CN112926767A - Annular fog flow gas phase apparent flow velocity prediction method based on particle swarm BP neural network - Google Patents
Annular fog flow gas phase apparent flow velocity prediction method based on particle swarm BP neural network Download PDFInfo
- Publication number
- CN112926767A CN112926767A CN202110109081.9A CN202110109081A CN112926767A CN 112926767 A CN112926767 A CN 112926767A CN 202110109081 A CN202110109081 A CN 202110109081A CN 112926767 A CN112926767 A CN 112926767A
- Authority
- CN
- China
- Prior art keywords
- neural network
- gas phase
- prediction
- flow velocity
- prediction model
- 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
- 239000002245 particle Substances 0.000 title claims abstract description 39
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 28
- 238000012549 training Methods 0.000 claims abstract description 24
- 238000001514 detection method Methods 0.000 claims abstract description 13
- 238000003062 neural network model Methods 0.000 claims abstract description 12
- 238000012360 testing method Methods 0.000 claims abstract description 12
- 230000008569 process Effects 0.000 claims abstract description 8
- 239000007788 liquid Substances 0.000 claims description 19
- 230000003287 optical effect Effects 0.000 claims description 5
- 239000012530 fluid Substances 0.000 claims description 4
- 238000010183 spectrum analysis Methods 0.000 claims description 4
- 238000007619 statistical method Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 238000007689 inspection Methods 0.000 claims description 2
- 239000012071 phase Substances 0.000 description 26
- 238000004422 calculation algorithm Methods 0.000 description 10
- 238000005259 measurement Methods 0.000 description 6
- 230000005514 two-phase flow Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000002834 transmittance Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000000889 atomisation Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 239000007791 liquid phase Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000010297 mechanical methods and process Methods 0.000 description 1
- 238000005272 metallurgy Methods 0.000 description 1
- 239000003595 mist Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012634 optical imaging Methods 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 238000003825 pressing Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 239000012808 vapor phase Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
- G06F30/28—Design optimisation, verification or simulation using fluid dynamics, e.g. using Navier-Stokes equations or computational fluid dynamics [CFD]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Abstract
The invention relates to a particle swarm BP neural network-based method for predicting the apparent flow velocity of an annular atomized flow gas phase, which comprises the following steps: (1) the prediction of the gas phase superficial flow rate is output based on the prediction model. The steps of establishing the prediction model are as follows: determining an input layer variable of the prediction model, and taking the gas phase apparent flow velocity as an output layer variable; acquiring variable data required by a prediction model; determining a BP neural network topological structure, and establishing a BP neural network model; training the model through training set data, bringing the test sample into a BP neural network model after training convergence, and calculating the gas phase apparent flow velocity UgsThe prediction error of the method, the number of hidden layers is gradually increased, the training process is repeated,stopping training until a prediction error occurs, determining the number of hidden layers, and establishing a prediction model based on a PSO-BP neural network; (2) and substituting the data of the input layer parameters detected by the detection device into the established prediction model to predict the gas phase apparent flow velocity.
Description
Technical Field
The invention belongs to the technical field of multiphase flow parameter detection, and particularly relates to a prediction method of gas-phase apparent flow velocity of gas-liquid two-phase flow.
Background
The gas-liquid two-phase flow widely exists in the industrial processes of chemical industry, metallurgy, petroleum, energy and the like, has important effects on industrial production and scientific research, brings many safety and economic problems, and has important significance on the accurate detection of flow characteristic parameters of the annular atomized flow as an important two-phase flow pattern[1]. In order to obtain the flow parameters of each phase, the traditional detection method is to separate the mixed multiphase flow on an operation platform and measure the multiphase flow in a split-phase manner by using a conventional measurement means, however, in an actual industrial field, a large-scale separator has high requirements on space and high energy consumption, and is not easy to realize. Therefore, the method has practical significance for establishing the gas phase apparent flow velocity prediction model by combining measurable parameters in a laboratory.
In two-phase flow, fluid flowing through the flow-resisting generator forms vortex at the downstream, and vortex signals are influenced by flow parameters such as pressure, flow and liquid phase content to reflect the flow characteristics to a certain extent. The characteristic parameters of the liquid drops comprise flow field information and reflect the characteristics of the internal flow field[2]The measurement of the characteristic parameters of the liquid drops mainly comprises a mechanical method, an electronic method and an optical method, the optical method is widely applied due to the characteristics of non-invasiveness, simple structure and high precision, and the digital image processing technology is combined to extract the parameters of the liquid drops such as the particle size, the speed and the concentration. The characteristic parameters are combined with two-phase working conditions to comprehensively reflect the flow characteristics, and a relation model can be established for the characteristic parameters to realize prediction.
The common prediction algorithm mainly comprises a regression model method, a grey prediction method, a time sequence method, a neural network and the like, but the traditional nonlinear system prediction has high dependence on qualitative relations among variables and low prediction precision, and in comparison, the neural network can complete nonlinear modeling on the premise of not solving the relation among input and output variables, so that the demand of predicting the gas phase flow rate is met. The BP neural network is a neural network model with wide application, has a simple structure and good self-learning capability, can approach any function with high precision, and has the defects of falling into a local minimum value, poor qualitative property, low efficiency and the like when the structure is complex. The Particle Swarm Optimization (PSO) algorithm is a machine learning algorithm with strong global optimization capability, and is introduced into the BP neural network model, so that the convergence rate of the traditional BP neural network algorithm can be increased, and the efficiency of flow prediction is improved. The flow parameters can be measured by combining a PSO-BP neural network model with an experiment, so that the gas phase flow rate of the annular atomized flow can be effectively predicted.
Reference to the literature
[1] The detection technology of the process parameters of the Tan super, Dongfeng and multiphase flow is summarized in the technical paper of automated chemistry, 2013,39(11):1923-1932.
[2].Sun H,Luo Y,Ding H,et al.Research on Droplet Properties in Atomization using Optical Imaging Measurements[C]//2020IEEE International Instrumentation and Measurement Technology Conference(I2MTC).IEEE,2020.
Disclosure of Invention
The invention provides a method for predicting the gas phase apparent flow velocity in the annular atomized flow, which has high response speed and high accuracy. In a two-phase environment, the indicating flow rate measured by the vortex shedding flowmeter and the real gas phase apparent flow rate come in and go out, and therefore a prediction model is established by combining the characteristics of vortex shedding signals and the measurement results of two-phase working conditions and liquid drop parameters, and the gas phase apparent flow rate of the annular atomized flow is predicted. A Particle Swarm Optimization (PSO) algorithm is introduced into a BP neural network model, so that the problems of easiness in falling into local minimum, poor stability, low efficiency and the like of the traditional BP neural network model are solved well. The technical scheme is as follows:
a method for predicting the gas phase apparent flow velocity of an annular atomized flow based on a particle swarm BP neural network comprises the following steps:
(1) the prediction of the gas phase superficial flow rate is output based on the prediction model. The steps of establishing the prediction model are as follows:
1) pressing the working conditionForce P, operating temperature T, and Sortel mean droplet diameter dpThe arithmetic mean velocity v of the dropletspDroplet concentration cpVortex street dominant frequency fvsAmplitude of eddy signal AvsAs input layer variable of the prediction model, gas phase apparent flow velocity UgsAs an output layer variable;
2) collecting variable data required by a prediction model by using a detection device, and selecting multiple groups of data as training samples, inspection samples and test samples;
3) determining a BP neural network topological structure, wherein the BP neural network topological structure comprises an input layer hidden layer and an output layer, setting the initial layer number k of the hidden layer, and setting the prediction error precision e, wherein the node number M of the hidden layer is determined by a formula (1):
wherein M is the number of hidden layer nodes, I is the number of input layer nodes, O is the number of output layer nodes, and a is a constant between 1 and 10;
4) modeling the connection weight and threshold of each layer in the BP neural network as particles for encoding, setting the prediction error precision e and setting the maximum iteration number genmax；
5) Calculating the particle fitness value FiThe calculation formula of the fitness value is as follows (2):
wherein y isjRepresents the observed value, t, of sample jjRepresenting the predicted value of the sample j, wherein N is the number of samples;
6) and (3) comparing the fitness value to update the individual optimal position pbest and the group optimal position gbest, and updating the position and the speed according to the formulas (3) and (4):
vi＝vi+c1rand1(pbest-xi)+c2rand2(gbest-xi) (3)
xi＝xi+vi (4)
wherein i is 1,2 … N, N is the number of particles, viIs the velocity of the particles, rand1、rand2Are random numbers distributed in 0-1; x is the number ofiIs the current position of the particle, pbest is the individual optimal position, gbest is the group optimal position, c1And c2Is a learning factor;
7) repeating the steps 5) and 6), and continuously updating the individual optimal position and the group optimal position until the maximum iteration number gen is reachedmaxFinally, the optimal position of the population particles is obtained, the optimal position is decoded to generate an optimal solution which is used as the globally optimal connection weight and threshold of the BP neural network, and a BP neural network model is established;
8) training the model through training set data, bringing the test sample into a BP neural network model after training convergence, and calculating the gas phase apparent flow velocity UgsPrediction error of | tj-yjGradually increasing the number of hidden layers, and repeating the training process until the prediction error is tj-yjStopping training when the absolute value is less than e, determining the number of hidden layers, and establishing a prediction model based on a PSO-BP neural network;
(2) substituting the data of the input layer parameters detected by the detection device into the established prediction model to obtain the gas phase apparent flow velocity UgsAnd (6) performing prediction.
Further, the method is characterized in that the step (2) is as follows:
1) measuring the pressure P and the temperature T of fluid in the pipeline through a pressure sensor and a temperature sensor;
2) the piezoelectric sensor is arranged at the downstream of the vortex generating body to detect the lift force, and the vortex main frequency f is obtained by performing frequency spectrum analysis on the output signal of the piezoelectric sensorvsSum vortex signal amplitude Avs；
3) Collecting a liquid drop image by using an optical measuring device, extracting characteristic parameters based on a single-frame single exposure method, and obtaining the mean diameter d of the Sontar through the statistical analysis of the imagepAnd arithmetic mean velocity vpObtaining drop concentration c in combination with depth of fieldp；
4) Take the above data asThe input layer variable is brought into a prediction model, and the gas phase apparent flow velocity U is outputgsIs used as a prediction value.
Drawings
FIG. 1: schematic diagram of detection device
FIG. 2: algorithm structure diagram
FIG. 3: PSO-BP neural network prediction model establishment flow chart
Detailed Description
In order to further understand the features and technical means of the present invention and achieve specific objects and functions, the following detailed description of the embodiments of the present invention is provided with reference to the accompanying drawings.
The establishment of the prediction model requires detection of variables of an input layer and an output layer, and a large amount of data is extracted to form a training sample, a test sample and a test sample. The detection device is shown in fig. 1 and mainly comprises a CCD industrial camera 1, an LED light source 2, an acquisition window 3, a pressure sensor 4, a piezoelectric sensor 5 and a temperature sensor 6. In a two-phase environment, the pressure x in the pipeline is measured by a pressure sensor and a temperature sensor1With temperature x2The piezoelectric sensor is adopted to collect vortex signals, and vortex street dominant frequency x in two phases is extracted through frequency spectrum analysis3Amplitude x of eddy signal4。
The liquid drop parameter measuring device mainly comprises a CCD industrial camera, an LED light source and an acquisition window, wherein the industrial camera is provided with a telecentric lens, so that the imaging magnification does not change along with the change of an object distance, and the window adopts an embedded hollow cylindrical tube to isolate a liquid film. The method comprises the steps that a camera collects liquid drop images through a high-transmittance window, preprocessing is carried out on the liquid drop images to improve the accuracy of particle identification and characteristic parameter extraction, and then a single-frame single exposure method is combined to carry out Sortell average diameter x on liquid drops5Average speed x6And the droplet concentration x7Carrying out measurement; vapor phase flow velocity y of multi-parameter adjustable mist flow experimental system designed by patent 201810644726.71And (6) detecting.
The prediction model establishment process is as follows:
1) model variables were selected as in table 1. Selecting working condition pressure P, working condition temperature T and vortex street dominant frequency fvsAmplitude of eddy signal AvsSortel mean diameter of the droplets dpThe arithmetic mean velocity v of the dropletspDroplet concentration cpAs input layer variable, gas phase apparent flow rate UsgAs an output layer variable. The input layer variable comprises working condition and internal flow parameters, and the gas phase flow rate can be comprehensively and comprehensively predicted. Extracting a large amount of detected data to form a training sample, a test sample and a test sample;
TABLE 1 model variable selection
2) The BP neural network algorithm adopts a gradient descent algorithm, a learning process is carried out in training data, the minimum square of an output error is taken as a target, error back propagation is adopted, and a network node weight and a threshold are trained. The transfer function adopts an S function, and is shown as formula (5):
3) determining the BP neural network structure. The prediction model comprises an input layer, a hidden layer and an output layer, the initial number k of hidden layer is set, a prediction error e is set, and the number M of nodes of the hidden layer is designed by adopting a formula (6):
wherein M is the number of hidden layer nodes, I is the number of input layer nodes 7, O is the number of output layer nodes 1, and a is a constant between 1 and 10;
4) introducing a particle swarm optimization algorithm into the BP neural network. And coding by taking the weight and the threshold of each layer of the BP neural network as a particle swarm. The motion characteristics of the particles are determined by the positions,The method is characterized by speed and a fitness value, the position represents a potential optimal solution, the speed controls the direction and the displacement of the movement of the particles, the size of the fitness value represents the advantages and disadvantages of the positions of the particles, and the maximum iteration number gen is setmax；
5) The particle fitness value is calculated according to equation (7):
wherein y isjRepresents the observed value, t, of sample jjRepresenting the predicted value of the sample j, wherein N is the number of samples;
6) and judging the optimal position by comparing the fitness values of the particles. The fitness value F is compared with the fitness value of the individual optimal position pbestpbMaking a comparison if Fi＜FpbThen F ispb＝FiReplacing the position of the particle with pbest, and then matching the position of the particle with the fitness value F of the optimal group position gbestgbMaking a comparison if Fi＜FgbThen gb isi＝FiReplacing the particle with gbest; continuously updating the speed and position of the particles on the basis of the speed and position; updating the speed and the position of the user according to the equations (8) and (9):
vi＝vi+c1rand1(pbest-xi)+c2rand2(gbest-xi) (8)
xi＝xi+vi (9)
wherein i is 1,2 … N, N is the number of particles, viIs the velocity of the particles, rand1、rand2Are random numbers distributed in 0-1; x is the number ofiIs the current position of the particle, pbest is the individual optimal position, gbest is the group optimal position, c1And c2Is a learning factor, usually c1＝c 22, the maximum value V of speed and position is specifiedmax、XmaxIf v isi＞VmaxThen v isi＝Vmax；
7) Repeating the steps 5) and 6), continuously updating the speed and the position of the particles,until the number of iterations reaches the maximum number of iterations genmaxDecoding the position of the particle with the minimum fitness value into an optimal solution to serve as the optimal connection weight and the threshold of the BP neural network, and establishing a BP neural network model;
8) the model is learned and trained through the training sample, after the training is converged, the testing sample is substituted into the network model, and the gas phase apparent flow velocity U is calculatedgsPrediction error of | tj-yjGradually increasing the number of hidden layers, and repeating the training process until the prediction error is tj-yjStopping training when the absolute value is less than e, finally determining the number of hidden layers, and establishing a PSO-BP neural network algorithm model;
9) and testing the model by using the test sample, and verifying the effectiveness of the prediction model.
After the model is established, the gas phase apparent flow velocity can be predicted by detecting the relevant variable data of the input layer, and the specific steps are as follows:
1) and collecting parameters of an input layer. Measuring the pressure P and the temperature T of fluid in the pipeline by using a pressure sensor and a temperature sensor;
2) the lift force is detected by adopting a piezoelectric sensor, the generated charge output is amplified and filtered to be converted into an approximate sine signal, and then the vortex main frequency f is obtained through spectral analysisvsSum vortex signal amplitude Avs；
3) Measuring the parameters of the liquid drops by using an optical measuring device and combining with a single-frame single exposure method, collecting two-phase flow images by a camera through a window with high transmittance in the liquid drop detection process, generating a smear on the particle images of the liquid drops by using long exposure, preprocessing the images, extracting the particle diameters and the speeds of the liquid drops by combining with the single-frame single exposure method, and performing statistical analysis on the parameter data of a large number of images to obtain the mean diameter d of the liquid drops by the SotelpAnd the average velocity vpThe drop concentration c can be obtained in combination with the depth of fieldp；
4) The parameters are brought into a prediction model to be used as input layer variables, and then the gas phase apparent flow velocity U can be outputgsThe prediction is effective.
The above-mentioned embodiments are intended to explain the technical solutions of the theoretical innovations and embodiments of the present invention in detail, and the present invention is not limited to the above-mentioned implementation routines, but it should be understood that modifications and substitutions based on the above-mentioned principles and spirit by those skilled in the art are included in the protection scope of the present invention.
Claims (2)
1. A method for predicting the gas phase apparent flow velocity of an annular atomized flow based on a particle swarm BP neural network comprises the following steps:
(1) the prediction of the gas phase superficial flow rate is output based on the prediction model. The steps of establishing the prediction model are as follows:
1) the working condition pressure P, the working condition temperature T and the Sortel average diameter d of the liquid droppThe arithmetic mean velocity v of the dropletspDroplet concentration cpVortex street dominant frequency fvsAmplitude of eddy signal AvsAs input layer variable of the prediction model, gas phase apparent flow velocity UgsAs an output layer variable;
2) collecting variable data required by a prediction model by using a detection device, and selecting multiple groups of data as training samples, inspection samples and test samples;
3) determining a BP neural network topological structure, wherein the BP neural network topological structure comprises an input layer hidden layer and an output layer, setting the initial layer number k of the hidden layer, and setting the prediction error precision e, wherein the node number M of the hidden layer is determined by a formula (1):
wherein M is the number of hidden layer nodes, I is the number of input layer nodes, O is the number of output layer nodes, and a is a constant between 1 and 10;
4) modeling the connection weight and threshold of each layer in the BP neural network as particles for encoding, setting the prediction error precision e and setting the maximum iteration number genmax；
5) Calculating the particle fitness value FiThe calculation formula of the fitness value is as follows (2):
wherein y isjRepresents the observed value, t, of sample jjRepresenting the predicted value of the sample j, wherein N is the number of samples;
6) and (3) comparing the fitness value to update the individual optimal position pbest and the group optimal position gbest, and updating the position and the speed according to the formulas (3) and (4):
vi＝vi+c1rand1(pbest-xi)+c2rand2(gbest-xi) (3)
xi＝xi+vi (4)
wherein i is 1,2 … N, N is the number of particles, viIs the velocity of the particles, rand1、rand2Are random numbers distributed in 0-1; x is the number ofiIs the current position of the particle, pbest is the individual optimal position, gbest is the group optimal position, c1And c2Is a learning factor;
7) repeating the steps 5) and 6), and continuously updating the individual optimal position and the group optimal position until the maximum iteration number gen is reachedmaxFinally, the optimal position of the population particles is obtained, the optimal position is decoded to generate an optimal solution which is used as the globally optimal connection weight and threshold of the BP neural network, and a BP neural network model is established;
8) training the model through training set data, bringing the test sample into a BP neural network model after training convergence, and calculating the gas phase apparent flow velocity UgsPrediction error of | tj-yjGradually increasing the number of hidden layers, and repeating the training process until the prediction error is tj-yjStopping training when the absolute value is less than e, determining the number of hidden layers, and establishing a prediction model based on a PSO-BP neural network;
(2) substituting the data of the input layer parameters detected by the detection device into the established prediction model to obtain the gas phase apparent flow velocity UgsAnd (6) performing prediction.
2. The method for predicting the apparent gas flow velocity of an atomized circular stream according to claim 1, wherein the step (2) is as follows:
1) measuring the pressure P and the temperature T of fluid in the pipeline through a pressure sensor and a temperature sensor;
2) the piezoelectric sensor is arranged at the downstream of the vortex generating body to detect the lift force, and the vortex main frequency f is obtained by performing frequency spectrum analysis on the output signal of the piezoelectric sensorvsSum vortex signal amplitude Avs；
3) Collecting a liquid drop image by using an optical measuring device, extracting characteristic parameters based on a single-frame single exposure method, and obtaining the mean diameter d of the Sontar through the statistical analysis of the imagepAnd arithmetic mean velocity vpObtaining drop concentration c in combination with depth of fieldp；
4) The data are taken as input layer variables and are brought into a prediction model, and the gas phase apparent flow velocity U is outputgsIs used as a prediction value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110109081.9A CN112926767A (en) | 2021-01-27 | 2021-01-27 | Annular fog flow gas phase apparent flow velocity prediction method based on particle swarm BP neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110109081.9A CN112926767A (en) | 2021-01-27 | 2021-01-27 | Annular fog flow gas phase apparent flow velocity prediction method based on particle swarm BP neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112926767A true CN112926767A (en) | 2021-06-08 |
Family
ID=76166780
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110109081.9A Pending CN112926767A (en) | 2021-01-27 | 2021-01-27 | Annular fog flow gas phase apparent flow velocity prediction method based on particle swarm BP neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112926767A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111649779A (en) * | 2020-05-31 | 2020-09-11 | 天津大学 | Oil well oil content and total flow rate measuring method based on dense neural network and application |
| CN113392961A (en) * | 2021-08-17 | 2021-09-14 | 中国海洋大学 | Method for extracting mesoscale eddy track stable sequence and predicting cyclic neural network |
| WO2023005302A1 (en) * | 2021-07-26 | 2023-02-02 | 东南大学 | Method for predicting biomass coke gasification reaction characteristics |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050081643A1 (en) * | 2003-02-10 | 2005-04-21 | Mattar Wade M. | Multiphase coriolis flowmeter |
| CN101688856A (en) * | 2007-04-20 | 2010-03-31 | 因万西斯系统股份有限公司 | Wet gas measurement |
| CN106920008A (en) * | 2017-02-28 | 2017-07-04 | 山东大学 | A kind of wind power forecasting method based on Modified particle swarm optimization BP neural network |
| CN108469281A (en) * | 2018-01-18 | 2018-08-31 | 天津大学 | Two-phase Research on vortex signal processing based on EMD and Spectrum Correction |
| CN110186522A (en) * | 2019-05-31 | 2019-08-30 | 天津大学 | Reading compensation and flow-measuring method are crossed in conjunction with the moisture of vortex street amplitude characteristic |
| CN111222229A (en) * | 2019-12-27 | 2020-06-02 | 清华大学深圳国际研究生院 | Method for constructing instantaneous flow measurement model in gas-liquid two-phase flow dynamic flow process |
| CN111370080A (en) * | 2020-03-05 | 2020-07-03 | 中国工程物理研究院上海激光等离子体研究所 | Radiation temperature inversion method based on artificial neural network algorithm |
| CN111380581A (en) * | 2020-02-17 | 2020-07-07 | 天津大学 | Image method-based fog annular flow phase separation flow parameter measurement method |
-
2021
- 2021-01-27 CN CN202110109081.9A patent/CN112926767A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050081643A1 (en) * | 2003-02-10 | 2005-04-21 | Mattar Wade M. | Multiphase coriolis flowmeter |
| CN101688856A (en) * | 2007-04-20 | 2010-03-31 | 因万西斯系统股份有限公司 | Wet gas measurement |
| CN106920008A (en) * | 2017-02-28 | 2017-07-04 | 山东大学 | A kind of wind power forecasting method based on Modified particle swarm optimization BP neural network |
| CN108469281A (en) * | 2018-01-18 | 2018-08-31 | 天津大学 | Two-phase Research on vortex signal processing based on EMD and Spectrum Correction |
| CN110186522A (en) * | 2019-05-31 | 2019-08-30 | 天津大学 | Reading compensation and flow-measuring method are crossed in conjunction with the moisture of vortex street amplitude characteristic |
| CN111222229A (en) * | 2019-12-27 | 2020-06-02 | 清华大学深圳国际研究生院 | Method for constructing instantaneous flow measurement model in gas-liquid two-phase flow dynamic flow process |
| CN111380581A (en) * | 2020-02-17 | 2020-07-07 | 天津大学 | Image method-based fog annular flow phase separation flow parameter measurement method |
| CN111370080A (en) * | 2020-03-05 | 2020-07-03 | 中国工程物理研究院上海激光等离子体研究所 | Radiation temperature inversion method based on artificial neural network algorithm |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111649779A (en) * | 2020-05-31 | 2020-09-11 | 天津大学 | Oil well oil content and total flow rate measuring method based on dense neural network and application |
| CN111649779B (en) * | 2020-05-31 | 2022-04-08 | 天津大学 | Oil well oil content and total flow rate measuring method based on dense neural network and application |
| WO2023005302A1 (en) * | 2021-07-26 | 2023-02-02 | 东南大学 | Method for predicting biomass coke gasification reaction characteristics |
| CN113392961A (en) * | 2021-08-17 | 2021-09-14 | 中国海洋大学 | Method for extracting mesoscale eddy track stable sequence and predicting cyclic neural network |
| CN113392961B (en) * | 2021-08-17 | 2021-11-09 | 中国海洋大学 | Method for extracting mesoscale eddy track stable sequence and predicting cyclic neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112926767A (en) | Annular fog flow gas phase apparent flow velocity prediction method based on particle swarm BP neural network | |
| CN105678332B (en) | Converter steelmaking end point judgment method and system based on flame image CNN recognition modeling | |
| Zheng et al. | Gas–liquid two phase flow measurement method based on combination instrument of turbine flowmeter and conductance sensor | |
| AU2020102518A4 (en) | A method of gas sensor calibration based on linear optimization | |
| CN111896616B (en) | Gas-liquid two-phase flow pattern identification method based on acoustic emission-BP neural network | |
| Jiang et al. | A multisensor cycle-supervised convolutional neural network for anomaly detection on magnetic flux leakage signals | |
| CN101464172A (en) | Soft measuring method for power boiler breeze concentration mass flow | |
| CN109060023A (en) | A kind of the data quality control method and system of miniature environmental monitoring | |
| CN111537032A (en) | Oil-gas two-phase flow parameter measuring device and method | |
| CN106770170A (en) | Can modularization apolegamy assembling Raman microscratch fast detector | |
| Xie et al. | An online flow pattern identification system for gas–oil two-phase flow using electrical capacitance tomography | |
| CN111275744B (en) | Non-contact vibration frequency measurement method based on deep learning and image processing | |
| CN104062409B (en) | Intelligence soot gas analyzes method and apparatus | |
| Shen et al. | SSCT-Net: A semisupervised circular teacher network for defect detection with limited labeled multiview MFL samples | |
| Zuo et al. | Active learning-based complex pipeline weld defect detection with lightweight neural network | |
| Bai et al. | Online recognition of the multiphase flow regime | |
| WO2022241883A1 (en) | Pollution source recognition method and apparatus based on three-dimensional liquid chromatography fingerprint | |
| Grace et al. | Analysis and visualization of air quality using real time pollutant data | |
| Teng et al. | Vibration signal-based structural damage detection through deep learning and digital image correlation | |
| OuYang et al. | Soft measurement of oil–water two-phase flow using a multi-task sequence-based CapsNet | |
| CN114462717A (en) | Small sample gas concentration prediction method based on improved GAN and LSTM | |
| CN111814578A (en) | Method for extracting frequency of ultralow frequency Doppler signal | |
| Li et al. | Intelligent diagnosis and recognition method of GIS partial discharge data map based on deep learning | |
| Zhao et al. | Research on Transformer Oil Multi-frequency Ultrasonic Monitoring Technology Based on Convolutional Neural Network | |
| Fan et al. | Wavelet entropy applied in gas-liquid two-phase flow |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210608 |