CN106650929B - Deep learning model based on recurrence plot and the application in the measurement of grease phase content - Google Patents
Deep learning model based on recurrence plot and the application in the measurement of grease phase content Download PDFInfo
- Publication number
- CN106650929B CN106650929B CN201610888490.2A CN201610888490A CN106650929B CN 106650929 B CN106650929 B CN 106650929B CN 201610888490 A CN201610888490 A CN 201610888490A CN 106650929 B CN106650929 B CN 106650929B
- Authority
- CN
- China
- Prior art keywords
- training
- recurrence plot
- layer
- phase
- boltzmann machine
- 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
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
- G06N3/088—Non-supervised learning, e.g. competitive learning
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/85—Investigating moving fluids or granular solids
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N27/00—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means
- G01N27/02—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating impedance
- G01N27/04—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating impedance by investigating resistance
- G01N27/06—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating impedance by investigating resistance of a liquid
- G01N27/08—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating impedance by investigating resistance of a liquid which is flowing continuously
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N2021/8405—Application to two-phase or mixed materials, e.g. gas dissolved in liquids
Landscapes
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Theoretical Computer Science (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Electrochemistry (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Image Analysis (AREA)
Abstract
A kind of deep learning model based on recurrence plot and the application in the measurement of grease phase content, recurrence plot is constructed based on four sector distributing triggers reorganization data: unitary time series is regarded as each channel data of four sector distributing triggers reorganizations, phase space reconfiguration is carried out to unitary time series, obtains sequence vector；Recurrence plot is defined for sequence vector；Deep learning model training: by recurrence plot, one-dimensional obtains depth confidence network training to generate model parameter, training obtains limited Boltzmann machine parameter using greedy layer-by-layer training algorithm；Using limited Boltzmann machine parameter as initial value, it can be used for the neural network parameter of Accurate classification for resetting the initial weight of deep neural network and finally obtaining.Vertical oil-water two-phase flow experiment is carried out using four sector distributing triggers reorganizations, the proportion of fixing oil phase and water phase changes oil and mutually tested with the flow of water phase.The present invention can obtain the deep learning model based on recurrence plot.
Description
Technical field
The present invention relates to a kind of deep learning models.It is obtained more particularly to one kind for four sector distributing triggers reorganizations
To multi-channel data building deep learning model based on recurrence plot and the application in the measurement of grease phase content.
Background technique
Oil-water two-phase flow is widely present in oil exploitation and transportation industry.In water-oil phase streaming system, the distribution of each phase
As time and space are constantly changing, different nowed formings, referred to as flow pattern are formd.The flow pattern of two phase flow is complicated more
Become, local flow information is difficult to accurately capture, so that there are many difficult points for the measurement of the two-phase flow parameters such as phase content.This is right
Many influences are caused in oil exploitation and technological transformation.Currently, the research for flow pattern mainly uses observation method, small baud
Sign analysis and fuzzy C-means clustering, fuzzy logic and genetic algorithm, Digital Image Processing algorithm etc..The measurement of phase content is adopted more
With conductance method, capacitance method, optical method and ray method etc..The annular conductivity sensor and double helix capacitor that traditional measurement uses pass
Sensor etc. is single channel sensor, is easily lost microcosmic local flow information.And distributed conductivity sensor and excitation follow
Ring stimulus sensor etc. can then acquire multi channel signals simultaneously, capture richer micro flow information, be flow pattern and phase
Research containing rate provides important technology support.
Neural network has obtained tremendous development as a kind of classifier in the past few decades.It is based primarily upon logarithm
According to feature learnt with realize classification.But most of past training pattern is shallow-layer, classifying quality is unable to reach
Very accurate stage.As the problem of gradient disappears in neural network is effectively solved, it is such as deep to build profound neural network
Degree confidence network is possibly realized.Deep learning model obtains significantly compared to shallow-layer network in feature extraction and classificatory performance
It is promoted.
Recurrence plot is an important and effective tool in Nonlinear Time Series Analysis, and is obtained extensively multi-field
Using especially for unstable, short time series, analytical effect is very significant.It may be implemented to unitary signal
Reconstruct the visualization of the dynamic characteristic of obtained trajectory of phase space.Different time sequence is disclosed from intuitive angle using picture
Between intrinsic characteristic.
Summary of the invention
The deep learning model that the technical problem to be solved by the invention is to provide a kind of based on recurrence plot and in grease phase
Containing the application in rate measurement.
The technical scheme adopted by the invention is that: a kind of foundation of the deep learning model based on recurrence plot, including it is as follows
Step:
1) recurrence plot is constructed based on four sector distributing triggers reorganization data, comprising:
(1) unitary time series u is regarded as each channel data of four sector distributing triggers reorganizationsk, to described
Unitary time series ukPhase space reconfiguration is carried out, a sequence vector is obtainedAny one vector in the sequence vector
It indicates are as follows:
Wherein, m is Embedded dimensions, is determined using wrong nearest neighbor method；τ is delay time, is determined using mutual information method；
(2) for unitary time series ukObtain a sequence vectorFor sequence vectorDefine recurrence plot:
Wherein, Ri,jIndicate any point in recursion matrix corresponding to recurrence plot, ε is threshold value, using 15% unitary
The standard deviation of time series is determined,Indicate the distance between two vectors；Indicate ifThen its value is 1, ifThen its value is 0；So, if Ri,jValue be 1, then in recurrence plot
In be black, if Ri,jValue be 0, then be in recurrence plot white；
2) deep learning model training, comprising:
(1) by recurrence plot obtained in step 1), after carrying out one-dimensional, as the depth confidence net in deep learning model
The input of network；
(2) using greedy layer-by-layer training algorithm, depth confidence network training is obtained to generate model parameter, described is greedy
Greedy layer-by-layer training is the training algorithm using unsupervised learning, and in training from the bottom to top, one limited Bohr of every two layers of composition is hereby
Graceful machine obtains limited Boltzmann machine parameter with contrast divergence algorithm training, then fixes the layer parameter, continue to upper one layer of structure
The limited Boltzmann machine made is trained, until top terminates；
(3) the limited Boltzmann machine parameter obtained using (2) step is initial value, for resetting the first of deep neural network
Beginning weight finally obtains using there is the method for supervision to carry out global training to whole network parameter and can be used for Accurate classification
Neural network parameter.
Depth confidence network described in step 2) is to imply the generative probabilistic model that stochastic variable is constituted by multilayer, by one
Fixed number purpose is limited Boltzmann machine and stacks, and then Down-Up carries out layer-by-layer pre-training；Stacking process is as follows: training one
After a Bernoulli Jacob-Bernoulli Jacob is limited Boltzmann machine, it is limited the activation probability of hidden unit as next layer of Bernoulli Jacob-Bernoulli Jacob
The input data of Boltzmann machine；Second layer Bernoulli Jacob-Bernoulli Jacob is limited the activation primitive of Boltzmann machine as third layer primary
Nu Li-Bernoulli Jacob is limited the visible input data of Boltzmann machine, later each layer and so on；The limited Boltzmann machine
It is that the theory based on layer-by-layer greedy learning strategy is constructed.
The successively training of greed described in step 2) is the training algorithm using unsupervised learning, i.e. training process does not need
Know classification belonging to training sample, is very easy to obtain a large amount of training sample in this way；In training from the bottom to top, every two layers of structure
At a limited Boltzmann machine, limited Boltzmann machine parameter is obtained with contrast divergence algorithm training, then fixes this layer ginseng
Number, the limited Boltzmann machine for continuing to construct upper one layer is trained, until top terminates.
The training of the overall situation described in (3) step of step 2) is the training algorithm using supervised learning, with band phase content
The training sample of label is trained, at the beginning of being the limited Boltzmann machine parameter as deep neural network obtained using (2) step
Beginning parameter, and training is finely adjusted with back-propagation algorithm.
The application of a kind of deep learning model based on recurrence plot in the measurement of water-oil phase flow containing rate, using four sectors
Distributing triggers reorganization carries out vertical oil-water two-phase flow experiment, and the proportion of fixing oil phase and water phase changes oil phase and water phase
Flow is tested；Experiment includes the following steps:
1) proportion of fixed water phase and oily phase, a certain amount of water is passed through into vertical ascent pipeline, then gradually to pipeline
In be passed through oily phase, after water-oil phase sufficiently merges and is gradually stable, using four sector distributing triggers reorganizations acquire multi-pass
Road measuring signal, and flow pattern is recorded with high-speed camera instrument；After one acquisition, change the flow of oil phase and water phase, by above-mentioned
Process continues to acquire, until operating condition designed under fixed proportion is all completed, then changes the proportion of oily phase and water phase, repeats
The above process completes the measurement under different water phases and oily matched, until all operating conditions of design are all measured；
2) it is based on the multi-channel measurement signal of step 1), four sectors distribution is based on according to above-mentioned for each channel signal
The method that conductivity sensor data construct recurrence plot constructs recurrence plot；
3) after the recurrence plot for obtaining all operating conditions, using the recurrence plot of a portion as the training of depth confidence network
Collection, test set of the recurrence plot of another part as depth confidence network, using recurrence plot as the input of depth confidence network, institute
It states depth confidence network Boltzmann machine is limited by multilayer and form, the input that each layer is limited Boltzmann machine is upper one layer defeated
Out, it is trained by the successively trained unsupervised weight to each layer of Boltzmann machine of mechanism；
4) with the weight of the weights initialisation deep neural network of each layer of limited Boltzmann machine after training；
5) recurrence plot as training set is inputted into deep neural network, using to the weight of deep neural network has supervision
Training be finely adjusted；
6) by the unsupervised learning and supervised learning to sample, the depth confidence network based on recurrence plot is obtained
Model；By the model realization in oilfield exploitation under unknown operating condition to effective measurement of phase content.
Deep learning model based on recurrence plot and the application in the measurement of grease phase content of the invention, by being fanned to four
The multichannel sensor data that area's distributing triggers reorganization obtains are visualized using the method for recurrence plot, and recurrence plot is made
For the input of deep learning model, allows model to be trained by the unsupervised and supervised learning to great amount of samples, obtain one
A deep learning model based on recurrence plot, and it is used for the measurement to oil-water two-phase flow flow parameter such as phase content.
Detailed description of the invention
Fig. 1 is the deep learning model schematic the present invention is based on recurrence plot.
Specific embodiment
Below with reference to embodiment and attached drawing to the deep learning model of the invention based on recurrence plot and in grease phase content
Application in measurement is described in detail.
Deep learning model based on recurrence plot and the application in the measurement of grease phase content of the invention, provides a kind of base
In recurrence plot deep learning model and be used in the measurement of water-oil phase flow containing rate, by four sector distributed electricals
The multi-channel data that derivative sensor obtains is visualized using the method for recurrence plot, using recurrence plot as deep learning model
Input, allows the model to be trained by the unsupervised and supervised learning to great amount of samples, obtains one based on recurrence plot
Deep learning model, and it is used for the measurement of oil-water two-phase flow flow parameter such as phase content.
As shown in Figure 1, the foundation of the deep learning model of the invention based on recurrence plot, includes the following steps:
1) recurrence plot is constructed based on four sector distributing triggers reorganization data, comprising:
(1) unitary time series u is regarded as each channel data of four sector distributing triggers reorganizationsk, to described
Unitary time series ukPhase space reconfiguration is carried out, a sequence vector is obtainedAny one vector in the sequence vector
It may be expressed as:
Wherein, m is Embedded dimensions, is determined using wrong nearest neighbor method；τ is delay time, is determined using mutual information method；
(2) for unitary time series ukObtain a sequence vectorFor sequence vectorDefine recurrence plot:
Wherein, Ri,jIndicate any point in recursion matrix corresponding to recurrence plot, ε is threshold value, using 15% unitary
The standard deviation of time series is determined,Indicate the distance between two vectors；Indicate ifThen its value is 1, ifThen its value is 0；So, if Ri,jValue be 1, then in recurrence plot
In be black, if Ri,jValue be 0, then be in recurrence plot white；
2) deep learning model training, comprising:
(1) by recurrence plot obtained in step 1), after carrying out one-dimensional, as the depth confidence net in deep learning model
The input of network；
The depth confidence network is to imply the generative probabilistic model that stochastic variable is constituted by multilayer, by certain amount
Limited Boltzmann machine (RBM) is stacked, and then Down-Up carries out layer-by-layer pre-training；Stacking process is as follows: training one
After Bernoulli Jacob-Bernoulli Jacob RBM, using the activation probability of hidden unit as next layer of Bernoulli Jacob-Bernoulli Jacob RBM input data；The
Two layers of Bernoulli Jacob-Bernoulli Jacob RBM activation primitive is later each as the visible input data of third layer Bernoulli Jacob-Bernoulli Jacob RBM
Layer and so on；The limited Boltzmann machine is that the theory based on layer-by-layer greedy learning strategy is constructed.
(2) using greedy layer-by-layer training algorithm, depth confidence network training is obtained to generate model parameter；
The greediness successively training is the training algorithm using unsupervised learning, i.e. training process requires no knowledge about training
Classification belonging to sample is very easy to obtain a large amount of training sample in this way；In training from the bottom to top, one is constituted every two layers
RBM obtains RBM parameter with contrast divergence algorithm training, then fixes the layer parameter, continues the RBM constructed to upper one layer and carry out
Training, until top terminates.
(3) the RBM parameter obtained using (2) step is initial value, for resetting the initial weight of deep neural network, uses
There is the method for supervision to carry out global training to whole network parameter, finally obtains the neural network ginseng that can be used for Accurate classification
Number；
The global training is the training algorithm using supervised learning, is carried out with the training sample with phase content label
Training is the RBM parameter that is obtained using (2) step as deep neural network initial parameter, and is carried out with back-propagation algorithm micro-
Adjust training.The layer-by-layer pre-training of greed due to have passed through the first step can well solve biography when carrying out global training
On system is easy to deep layer network training mode the problem of converging to local best points.
Application of the deep learning model based on recurrence plot of the invention in the measurement of water-oil phase flow containing rate, using four
Sector distributing triggers reorganization carries out vertical oil-water two-phase flow experiment, and the proportion of fixing oil phase and water phase changes oil phase and water
The flow of phase is tested；Experiment includes the following steps:
1) proportion of fixed water phase and oily phase, a certain amount of water is passed through into vertical ascent pipeline, then gradually to pipeline
In be passed through oily phase, after water-oil phase sufficiently merges and is gradually stable, using four sector distributing triggers reorganizations acquire multi-pass
Road measuring signal, and flow pattern is recorded with high-speed camera instrument；Sample frequency is 4000Hz, sampling time 30s.One acquisition terminates
Afterwards, change oil mutually with the flow of water phase, continue to acquire by the above process, until operating condition designed under fixed proportion is all complete
At, then change the proportion of oily phase and water phase, the measurement completed under different water phases and oily matched is repeated the above process, up to design
All operating conditions be all measured；
2) it is based on the multi-channel measurement signal of step 1), four sectors distribution is based on according to above-mentioned for each channel signal
The method that conductivity sensor data construct recurrence plot constructs recurrence plot；
3) after the recurrence plot for obtaining all operating conditions, using the recurrence plot of a portion as the training of depth confidence network
Collection, test set of the recurrence plot of another part as depth confidence network, using recurrence plot as the input of depth confidence network, institute
It states depth confidence network Boltzmann machine is limited by multilayer and form, the input that each layer is limited Boltzmann machine is upper one layer defeated
Out, it is trained by the successively trained unsupervised weight to each layer of Boltzmann machine of mechanism；
4) with the weight of the weights initialisation deep neural network of each layer of limited Boltzmann machine after training；
5) recurrence plot as training set is inputted into deep neural network, using to the weight of deep neural network has supervision
Training be finely adjusted；
6) by the unsupervised learning and supervised learning to sample, the depth confidence network based on recurrence plot is obtained
Model；By the model realization in oilfield exploitation under unknown operating condition to effective measurement of phase content.
Using the above method, the inherence of the sensor electric signal of the inherent texture structure reflection non-stationary of recurrence plot can be used
Nonlinear dynamic characteristic.Deep learning model then can realize that classification is distinguished by extracting the feature of recurrence plot and being trained
Know.Using recurrence plot as bridge, effective identification of the deep learning model to non-stationary electric signal is realized.
The multi channel signals obtained for four sector distributing triggers reorganization measurements construct recurrence plot, using recurrence plot as
Deep learning mode input obtains the depth of an achievable phase content measurement by the training and study to great amount of samples
Model is practised, realization is in oilfield exploitation to the accurate measurement of phase content under unknown operating condition.
Above to the description of the present invention and embodiment, it is not limited to which this, the description in embodiment is only reality of the invention
One of mode is applied, it is without departing from the spirit of the invention, any not inventively to design and the technical solution
Similar structure or embodiment, category protection scope of the present invention.
Claims (5)
1. a kind of foundation of the deep learning model based on recurrence plot, which comprises the steps of:
1) recurrence plot is constructed based on four sector distributing triggers reorganization data, comprising:
(1) unitary time series u is regarded as each channel data of four sector distributing triggers reorganizationsk, to the unitary
Time series ukPhase space reconfiguration is carried out, a sequence vector is obtainedAny one vector in the sequence vectorIt indicates
Are as follows:
Wherein, m is Embedded dimensions, is determined using wrong nearest neighbor method；τ is delay time, is determined using mutual information method；
(2) for unitary time series ukObtain a sequence vectorFor sequence vectorDefine recurrence plot:
Wherein, Ri,jIndicate any point in recursion matrix corresponding to recurrence plot, ε is threshold value, using 15% elementary time
The standard deviation of sequence is determined,Indicate the distance between two vectors；Indicate ifThen its value is 1, ifThen its value is 0；So, if Ri,jValue be 1, then in recurrence plot
In be black, if Ri,jValue be 0, then be in recurrence plot white；
2) deep learning model training, comprising:
(1) by recurrence plot obtained in step 1), after carrying out one-dimensional, as the depth confidence network in deep learning model
Input；
(2) using greedy layer-by-layer training algorithm, obtain generating model parameter to depth confidence network training, it is described it is greedy by
Layer training is the training algorithm using unsupervised learning, it is trained in from the bottom to top, one limited Boltzmann machine of every two layers of composition,
Limited Boltzmann machine parameter is obtained with contrast divergence algorithm training, then fixes the limited Boltzmann machine parameter, is continued pair
The limited Boltzmann machine of next construction is trained, until top terminates；
(3) the limited Boltzmann machine parameter obtained using (2) step is initial value, for resetting the initial power of deep neural network
Weight finally obtains the nerve that can be used for Accurate classification using there is the method for supervision to carry out global training to whole network parameter
Network parameter.
2. the foundation of the deep learning model according to claim 1 based on recurrence plot, which is characterized in that institute in step 2)
The depth confidence network stated is to imply the generative probabilistic model that stochastic variable is constituted by multilayer, by setting number limited Bohr hereby
Graceful machine stacks, and then Down-Up carries out layer-by-layer pre-training；Stacking process is as follows: training one Bernoulli Jacob-Bernoulli Jacob by
After limiting Boltzmann machine, the input of Boltzmann machine is limited using the activation probability of hidden unit as next layer of Bernoulli Jacob-Bernoulli Jacob
Data；The activation primitive that second layer Bernoulli Jacob-Bernoulli Jacob is limited Boltzmann machine is limited glass as third layer Bernoulli Jacob-Bernoulli Jacob
The visible input data of the graceful machine of Wurz, later each layer and so on；The limited Boltzmann machine is learned based on successively greedy
The theory for practising strategy is constructed.
3. the foundation of the deep learning model according to claim 1 based on recurrence plot, which is characterized in that institute in step 2)
The greediness successively training stated is the training algorithm using unsupervised learning, i.e. training process requires no knowledge about belonging to training sample
Classification is very easy to obtain a large amount of training sample in this way；In training from the bottom to top, one limited Boltzmann of every two layers of composition
Machine obtains limited Boltzmann machine parameter with contrast divergence algorithm training, then fixes the layer parameter, continues to construct upper one layer
Limited Boltzmann machine be trained, until top terminates.
4. the foundation of the deep learning model according to claim 1 based on recurrence plot, which is characterized in that the of step 2)
(3) training of the overall situation described in step is the training algorithm using supervised learning, is carried out with the training sample with phase content label
Training is the limited Boltzmann machine parameter that is obtained using (2) step as deep neural network initial parameter, and uses backpropagation
Algorithm is finely adjusted training.
5. a kind of deep learning model described in claim 1 based on recurrence plot answering in the measurement of water-oil phase flow containing rate
With, which is characterized in that vertical oil-water two-phase flow experiment, fixing oil phase and water phase are carried out using four sector distributing triggers reorganizations
Proportion, change oil and mutually tested with the flow of water phase；Experiment includes the following steps:
1) proportion of fixed water phase and oily phase, a certain amount of water is passed through into vertical ascent pipeline, is then gradually led into pipeline
Enter oily phase, after water-oil phase sufficiently merges and is gradually stable, is surveyed using four sector distributing triggers reorganizations acquisition multichannel
Signal is measured, and records flow pattern with high-speed camera instrument；After one acquisition, change the flow of oil phase and water phase, by the above process
Continue to acquire, until operating condition designed under fixed proportion is all completed, then changes the proportion of oily phase and water phase, repeat above-mentioned
Process completes the measurement under different water phases and oily matched, until all operating conditions of design are all measured；
2) it is based on the multi-channel measurement signal of step 1), four sector distribution conductances are based on according to above-mentioned for each channel signal
The method that sensing data constructs recurrence plot constructs recurrence plot；
3) after the recurrence plot for obtaining all operating conditions, using the recurrence plot of a portion as the training set of depth confidence network, separately
Test set of the recurrence plot of a part as depth confidence network, using recurrence plot as the input of depth confidence network, the depth
Degree confidence network is limited Boltzmann machine by multilayer and forms, and the input that each layer is limited Boltzmann machine is upper one layer of output,
It is trained by the successively trained unsupervised weight to each layer of Boltzmann machine of mechanism；
4) with the weight of the weights initialisation deep neural network of each layer of limited Boltzmann machine after training；
5) recurrence plot as training set is inputted into deep neural network, the instruction for having supervision is used to the weight of deep neural network
White silk is finely adjusted；
6) by the unsupervised learning and supervised learning to sample, the depth confidence network mould based on recurrence plot is obtained
Type；By the model realization in oilfield exploitation under unknown operating condition to effective measurement of phase content.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610888490.2A CN106650929B (en) | 2016-10-11 | 2016-10-11 | Deep learning model based on recurrence plot and the application in the measurement of grease phase content |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610888490.2A CN106650929B (en) | 2016-10-11 | 2016-10-11 | Deep learning model based on recurrence plot and the application in the measurement of grease phase content |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106650929A CN106650929A (en) | 2017-05-10 |
| CN106650929B true CN106650929B (en) | 2019-02-26 |
Family
ID=58855288
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610888490.2A Active CN106650929B (en) | 2016-10-11 | 2016-10-11 | Deep learning model based on recurrence plot and the application in the measurement of grease phase content |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106650929B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111479982B (en) * | 2017-11-15 | 2023-01-17 | 吉奥奎斯特系统公司 | In-situ operating system with filter |
| CN108304924B (en) * | 2017-12-21 | 2021-10-12 | 内蒙古工业大学 | Pipelined pre-training method for deep belief network |
| CN108445751B (en) * | 2018-02-28 | 2021-03-16 | 天津大学 | Multi-target SSVEP idea control method fusing recursive graph and deep learning and application |
| CN108897354B (en) * | 2018-07-13 | 2020-10-20 | 广西大学 | Aluminum smelting process hearth temperature prediction method based on deep belief network |
| CN109670539A (en) * | 2018-12-03 | 2019-04-23 | 中国石油化工股份有限公司 | A kind of silt particle layer detection method based on log deep learning |
| CN110243886A (en) * | 2019-07-09 | 2019-09-17 | 邓博洋 | A kind of low yield gas well mouth of oil well hydrated comples ion method based on nonlinear characteristic |
| CN110243885A (en) * | 2019-07-09 | 2019-09-17 | 东营智图数据科技有限公司 | A kind of low yield gas well mouth of oil well hydrated comples ion method based on time-frequency characteristics |
| CN110630256B (en) * | 2019-07-09 | 2022-12-02 | 吴晓南 | Low-gas-production oil well wellhead water content prediction system and method based on depth time memory network |
| CN111189882B (en) * | 2020-01-14 | 2022-03-08 | 天津大学 | Two-phase flow instantaneous phase holdup prediction method based on phase space topological causal effect |
| CN111638249B (en) * | 2020-05-31 | 2022-05-17 | 天津大学 | Water content measuring method based on deep learning and application of water content measuring method in oil well exploitation |
| CN114298278A (en) * | 2021-12-28 | 2022-04-08 | 河北工业大学 | Electric equipment performance prediction method based on pre-training model |
| CN117092445A (en) * | 2023-10-19 | 2023-11-21 | 盛隆电气集团有限公司 | Fault detection method and system of power distribution system based on big data |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2013012312A2 (en) * | 2011-07-19 | 2013-01-24 | Jin Hem Thong | Wave modification method and system thereof |
| CN103760197A (en) * | 2014-01-23 | 2014-04-30 | 天津大学 | Two-phase flow measuring system based on distributed conductivity sensor |
| CN103776876A (en) * | 2014-01-23 | 2014-05-07 | 天津大学 | Structural parameter optimization method of distributed conductivity sensor |
| CN104049000A (en) * | 2014-05-27 | 2014-09-17 | 天津大学 | Gas-liquid phase content measurement method based on modal migration complex network and verification method thereof |
| CN104112138A (en) * | 2013-12-17 | 2014-10-22 | 深圳市华尊科技有限公司 | Object color classification method and device |
| CN104408483A (en) * | 2014-12-08 | 2015-03-11 | 西安电子科技大学 | Deep neural network-based SAR texture image classification method |
| CN105004763A (en) * | 2015-06-10 | 2015-10-28 | 天津大学 | Insert-type four-sector arc-shaped wall conductivity sensor of oil-water two-phase flow |
-
2016
- 2016-10-11 CN CN201610888490.2A patent/CN106650929B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2013012312A2 (en) * | 2011-07-19 | 2013-01-24 | Jin Hem Thong | Wave modification method and system thereof |
| CN104112138A (en) * | 2013-12-17 | 2014-10-22 | 深圳市华尊科技有限公司 | Object color classification method and device |
| CN103760197A (en) * | 2014-01-23 | 2014-04-30 | 天津大学 | Two-phase flow measuring system based on distributed conductivity sensor |
| CN103776876A (en) * | 2014-01-23 | 2014-05-07 | 天津大学 | Structural parameter optimization method of distributed conductivity sensor |
| CN104049000A (en) * | 2014-05-27 | 2014-09-17 | 天津大学 | Gas-liquid phase content measurement method based on modal migration complex network and verification method thereof |
| CN104408483A (en) * | 2014-12-08 | 2015-03-11 | 西安电子科技大学 | Deep neural network-based SAR texture image classification method |
| CN105004763A (en) * | 2015-06-10 | 2015-10-28 | 天津大学 | Insert-type four-sector arc-shaped wall conductivity sensor of oil-water two-phase flow |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106650929A (en) | 2017-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106650929B (en) | Deep learning model based on recurrence plot and the application in the measurement of grease phase content | |
| Xie et al. | A deep-learning-based real-time detector for grape leaf diseases using improved convolutional neural networks | |
| CN106503800B (en) | Deep learning model based on complex network and the application in measuring signal analysis | |
| Kukreja et al. | A Deep Neural Network based disease detection scheme for Citrus fruits | |
| CN106874956B (en) | The construction method of image classification convolutional neural networks structure | |
| CN104102929B (en) | Hyperspectral remote sensing data classification method based on deep learning | |
| CN108182450A (en) | A kind of airborne Ground Penetrating Radar target identification method based on depth convolutional network | |
| CN106096561A (en) | Infrared pedestrian detection method based on image block degree of depth learning characteristic | |
| CN107945153A (en) | A kind of road surface crack detection method based on deep learning | |
| CN110232419A (en) | A kind of method of side slope rock category automatic identification | |
| CN111639587B (en) | Hyperspectral image classification method based on multi-scale spectrum space convolution neural network | |
| CN108052966A (en) | Remote sensing images scene based on convolutional neural networks automatically extracts and sorting technique | |
| CN109766835A (en) | The SAR target identification method of confrontation network is generated based on multi-parameters optimization | |
| CN109657610A (en) | A kind of land use change survey detection method of high-resolution multi-source Remote Sensing Images | |
| CN104021396A (en) | Hyperspectral remote sensing data classification method based on ensemble learning | |
| Wang et al. | When low rank representation based hyperspectral imagery classification meets segmented stacked denoising auto-encoder based spatial-spectral feature | |
| CN108445752A (en) | A kind of random weight Artificial neural network ensemble modeling method of adaptively selected depth characteristic | |
| CN106485325B (en) | Two phase flow multivariate information fusion method based on complex network and deep learning and application | |
| CN103886342A (en) | Hyperspectral image classification method based on spectrums and neighbourhood information dictionary learning | |
| CN105989336A (en) | Scene recognition method based on deconvolution deep network learning with weight | |
| CN104239902A (en) | Hyper-spectral image classification method based on non-local similarity and sparse coding | |
| CN112766161A (en) | Hyperspectral target detection method based on integrated constraint multi-example learning | |
| CN104809471A (en) | Hyperspectral image residual error fusion classification method based on space spectrum information | |
| Dong et al. | Progressive cascaded convolutional neural networks for single tree detection with google earth imagery | |
| Vaidya et al. | Leveraging YOLOv7 for plant disease detection |
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 | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210105 Address after: No.4, Keji Avenue, Daqiuzhuang Industrial Park, Jinghai District, Tianjin Patentee after: TIANJIN FURUILONG METAL PRODUCTS Co.,Ltd. Address before: 300072 Tianjin City, Nankai District Wei Jin Road No. 92 Patentee before: Tianjin University |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210310 Address after: No.2301, building 1, Vanke times center, building 1, Huike building, intersection of Anshan West Road and Baidi Road, Nankai District, Tianjin Patentee after: Junsheng (Tianjin) Technology Development Co.,Ltd. Address before: No.4, Keji Avenue, Daqiuzhuang Industrial Park, Jinghai District, Tianjin Patentee before: TIANJIN FURUILONG METAL PRODUCTS Co.,Ltd. |