CN110489717A - Barge heave movement prediction technique during offshore platform floating support mounting - Google Patents
Barge heave movement prediction technique during offshore platform floating support mounting Download PDFInfo
- Publication number
- CN110489717A CN110489717A CN201910597911.XA CN201910597911A CN110489717A CN 110489717 A CN110489717 A CN 110489717A CN 201910597911 A CN201910597911 A CN 201910597911A CN 110489717 A CN110489717 A CN 110489717A
- Authority
- CN
- China
- Prior art keywords
- barge
- sequence
- heave movement
- data
- heave
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Health & Medical Sciences (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses barge heave movement prediction techniques during offshore platform floating support mounting, comprising the following steps: using MRU attitude transducer heave shift values several to barge continuous acquisition as heave movement data sample;Zero-mean is carried out to heave movement data sample and standardizes pretreatment operation, forms barge heave movement sequence;Data in heave movement sequence are divided into training set and test set;Discrete wavelet transformation and single branch reconstruct, low frequency component sequence and several high fdrequency component sequences after being reconstructed are carried out to training set;Support vector regression prediction model based on particle swarm optimization algorithm is established to respective vector sequence and carries out jump ahead prediction;The superposition of respective predicted value is merged and obtains final predicted value;It updates training set data repetition corresponding steps and obtains multiple predicted values;Predicted value and test set are subjected to data comparison.The method of the present invention can effectively improve the heave movement prediction step and accuracy of barge, guarantee the stability of prediction.
Description
Technical field
The present invention relates to barges during a kind of motion forecast method more particularly to a kind of offshore platform floating support mounting to heave
Motion forecast method.
Background technique
Floating support mounting method is a kind of huge buoyancy provided using barge container by offshore platform top module integral installation
Method on to bottom bracing structure.Due to will receive the influence of the complex environments load such as wind, wave, stream, tidal bore in installation process,
It translates, wave to which the upper module for inevitably leading to barge and on barge generates, upper and lower heave movement, wherein barge
Heave movement up and down be affected to the installation of upper module and more difficult to control.Using such as Fig. 1 institute during floating support mounting
The mode shown installs array Active Compensation oil cylinder 4 that is, between barge 1 and upper module 7, and Active Compensation oil cylinder 4 is connected to oil
On cylinder support 3, intermediate barge support construction 5 and deck support unit 6 are mainly used to support firm upper module 7, in floating support
The heave that barge generation can be compensated when installation in time by controlling the stretching motion of several Active Compensation oil cylinders 4 is displaced, thus
Teetertottering for upper module 7 is efficiently reduced, the safety during floating support mounting is improved, reduces duty cycle and cost.
In order to effectively carry out heaving movement compensation to upper module during floating support mounting, needing will by movement shortest time prediction
Heave movement posture look-ahead within barge following more than ten seconds comes out, and makes to carry out heaving bit shift compensation accordingly in time
Upper module held stationary during the installation process, wherein the heave displacement of barge can use MRU attitude transducer and be acquired,
Its installation site is as shown in Figure 1.It is currently used for heave movement prediction technique mainly to have statistical fluctuation method, convolution method, card
Kalman Filtering method, time series analysis method etc., wherein statistical fluctuation method needs to pre-process mass data, and with forecast
The increase of time will lead to prediction error and significantly increase, and practice is difficult;Convolution method needs accurate measurement certain distance
Ariyoshi wave height and ship itself response kernel function, this two o'clock is extremely difficult under the sea situation complicated in this way of deep-sea, because
This algorithm practical application difficulty is also very big;Kalman filtering method of prediction also has certain limitation, calls time in advance shorter;When
Between sequence analysis be a kind of more commonly used method, but the prediction technique is predicted for non-linear, non-stationary heave movement
With certain limitation, and as there are serious lag and wild effect for the increase of predicted time.
Summary of the invention
It is an object of the invention to overcome the shortcomings of prior art, provide a kind of for during offshore platform floating support mounting
Barge heave movement prediction technique, this method can effectively improve the heave movement prediction duration and accuracy of barge, protect
Demonstrate,prove the stability of prediction.
In order to achieve the above object, the technical solution adopted by the present invention is that:
Barge heave movement prediction technique during offshore platform floating support mounting of the invention, comprising the following steps:
Step 1: interior for a period of time to barge using the MRU attitude transducer on barge in the sea area for carrying out floating support mounting
N heave shift value of continuous acquisition is as heave movement data sample;
Step 2: carrying out zero-mean to the heave movement data sample of barge standardizes pretreatment operation, pretreatment is formed
Barge heave movement sequence X afterwards, is represented by X={ x1,x2,…,xi,…,xn, wherein xiIndicate i-th of heave displacement
Value standardizes pretreated result by zero-mean;
Step 3: by barge heave movement sequence X={ x1,x2,…,xi,…,xnIn data be divided into training set { x1,
x2,…,xmAnd test set { xm+1,xm+2,…,xn, wherein m < n, and the data amount check m in training set accounts for about total data n's
80%;
Step 4: choosing wavelet basis function, the wavelet decomposition number of plies is determined, to the training set { x in heave movement sequence1,
x2,…,xmDiscrete wavelet transformation and single branch reconstruct are carried out, low frequency component sequence and several high frequencies after single branch reconstruct can be obtained
Vector sequence.
Wherein preferentially, dmey wavelet basis function is selected, Decomposition order is 3 layers, after carrying out discrete wavelet transformation and reconstruct
Just first layer high fdrequency component sequence D is obtained1, second layer high fdrequency component sequence D2, third layer high fdrequency component sequence D3It is low with third layer
Frequency component sequence C3, high frequency components sequence and low frequency component sequence can respectively indicate are as follows: D1={ d1,1,d1,2,…,d1,m,
D2={ d2,1,d2,2,…,d2,m, D3={ d3,1,d3,2,…,d3,mAnd C3={ c3,1,c3,2,…,c3,m, d1,mIndicate first layer
Than the m-th data in high fdrequency component sequence, similarly d2,m、d3,m;c3,mIndicate than the m-th data in third layer low frequency component sequence;
Step 5: being established respectively to the low frequency component sequence and high fdrequency component sequence that are obtained after reconstruct respective based on particle
The Support vector regression prediction model (abbreviation PSO-SVR prediction model) of colony optimization algorithm, then using respectively established
PSO-SVR prediction model carries out jump ahead prediction to the data in each vector sequence, thus under obtaining in respective vector sequence
The predicted value of one dataWith
Step 6: the predicted value in the predicted value and each high fdrequency component sequence in obtained low frequency component sequence is carried out
Superposition merges, and finally obtains the next data predicted value of barge heave movement, it may be assumed that
Step 7: updating training set data, that is, remove an oldest data x1, add newest prediction data
Form new training set
Step 8: repeating step 4 to step 7, x in barge heave movement sequence can be successively obtainedm+1,…,xnIt is pre-
Measured value
Step 9: by test set { xm+1,xm+2,…,xnIn data and corresponding predicted value carry out data comparison, to test
Demonstrate,prove the accuracy of the barge heave movement prediction technique.
Compared with prior art, the invention has the following advantages:
The method of the present invention is by carrying out decomposition and reconstruction using discrete data of the wavelet analysis to barge heave movement, then
It carries out the Support vector regression prediction based on particle swarm optimization algorithm respectively to each component, is superimposed and closes by each component predicted value
And final predicted value is obtained, this method can barge be non-linear unstable when floating support mounting preferably suitable for practical sea situation
The prediction of heave movement situation, and the convergence rate and prediction step of prediction algorithm are effectively improved, in barge heave movement
Accuracy and stability with higher in prediction.
Detailed description of the invention
Fig. 1 is offshore platform floating support mounting process heave compensation model structure schematic diagram;
Fig. 2 is barge heave movement prediction technique flow chart during offshore platform floating support mounting of the invention.
Specific embodiment
The present invention will be described in detail in the following with reference to the drawings and specific embodiments.
As shown in Fig. 2, barge heave movement prediction technique during offshore platform floating support mounting of the invention, including it is following
Step:
Step 1: interior for a period of time to barge using the MRU attitude transducer 2 on barge in the sea area for carrying out floating support mounting
N heave shift value of continuous acquisition is as heave movement data sample;
Step 2: carrying out zero-mean to the heave movement data sample of barge standardizes pretreatment operation, pretreatment is formed
Barge heave movement sequence X afterwards, is represented by X={ x1,x2,…,xi,…,xn, wherein xiIndicate i-th of heave displacement
Value standardizes pretreated result by zero-mean;
Step 3: by barge heave movement sequence X={ x1,x2,…,xi,…,xnIn data be divided into training set { x1,
x2,…,xmAnd test set { xm+1,xm+2,…,xn, wherein m < n, the data amount check m in preferred training set accounts for about total data n
80%;
Step 4: choosing wavelet basis function, the wavelet decomposition number of plies is determined, to the training set { x in heave movement sequence1,
x2,…,xmDiscrete wavelet transformation and single branch reconstruct are carried out, the low frequency component sequence and several high fdrequency components after obtaining single branch reconstruct
Sequence.
Wherein preferentially, dmey wavelet basis function is selected, Decomposition order is 3 layers, is just obtained after carrying out discrete wavelet transformation
First layer high fdrequency component sequence d1, second layer high fdrequency component sequence d2, third layer high fdrequency component sequence d3With third layer low frequency component
Sequence c3, but due to one layer of wavelet decomposition of every execution, the length of sequence will shorten to the half before decomposing, and Decomposition order is got over
Greatly, the sequence length of acquisition is shorter, and after the reduction of data is unfavorable in sequence heave movement prediction model foundation, institute
Become the data of each layer component as data amount check in original training set to need to carry out the reconstruct of small echo list branch, through too small
After the reconstruct of wave list branch, first layer high fdrequency component sequence D can be obtained1, second layer high fdrequency component sequence D2, third layer high fdrequency component
Sequence D3With third layer low frequency component sequence C3, high frequency components sequence and low frequency component sequence can respectively indicate are as follows: D1=
{d1,1,d1,2,…,d1,m, D2={ d2,1,d2,2,…,d2,m, D3={ d3,1,d3,2,…,d3,mAnd C3={ c3,1,c3,2,…,
c3,m, d1,mIndicate the than the m-th data in first layer high fdrequency component sequence, similarly d2,m、d3,m;c3,mIndicate third layer low frequency component
Than the m-th data in sequence, and training set { x1,x2,…,xm}=C3+D1+D2+D3;Discrete wavelet transformation and the principle of reconstruct are specific
Referring to " wavelet analysis and its application " of Sun Yankui.
Step 5: being established respectively to the low frequency component sequence and high fdrequency component sequence that are obtained after reconstruct respective based on particle
The Support vector regression prediction model (abbreviation PSO-SVR prediction model) of colony optimization algorithm, then using respectively established
PSO-SVR prediction model carries out jump ahead prediction to the data in each vector sequence, thus under obtaining in respective vector sequence
The predicted value of one data.With first layer high frequency series component D1={ d1,1,d1,2,…,d1,mFor, it is predicted by PSO-SVR
The predicted value of the m+1 data in first layer high fdrequency component sequence can be obtained after modelSimilarly obtainWithWherein, the foundation of PSO-SVR prediction model can be found in Wang Qian et al. " based on PSO-SVR's
Danjiangkou annual flow forecast ".
Step 6: the predicted value in the predicted value and each high fdrequency component sequence in obtained low frequency component sequence is carried out
Superposition merges, and finally obtains the next data predicted value of barge heave movement, it may be assumed that
Step 7: updating training set data, that is, remove an oldest data x1, add newest prediction data
Form new training set
Step 8: repeating step 4 to step 7, x in barge heave movement sequence can be successively obtainedm+1,…,xnIt is pre-
Measured value
Step 9: by test set { xm+1,xm+2,…,xnIn data and corresponding predicted value carry out data comparison, to test
Demonstrate,prove the accuracy of the barge heave movement prediction technique.Wherein, data comparison method uses existing method: as using equal
Square error (RMSE) method, i.e.,
Claims (2)
1. barge heave movement prediction technique during offshore platform floating support mounting, it is characterised in that the following steps are included:
Step 1: continuous in for a period of time to barge using the MRU attitude transducer on barge in the sea area for carrying out floating support mounting
N heave shift value is acquired as heave movement data sample;
Step 2: carrying out zero-mean to the heave movement data sample of barge standardizes pretreatment operation, formed pretreated
Barge heave movement sequence X, is expressed as X={ x1,x2,…,xi,…,xn, wherein xiIndicate i-th of heave shift value through zero passage
Mean value standardizes pretreated result;
Step 3: by barge heave movement sequence X={ x1,x2,…,xi,…,xnIn data be divided into training set { x1,
x2,…,xmAnd test set { xm+1,xm+2,…,xn, wherein m < n;
Step 4: choosing wavelet basis function, the wavelet decomposition number of plies is determined, to the training set { x in heave movement sequence1,x2,…,
xmDiscrete wavelet transformation and single branch reconstruct are carried out, the low frequency component sequence and several high fdrequency component sequences after obtaining single branch reconstruct;
Step 5: being established respectively to the low frequency component sequence and high fdrequency component sequence that are obtained after reconstruct respective excellent based on population
Change the Support vector regression prediction model of algorithm, then using respective established PSO-SVR prediction model to each component sequence
Data in column carry out jump ahead prediction, to obtain the predicted value of next data in respective vector sequence;
Step 6: the predicted value in the predicted value and each high fdrequency component sequence in obtained low frequency component sequence is overlapped
Merge, obtains the next data predicted value of barge heave movement;
Step 7: updating training set data, remove an oldest data x1, add newest prediction dataIt is formed new
Training set
Step 8: repeating step 4 to step 7, x in barge heave movement sequence can be successively obtainedm+1,…,xnPredicted value
Step 9: by test set { xm+1,xm+2,…,xnIn data and corresponding predicted value carry out data comparison, with verifying should
The accuracy of barge heave movement prediction technique.
2. barge heave movement prediction technique, feature exist during offshore platform floating support mounting according to claim 1
In: the wavelet basis function chosen in the step four is dmey wavelet basis function, and Decomposition order is 3 layers, carries out discrete wavelet
It decomposes and obtains first layer high fdrequency component sequence D after reconstructing1, second layer high fdrequency component sequence D2, third layer high fdrequency component sequence D3
With third layer low frequency component sequence C3。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910597911.XA CN110489717A (en) | 2019-07-04 | 2019-07-04 | Barge heave movement prediction technique during offshore platform floating support mounting |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910597911.XA CN110489717A (en) | 2019-07-04 | 2019-07-04 | Barge heave movement prediction technique during offshore platform floating support mounting |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110489717A true CN110489717A (en) | 2019-11-22 |
Family
ID=68546567
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910597911.XA Pending CN110489717A (en) | 2019-07-04 | 2019-07-04 | Barge heave movement prediction technique during offshore platform floating support mounting |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110489717A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112762935A (en) * | 2020-12-23 | 2021-05-07 | 赫星(厦门)电子有限公司 | Positioning method and system based on hull attitude monitoring |
CN113942622A (en) * | 2021-11-19 | 2022-01-18 | 博迈科海洋工程股份有限公司 | Motion compensation method suitable for FPSO upper module lifting installation process |
CN114894043A (en) * | 2022-03-30 | 2022-08-12 | 北京航天飞腾装备技术有限责任公司 | Precise guidance ammunition attitude filtering method and system based on wavelet packet transformation |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103778482A (en) * | 2014-02-12 | 2014-05-07 | 中国农业大学 | Aquaculture dissolved oxygen short-term prediction method based on multi-scale analysis |
CN105117593A (en) * | 2015-08-18 | 2015-12-02 | 河海大学 | Wavelet transform and particle swarm optimized grey model-based short-term wind speed forecasting method |
CN106384170A (en) * | 2016-09-24 | 2017-02-08 | 华北电力大学(保定) | Wavelet decomposition and reconstruction-based time sequence wind speed prediction method |
CN107316101A (en) * | 2017-06-02 | 2017-11-03 | 西南交通大学 | A kind of wind speed forecasting method selected in advance based on wavelet decomposition and component |
CN109657709A (en) * | 2018-12-06 | 2019-04-19 | 湖北博华自动化系统工程有限公司 | A kind of equipment fault prediction technique based on particle group optimizing support vector regression |
-
2019
- 2019-07-04 CN CN201910597911.XA patent/CN110489717A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103778482A (en) * | 2014-02-12 | 2014-05-07 | 中国农业大学 | Aquaculture dissolved oxygen short-term prediction method based on multi-scale analysis |
CN105117593A (en) * | 2015-08-18 | 2015-12-02 | 河海大学 | Wavelet transform and particle swarm optimized grey model-based short-term wind speed forecasting method |
CN106384170A (en) * | 2016-09-24 | 2017-02-08 | 华北电力大学(保定) | Wavelet decomposition and reconstruction-based time sequence wind speed prediction method |
CN107316101A (en) * | 2017-06-02 | 2017-11-03 | 西南交通大学 | A kind of wind speed forecasting method selected in advance based on wavelet decomposition and component |
CN109657709A (en) * | 2018-12-06 | 2019-04-19 | 湖北博华自动化系统工程有限公司 | A kind of equipment fault prediction technique based on particle group optimizing support vector regression |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112762935A (en) * | 2020-12-23 | 2021-05-07 | 赫星(厦门)电子有限公司 | Positioning method and system based on hull attitude monitoring |
CN113942622A (en) * | 2021-11-19 | 2022-01-18 | 博迈科海洋工程股份有限公司 | Motion compensation method suitable for FPSO upper module lifting installation process |
CN113942622B (en) * | 2021-11-19 | 2023-11-07 | 博迈科海洋工程股份有限公司 | Motion compensation method suitable for FPSO upper module lifting and installing process |
CN114894043A (en) * | 2022-03-30 | 2022-08-12 | 北京航天飞腾装备技术有限责任公司 | Precise guidance ammunition attitude filtering method and system based on wavelet packet transformation |
CN114894043B (en) * | 2022-03-30 | 2023-12-22 | 北京航天飞腾装备技术有限责任公司 | Accurate guided ammunition attitude filtering method and system based on wavelet packet transformation |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110489717A (en) | Barge heave movement prediction technique during offshore platform floating support mounting | |
Peri et al. | Multidisciplinary design optimization of a naval surface combatant | |
CN111199270B (en) | Regional wave height forecasting method and terminal based on deep learning | |
CN112213771A (en) | Seismic wave impedance inversion method and device | |
CN107515994A (en) | A kind of more fidelity design optimization methods of the shell structure of Autonomous Underwater Vehicle | |
CN103926932A (en) | Intelligent ship moving posture decomposition field forecasting method | |
CN104900063A (en) | Short distance driving time prediction method | |
Borg et al. | Elastic deformations of floaters for offshore wind turbines: dynamic modelling and sectional load calculations | |
Chen et al. | A layer decomposition-recomposition framework for neuron pruning towards accurate lightweight networks | |
CN114663439A (en) | Remote sensing image land and sea segmentation method | |
Cheng et al. | Inverting Rayleigh surface wave velocities for crustal thickness in eastern Tibet and the western Yangtze craton based on deep learning neural networks | |
CN107944100B (en) | Linear design method for comprehensively producing umbilical cable | |
King et al. | Tidal current variability in the central Great Barrier Reef | |
Ormberg et al. | Integrated vessel motion and mooring analysis applied in hybrid model testing | |
CN116993821A (en) | Ship attitude real-time prediction method based on transducer-AdaRNN model | |
Huang et al. | A new quasi-dynamic method for the prediction of tendon tension of TLP platform | |
CN111062476A (en) | Water quality prediction method based on gated circulation unit network integration | |
Reyes-Casimiro et al. | Optimization of a production semi-submersible hull configuration system based on the hydrodynamic response for a location in Mexican waters of the Gulf of Mexico | |
Chen et al. | CFD simulation of wave energy converters in focused wave groups using overset mesh | |
Makarynskyy | Neural pattern recognition and prediction for wind wave data assimilation | |
CN104794346A (en) | Quantitative method for quickly recognizing influences of human activities on lake water bloom | |
Isnaini et al. | Real-time wave prediction for floating offshore wind turbine based on the kalman filter | |
Haji et al. | Cost Analysis of Wind and Uranium from Seawater Acquisition symBiotic Infrastructure using Shell Enclosures | |
CN112948969A (en) | Ship rolling prediction method based on LSTMC hybrid network | |
Hlophe | Efficient real-time deterministic wave prediction for marine renewable energy |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191122 |
|
RJ01 | Rejection of invention patent application after publication |