systems that can only monitor real-time stress, this invention obtains real-time single-point structural stress based on real-time hull draft and environmental data, and simultaneously predicts single-point structural stress based on hull draft estimated by production plan and forecast environmental data.

[0039] It boasts fast prediction speed and strong stability. Based on neural networks, it predicts stress at a single point in a structure without repeatedly modeling the coupled model, resulting in rapid prediction. Furthermore, the single-point structural stress prediction proxy model, mounted on a computer, is free from external interference and exhibits strong operational stability. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the structural safety forecasting and early warning method for a single-point mooring system with a soft steel boom as described in this invention.

[0041] Figure 2 This is a schematic diagram of steps S1, S2, and S3 in the method of the present invention;

[0042] Figure 3 This is a schematic diagram of steps S4 and S5 in the method of the present invention;

[0043] Figure 4 This is a schematic diagram of steps S6 and S7 in the method of the present invention;

[0044] Figure 5 This is a schematic diagram of the input and output of the single-point structural stress prediction proxy model in the method of this invention. Detailed Implementation

[0045] The following is in conjunction with the appendix Figures 1 to 5 Specific examples will be used to further illustrate the present invention. It should be understood that the specific examples described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0046] like Figure 2 Based on the FPSO type value table and typical loading conditions, a hydrodynamic model of FPSO covering different drafts was established in AQWA for hydrodynamic analysis. The hydrodynamic analysis results were then imported into Orcaflex to establish a coupled model of the FPSO and the soft steel boom single-point mooring system.

[0047] Based on the single-point structural drawings, a single-point structural finite element model was established in ANSYS. The established single-point structural finite element model only includes the left and right mooring legs and the yoke, simplifying the connection between the yoke head and the jacket as a fixed-end constraint. The loading method for the single-point structural finite element model is to apply displacement loads to the upper ends of the left and right mooring legs.

[0048] Trial calculations were performed on the coupled model and the finite element model, and the linear correlation coefficient was calculated using the following formula: The correlation between different hull parameters, marine environmental parameters, and single-point structural stress was analyzed. Parameters with strong correlations were selected as input features for the neural network model. The correlation coefficients are shown in Table 1.

[0049] Table 1 Correlation Coefficient Table

[0050]

[0051] The initial neural network input features were: bow draft, stern draft, significant wave height, spectral peak period, wind speed, relative wind direction, current velocity, and relative current direction. Based on the value range of the features and the interaction between wind, waves, and current, the features were combined into different calculation conditions. Table 2 shows the combined conditions of significant wave height, spectral peak period, and wind speed, totaling 290 groups. Three current velocity values ​​were selected: 0.5, 1.5, and 2.5 m / s. The relative wind direction ranged from ±45°, with 15° intervals, for a total of 7 values. The relative current direction ranged from ±90°, with 15° intervals, for a total of 13 values. Time-domain simulations were performed on the coupled model, totaling 633,360 sets of conditions.

[0052] Table 2. Combination of significant wave height, spectral peak period, and wind speed.

[0053]

[0054]

[0055] For a single operating condition, time-domain data of displacement loads at the upper ends of the port and starboard mooring legs are extracted from the time-domain simulation results of the coupled model. In the single-point structural finite element model, the time-domain data of displacement loads are applied as displacement loads to the upper ends of the port and starboard mooring legs for finite element calculation. The maximum single-point structural stress is extracted, namely the maximum stress at the mooring leg hinge point and the maximum stress at the YOKE critical location. This step is repeated to obtain the maximum stress at the mooring leg hinge point and the maximum stress at the YOKE critical location under all operating conditions. Numerical simulation and finite element calculations in S3 are used to establish a training set for the neural network. Input features include bow draft, stern draft, significant wave height, spectral peak period, wind speed, relative wind direction, current velocity, and relative current direction. Output variables include the maximum stress at the mooring leg hinge point and the maximum stress at the YOKE critical location. Randomly selected hull draft and marine environment data are used for numerical simulation to establish samples for the neural network.

[0056] S4, as shown in Figure 3, establish an initial fully connected layer neural network model. The initial fully connected layer neural network has 8 input layer neurons, only one hidden layer with 8 hidden layer neurons, and 2 output layer neurons.

[0057] S5, the sample set in S3 is processed using the formula After standardization, the data is input into the neural network. The sample set is divided into a training set and a test set. The neural network model is trained, and the features of the test set are standardized and then input into the neural network to calculate the mean squared error of the test set. Adjusting the neural network, the method is as follows:

[0058] The initial batch size is 5, the initial number of hidden layer neurons is 8, and the initial number of hidden layers is 1.

[0059] The first step is to increase the batch size in increments of 5 until the mean squared error of the test set no longer decreases, at which point the batch size is no longer changed. The second step is to increase the number of hidden layer neurons in increments of 8 until the mean squared error of the test set no longer decreases, at which point the number of hidden layer neurons is no longer increased.

[0060] The third step is to increase the number of hidden layers by one layer and increase the number of neurons in the last hidden layer until the mean square error of the test set no longer decreases.

[0061] Repeat step 3 until the mean square error no longer decreases after adding hidden layers, the network structure reaches the optimal level, and a surrogate model for predicting structural stress of a single-point mooring system is obtained.

[0062] S6, such as Figure 4 , Figure 5 Based on the single-point mooring system structural safety prediction proxy model, the predicted effective wave height, spectral peak period, wind speed, relative wind direction, current speed, relative current direction, and the expected bow and stern drafts according to the production plan are input to obtain the maximum predicted stress value of the mooring leg hinge point and the maximum predicted stress value of the YOKE key position.

[0063] S7 uses 80% of the yield strength of a single-point structure as the stress threshold for the single-point structure. It compares the predicted value of the maximum stress at the mooring leg hinge point with the threshold, and compares the predicted value of the maximum stress at the YOKE key location with the threshold. When the predicted value reaches or exceeds the threshold, an early warning is issued.

[0064] The above description is merely one example of the present invention and is not intended to limit its other uses. Any person skilled in the art can make slight modifications to the content of this invention and apply them accordingly. However, any improvements made under the premise of the principles of this invention should also be considered within the scope of protection of this invention.

Claims

1. A method for predicting and warning the structural safety of a single-point mooring system with a soft steel boom, characterized in that, Includes the following steps: S1. Based on the FPSO type value table and typical loading conditions, establish an FPSO hydrodynamic model covering different drafts in AQWA and perform hydrodynamic analysis; import the hydrodynamic analysis results into Orcaflex to establish a coupled model of the FPSO and flexible boom single-point mooring system; S2. Based on the single-point structure drawings, establish a single-point structure finite element model in ANSYS; wherein: the single-point structure finite element model includes the left and right mooring legs and the yoke, simplifying the connection position of the yoke head to the jacket as a fixed end constraint; S3. Perform time-domain simulation on the coupled model of the FPSO and flexible boom single-point mooring system and finite element calculation on the single-point structure finite element model, wherein, for a single loading condition, extract the time-domain data of the displacement load at the upper end of the left and right mooring legs in the coupled model; in the single-point structure finite element model, apply the displacement load time-domain data as the displacement load to the... S4. Finite element analysis is performed on the upper part of the port and starboard mooring legs; S5. A sample set of marine environment, hull characteristics, and single-point structural stress is established based on the results of time-domain simulation and finite element analysis; S6. The neural network model is trained and tested using the sample set to obtain a single-point mooring system structural safety prediction proxy model; S7. The predicted values ​​of the input features are used to obtain the predicted values ​​of single-point structural stress; wherein: the features include FPSO bow draft, FPSO stern draft, significant wave height, spectral peak period, wind speed, relative wind direction, current speed and relative current direction, and the single-point structural stress includes the maximum stress value at the mooring leg hinge point and the maximum stress value at the YOKE key position; S8. The predicted values ​​of single-point structural stress are compared with a threshold to achieve a forecast and early warning, wherein the threshold is a preset percentage of the yield strength of the single-point structural material; if the predicted value reaches or exceeds the threshold, the safety prediction proxy model issues an early warning.

2. The structural safety prediction and early warning method for a single-point mooring system with a soft steel boom according to claim 1, characterized in that, In step S3, trial calculations are performed on the coupled model and the finite element model, and the correlation between different hull parameters, marine environment parameters and single-point structural stress is analyzed according to the following formula. Where: X is the environmental parameter or hull parameter, Y is the single-point structural stress, Cov(X,Y) is the covariance, and Var[X] is the variance. r(X,Y) is the linear correlation coefficient; the features are: FPSO bow draft, FPSO stern draft, significant wave height, spectral peak period, wind speed, relative wind direction, current velocity and relative current direction; based on the value range of the features and the mutual influence between significant wave height, wind speed and current velocity, the features are arranged and combined into different calculation conditions to perform time-domain simulation of the coupled model.

3. The structural safety prediction and early warning method for a single-point mooring system with a soft steel boom according to claim 1, characterized in that, Step 5 involves training and testing the neural network model using a sample set to obtain a proxy model for predicting the structural safety of a single-point mooring system: A preliminary neural network model is established based on the features and the stress variables of the single-point structure to be predicted; the sample set is then further standardized according to the following formula. Where: X is the original value, x is the standardized value, μ is the mean score of the sample space, and σ is the standard deviation of the sample space; the sample set is divided into a training set and a test set; the batch size, the number of hidden layer neurons, and the number of hidden layers are adjusted to train the neural network; the mean square error of the test set is calculated using the following formula, and when the mean square error reaches its minimum, the single-point mooring system structural safety prediction proxy model is obtained: Where: Op is the model output value, Tp is the expected output value, and n is the number of input values.

Images