CN106021880A - Jacket platform structure response computing method based on BP neural network - Google Patents
Jacket platform structure response computing method based on BP neural network Download PDFInfo
- Publication number
- CN106021880A CN106021880A CN201610312508.4A CN201610312508A CN106021880A CN 106021880 A CN106021880 A CN 106021880A CN 201610312508 A CN201610312508 A CN 201610312508A CN 106021880 A CN106021880 A CN 106021880A
- Authority
- CN
- China
- Prior art keywords
- network system
- layer
- input
- neural network
- nerve network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16Z—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS, NOT OTHERWISE PROVIDED FOR
- G16Z99/00—Subject matter not provided for in other main groups of this subclass
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2219/00—Indexing scheme relating to application aspects of data processing equipment or methods
- G06F2219/10—Environmental application, e.g. waste reduction, pollution control, compliance with environmental legislation
Abstract
The invention relates to the technical field of structure design of ocean platforms, and specifically discloses a jacket platform structure response computing method based on a BP neural network. The computing method comprises the following steps: step one, establishing a BP neural network system, wherein the BP neural network system comprises an input layer, a hidden layer and an output layer, the input layer and the hidden layer are interconnected as the first stage of the jacket platform stress, and the hidden layer and the output layer are interconnected as the second stage of the jacket platform stress; step two, training the BP neural network system; step three, applying the BP neural network system, and inputting any wind, wave and stream environment load combined numerical value into the BP neural network system as an input vector to obtain the corresponding platform structure response. An artificial neural network system established by the invention is applied to compute the structure response of the jacket platform, the structure response quite close to numerical simulation can be immediately obtained by inputting any environment load combination into a neural network system, and the operation is accurate and fast.
Description
Technical field
The present invention relates to the structure-design technique field of ocean platform, particularly relate to this kind of ocean of jacket platform
The structural response computational methods of platform.
Background technology
Ocean platform is to produce for offshore oil and gas and the marine heavy construction works of operation.Ocean platform is tied
Structure is complicated, bulky, involve great expense, and its Service Environment condition is sufficiently complex and severe, not only wind-engaging,
Wave, stream and the synergy of ice, the most also threatened by extreme loads such as earthquakes, accident, institute once occurred
The economy, the environmental hazard that produce are the hugest.
In the structure design of ocean platform and relevant fail-safe analysis, need to carry out the rings such as substantial amounts of stormy waves stream
Border Load Simulation calculates, to obtain the structural response of platform.Computer and finite element despite modernization are soft
Part carries out computational analysis, but software analysis simulation process is the most loaded down with trivial details, calculates process the most long.
Specifically, for the structural response analysis of ocean platform, the most the most frequently used technological means is for using
The Large-scale professional finite element softwares such as ANSYS, SEASAM carry out numerical simulation, and first it analyzes process needs
Ocean platform uses the 3D graphics softwares such as AutoCAD carry out simplifying modeling, then imports finite element software
Calculating, the major defect of this technological means has:
(1) to modeling sufficiently complex as this kind of large complicated marine structure of ocean platform carries out 3D, tool
The shortcoming having modeling difficulty;
(2) FEM software is run and computer hardware is required height, buys specialty finite element software simultaneously
Required costly and need could grasp through system training how to operate.For same structure, use not
Same cell type, Meshing Method, obtained result of calculation there may be a great difference, incorrect
Operation likely result in calculating and occur not restraining and causing running unsuccessfully, the early stage of program is debugged the most loaded down with trivial details,
Related personnel needs the comprehensive finite element of suitable system and software operative knowledge just can complete correlation computations, has
The shortcoming that technical threshold is high;
(3) finite element algorithm belongs to iterative algorithm, for large scale structure, even if using senior finite elements
Division methods, during single iteration, required amount of calculation is the hugest, although using this technological means permissible
Obtain accurate structural response, but have the calculating speed of service slow, the shortcoming of inefficiency;
(4) for the ocean platform that catheterostat is this kind of, the environmental variable such as numerical simulation calculation such as wave acts on
On works, the computational methods simplified in theory are the sizes first calculating wave force, then by wave force
Act on works, analyze the load effect that a certain key position is produced.Need to analyze fluid and structure
Related software co-ordination just can complete numerical analysis, increases difficulty in computation further, it is often necessary to simultaneously
Calling two or more finite element program, it is the most also a relatively difficult problem that its structural response calculates.
In a word, owing to ocean platform environmental load combined situation is complicated, need to consider the most different operating mode groups
Closing, load cases combination is the most up to ten thousand, calculates offshore platform structure as used numerical simulation technology means to be applied to and rings
Should analyze, relevant design department generally requires one or two months even more long-time just can completing and calculates.
And using the structural response of physical experiments Measuring Oceanic platform, practical operation is wasted time and energy, at present
It is only applied to compare with numerical simulation result.
For the inefficient shortcoming of numerical simulation calculation, it is public that a lot of experts and scholars both at home and abroad propose a lot of experiences
Formula is for calculating the structural response of platform.Its technological means is for concrete ocean platform, its structure is rung
(such as base shear, tilting moment, deck displacement etc.) environment such as wave height, wind speed and flow velocity should be considered as
The function of key element.But due to the complex nonlinear of ocean platform self structure, use this technological means to obtain
Structural response and truth error are relatively big, and general relative error can control to be more satisfactory knot 30%
Really, reality cannot extensively be applied.
Summary of the invention
The technical problem to be solved in the present invention find exactly a kind of computational accuracy close to numerical simulation, but the most just
In operation, low technical threshold, calculate technological means rapid, efficient and ring to the structure realizing jacket platform
Should calculate.
In order to solve above-mentioned technical problem, the technical solution adopted in the present invention is: a kind of neural based on BP
The jacket platform structural response computational methods of network, it comprises the steps:
Step one, sets up BP nerve network system, and this BP nerve network system includes: input layer, hidden layer
And output layer, it is linked as the first stage of jacket platform stress, hidden layer and output between input layer and hidden layer mutually
The second stage of jacket platform stress it is linked as mutually, if every layer comprises dry contact, between every node layer between Ceng
Being not attached to, if the input vector of input layer is wind, wave and the combination of stream, the output vector of output layer is platform
Structural response FX, FY, FZ and MX, MY and MZ, the number of input layer is input vector
Dimension, the number of output layer node is the dimension of output vector;
Step 2, train BP nerve network system, utilize collect obtain sea area survey wind, wave for a long time
And flow data, according to wind, wave and the distribution situation of flow data, respectively take several eigenvalues and be combined, obtain
Obtain feature load cases combination X to obtain as input vector, employing physical experiments or employing method for numerical simulation
Expectation platform structure response data W in the case of each feature load cases combination X, by X and corresponding W input
BP nerve network system to be trained, with BP nerve network system output vector actual platform structural response number
According to the root-mean-square error of Y Yu W as network performance function, during training in BP nerve network system
Weights and deviation are adjusted according to the error performance of network, repeatedly revise what BP nerve network system obtained
Actual platform structural response data Y so that it is finally the root-mean-square error with W minimizes;
Step 3, applies BP nerve network system, by arbitrary wind, wave and stream environmental load combined value
As input vector input BP nerve network system, then obtain corresponding output vector platform structure response FX,
FY, FZ and MX, MY and MZ.
Preferably, this BP nerve network system only has 1 hidden layer, and this hidden layer comprises 6 nodes.
Preferably, input layer excitation function is tansig,
Hidden layer excitation function is logsig,
Output layer excitation function is purelin, purelin (x)=x.
Technical solution of the present invention has the benefit that
(1) the BP neural network topology structure that the present invention is applied decomposes division according to two benches stress, is subject to
Structural mechanics ultimate principle inspires, it is contemplated that when various environmental loads act on jacket platform jointly, combination
Non-linear relation complicated between load and structural response;Avoid lengthy and jumbled structural mechanics theoretical, by conduit
The structural response of body panel calculates concise, greatly reduces the professional threshold of relevant design personnel.Should simultaneously
The determination of topological structure, had both avoided the excessively complicated training caused of neutral net slow, had also been effectively ensured
Accuracy for structural response simulation;
(2) extensively it is built in the form of program bag at present due to the relevant base program of artificial neural network
In the conventional numerical computations softwares such as Matlab, relevant design personnel specify that neural network topology structure
After, can be directly utilized it and build BP nerve network system, provide convenience condition for the popularization of the present invention;
(3) artificial neural network system utilizing the present invention to be built is applied to calculate the structure of jacket platform
Response, combines for any environmental load, is inputted nerve network system and then can be instantly available and Numerical-Mode
Intend the structural response being sufficiently close to, accurately, quickly.
Accompanying drawing explanation
The BP neural network topology structure figure of the mono-hidden layer of Fig. 1.
Fig. 2 one embodiment of the invention BP neural network topology structure figure.
Fig. 3-1 one embodiment of the invention N orientation FX--BP neutral net fitting result schematic diagram.
Fig. 3-2 one embodiment of the invention N orientation FY--BP neutral net fitting result schematic diagram.
Fig. 3-3 one embodiment of the invention N orientation FZ--BP neutral net fitting result schematic diagram.
Fig. 3-4 one embodiment of the invention N orientation MX--BP neutral net fitting result schematic diagram.
Fig. 3-5 one embodiment of the invention N orientation MY--BP neutral net fitting result schematic diagram.
Fig. 3-6 one embodiment of the invention N orientation MZ--BP neutral net fitting result schematic diagram.
Detailed description of the invention
The invention will be further described with specific embodiment below in conjunction with the accompanying drawings, so that those skilled in the art
Member can be better understood from the present invention and can be practiced, but illustrated embodiment is not as a limitation of the invention.
The present invention is based on artificial neural network technology.Artificial neural network is the most ripe a kind of engineering
Habit means, all obtain good effect in the application in other a lot of fields, therefore the present invention calculate at model
Aspect has solid theory for supporting.
Artificial neural network is on the basis of the modern biotechnology successes achieved in research, by nervous physiology science,
The achievement in research of the related sciences such as information science, mathematical and physical science and computational science is set up, artificial neural network
System is made up of a large amount of processing units (i.e. neuron) interconnection, it is possible to imitate human brain information processing mechanism.Its
Main Types includes: BP, radially base, self-organizing and feedback neural network, the most most widely used is
Use the MLFFANN of error backpropagation algorithm, i.e. BP neutral net.
BP neutral net has good None-linear approximation ability, generalization ability and easy adaptive, and it is the most special
Point includes:
(1) distributed information storage means
BP neutral net is the state with each processor itself and the storage information of the type of attachment between them
, an information is not stored in a place, but presses distribution of content over the entire network.On network a certain
Place is not only to store an external information, but stores the partial content of multiple information.Whole network is to many
After individual Information procession, just storage arrives network everywhere, and therefore, it is a kind of distributed storage mode.
(2) MPP
Owing to BP neutral net is made up of in a unique way a large amount of artificial neurons, it can connect simultaneously
Receiving multiple input information, and also can transmit simultaneously, artificial neuron can respond with the form of voting,
Therefore the output result of artificial neural network is decided by vote by most artificial neurons simultaneously, can automatically achieve " few
Number obey majority " effect.This function has substantially maximally utilised space complexity, and non-
Often significantly reduce time complexity.
(3) self study and adaptivity
Artificial neural network, by posteriori study and training, can develop various function, similar
In biological neural network.The each layer of BP neutral net directly connects weights and has certain adjustability, and network can
To determine the weights of network by training and study, present the strongest self adaptation to environment and to external world
The self-learning capability of things.
(4) stronger robustness and fault-tolerance
The distributed information storage means of BP neutral net so that it is there is stronger fault-tolerance and associative memory merit
Can, if the so information dropout of certain part or damage, network remains to recover the most complete information, system
Remain to run.
BP neural networks principles
BP neutral net is generally made up of input layer, hidden layer and output layer, the most totally interconnected, every layer
It is not attached between node.The number of its input layer generally takes the dimension of input vector, output layer node
Number generally take the dimension of output vector, hidden node number there is no the standard determined at present, need to be by repeatedly
The method that examination is gathered, then obtains final result.According to Kolmogor theorem, there is a hidden layer (hidden node
Abundant) three layers of BP neutral net in closed set, any non-linear continuous function can be approached with arbitrary accuracy.
As a example by the BP neutral net of single hidden layer, its topological structure is as shown in Figure 1.
BP neutral net can regard as one from be input to output mapping, i.e. F:Rn→Rm,
F (X)=Y.For sample set: input value is xi(∈Rn) and output valve be yi(∈Rm), it is believed that
There is a certain mapping g and make g (xi)=yi(i=1,2 ..., n).Neutral net is by carrying out simple function
Matching for several times, show that the function f of approximation is most preferably approaching of g.
The learning process of BP algorithm is made up of forward-propagating and back propagation.During forward-propagating, input mould
Formula successively processes through hidden layer from input layer, and is transmitted to output layer, and the state of each layer of neuron only affects next
The state of layer neuron.If output layer can not obtain desired output, then proceed to back propagation, by error
Signal returns along original connecting path, by revising the connection weights of each neuron so that error signal is
Little.Iterate desired value Y making neural computing obtainLMinimum with real response value T root-mean-square error.
General L layer BP neutral net, note input layer is the 0th layer, and output layer is L layer, and intermediate layer is (i.e.
Hidden layer) it is followed successively by the 1st layer to L mono-1 layers.The neuron number of kth layer is nk, kth one 1 layers arrives kth
The weight matrix of layer isWhereinRepresent-1 layer of i-th neuron of kth and kth layer jth
The connection weights of individual neuron.
The input vector assuming artificial neural network is X=(X1,X2,…,Xn0)T, then its 1st layer connects
Receiving vector is Z1=W1 TX, output vector is Y1=(Y1 1,Y2 1,…,Yn1 1)T;Its kth layer (k >=2) can be obtained equally
Reception vector beOutput vector isWherein:
In formula, fi k() is the excitation function of kth layer i-th neuron.Excitation function has various ways, as conventional
S type function or sigmoid function be called:
The study of network seeks to determine weight matrix Wk(k=1,2 ..., L) so that network ideal exports
YL=(Y1 L,Y2 L,…,YnL L)T, reach the effect minimum with the error of actual output T.
The present invention is BP artificial neural network to be applied to jacket platform structural response calculate.
One embodiment of present invention jacket platform based on BP neutral net structural response computational methods is as follows.
Consider that the load effect of jacket platform is mainly produced by wind, wave, stream effect, need during actual design
Calculate given one group of wind, wave, stream compound action when jacket platform, if the response at platform mud face be FX,
FY, FZ and MX, MY and MZ.
Step1: set up BP nerve network system
The BP nerve network system of this jacket platform has 3 inputs, and 6 outputs, in conjunction with classical knot
Structure mechanical knowledge, is decomposed into two stages by the true stress physical process of jacket platform:
(1) when the environmental load such as wind, wave, stream acts solely on works, due to its active position, effect
Varying in size of power, can produce different platform response FX, FY, FZ and MX, MY and MZ, and
When environmental load simultaneously acts on jacket platform, the same type structural response meeting that each environmental load produces
Influence each other, be finally changed into the structural response that jacket platform is actual;
(2) there is also the relation of a kind of mutual restriction between each response of jacket platform, i.e. FX can to FY,
FZ and MX, MY and MZ have an impact, in like manner with other kinds of structural response.
Decompose based on above two benches stress, BP nerve network system be designed to the topological structure such as Fig. 2,
It is the first stage of jacket platform stress between input layer and hidden layer, considers between hidden layer to output layer
The stress relation of second stage.
For excitation function, tansig Yu logsig function is S type function, when it is differentiated, and can
To represent by certain form of self.This point is critically important when doing numerical experimentation, in training neutral net
Time, the back propagation of weights needs to use the derivative of excitation function, and multilamellar then needs to use multiple derivative, this
Time use S type function then can reduce the storage area of computer as excitation function, improve computing and convergence
Speed.Additionally, excitation function all has significant impact for discrimination or convergence rate.Approaching height
During secondary curve, sigmoid function ratio of precision linear function is much higher, but amount of calculation is the most much greater.To sum up institute
Stating, input layer excitation function is set to tansig by the present invention, plays the effect expanding codomain, makes in hidden layer each
Non-linear joint between branch is more extensive, and hidden layer uses logsig, and output layer uses purelin.
Net=newff (in, ou, [6 6], { ' tansig ' ' logsig ' ' purelin ' }, ' trainlm ');
The expression formula of each excitation function is as follows;
Purelin (x)=x (5)
Step2: training BP nerve network system
BP neutral net belongs to the learning process having supervision, needs to utilize the long-term actual measurement collecting the sea area obtained
Wind, wave and flow data, according to wind, wave and the distribution situation of flow data, respectively take several eigenvalues and carry out group
Close, it is thus achieved that feature load cases combination X, physical experiments can be used or use method for numerical simulation to record respectively
Platform structure response data W in the case of load cases combination, training process is by input vector X and corresponding phase
Hope output response W input BP nerve network system to be trained, with BP neutral net output response Y with
The root-mean-square error of W as network performance function, the weights in BP neutral net and deviation during training
Error performance according to network is adjusted, and repeatedly revises the platform structure sound that BP nerve network system obtains
Answer Y so that it is the root-mean-square error of Expected Response W minimizes the most finally;
Net=train (net, input, output2);%input is input data X, and output2 is that target exports W
In the present embodiment, utilize long-term actual measurement 10min mean wind speed, the significant wave collecting certain sea area obtained
High, surface current speed data, according to wind speed, wave height, the distribution situation of flow velocity, respectively take 8,6,8 features
Value, see table 1, it is considered to stormy waves stream is all from same orientation, has 384 kinds of composite conditions, can use physics
Model test records 6 structural responses that often group operating mode is corresponding, may be used without number in the case of conditions permit
The technological means of value simulation simulates its result.
Table 1 stormy waves stream eigenvalue
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | |
Wind speed | 5 | 10 | 20 | 30 | 40 | 50 | 60 | 65 |
Significant wave height | 2 | 5 | 8 | 11 | 14 | 17 | ||
Surface current speed | 0.6 | 0.9 | 1.2 | 1.5 | 1.8 | 2.1 | 2.4 | 2.7 |
Note: wind speed, flow rate take m/s, significant wave height unit is m.
Step3: application BP neutral net
Utilize the BP neutral net that trains, arbitrary environmental load combined, only need to by wind, wave,
Fluxion value input BP nerve network system, then can be instantly available corresponding structural response.
Simout=sim (net, input);%input is input data X, and simout is neural computing response
In order to verify the accuracy of the present invention, Fig. 3-1 to 3-6 is that N intends to structural response BP neutral net
Close figure.Provide this catheterostat using BP neutral net matching to obtain when stormy waves stream is all from N orientation to put down
Platform structural response.Result display uses BP neutral net matching platform response effect very good, and networking is defeated
Go out value almost identical with actual value.
Table 2 N is to Fitness Test table
RMSE | MaxRSE | MinRSE | MRSE | DRSE | PSE (%) | |
FX | 0.0115 | 0.0528 | 1.8897e-5 | 0.0071 | 0.0034 | 0 |
FY | 0.0027 | 0.0300 | 1.8728e-6 | 0.0014 | 8.1096e-4 | 0 |
FZ | 5.8882e-5 | 9.8062e-5 | 8.9855e-7 | 5.1852e-5 | 5.9299e-5 | 0 |
MX | 0.0071 | 0.1276 | 1.1146e-5 | 0.0015 | 5.9918e-4 | 0.2604 |
MY | 6.7219e-4 | 0.0014 | 1.3212e-6 | 5.7532e-4 | 5.7648e-4 | 0 |
MZ | 0.0686 | 1.2822 | 8.2276e-7 | 0.0087 | 0.0015 | 1.0417 |
Note: RMSE is that average deviation quadratic sum opens radical sign, and MaxRSE is relative error maximum, MinRSE
For relative error minimum, MRSE is relative error average, and DRSE is relative error median, PSE
For the relative error percentage ratio more than 10%.
Embodiment described above is only the preferred embodiment lifted by absolutely proving the present invention, the present invention's
Protection domain is not limited to this.Equivalent that those skilled in the art are made on the basis of the present invention substitute or
Conversion, all within protection scope of the present invention.Protection scope of the present invention is as the criterion with claims.
Claims (3)
1. jacket platform structural response computational methods based on BP neutral net, it is characterised in that its
Comprise the steps:
Step one, sets up BP nerve network system, and this BP nerve network system includes: input layer, hidden layer
And output layer, it is linked as the first stage of jacket platform stress, hidden layer and output between input layer and hidden layer mutually
The second stage of jacket platform stress it is linked as mutually, if every layer comprises dry contact, between every node layer between Ceng
Being not attached to, if the input vector of input layer is wind, wave and the combination of stream, the output vector of output layer is platform
Structural response FX, FY, FZ and MX, MY and MZ, the number of input layer is input vector
Dimension, the number of output layer node is the dimension of output vector;
Step 2, train BP nerve network system, utilize collect obtain sea area survey wind, wave for a long time
And flow data, according to wind, wave and the distribution situation of flow data, respectively take several eigenvalues and be combined, obtain
Obtain feature load cases combination X to obtain as input vector, employing physical experiments or employing method for numerical simulation
Expectation platform structure response data W in the case of each feature load cases combination X, by X and corresponding W input
BP nerve network system to be trained, with BP nerve network system output vector actual platform structural response number
According to the root-mean-square error of Y Yu W as network performance function, during training in BP nerve network system
Weights and deviation are adjusted according to the error performance of network, repeatedly revise what BP nerve network system obtained
Actual platform structural response data Y so that it is finally the root-mean-square error with W minimizes;
Step 3, applies BP nerve network system, by arbitrary wind, wave and stream environmental load combined value
As input vector input BP nerve network system, then obtain corresponding output vector platform structure response FX,
FY, FZ and MX, MY and MZ.
2. jacket platform structural response computational methods based on BP neutral net as claimed in claim 1,
It is characterized in that, this BP nerve network system only has 1 hidden layer, and this hidden layer comprises 6 nodes.
3. jacket platform structural response computational methods based on BP neutral net as claimed in claim 1,
It is characterized in that,
Input layer excitation function is tansig,
Hidden layer excitation function is logsig,
Output layer excitation function is purelin, purelin (x)=x.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610312508.4A CN106021880B (en) | 2016-05-12 | 2016-05-12 | Jacket platform structural response calculation method based on BP neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610312508.4A CN106021880B (en) | 2016-05-12 | 2016-05-12 | Jacket platform structural response calculation method based on BP neural network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106021880A true CN106021880A (en) | 2016-10-12 |
CN106021880B CN106021880B (en) | 2019-03-26 |
Family
ID=57100066
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610312508.4A Active CN106021880B (en) | 2016-05-12 | 2016-05-12 | Jacket platform structural response calculation method based on BP neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106021880B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107908817A (en) * | 2017-10-17 | 2018-04-13 | 江苏大学 | A kind of sea transfer platform device structure stress monitoring and diagnostic system |
CN109480872A (en) * | 2018-11-08 | 2019-03-19 | 哈尔滨工业大学 | Driving fatigue detection method based on EEG signals frequency band energy than feature |
CN109871609A (en) * | 2019-02-18 | 2019-06-11 | 中国海洋大学 | The prediction technique that marine floating type platform mooring system is responded based on BP-FEM |
CN113283138A (en) * | 2021-05-25 | 2021-08-20 | 大连理工大学 | Deep sea culture platform dynamic response analysis method based on deep learning |
-
2016
- 2016-05-12 CN CN201610312508.4A patent/CN106021880B/en active Active
Non-Patent Citations (4)
Title |
---|
渠基顺 等: "基于SESAM的导管架平台结构动力响应分析", 《江苏船舶》 * |
王春 等: "基于神经网络算法的海上风机结构状态监测研究", 《中国水利水电科学研究院学报》 * |
郭剑虹: "BP网络的结构动力响应预测", 《泰州职业技术学院学报》 * |
韩卫华 等: "神经网络在结构响应预测中的应用研究", 《信息技术》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107908817A (en) * | 2017-10-17 | 2018-04-13 | 江苏大学 | A kind of sea transfer platform device structure stress monitoring and diagnostic system |
CN109480872A (en) * | 2018-11-08 | 2019-03-19 | 哈尔滨工业大学 | Driving fatigue detection method based on EEG signals frequency band energy than feature |
CN109871609A (en) * | 2019-02-18 | 2019-06-11 | 中国海洋大学 | The prediction technique that marine floating type platform mooring system is responded based on BP-FEM |
CN109871609B (en) * | 2019-02-18 | 2020-10-27 | 中国海洋大学 | Method for predicting response of marine floating platform mooring system based on BP-FEM |
CN113283138A (en) * | 2021-05-25 | 2021-08-20 | 大连理工大学 | Deep sea culture platform dynamic response analysis method based on deep learning |
CN113283138B (en) * | 2021-05-25 | 2024-03-26 | 大连理工大学 | Deep-learning-based dynamic response analysis method for deep-sea culture platform |
Also Published As
Publication number | Publication date |
---|---|
CN106021880B (en) | 2019-03-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Huang et al. | Development of a regional neural network for coastal water level predictions | |
Giannakoglou et al. | Low-cost genetic optimization based on inexact pre-evaluations and the sensitivity analysis of design parameters | |
Chen et al. | Prediction of pressure coefficients on roofs of low buildings using artificial neural networks | |
CN106021880A (en) | Jacket platform structure response computing method based on BP neural network | |
CN104077438B (en) | Power network massive topologies structure construction method and system | |
Kalra et al. | RBF network for spatial mapping of wave heights | |
CN115577436B (en) | Combined deep learning method for solving wind-induced vibration response of uncertain structure | |
Goel | ANN-based approach for predicting rating curve of an Indian River | |
CN103559340A (en) | Irrelevant multisource frequency domain load identification method based on COMI-PSO (comprehensive improved particle swarm optimization) algorithm | |
Cintra et al. | Tracking the model: Data assimilation by artificial neural network | |
CN104361399A (en) | Solar irradiation intensity minute-scale predication method | |
CN101540504B (en) | Current analytical device and method on basis of step-length variable neural network | |
Shi et al. | A fault location method for distribution system based on one-dimensional convolutional neural network | |
CN115048857A (en) | CNN-based transient stability limit transmission power determination method | |
Chen et al. | Predictive model of artificial neural network for disaster prevention | |
CN115099066A (en) | Kriging-based power transmission tower structure wind disaster vulnerability modeling method | |
Song et al. | Neural network application based on GIS and Matlab to evaluation of flood risk | |
Guo et al. | Force finding of suspended-domes using back propagation (BP) algorithm | |
Ji et al. | Data preprocessing method and fault diagnosis based on evaluation function of information contribution degree | |
Meng et al. | Nonlinear system simulation based on the BP neural network | |
Alsulami | Fast and Accurate Load Flow Solution for On-line Applications Using ANN | |
Shi et al. | The interval-valued intuitionistic fuzzy optimized weighted bonferroni means and their application | |
Wang et al. | A new definition of sensitivity for RBFNN and its applications to feature reduction | |
Singh et al. | Degree of insecurity estimation in a power system using functional link neural network | |
Zhu et al. | Application of artificial neural network in the prediction of output in oilfield |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |