CN112541292B - Submarine cable burial depth estimation algorithm based on distributed optical fiber temperature measurement principle - Google Patents
Submarine cable burial depth estimation algorithm based on distributed optical fiber temperature measurement principle Download PDFInfo
- Publication number
- CN112541292B CN112541292B CN202011523132.4A CN202011523132A CN112541292B CN 112541292 B CN112541292 B CN 112541292B CN 202011523132 A CN202011523132 A CN 202011523132A CN 112541292 B CN112541292 B CN 112541292B
- Authority
- CN
- China
- Prior art keywords
- data
- model
- optical fiber
- factor
- distributed
- 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.)
- Active
Links
- 239000013307 optical fiber Substances 0.000 title claims abstract description 46
- 238000009933 burial Methods 0.000 title claims abstract description 43
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 32
- 238000009529 body temperature measurement Methods 0.000 title claims abstract description 18
- 238000012549 training Methods 0.000 claims abstract description 45
- 238000012545 processing Methods 0.000 claims abstract description 14
- 238000004364 calculation method Methods 0.000 claims abstract description 13
- 238000007405 data analysis Methods 0.000 claims abstract description 8
- 238000000034 method Methods 0.000 claims description 29
- 238000005070 sampling Methods 0.000 claims description 21
- 238000011478 gradient descent method Methods 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 13
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 11
- 230000006870 function Effects 0.000 claims description 8
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000013145 classification model Methods 0.000 claims description 6
- 238000011946 reduction process Methods 0.000 claims description 6
- 238000011161 development Methods 0.000 claims description 4
- 238000012821 model calculation Methods 0.000 claims description 4
- 238000012952 Resampling Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 claims description 3
- 230000008859 change Effects 0.000 description 5
- 239000000835 fiber Substances 0.000 description 5
- 230000002093 peripheral effect Effects 0.000 description 4
- 238000012546 transfer Methods 0.000 description 4
- 241000282414 Homo sapiens Species 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 3
- 238000003066 decision tree Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000010438 heat treatment Methods 0.000 description 2
- 230000000704 physical effect Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 229910052709 silver Inorganic materials 0.000 description 2
- 239000004332 silver Substances 0.000 description 2
- 230000001052 transient effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000004020 conductor Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000004134 energy conservation Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000013049 sediment Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01K—MEASURING TEMPERATURE; MEASURING QUANTITY OF HEAT; THERMALLY-SENSITIVE ELEMENTS NOT OTHERWISE PROVIDED FOR
- G01K11/00—Measuring temperature based upon physical or chemical changes not covered by groups G01K3/00, G01K5/00, G01K7/00 or G01K9/00
- G01K11/32—Measuring temperature based upon physical or chemical changes not covered by groups G01K3/00, G01K5/00, G01K7/00 or G01K9/00 using changes in transmittance, scattering or luminescence in optical fibres
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/27—Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/08—Thermal analysis or thermal optimisation
Abstract
The invention provides a submarine cable burial depth estimation algorithm based on a distributed optical fiber temperature measurement principle, which is an algorithm of an integrated learning framework based on bagging and acquired optical fiber temperature data, wherein the integrated learning framework comprises a data analysis module, a database IO module and a model training module; the data analysis module is used for estimating the relative temperature and storing the relative temperature into the distributed database to be used as a data source for subsequent calculation; the database IO module is used for accessing the original data and the preprocessed factor data; the model training module is used for training the model of the existing factor data and constructing a learning network by utilizing the correlation degree and the weight of the factor data; the invention can save labor force, reduce data processing time and improve data processing efficiency.
Description
Technical Field
The invention relates to the field of submarine cable buried depth estimation algorithms, in particular to a submarine cable buried depth estimation algorithm based on a distributed optical fiber temperature measurement principle.
Background
With the continuous development of economy and society, the construction of various offshore and submarine engineering projects is orderly progressed, but the difficulty and the danger degree of the human work in the underwater environment are higher compared with the land, particularly the marine environment is more complex, and the accident risk is more easily encountered; the currently used submarine burial methods are as follows: the finite element technology and the engineering analogy method are used for calculating the submarine burial depth by obtaining water conservancy parameter values through weather, hydrology and environmental conditions, but the dynamics calculation is complex to build an analysis model according to the submarine burial depth.
Disclosure of Invention
In order to solve the problems, the invention provides a submarine cable burial depth estimation algorithm based on a distributed optical fiber temperature measurement principle, which can save labor force, reduce data processing time and improve data processing efficiency.
In order to solve the technical problems, the embodiment of the invention provides a submarine cable burial depth estimation algorithm based on a distributed optical fiber temperature measurement principle, wherein the estimation algorithm is an algorithm of an integrated learning framework based on bagging and by collecting optical fiber temperature data, and the integrated learning framework comprises a data analysis module, a database IO module and a model training module;
the data analysis module is used for collecting an original binary file generated by an application program interface API of the temperature equipment on line, analyzing and generating original frequency data, performing curve fitting on frequency scattered points through a Lorentz formula, generating a fitted curve through a gradient descent method, finally comparing the estimated center frequency with a baseline frequency, estimating relative temperature, and storing the estimated relative temperature into a distributed database as a data source of subsequent calculation;
the database IO module is used for accessing the original data and the preprocessed factor data;
the model training module is used for training the model of the existing factor data and constructing a learning network by utilizing the correlation degree and the weight of the factor data.
The preprocessing factor data quantizes the correlation degree of each equipment attribute feature by using the correlation calculation rule set by the feature engineering through the existing original data in the database, and generates a series of model factors and weights of each factor, wherein the model factors are obtained based on the original data.
The factors and the weight data are simultaneously applied to background model calculation and backup in a database.
The model training module mainly comprises a classification engine and a regression engine, wherein the classification engine is used for training the model of the existing factor data, an integrated learning network is constructed by utilizing the correlation degree and the weight value of the factor data, the integrated learning network comprises 7 classification models, each model is combined in a bagging mode for each classification scene to generate a burial depth value, the regression engine takes a thermodynamic finite element analysis model as a kernel function, and the gradient descent method is used for carrying out distributed computation and combining with characteristic factor regression to estimate the corresponding burial depth and comparing with the burial depth value of the classification engine.
The training of each model is carried out simultaneously, data generated by each model are stored in respective tables of a database in a distributed mode, and the model training result is used for predicting the development trend and the movement of the cable state and the peripheral sea state.
The distributed computing unit of the gradient descent method adopts a distributed processing function module Hadoop platform to realize distributed processing on temperature data, a MapReduce mechanism in the Hadoop platform comprises a Map process and a reduced reduction process, and the Map process and the reduced reduction process effectively divide and recombine original data.
The gradient descent method distributed calculation comprises the following steps:
(one) obtaining combined silver sub-data of sampling points: the Map operation is used for converting the original factors into readable data formats in parallel, sampling points which are not repeated mutually are distributed to different servers, the sampling points are numbered according to sequence, factor data of the sampling points are preprocessed and feature engineering is carried out, obvious missing values of the missing values are removed, and combined factor matrix data with equal length and equal width are obtained;
(II) iterative operation of the combination factor data of each sampling point: each server reads out the combination factor data in the database in the Map stage, obtains an iteration value of model parameter estimation by using a gradient descent method, and returns the result in the Reduced process;
and (III) repeatedly executing Map and Reduce operations until the fitting residual value is not changed or the maximum iteration number is reached, ending the iteration, and outputting the estimated value of each parameter of the regression model combination factor.
In the model training process, factor data is divided into a plurality of small data sets, a plurality of models are learned and combined, the divided data sets are sampled with a put back through a Bootstrap method, and the distribution and the confidence interval of each factor are obtained, wherein the specific steps are as follows:
sampling a certain number of samples from the original samples by adopting a resampling method (with a put-back sampling);
secondly, calculating a desired statistic T according to the extracted samples;
repeating the steps for N times to obtain N statistics T;
and (IV) calculating the confidence interval of the statistic according to the N statistic.
The algorithm of the integrated learning framework based on bagging is to adopt a sampling-back method from the whole data set to obtain N data sets, learn a model on each data set, and finally generate a prediction result in a voting mode by utilizing the output of the N models, wherein the method comprises the following steps of:
extracting a training set from an original sample set: extracting N training samples from an original sample set by using a bootstrap method in each round, and carrying out k rounds of extraction to obtain k independent training sets;
(II) obtaining a model: obtaining a model by using one training set each time, and obtaining k models by k training sets;
(III) obtaining a classification result: and obtaining classification results by voting the obtained k models.
Wherein the factors include fiber optic temperature data, cable current capacity data, local air temperature data, cable burial depth data, and cable routing water depth data.
The technical scheme of the invention has the following beneficial effects:
1. in the invention, factors highly related to equipment attributes and peripheral sea conditions are extracted by utilizing characteristic engineering, so that the accuracy and the interpretability of the model are greatly improved;
2. according to the invention, the submarine cable burial depth is estimated by utilizing the temperature-based integrated learning algorithm, so that direct submarine operation of projects is avoided, labor force is reduced, and various costs are saved;
3. in the invention, a distributed computing platform is adopted to distribute large-scale temperature data IO and computation on different servers, thereby reducing data processing time and improving data processing efficiency.
Drawings
FIG. 1 is a schematic diagram of an integrated learning algorithm architecture for estimating submarine cable burial depth through optical fiber temperature measurement;
FIG. 2 is a schematic flow chart of regression model calculation by MapReduce;
FIG. 3 is a schematic flow chart of the integrated learning classification model trained by temperature data according to the present invention;
FIG. 4 is a schematic diagram of the structure of decision tree components in the framework of the integrated algorithm of the present invention.
Detailed Description
In order to make the technical problems, technical solutions and advantages to be solved by the present invention more apparent, the following detailed description will be given with reference to the accompanying drawings and specific embodiments, it being apparent that the described embodiments are only some of the embodiments of the present invention, but not all of the embodiments.
As shown in fig. 1, the invention provides a submarine cable burial depth estimation algorithm based on a distributed optical fiber temperature measurement principle, which is an algorithm of an integrated learning framework based on bagging and acquired optical fiber temperature data, wherein the integrated learning framework comprises a data analysis module, a database IO module and a model training module;
the data analysis module is used for collecting an original binary file generated by the temperature equipment application program interface API on line, analyzing and generating original frequency data, performing curve fitting on the frequency scattered points through a Lorentz formula, generating a fitted curve through a gradient descent method, finally comparing the estimated center frequency with a baseline frequency, estimating the relative temperature, and storing the relative temperature into a distributed database as a data source for subsequent calculation;
the database IO module is used for accessing the original data and the preprocessed factor data;
and the model training module is used for training the model of the existing factor data and constructing a learning network by utilizing the correlation degree and the weight of the factor data.
The preprocessing factor data quantizes the correlation degree of each equipment attribute feature by using the correlation calculation rule set by the feature engineering through the existing original data in the database, and generates a series of model factors and weight values of each factor, wherein the model factors are obtained based on the original data.
The factors and the weight data are simultaneously applied to background model calculation and backup in a database.
The model training module mainly comprises a classification engine and a regression engine, wherein the classification engine is used for training the models of the existing factor data, an integrated learning network is constructed by utilizing the correlation degree and the weight value of the factor data, the integrated learning network comprises 7 classification models, each model is combined in a bagging mode for each classification scene to generate a burial depth value, the regression engine takes a thermodynamic finite element analysis model as a kernel function, and the gradient descent method is used for carrying out distributed computation and combining with a characteristic factor regression to estimate the corresponding burial depth and comparing with the burial depth value of the classification engine.
The regression engine is intended to make a quantitative estimate of the state of sea cable burial depth. The engine uses a thermodynamic finite element analysis model as a kernel function, and carries out regression estimation on the burial depth by using a gradient descent method in combination with other data measured by equipment, so as to provide a quantitative result; the basic variable of the heat transfer process is temperature, which is a function of the geometric position in the object and time. According to Fourier heat transfer law and energy conservation law, a control equation of the heat transfer problem can be established. I.e. the transient temperature field u (x, y, t) of the two-dimensional object should satisfy the following equation:
where u denotes the temperature, kx, ky are the thermal conductivities of the object in the x, y directions, ρ is the density of the object (kg/m 3), c is the specific heat of the object (J/(kg. K)), and Q is the heat source density in the object (W/(m. K)). Since the burial depth is only a change in the y-direction, the two-dimensional heat conduction problem can be reduced to one dimension, i.e. the partial derivative of temperature in the x-direction is 0. To sum up, a transient temperature field equation of the submarine cable is established:
in addition, in a specific project environment, the heat transfer boundary is the interface between the sea bottom water and the sediment, and the equation satisfies the boundary condition:
where ny is the direction cosine of the boundary in the direction of the external normal of the surface, h represents the heat exchange coefficient of the object and the surrounding medium, and u0 represents the ambient temperature.
In the thermodynamic finite element analysis model, parameter estimation mainly relates to physical attribute factors of equipment and external environment factors. The physical properties of the equipment can be directly measured in a laboratory environment, and the external environment influence is basically a nonlinear combination of original factors such as local air temperature, water depth, burial depth and the like, so that the burial depth of the cable is reversely pushed out under the condition that the temperature of the optical fiber and other environmental factors are known. In addition, the model involves more nonlinear operation, and consumes more memory and calculation power.
The training of each model is carried out simultaneously, data generated by each model are stored in respective tables of a database in a distributed mode, and the model training results are used for predicting the development trend and the movement of the cable state and the peripheral sea state.
In the initial temperature-burial depth supervised learning algorithm training, the aim is to learn a stable classification model with better performance in all aspects, but only a plurality of weak classification models with better performance in certain periods can be obtained.
As shown in fig. 2, the gradient descent method distributed computing unit adopts a distributed processing function module Hadoop platform to implement distributed processing on temperature data, and a MapReduce mechanism in the Hadoop platform comprises a Map process and a reduced reduction process, wherein the Map process and the reduced reduction process are used for effectively dividing and recombining original data.
The gradient descent method distributed calculation steps are as follows:
(one) obtaining combined silver sub-data of sampling points: the Map operation is used for converting the original factors into readable data formats in parallel, sampling points which are not repeated mutually are distributed to different servers, the sampling points are numbered according to sequence, factor data of the sampling points are preprocessed and feature engineering is carried out, obvious missing values of the missing values are removed, and combined factor matrix data with equal length and equal width are obtained;
(II) iterative operation of the combination factor data of each sampling point: each server reads out the combination factor data in the database in the Map stage, obtains an iteration value of model parameter estimation by using a gradient descent method, and returns the result in the Reduced process;
and (III) repeatedly executing Map and Reduce operations until the fitting residual value is not changed or the maximum iteration number is reached, ending the iteration, and outputting the estimated value of each parameter of the regression model combination factor.
In the model training process, factor data are divided into a plurality of small data sets, a plurality of models are learned and combined, the divided data sets are sampled with a put back through a Bootstrap method, and the distribution and the confidence interval of each factor are obtained, wherein the specific steps are as follows:
sampling a certain number of samples from the original samples by adopting a resampling method (with a put-back sampling);
secondly, calculating a desired statistic T according to the extracted samples;
repeating the steps for N times to obtain N statistics T;
and (IV) calculating the confidence interval of the statistic according to the N statistic.
The algorithm of the integrated learning framework based on bagging is to adopt a sampling-back method from the whole data set to obtain N data sets, learn a model on each data set, and finally generate a prediction result in a voting mode by utilizing the output of the N models, wherein the method comprises the following steps of:
extracting a training set from an original sample set: extracting N training samples from an original sample set by using a bootstrap method in each round, and carrying out k rounds of extraction to obtain k independent training sets;
(II) obtaining a model: obtaining a model by using one training set each time, and obtaining k models by k training sets;
(III) obtaining a classification result: and obtaining classification results by voting the obtained k models.
As shown in fig. 3, the factors include fiber optic temperature data, cable current capacity data, local air temperature data, cable burial depth data, and cable routing water depth data, and are specifically as follows:
(1) Current-carrying capacity of the cable: the temperature of the conductor is mainly influenced by the external environment temperature and self heating, the self heating of the cable is mainly determined by the current-carrying capacity of the cable, and the self influence and the environment influence in the optical fiber temperature factor can be distinguished through the current-carrying capacity of the incoming cable;
(2) Maximum daily fiber temperature: the captured outside environment temperature only has data of daily granularity, so the optical fiber temperature also adopts the data of daily granularity; the maximum value of the daily temperature of the optical fiber reflects the limit of the daily temperature of the optical fiber, removes the current-carrying capacity of the cable, and mainly influences the maximum value of the daily temperature of the optical fiber, namely the air temperature of the external environment; one obvious fact is: the temperature of the optical fiber at the position with larger burial depth is influenced by the outside environment temperature less, and the temperature of the optical fiber at the position with smaller burial depth is influenced by the outside environment temperature more;
(3) Minimum daily temperature of the fiber: besides the maximum value of the daily temperature of the optical fiber, the minimum value of the daily temperature of the optical fiber is combined for investigation, and in the training of the model, the model automatically combines and derives other explicit indexes such as the fluctuation range of the daily temperature of the optical fiber, the daily temperature stability of the optical fiber and the like or implicit indexes which cannot be directly understood by some human beings through the maximum value and the minimum value of the daily temperature of the optical fiber, so that the judgment of the model is helped;
(4) Median daily fiber temperature: the maximum value and the minimum value of the daily temperature of the optical fiber show the change range of the daily temperature of the optical fiber, and the median of the daily temperature of the optical fiber shows the distribution state of the daily temperature change of the optical fiber, so that the change trend of the daily temperature of the optical fiber can be seen more clearly through the median of the daily temperature of the optical fiber;
(5) Cable depth: the devices in different positions and structures have different physical properties, and one factor similar to the buried depth state is the water depth state. The relative state difference of the cable and the environment is influenced by the water depth besides the burial depth. But the measurement of water depth has a more convenient and mature scheme relative to the depth of burial. The influence of water depth factors on the cable state can be eliminated by the water depth data, so that the estimation of the embedded depth is more accurate;
(6) Local daily maximum air temperature: one fact mentioned above is: the temperature of the optical fiber at the position with larger burial depth is less influenced by the outside environment air temperature, and the temperature of the optical fiber at the position with smaller burial depth is more influenced by the outside environment air temperature. The highest local daily air temperature reflects the influence of external environmental factors on the temperature of the optical fiber, so that the change of the burial depth of the cable is reflected on the side surface, and the theoretical optical fiber temperature threshold is expected by combining the current capacity of the cable and the highest local daily air temperature;
(7) Local daily air temperature minimum: in addition to the local daily air temperature maximum, it is also necessary to incorporate a local daily air temperature minimum to investigate the local daily air temperature variation. In the training of the model, the maximum value and the minimum value of the local daily temperature are automatically combined to derive other implicit indexes such as the local daily temperature fluctuation range, the local daily temperature stability and the like, which are explicit or cannot be directly understood by some human beings, so that the judgment of the model is facilitated.
As shown in fig. 4, in the decision tree component of ensemble learning, each factor plays a role in the final judgment of the model.
The working principle of the invention is as follows: .
1. In the invention, factors highly related to equipment attributes and peripheral sea conditions are extracted by utilizing characteristic engineering, so that the accuracy and the interpretability of the model are greatly improved;
2. according to the invention, the submarine cable burial depth is estimated by utilizing the temperature-based integrated learning algorithm, so that direct submarine operation of projects is avoided, labor force is reduced, and various costs are saved;
3. in the invention, a distributed computing platform is adopted to distribute large-scale temperature data IO and computation on different servers, thereby reducing data processing time and improving data processing efficiency.
While the preferred embodiments of the present invention have been described, the scope of the present invention is not limited thereto, and any person skilled in the art, who is skilled in the art, should make equivalents and modifications within the scope of the present invention according to the technical scheme and the inventive concept thereof.
Claims (10)
1. The submarine cable burial depth estimation algorithm based on the distributed optical fiber temperature measurement principle is characterized in that the estimation algorithm is an algorithm of an integrated learning framework based on bagging through optical fiber temperature data acquired, and the integrated learning framework comprises a data analysis module, a database IO module and a model training module;
the data analysis module is used for collecting an original binary file generated by an application program interface API of the temperature equipment on line, analyzing and generating original frequency data, performing curve fitting on frequency scattered points through a Lorentz formula, generating a fitted curve through a gradient descent method, finally comparing the estimated center frequency with a baseline frequency, estimating relative temperature, and storing the estimated relative temperature into a distributed database as a data source of subsequent calculation;
the database IO module is used for accessing the original data and the preprocessed factor data;
the model training module is used for training the model of the existing factor data and constructing a learning network by utilizing the correlation degree and the weight of the factor data.
2. The submarine cable buried depth estimation algorithm based on the distributed optical fiber temperature measurement principle according to claim 1, wherein the preprocessed factor data quantizes the correlation degree of each equipment attribute feature by using a correlation calculation rule set by feature engineering through existing original data in a database, and generates a series of model factors and weights of each factor, wherein the model factors are obtained based on the original data.
3. The submarine cable burial depth estimation algorithm based on the distributed optical fiber temperature measurement principle according to claim 2, wherein the factors and the weight data are simultaneously applied to background model calculation and backup in a database.
4. The submarine cable burial depth estimation algorithm based on the distributed optical fiber temperature measurement principle according to claim 1, wherein the model training module mainly comprises a classification engine and a regression engine, the classification engine is used for training models of existing factor data, an integrated learning network is constructed by using the correlation degree and the weight value of the factor data, the integrated learning network comprises 7 classification models, each model is combined in a bagging mode for each classification scene to generate a burial depth value, the regression engine takes a thermodynamic finite element analysis model as a kernel function, and the gradient descent method is used for carrying out distributed calculation and combining with characteristic factor regression to estimate corresponding burial depths, and the burial depth value is compared with the burial depth value of the classification engine.
5. The submarine cable burial depth estimation algorithm based on the distributed optical fiber temperature measurement principle according to claim 4, wherein training of the models is performed simultaneously, data generated by each model are stored in a table of a database in a distributed mode, and the result of model training is used for predicting the development trend and the movement of the cable state and the surrounding sea state.
6. The submarine cable buried depth estimation algorithm based on the distributed optical fiber temperature measurement principle according to claim 4, wherein the gradient descent method distributed computing unit adopts a distributed processing function module Hadoop platform to realize distributed processing on temperature data, and a MapReduce mechanism in the Hadoop platform comprises a Map process and a reduced reduction process, wherein the Map process and the reduced reduction process are used for effectively dividing and recombining original data.
7. The submarine cable buried depth estimation algorithm based on the distributed optical fiber temperature measurement principle according to claim 6, wherein the gradient descent method distributed calculation steps are as follows:
(one) obtaining combination factor data of sampling points: the Map operation is used for converting the original factors into readable data formats in parallel, sampling points which are not repeated mutually are distributed to different servers, the sampling points are numbered according to sequence, factor data of the sampling points are preprocessed and feature engineering is carried out, obvious missing values of the missing values are removed, and combined factor matrix data with equal length and equal width are obtained;
(II) iterative operation of the combination factor data of each sampling point: each server reads out the combination factor data in the database in the Map stage, obtains an iteration value of model parameter estimation by using a gradient descent method, and returns the result in the Reduced process;
and (III) repeatedly executing Map and Reduce operations until the fitting residual value is not changed or the maximum iteration number is reached, ending the iteration, and outputting the estimated value of each parameter of the regression model combination factor.
8. The submarine cable buried depth estimation algorithm based on the distributed optical fiber temperature measurement principle according to claim 4, wherein in the model training process, factor data is divided into a plurality of small data sets, a plurality of models are learned and combined, the divided data sets are sampled with a put back by a Bootstrap method, and the distribution and confidence intervals of each factor are obtained, and the specific steps are as follows:
sampling a certain number of samples from the original samples by adopting a resampling method;
secondly, calculating a desired statistic T according to the extracted samples;
repeating the steps for N times to obtain N statistics T;
and (IV) calculating the confidence interval of the statistic according to the N statistic.
9. The submarine cable buried depth estimation algorithm based on the distributed optical fiber temperature measurement principle according to claim 8, wherein the algorithm of the integrated learning framework based on bagging is to take a set of N data sets obtained by sampling back from the whole data set by using a bootstrap method, learn a model on each data set, and finally predict the result to be generated in a voting way by using the output of the N models, and the steps are as follows:
extracting a training set from an original sample set: extracting N training samples from an original sample set by using a bootstrap method in each round, and carrying out k rounds of extraction to obtain k independent training sets;
(II) obtaining a model: obtaining a model by using one training set each time, and obtaining k models by k training sets;
(III) obtaining a classification result: and obtaining classification results by voting the obtained k models.
10. A submarine cable burial depth estimation algorithm based on the distributed optical fiber temperature measurement principle according to claim 1, wherein the factors include optical fiber temperature data, cable current capacity data, local air temperature data, cable burial depth data, and cable routing water depth data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011523132.4A CN112541292B (en) | 2020-12-22 | 2020-12-22 | Submarine cable burial depth estimation algorithm based on distributed optical fiber temperature measurement principle |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011523132.4A CN112541292B (en) | 2020-12-22 | 2020-12-22 | Submarine cable burial depth estimation algorithm based on distributed optical fiber temperature measurement principle |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112541292A CN112541292A (en) | 2021-03-23 |
CN112541292B true CN112541292B (en) | 2023-12-12 |
Family
ID=75019507
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011523132.4A Active CN112541292B (en) | 2020-12-22 | 2020-12-22 | Submarine cable burial depth estimation algorithm based on distributed optical fiber temperature measurement principle |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112541292B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113776696A (en) * | 2021-09-17 | 2021-12-10 | 浙江师范大学 | Temperature sensitivity testing method based on seven-core photonic crystal fiber |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103424210A (en) * | 2013-08-21 | 2013-12-04 | 国家电网公司 | Long-term temperature forecasting method for power cable tunnel |
CN106294963A (en) * | 2016-08-04 | 2017-01-04 | 国网上海市电力公司 | Direct-buried cable carrying current calculation method |
KR20180112653A (en) * | 2017-03-30 | 2018-10-12 | 삼성전자주식회사 | Data learning server and method for generating and using thereof |
EP3514488A1 (en) * | 2018-01-23 | 2019-07-24 | Fluves NV | Method for monitoring a burial depth of a submarine power cable |
CN110083908A (en) * | 2019-04-19 | 2019-08-02 | 陕西科技大学 | Cable conductor temperature predicting method based on finite element analysis |
CN110135019A (en) * | 2019-04-26 | 2019-08-16 | 广东工业大学 | A kind of loss of power cable and core temperature prediction technique |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2002048923A1 (en) * | 2000-12-12 | 2002-06-20 | Chuo Hatsujo Kabushiki Kaisha | Method and apparatus for calculating wiring route of control cable |
CN108932480B (en) * | 2018-06-08 | 2022-03-15 | 电子科技大学 | Distributed optical fiber sensing signal feature learning and classifying method based on 1D-CNN |
-
2020
- 2020-12-22 CN CN202011523132.4A patent/CN112541292B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103424210A (en) * | 2013-08-21 | 2013-12-04 | 国家电网公司 | Long-term temperature forecasting method for power cable tunnel |
CN106294963A (en) * | 2016-08-04 | 2017-01-04 | 国网上海市电力公司 | Direct-buried cable carrying current calculation method |
KR20180112653A (en) * | 2017-03-30 | 2018-10-12 | 삼성전자주식회사 | Data learning server and method for generating and using thereof |
EP3514488A1 (en) * | 2018-01-23 | 2019-07-24 | Fluves NV | Method for monitoring a burial depth of a submarine power cable |
CN110083908A (en) * | 2019-04-19 | 2019-08-02 | 陕西科技大学 | Cable conductor temperature predicting method based on finite element analysis |
CN110135019A (en) * | 2019-04-26 | 2019-08-16 | 广东工业大学 | A kind of loss of power cable and core temperature prediction technique |
Also Published As
Publication number | Publication date |
---|---|
CN112541292A (en) | 2021-03-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110427654B (en) | Landslide prediction model construction method and system based on sensitive state | |
GB2547816A (en) | Actually-measured marine environment data assimilation method based on sequence recursive spare filtering three-dimensional variation | |
Song et al. | Integration of a statistical emulator approach with the SCE-UA method for parameter optimization of a hydrological model | |
Feigl et al. | Function space optimization: A symbolic regression method for estimating parameter transfer functions for hydrological models | |
Li et al. | Real-time flood simulations using CA model driven by dynamic observation data | |
CN116306318B (en) | Three-dimensional ocean thermal salt field forecasting method, system and equipment based on deep learning | |
Fisher et al. | Low rank updates in preconditioning the saddle point systems arising from data assimilation problems | |
CN112541292B (en) | Submarine cable burial depth estimation algorithm based on distributed optical fiber temperature measurement principle | |
Üneş et al. | Determination of groundwater level fluctuations by artificial neural networks | |
Alarie et al. | Snow water equivalent estimation using blackbox optimization | |
Shao et al. | Ocean reanalysis data‐driven deep learning forecast for sea surface multivariate in the South China Sea | |
Lei et al. | Temperature field reconstruction from the partial measurement data using the gappy proper orthogonal decomposition | |
Altaf et al. | A study of reduced‐order 4DVAR with a finite element shallow water model | |
Güntürkün | Sequential reconstruction of driving-forces from nonlinear nonstationary dynamics | |
CN114692529B (en) | CFD high-dimensional response uncertainty quantification method and device, and computer equipment | |
Shi et al. | Uncertainty quantification of contaminant transport and risk assessment with conditional stochastic collocation method | |
Khan et al. | Forecasting renewable energy for environmental resilience through computational intelligence | |
Abouali et al. | Stability and performance analysis of the Castillo-Grone mimetic operators in conjunction with RK3 time discretization in solving advective equations | |
Chen et al. | Control-oriented low-order approximation and reconstruction of yaw-excited wind turbine wake dynamics | |
Fang et al. | Prediction accuracy improvement of pressure pulsation signals of reversible pump‐turbine: A LSTM and VMD‐based optimization approach | |
Yuan et al. | A parallel sensitive area selection-based particle swarm optimization algorithm for fast solving CNOP | |
Lawal et al. | Assessing the Feasibility of Wave Energy Harvesting through Coastal Wave Modeling and Forecasting with Long Short-Term Memory Optimization using Univariate Time Series Approach | |
CN117421562B (en) | Ocean dissolved oxygen content space-time distribution prediction method, system, medium and equipment | |
Yin et al. | An Improved Finite Element Model Updating Method Based on the Singular Values of Frequency Response Functions | |
Pandhiani et al. | A hybrid model for monthly time series forecasting |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |