CN111950358A - Valve viscosity detection method based on image recognition - Google Patents
Valve viscosity detection method based on image recognition Download PDFInfo
- Publication number
- CN111950358A CN111950358A CN202010627137.5A CN202010627137A CN111950358A CN 111950358 A CN111950358 A CN 111950358A CN 202010627137 A CN202010627137 A CN 202010627137A CN 111950358 A CN111950358 A CN 111950358A
- Authority
- CN
- China
- Prior art keywords
- valve
- neural network
- convolutional neural
- network model
- data
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 34
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 44
- 238000012360 testing method Methods 0.000 claims abstract description 27
- 238000012549 training Methods 0.000 claims abstract description 24
- 238000000034 method Methods 0.000 claims abstract description 17
- 230000000694 effects Effects 0.000 claims abstract description 6
- 238000009776 industrial production Methods 0.000 claims abstract description 6
- 238000012545 processing Methods 0.000 abstract description 3
- 238000011156 evaluation Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000007792 addition Methods 0.000 description 1
- 238000003889 chemical engineering Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the field of process industrial performance evaluation, in particular to a valve viscosity detection method based on image recognition, which comprises the following steps: acquiring input data of a plurality of valves and output data of an industrial production process acted by the valves; converting input data and output data of a plurality of valves into a planar two-dimensional curve image and dividing the planar two-dimensional curve image into a training set and a testing set; using the training set for training a convolutional neural network model; the test set is used for testing the convolutional neural network model; and performing valve viscosity detection by using the trained convolutional neural network model. By using the present invention, the following effects can be achieved: the convolutional neural network model is used for judging the valve viscosity, the capability of the convolutional neural network on abstracting data information from the image is utilized, manual processing on the data is reduced, and the convolutional neural network model is suitable for various data.
Description
Technical Field
The invention relates to the field of process industrial performance evaluation, in particular to a valve viscosity detection method based on image recognition.
Background
Valves are a large number of actuators found in control loops in the process industry. The method is widely applied to a plurality of industrial production fields, such as oil refining, chemical engineering, food, medicine and the like, and is used for controlling the flow. The valve has better performance and normal work in the initial use stage, and can ensure that the control loop has better control performance. However, with long-term operation, the internal structure of the valve is worn and aged, the performance of the valve is gradually deteriorated, and the flow cannot be accurately controlled. One of the typical problems is valve sticking, i.e. the output of the valve sometimes cannot follow the input and there are obstacles in the valve movement, such as the valve "sticking" and the output flow changing with sticking.
The existing valve viscosity detection method mainly searches the viscosity characteristics of the valve on numerical characteristics and mainly uses methods of fitting, regression, numerical statistical analysis and the like of the numerical characteristics. These methods have many process requirements, linear process assumptions and white noise assumptions, complex processes and low accuracy.
Disclosure of Invention
In order to solve the problems, the invention provides a valve viscosity detection method based on image recognition.
A valve sticking detection method based on image recognition comprises the following steps:
acquiring input data of a plurality of valves and output data of an industrial production process acted by the valves;
converting input data and output data of a plurality of valves into a planar two-dimensional curve image and dividing the planar two-dimensional curve image into a training set and a testing set;
using the training set for training a convolutional neural network model;
using the test set for testing the convolutional neural network model;
and performing valve viscosity detection by using the trained convolutional neural network model.
Preferably, the converting the input data and the output data of the plurality of valves into the planar two-dimensional curve image includes:
obtaining corresponding curves according to input data and output data of the valves;
dividing the period of each waveform into a data segment according to the waveform of the curve corresponding to the valve output data;
and for each waveform data segment, taking input data of the valve as a value of an x axis, taking output data of the valve as a value of a y axis, drawing data points according to the time sequence, and connecting the data points back and forth on a two-dimensional coordinate plane to form a plane two-dimensional curve image.
Preferably, the ratio of the training set to the test set is 8: 2.
Preferably, the test using the test set for the convolutional neural network model includes:
and adjusting the structure and parameters of the convolutional neural network model according to the test result, so that the detection effect of the convolutional neural network model is optimal.
Preferably, the performing valve sticking detection by using the trained convolutional neural network model includes:
acquiring a planar two-dimensional curve image corresponding to each waveform data segment of input data and output data of a valve to be detected;
and inputting the two-dimensional curve images of each plane into the trained convolutional neural network model to judge the viscosity, if the proportion of the viscosity of the two-dimensional curve images of the output plane is greater than a set threshold value, judging that the valve is viscous, and otherwise, judging that the valve is not viscous.
By using the present invention, the following effects can be achieved:
1. the convolutional neural network model is used for judging the valve viscosity, the capability of the convolutional neural network in abstracting data information from the image is utilized, the manual processing on the data is reduced, and the convolutional neural network model is suitable for various data;
2. the input data and output data of the valve are segmented according to the waveform period, and each segment is further processed. Therefore, the crossing or overlapping of curves in the generated image is reduced, the identification of the valve viscosity characteristic in the planar two-dimensional curve image is facilitated, and the accuracy of valve viscosity detection is improved;
3. the input data of the valve is used as the value of an x axis, and the output data of the valve is converted into a two-dimensional curve image as the value of a y axis. Therefore, the original input data and output data of the valve are converted into an image form which is convenient for the convolutional neural network model to process, and the efficiency of valve viscosity detection is improved.
Drawings
The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
Fig. 1 is a schematic flow chart of a valve sticking detection method based on image recognition according to an embodiment of the present invention;
fig. 2 is a schematic flowchart of step S2 in a valve sticking detection method based on image recognition according to an embodiment of the present invention;
fig. 3 is a schematic diagram of curves of input data and output data of a valve in a valve viscosity detection method based on image recognition according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a planar two-dimensional curve image in a valve sticking detection method based on image recognition according to an embodiment of the present invention;
fig. 5 is a flowchart illustrating step S5 of a valve sticking detection method based on image recognition according to an embodiment of the present invention.
Detailed Description
The technical solutions of the present invention will be further described below with reference to the accompanying drawings, but the present invention is not limited to these embodiments.
The basic idea of the invention is to convert the input data of the valve and the output data of the industrial production process acted by the valve into a planar two-dimensional curve image, then identify the planar two-dimensional curve image of the valve through a convolution neural network model, and judge whether the valve has the problem of adhesion by using the result of image identification.
Based on the above inventive concept, the present embodiment provides a valve sticking detection method based on image recognition, as shown in fig. 1, including the following steps:
s1: and acquiring input data of a plurality of valves and output data of the industrial production process acted by the valves.
A large amount of data, including input data and output data of the valve, is acquired in advance from an actual industrial process or a simulation process.
S2: and converting input data and output data of a plurality of valves into a planar two-dimensional curve image and dividing the planar two-dimensional curve image into a training set and a testing set.
As shown in fig. 2, the method specifically includes the following steps:
s21: obtaining corresponding curves according to input data and output data of the valves;
s22: dividing the period of each waveform into a data segment according to the waveform of the curve corresponding to the valve output data;
s23: and for each waveform data segment, taking input data of the valve as a value of an x axis, taking output data of the valve as a value of a y axis, drawing data points according to the time sequence, and connecting the data points back and forth on a two-dimensional coordinate plane to form a plane two-dimensional curve image.
As shown in fig. 3, OP represents the input data curve of the valve, and PV represents the output data curve of the valve. The input data and output data of the valve are segmented according to the waveform period, and each segment is further processed. Therefore, the intersection or the superposition of curves in the generated image is reduced, the identification of the valve viscosity characteristic in the planar two-dimensional curve image is facilitated, and the accuracy of valve viscosity detection is improved.
As shown in fig. 4, the input data of the valve is converted into a two-dimensional curve image as the value of the x-axis and the output data of the valve is converted into the value of the y-axis. Therefore, the original input data and output data of the valve are converted into an image form which is convenient for the convolutional neural network model to process, and the efficiency of valve viscosity detection is improved.
Preferably, the ratio of training set to test set is 8: 2. The training set is used for training the convolutional neural network model, and the testing set is used for testing the convolutional neural network model.
S3: the training set is used for training the convolutional neural network model.
Before training, whether each section of data has valve viscosity is manually judged according to general experience, and the given mark has viscosity or no viscosity. A convolutional neural network model is then prepared for detecting whether the image has sticky properties. The training set of images converted in step S2 is used for training the convolutional neural network model. In the training process, the structure and parameters of the convolutional neural network model are adjusted, so that the convolutional neural network model achieves the best detection effect on the viscosity on a training set, namely the detection accuracy or recall rate is highest, and the convolutional neural network model which can be used for valve viscosity detection after the training is finished is obtained.
S4: the test set is used for testing the convolutional neural network model.
Unlike training, a test set does not require a large amount of data, only a small amount of data to test. And judging the detection accuracy of the convolutional neural network model in the test process, and adjusting the structure and parameters of the convolutional neural network model according to the test result so that the detection effect of the convolutional neural network model reaches the best.
S5: and performing valve viscosity detection by using the trained convolutional neural network model.
As shown in fig. 5, the method specifically includes the following steps:
s51: acquiring a planar two-dimensional curve image corresponding to each waveform data segment of input data and output data of a valve to be detected;
s52: and inputting the two-dimensional curve images of each plane into the trained convolutional neural network model to judge the viscosity, if the proportion of the viscosity of the two-dimensional curve images of the output plane is greater than a set threshold value, judging that the valve is viscous, and otherwise, judging that the valve is not viscous.
The input data and the output data of the valve with the viscosity characteristic are represented by images, and the images can show a specific shape mode which can be used for detecting the viscosity of the valve. And (5) judging the valve viscosity by using a convolutional neural network model. The convolutional neural network is good at the capability of abstracting data information from the image, manual processing of the data is reduced, and the convolutional neural network is suitable for various data.
Various modifications or additions may be made to the described embodiments or alternatives may be employed by those skilled in the art without departing from the spirit or ambit of the invention as defined in the appended claims.
Claims (5)
1. A valve viscosity detection method based on image recognition is characterized by comprising the following steps:
acquiring input data of a plurality of valves and output data of an industrial production process acted by the valves;
converting input data and output data of a plurality of valves into a planar two-dimensional curve image and dividing the planar two-dimensional curve image into a training set and a testing set;
using the training set for training a convolutional neural network model;
using the test set for testing the convolutional neural network model;
and performing valve viscosity detection by using the trained convolutional neural network model.
2. The method of claim 1, wherein converting input data and output data of the plurality of valves into the planar two-dimensional curve image comprises:
obtaining corresponding curves according to input data and output data of the valves;
dividing the period of each waveform into a data segment according to the waveform of the curve corresponding to the valve output data;
and for each waveform data segment, taking input data of the valve as a value of an x axis, taking output data of the valve as a value of a y axis, drawing data points according to the time sequence, and connecting the data points back and forth on a two-dimensional coordinate plane to form a plane two-dimensional curve image.
3. The valve sticking detection method based on image recognition is characterized in that the ratio of the training set to the test set is 8: 2.
4. The valve sticking detection method based on image recognition is characterized in that the test set is used for testing a convolutional neural network model and comprises the following steps:
and adjusting the structure and parameters of the convolutional neural network model according to the test result, so that the detection effect of the convolutional neural network model is optimal.
5. The valve sticking detection method based on image recognition according to claim 1, wherein the valve sticking detection by using the trained convolutional neural network model comprises:
acquiring a planar two-dimensional curve image corresponding to each waveform data segment of input data and output data of a valve to be detected;
and inputting the two-dimensional curve images of each plane into the trained convolutional neural network model to judge the viscosity, if the proportion of the viscosity of the two-dimensional curve images of the output plane is greater than a set threshold value, judging that the valve is viscous, and otherwise, judging that the valve is not viscous.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010627137.5A CN111950358A (en) | 2020-07-01 | 2020-07-01 | Valve viscosity detection method based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010627137.5A CN111950358A (en) | 2020-07-01 | 2020-07-01 | Valve viscosity detection method based on image recognition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111950358A true CN111950358A (en) | 2020-11-17 |
Family
ID=73337381
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010627137.5A Pending CN111950358A (en) | 2020-07-01 | 2020-07-01 | Valve viscosity detection method based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111950358A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113494999A (en) * | 2021-09-07 | 2021-10-12 | 南通索星流体设备有限公司 | Valve viscosity data detection processing method based on artificial intelligence |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103970129A (en) * | 2014-05-23 | 2014-08-06 | 东南大学 | Control valve adhesion detecting method |
| CN109063741A (en) * | 2018-07-05 | 2018-12-21 | 南京航空航天大学 | A kind of energy spectrum analysis method based on hibert curve transformation and deep learning |
| CN109407654A (en) * | 2018-12-20 | 2019-03-01 | 浙江大学 | A kind of non-linear causality analysis method of industrial data based on sparse depth neural network |
| CN110177122A (en) * | 2019-06-18 | 2019-08-27 | 国网电子商务有限公司 | A kind of method for establishing model and device identifying network security risk |
| US10402699B1 (en) * | 2015-12-16 | 2019-09-03 | Hrl Laboratories, Llc | Automated classification of images using deep learning—back end |
| CN111079861A (en) * | 2019-12-31 | 2020-04-28 | 国网北京市电力公司 | Power distribution network voltage abnormity diagnosis method based on image rapid processing technology |
-
2020
- 2020-07-01 CN CN202010627137.5A patent/CN111950358A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103970129A (en) * | 2014-05-23 | 2014-08-06 | 东南大学 | Control valve adhesion detecting method |
| US10402699B1 (en) * | 2015-12-16 | 2019-09-03 | Hrl Laboratories, Llc | Automated classification of images using deep learning—back end |
| CN109063741A (en) * | 2018-07-05 | 2018-12-21 | 南京航空航天大学 | A kind of energy spectrum analysis method based on hibert curve transformation and deep learning |
| CN109407654A (en) * | 2018-12-20 | 2019-03-01 | 浙江大学 | A kind of non-linear causality analysis method of industrial data based on sparse depth neural network |
| CN110177122A (en) * | 2019-06-18 | 2019-08-27 | 国网电子商务有限公司 | A kind of method for establishing model and device identifying network security risk |
| CN111079861A (en) * | 2019-12-31 | 2020-04-28 | 国网北京市电力公司 | Power distribution network voltage abnormity diagnosis method based on image rapid processing technology |
Non-Patent Citations (2)
| Title |
|---|
| 李潇: "基于机器视觉数据采集的MES关键技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 郑丽丽: "基于模糊聚类的阀门粘滞检测与量化", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113494999A (en) * | 2021-09-07 | 2021-10-12 | 南通索星流体设备有限公司 | Valve viscosity data detection processing method based on artificial intelligence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110503251B (en) | Non-holiday load prediction method based on Stacking algorithm | |
| CN107622276B (en) | Deep learning training method based on combination of robot simulation and physical sampling | |
| CN109815855B (en) | Electronic equipment automatic test method and system based on machine learning | |
| CN104234695A (en) | Oil well fault diagnosis method based on neural network | |
| CN114199892B (en) | Plate measuring method and system based on machine vision | |
| CN111950358A (en) | Valve viscosity detection method based on image recognition | |
| CN109814379A (en) | The different factor full format non-model control method of MIMO | |
| CN112231971A (en) | Blast furnace fault diagnosis method based on relative integral trend diffusion fault sample generation | |
| CN116913435A (en) | High-strength engineering plastic evaluation method and system based on component analysis | |
| CN114217025B (en) | Analysis method for evaluating influence of meteorological data on air quality concentration prediction | |
| CN111160428A (en) | Automatic vegetable identification method based on CNN-SVM algorithm | |
| CN112183469B (en) | Method for identifying congestion degree of public transportation and self-adaptive adjustment | |
| CN109508639A (en) | Road scene semantic segmentation method based on multiple dimensioned convolutional neural networks with holes | |
| CN110826258B (en) | Space-time convolutional neural network-based data center machine room simulation environment construction method | |
| KR102151733B1 (en) | Automated method of reflecting changes in process safety management drawings when changes occur due to high-tech manufacturing operations | |
| CN110320802B (en) | Complex system signal time sequence identification method based on data visualization | |
| CN116051496A (en) | Real-time sewer defect detection method | |
| WO2023085992A1 (en) | Image analysis of cut flowers | |
| CN115345846A (en) | Intelligent grading method and system for grain size of medium and low carbon steel | |
| CN116090390A (en) | FINFET device direct current characteristic prediction method based on deep learning | |
| CN113486599A (en) | Method for calculating effective stroke of oil pumping unit | |
| CN108009565B (en) | Method for calculating design flood level of river network area under variable environment | |
| CN111310907A (en) | Microwave assembly fault diagnosis method, device and equipment | |
| CN103116272A (en) | Online adaptive modeling method for hydrocyanic acid production process | |
| CN116126827B (en) | Method for simultaneously realizing modeling and visual analysis of power production data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201117 |
|
| RJ01 | Rejection of invention patent application after publication |