CN105139025B - Gas-solid fluidized bed flow pattern online intelligent recognition method based on nonlinear analysis method - Google Patents
Gas-solid fluidized bed flow pattern online intelligent recognition method based on nonlinear analysis method Download PDFInfo
- Publication number
- CN105139025B CN105139025B CN201510461151.1A CN201510461151A CN105139025B CN 105139025 B CN105139025 B CN 105139025B CN 201510461151 A CN201510461151 A CN 201510461151A CN 105139025 B CN105139025 B CN 105139025B
- Authority
- CN
- China
- Prior art keywords
- cluster
- gas
- fluidized bed
- pressure fluctuation
- solid fluidized
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Fluidized-Bed Combustion And Resonant Combustion (AREA)
Abstract
The gas-solid fluidized bed flow pattern online intelligent recognition method based on nonlinear analysis method that the invention discloses a kind of, by gas-solid fluidized bed progress pressure fluctuation signal sampling analysis and extracting characteristic value；Objective cluster is carried out by fuzzy clustering algorithm according to characteristic value and pressure fluctuation signal；It is established according to the result of pressure fluctuation signal, characteristic value and objective cluster and trains Flow Regime Ecognition system, then the system embedment computer is realized to the online intelligent recognition of gas-solid fluidized bed flow pattern.This method can be avoided the influence of subjective factor convection identification accuracy, while can carry out dynamic analysis and Flow Regime Ecognition to instantaneous state parameter.
Description
Technical field
The present invention relates to a kind of gas-solid fluidized bed flow type identification methods, belong to gas-solid fluidized bed technical field.
Background technique
It is gas-solid fluidized bed to be obtained extensively with good flowing heat transfer characteristic in multiple industrial circles such as chemical industry, the energy because of it
Using.The measurement of gas-solid fluidized bed parameter and flow pattern are closely related, how to be accurately identified in real time to gas-solid fluidized bed flow pattern
It is an important topic of current gas-solid fluidized bed area research.
Mainly there are direct observational method and Parameter analysis method to the method for gas-solid fluidized bed Flow Regime Ecognition at present.Direct observational method
Need reaction unit to visualize or using image documentation equipment progress Image Acquisition, the gas-solid fluidized bed apparatus in industrial application because by
The limitation of rapidoprint is difficult to adopt direct observational method and carries out accurate Flow Regime Ecognition.Parameter analysis method need by capacitance method,
The methods of optical fiber probe mensuration, pressure fluctuation signal acquisition obtain relevant parameter, then in conjunction with existing flow pattern decision criteria
Acquisition parameter is handled and analyzed with method.Both methods not can avoid subjective factor convection identification accuracy
It influences, it is also difficult to which dynamic analysis and Flow Regime Ecognition are carried out to instantaneous state parameter.
Summary of the invention
The purpose of the present invention: in order to overcome the deficiencies in the prior art, the present invention provides a kind of based on non-linear point
The gas-solid fluidized bed flow pattern online intelligent recognition method of analysis method, this method can be avoided subjective factor convection identification accuracy
Influence, while dynamic analysis and Flow Regime Ecognition can be carried out to instantaneous state parameter.
Technical solution: to achieve the above object, the technical solution adopted by the present invention are as follows:
A kind of gas-solid fluidized bed flow pattern online intelligent recognition method based on nonlinear analysis method, by gas-solid fluidized
Bed carries out pressure fluctuation signal sampling analysis and extracts characteristic value；Passed through according to the corresponding characteristic value of pressure fluctuation signal fuzzy poly-
Class algorithm carries out objective cluster；It is established according to the result of pressure fluctuation signal, characteristic value and objective cluster and training flow pattern is known
Other system, then the system embedment computer is realized to the online intelligent recognition of gas-solid fluidized bed flow pattern.
Specifically includes the following steps:
Step 1, gas-solid fluidized bed middle pressure fluctuation signal is acquired；
Step 2, step 1 pressure fluctuation signal collected is handled by nonlinear analysis method, extracts feature
Value；
Step 3, the characteristic value extracted in step 2 is introduced into fuzzy c clustering algorithm, to pressure arteries and veins collected in step 1
Dynamic signal carries out objective cluster；
Step 4, the cluster result of step 3 neural network is introduced to carry out the building of Flow Regime Ecognition system, train；
Step 5, trained Flow Regime Ecognition system is loaded into Flow Regime Ecognition and output system, realized to gas-solid fluidized bed visitor
It sees, is accurate, real-time Flow Regime Ecognition.
The acquisition method of pressure fluctuation signal in the step 1: it is gas-solid fluidized bed it is middle utilize differential pressure transmitter (7) carry out
The acquisition of pressure fluctuation signal；The signal of differential pressure transmitter (7) acquisition is carrying out step 2 after the conversion of A/D converter (8)
Characteristics extraction.
Nonlinear analysis method in the step 2 is Hilbert-Huang transform analysis method.
The Hilbert-Huang transform analysis method includes empirical mode decomposition and Hilbert transform, specifically include with
Lower step:
Step 21, first by empirical mode decomposition to the frequency point of the pressure fluctuation signal pressing force fluctuating signal of acquisition
Amount gradually decomposes itself intrinsic family's Intrinsic mode function from high to low, which embodies pressure fluctuation letter
Number characteristic information, and the most high frequency section of first Intrinsic mode function representative pressure fluctuating signal；
Step 22, Hilbert transform is carried out to each Intrinsic mode function, obtains its instantaneous frequency and instantaneous amplitude, into
And obtain its characteristic value.
Intrinsic mode function number is more than one in the step 21, and Intrinsic mode function is denoted as IMF, by IMF component
Energy definition is as follows:
In formula, E is the energy of IMF component, and n is sample total amount, aiIt (t) is the corresponding instantaneous amplitude of i-th of IMF；
Characteristic value refers to that the pressure difference signal flowed under different gas velocity to different bed materials is analyzed in the step 22, meter
The energy size for calculating each rank IMF component is classified as high band E according to the frequency range of each rank IMF componenth, Mid Frequency EmWith
Low-frequency range El, Eh、Em、ElAs extracted characteristic value.
Gas-solid fluidized bed flow pattern online intelligent recognition side according to claim 6 based on nonlinear analysis method
Method, it is characterised in that: the method that pressure fluctuation signal collected carries out objective cluster in step 1 in the step 3:
Step 31, characteristic value, the cluster numbers, cluster extracted according to pressure fluctuation signal collected in step 1, step 2
Center and pressure fluctuation signal and the cluster centre degree of membership establish cluster object module；
Step 32, the characteristic value and selection extracted according to pressure fluctuation signal collected in step 1, step 2 are initial to gather
Class center calculates the cluster target value of the initial cluster center by the cluster object module of step 3；Calculate what step 2 was extracted
Pressure fluctuation signal is included into corresponding class by the Euclidean distance of characteristic value and initial cluster center, according to the sample of every one kind
Mean value obtains new cluster centre, and the cluster target value of the new cluster centre is calculated by the cluster object module of step 3；And
The cluster target value of initial cluster center and the cluster target value of new cluster centre are judged, if the two cluster target values have difference
It is different, the Euclidean distance of each characteristic value Yu new cluster centre is calculated again, is then reclassified, and the poly- of step 3 is passed through
Class object module calculates the cluster target value of this new cluster centre, and judge the cluster target value of previous cluster centre with
The cluster target value of the latter cluster centre, until the two cluster target values are no longer changed, the cluster of complete paired samples.
The cluster object module established in the step 31:
Wherein, J is cluster objective function, X={ x1,x2…xNIt is sample, each xi=[xi1,xi2...xin] all wrap
Containing n dimensional feature value, c is cluster numbers, chooses c different xiAs initial cluster center vK, k=1,2 ..., c, U is degree of membership, V
For cluster centre, m is weight coefficient, dki=| | xi-vk| |, dji=| | xi-vj||。
Intrinsic mode function number is determined by frequency component in the step 21, generally 7 or 8, respectively IMF1-
IMF7 or IMF1-IMF8；The step 5 is by trained identifying system and gas-solid fluidized bed pressure fluctuation signal acquisition system
Connection, data are called directly by program and carry out data processing, shown in output system the moment it is gas-solid fluidized bed in
Flow pattern state.
The utility model has the advantages that a kind of gas-solid fluidized bed flow pattern on-line intelligence based on nonlinear analysis method provided by the invention is known
Other method has the advantages that compared with prior art
By to gas-solid fluidized bed progress pressure fluctuation signal sampling analysis and extracting characteristic value；Pass through mould according to characteristic value
It pastes clustering algorithm and carries out objective cluster；It is established according to the result of pressure fluctuation signal, characteristic value and objective cluster and training is flowed
Type identifying system, then the system embedment computer is realized to the online intelligent recognition of gas-solid fluidized bed flow pattern, therefore this method energy
The influence of subjective factor convection identification accuracy is enough avoided, while dynamic analysis and flow pattern can be carried out to instantaneous state parameter
Identification.
Detailed description of the invention
Fig. 1 is gas-solid fluidized bed flow pattern online recognition system equipment system diagram of the present invention.
Fig. 2 is that example of the present invention carries out the IMF component map obtained after data processing.
1 is air compressor in Fig. 1；2 be flow control system；3 be gas-solid fluidized bed ontology；4 be air compartment；5 be distribution
Plate；6 be pressure tap；7 be differential pressure transmitter；8 be A/D converter；9 be computer.
Specific embodiment
One detailed introduction done to the present invention in conjunction with attached drawing, typical case of the invention but non-limiting embodiment is as follows:
A kind of gas-solid fluidized bed flow pattern online intelligent recognition method based on nonlinear analysis method, by gas-solid fluidized
Bed carries out pressure fluctuation signal sampling analysis and extracts characteristic value；Objective gather is carried out by fuzzy clustering algorithm according to characteristic value
Class；It is established according to the result of pressure fluctuation signal, characteristic value and objective cluster and training Flow Regime Ecognition system, then by the system
Insertion computer realizes the online intelligent recognition of gas-solid fluidized bed flow pattern.
Wherein, identifying system include as shown in Figure 1: air compressor 1, flow adjusting be 2, gas-solid fluidized bed ontology 3, wind
Room 4, pressure tap 6, differential pressure transporting 7, A/D conversion 8, calculates 9 at distribution grid 5；The air compartment 4 is set to gas-solid fluidized bed ontology 3
Bottom, and the distribution grid 5 is set on air compartment 4, the pipeline of measurement all the way of the differential pressure transporting 7 measures gas by pressure tap 6
The pressure fluctuation signal of solid fluid bed ontology 3, another way measurement pipeline is set in air compartment 4, and is located at gas-solid fluidized bed ontology 3
Between distribution grid 5, and the differential pressure transporting 7 is connect by 8 progress signal conversion of A/D conversion with calculating 9, and the air pressure
It is 2 to enter in air compartment 4 that the air that contracting machine 1 compresses is adjusted by flow.
Recognition methods of the invention specifically includes the following steps:
Step 1, gas-solid fluidized bed middle pressure fluctuation signal is acquired；By it is gas-solid fluidized bed it is middle utilize differential pressure
The acquisition of transmitter (7) progress pressure fluctuation signal；The signal of differential pressure transmitter (7) acquisition passes through the conversion of A/D converter (8)
Afterwards in the characteristics extraction for carrying out step 2.
Step 2, step 1 pressure fluctuation signal collected is handled by nonlinear analysis method, extracts feature
Value.The nonlinear analysis method is Hilbert-Huang transform analysis method, mainly includes two aspect contents, empirical modal point
Solve (Empirical Mode Decomposition, abbreviation EMD) and Hilbert transform (Hilbert
Transportation, abbreviation HT).Wherein EMD is core, uses EMD method first, extracts from pressure difference fluctuating signal
Itself intrinsic family's Intrinsic mode function (Intrinsic Mode Function, abbreviation IMF), quantity is limited,
It include true physical significance.EMD extracts IMF from original signal, and obtained IMF embodies the feature letter of original signal
Breath, first IMF represent the most high frequency section of original signal.Then Hilbert transform is carried out to each IMF, it is instantaneous obtains it
Frequency and instantaneous amplitude.EMD decomposition is carried out to pressure difference fluctuating signal collected, gradually the frequency component of pressure fluctuation signal
It decomposites and from high to low, extract 7 or 8, respectively IMF1-IMF7 or IMF1-IMF8.Preferably to analyze different range
Each IMF component energy is defined as follows by the changing rule of instantaneous frequency:
In formula, aiIt (t) is the corresponding instantaneous amplitude of i-th of IMF.
The pressure difference signal flowed under different gas velocity to different bed materials is analyzed, and the energy for calculating each rank IMF component is big
It is small, according to the frequency range of each rank IMF component, it is classified as high band Eh(E1,2,3), Mid Frequency Em(E4~6) and low-frequency range
El(E7,8), Eh、Em、ElI.e. step 2 passes through extracted characteristic value after nonlinear analysis method processing pressure fluctuating signal.
Step 3, the characteristic value extracted in step 2 is introduced into fuzzy c clustering algorithm, to pressure arteries and veins collected in step 1
Dynamic signal carries out objective cluster.The characteristic value extracted to step 2 clusters, and compares gas velocity and image sampling determines that cluster corresponds to
Specific flow pattern.
The method that pressure fluctuation signal collected carries out objective cluster in step 1 in the step 3:
Step 31, characteristic value, the cluster numbers, cluster extracted according to pressure fluctuation signal collected in step 1, step 2
Center and the corresponding characteristic value of pressure fluctuation signal and the cluster centre degree of membership establish cluster object module；
The cluster object module established in the step 31
Wherein, J is cluster objective function, X={ x1,x2…xNIt is sample, each xi=[xi1,xi2...xin] all wrap
Containing n dimensional feature value, c is cluster numbers, chooses c different xiAs initial cluster center vK, k=1,2 ..., c, U is degree of membership, V
For cluster centre, m is weight coefficient,μki、viIt is calculate by the following formula really
It is fixed, dki=| | xi-vk| |, dji=| | xi-vj| |:
Step 32, the characteristic value and selection extracted according to pressure fluctuation signal collected in step 1, step 2 are initial to gather
Class center calculates the cluster target value of the initial cluster center by the cluster object module of step 3；Calculate what step 2 was extracted
Pressure fluctuation signal is included into corresponding class by the Euclidean distance of characteristic value and initial cluster center, according to the sample of every one kind
Mean value obtains new cluster centre, and the cluster target value of the new cluster centre is calculated by the cluster object module of step 3；And
The cluster target value of initial cluster center and the cluster target value of new cluster centre are judged, if the two cluster target values have difference
It is different, the Euclidean distance of each characteristic value Yu new cluster centre is calculated again, is then reclassified, and the poly- of step 3 is passed through
Class object module calculates the cluster target value of this new cluster centre, and judge the cluster target value of previous cluster centre with
The cluster target value of the latter cluster centre, until the two cluster target values are no longer changed, the cluster of complete paired samples,
I.e. by X={ x1,x2…xNIn xiBeing grouped into center is vkIt is all kinds of in.
Step 4, the cluster result of step 3 neural network is introduced to carry out the building of Flow Regime Ecognition system, train.
Building and training for Flow Regime Ecognition system is carried out using the characteristic value clustered as the training data of neural network.Step
Neural network algorithm used in rapid 4 is by input layer node, output layer neuron node and hidden layer neuron node
It constitutes.The step of standard genetic algorithm, is as follows:
1) number individual in population, filial generation number are determined；
2) fitness of individual is calculated；
3) individual is selected or eliminated according to individual adaptation degree；
4) according to certain intersection, mutation probability and cross method, new individual is generated；
5) population of a new generation is generated by intersecting and making a variation；
6) reach setting filial generation number and stop algorithm.
Step 5, trained Flow Regime Ecognition system is loaded into Flow Regime Ecognition and output system, realized to gas-solid fluidized bed visitor
It sees, is accurate, real-time Flow Regime Ecognition.
Above-mentioned steps 1 to step 4 is the training and foundation of convection identifying system, and trained identification is by step 5
System is connect with gas-solid fluidized bed pressure fluctuation signal acquisition system, is called directly data by program and is carried out data processing,
The moment gas-solid fluidized bed interior flow pattern state is shown in output system.
Example
Step 1: the acquisition of pressure fluctuation signal, as shown in Figure 1, gas-solid fluidized bed 3 pass through differential pressure transmitter 7 and A/D number
Word signal converts collector 8 and carries out data acquisition, and acquired data record is in the computer 9；Can also be used recording instrument without paper or
Other signals convert acquisition system.Guarantee sample frequency is 100Hz, using the period at 10 seconds or more.To different gas velocity and difference
The operating condition of thickness of bed layer carries out pressure fluctuation signal acquisition respectively, different using quartz sand and pelletized biomass in this example
The mixed material of ratio samples 120 groups altogether.
Step 2: being handled step 1 pressure fluctuation signal collected by nonlinear analysis method, is extracted special
Value indicative: being handled sampled data according to preceding method, obtains IMF energy interval value corresponding to each group of data.In this example
The corresponding three-dimensional array of every group of pressure fluctuation signal, the array are the energy of IMF low-frequency range, IMF Mid Frequency and IMF high band
Value.Fig. 2 is the IMF component map of a certain group of data.
Step 3: according to gas-solid fluidized bed classical taxonomy mode can be divided into fixed bed, bubbling bed, section gush bed, turbulent bed and
Fast bed, fast bed are generally realized in recirculating fluidized bed, are limited in this example by fluidized bed size, only to preceding 4 kinds of flow patterns into
Row classification, if cluster numbers are 4.Sampled data set is 120 groups, randomly selects 4 group data sets as initial cluster center, calculates every
The characteristic value of one group of data and the Euclidean distance of cluster centre, data are included into corresponding to the smallest cluster centre of Euclidean distance
In one kind, mean value is then carried out by the characteristic value of all data to every one kind, calculation is asked to obtain new cluster centre, counted again
It calculates the characteristic value of each group of data and the Euclidean distance of cluster centre and reclassifies data, until cluster centre is no longer sent out
Changing, cluster are completed.According to the corresponding flow velocity of sampled data, cluster and flow pattern can be corresponded.120 groups of data in this example
Cluster result are as follows: 31 groups of fixed bed, 38 groups of bubbling bed, section gush 24 groups of bed, 27 groups of turbulent flow group.
Step 4: the cluster result that step 3 is obtained is as population training data.If Population Size is 100, subalgebra
Mesh is 200, crossover probability 0.5, mutation probability 0.001.Neural network after training has neither part nor lot in training by remaining 20 groups
Data verified, have 18 groups of identifications correct, accuracy rate 90%.The selection of parameter can carry out reasonably according to the actual situation
Adjustment, parameter determines after test of many times herein.
Step 5: trained network is integrated as a module of online recognition system with data processing section,
It accesses in gas-solid fluidized bed system and runs again, every 10 seconds pressure fluctuation signals carry out online flow pattern knowledge as one group of data
Not.The display that recognition result directly passes through computer 10 is shown.
Claims (6)
1. a kind of gas-solid fluidized bed flow pattern online intelligent recognition method based on nonlinear analysis method, it is characterised in that: pass through
To gas-solid fluidized bed progress pressure fluctuation signal sampling analysis and extract characteristic value；According to the corresponding characteristic value of pressure fluctuation signal
Objective cluster is carried out by fuzzy clustering algorithm；According to the result of pressure fluctuation signal, characteristic value and objective cluster establish and
Flow Regime Ecognition system is trained, then the system embedment computer is realized to the online intelligent recognition of gas-solid fluidized bed flow pattern；Including with
Lower step:
Step 1, gas-solid fluidized bed middle pressure fluctuation signal is acquired；
Step 2, step 1 pressure fluctuation signal collected is handled by nonlinear analysis method, extracts characteristic value；
Characteristic value refers to that the pressure difference signal flowed under different gas velocity to different bed materials is analyzed, and calculates each rank IMF component
Energy size is classified as high band E according to the frequency range of each rank IMF componenth, Mid Frequency EmWith low-frequency range El, Eh、Em、El
As extracted characteristic value；
Step 3, the characteristic value extracted in step 2 is introduced into fuzzy c clustering algorithm, pressure fluctuation collected in step 1 is believed
Number carry out objective cluster；
The method that pressure fluctuation signal collected carries out objective cluster:
Step 31, characteristic value, the cluster numbers, cluster centre extracted according to pressure fluctuation signal collected in step 1, step 2
And the corresponding characteristic value of pressure fluctuation signal and the cluster centre degree of membership establish cluster object module；The cluster of the foundation
Object module:
；
Wherein,To cluster objective function,X={x 1 , x 2 …x N It is sample, eachx i = [x i1 , x i2 ... x in ] all include
HavenDimensional feature value, c are cluster numbers, and it is a different to choose cx i As initial cluster centerv k,k=1,2,…,c ,UFor degree of membership,VFor
Cluster centre,mIt is weight coefficient,；；,,；
Step 32, according in pressure fluctuation signal collected in step 1, the characteristic value of step 2 extraction and selection initial clustering
The heart calculates the cluster target value of the initial cluster center by the cluster object module of step 3；Calculate the feature that step 2 is extracted
The Euclidean distance of value and initial cluster center, pressure fluctuation signal is included into corresponding class, according to the sample average of every one kind
New cluster centre is obtained, the cluster target value of the new cluster centre is calculated by the cluster object module of step 3；And judge
The cluster target value of the cluster target value of initial cluster center and new cluster centre, if the two cluster target values are variant,
The Euclidean distance for calculating each characteristic value Yu new cluster centre again, is then reclassified, and the cluster mesh of step 3 is passed through
Mark model calculates the cluster target value of this new cluster centre, and judge the cluster target value of previous cluster centre with it is latter
The cluster target value of a cluster centre, until the two cluster target values are no longer changed, the cluster of complete paired samples；
Step 4, the cluster result of the characteristic value of step 3 neural network is introduced to carry out the building of Flow Regime Ecognition system, train；
Step 5, trained Flow Regime Ecognition system is loaded into Flow Regime Ecognition and output system, realized to gas-solid fluidized bed real-time
Flow Regime Ecognition.
2. the gas-solid fluidized bed flow pattern online intelligent recognition method according to claim 1 based on nonlinear analysis method,
It is characterized by: in the step 1 pressure fluctuation signal acquisition method: it is gas-solid fluidized bed it is middle utilize differential pressure transmitter (7)
Carry out the acquisition of pressure fluctuation signal；The signal of differential pressure transmitter (7) acquisition carries out again after the conversion of A/D converter (8)
The characteristics extraction of step 2.
3. the gas-solid fluidized bed flow pattern online intelligent recognition method according to claim 2 based on nonlinear analysis method,
It is characterized by: the nonlinear analysis method in the step 2 is Hilbert-Huang transform analysis method.
4. the gas-solid fluidized bed flow pattern online intelligent recognition method according to claim 3 based on nonlinear analysis method,
It is characterized by: the Hilbert-Huang transform analysis method includes empirical mode decomposition and Hilbert transform, specific packet
Include following steps:
Step 21, first by empirical mode decomposition to the frequency component of the pressure fluctuation signal pressing force fluctuating signal of acquisition by
Secondary to decompose itself intrinsic family's Intrinsic mode function from high to low, which embodies pressure fluctuation signal
Characteristic information, and the most high frequency section of first Intrinsic mode function representative pressure fluctuating signal；
Step 22, Hilbert transform is carried out to each Intrinsic mode function, obtains its instantaneous frequency and instantaneous amplitude, and then obtain
To its characteristic value.
5. the gas-solid fluidized bed flow pattern online intelligent recognition method according to claim 4 based on nonlinear analysis method,
It is characterized by: Intrinsic mode function number is more than one in the step 21, Intrinsic mode function is denoted as IMF, by IMF points
Energy is defined as follows:
In formula,For the energy of IMF component,For sample total amount,For the corresponding instantaneous amplitude of i-th of IMF.
6. the gas-solid fluidized bed flow pattern online intelligent recognition method according to claim 5 based on nonlinear analysis method,
It is characterized by: Intrinsic mode function number is determined by frequency component in the step 21, generally 7 or 8, respectively
IMF1-IMF7 or IMF1-IMF8；The step 5 adopts trained identifying system with gas-solid fluidized bed pressure fluctuation signal
Collecting system connection, calls directly data by program and carries out data processing, show that the moment is gas-solid fluidized in output system
Flow pattern state in bed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510461151.1A CN105139025B (en) | 2015-07-30 | 2015-07-30 | Gas-solid fluidized bed flow pattern online intelligent recognition method based on nonlinear analysis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510461151.1A CN105139025B (en) | 2015-07-30 | 2015-07-30 | Gas-solid fluidized bed flow pattern online intelligent recognition method based on nonlinear analysis method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105139025A CN105139025A (en) | 2015-12-09 |
| CN105139025B true CN105139025B (en) | 2019-03-05 |
Family
ID=54724368
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510461151.1A Active CN105139025B (en) | 2015-07-30 | 2015-07-30 | Gas-solid fluidized bed flow pattern online intelligent recognition method based on nonlinear analysis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105139025B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111829929A (en) * | 2020-07-30 | 2020-10-27 | 安徽工业大学 | Method for identifying binary particle separation or mixing state in fluidized bed |
| CN114417647B (en) * | 2021-11-19 | 2024-05-14 | 西南石油大学 | Dynamic compression river basin system airflow pulsation calculation method of reciprocating compressor |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5045240A (en) * | 1989-05-01 | 1991-09-03 | Westinghouse Electric Corp. | Contaminated soil restoration method |
| CN102338732A (en) * | 2011-07-06 | 2012-02-01 | 浙江大学 | Method for measuring fluidization parameter of gas-solid fluidized bed |
| CN103742901A (en) * | 2013-12-24 | 2014-04-23 | 广州市恒力安全检测技术有限公司 | Method for determining consumption difference analysis optimized target value system of circulating fluidized bed unit |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100158359A1 (en) * | 2008-12-18 | 2010-06-24 | Yue Qiao | Image data clustering and color conversion |
| CN103646354A (en) * | 2013-11-28 | 2014-03-19 | 国家电网公司 | Effective index FCM and RBF neural network-based substation load characteristic categorization method |
-
2015
- 2015-07-30 CN CN201510461151.1A patent/CN105139025B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5045240A (en) * | 1989-05-01 | 1991-09-03 | Westinghouse Electric Corp. | Contaminated soil restoration method |
| CN102338732A (en) * | 2011-07-06 | 2012-02-01 | 浙江大学 | Method for measuring fluidization parameter of gas-solid fluidized bed |
| CN103742901A (en) * | 2013-12-24 | 2014-04-23 | 广州市恒力安全检测技术有限公司 | Method for determining consumption difference analysis optimized target value system of circulating fluidized bed unit |
Non-Patent Citations (3)
| Title |
|---|
| 喷动流化床流动结构的SHANNON信息熵模糊聚类分析;钟文琪 等;《中国电机工程学报》;20050430;第25卷(第7期);第2.2节，第3节，第4.1节 |
| 基于现代非线性信息处理技术的气固流化床流型识别方法与实验研究;王晓萍;《中国优秀博硕士学位论文全文数据库(博士) 工程科技Ⅰ辑》;20040915(第3期);第4章 |
| 小波分析和神经网络理论及其在气-固流化床中的应用;周树孝;《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅰ辑》;20040315(第1期);第2.3.2.3节，第3.1节，第4.1节，第5.2节 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105139025A (en) | 2015-12-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107122790B (en) | Non-invasive load identification algorithm based on hybrid neural network and ensemble learning | |
| CN109685314B (en) | Non-intrusive load decomposition method and system based on long-term and short-term memory network | |
| Ullo et al. | Hybrid computerized method for environmental sound classification | |
| CN105678332A (en) | Converter steel-making endpoint determination method and system based on flame image CNN recognizing and modeling process | |
| CN112633317A (en) | CNN-LSTM fan fault prediction method and system based on attention mechanism | |
| CN108573225A (en) | A kind of local discharge signal mode identification method and system | |
| CN107480705A (en) | A kind of running water pipe leakage detection method | |
| CN110824586B (en) | Rainfall prediction method based on improved decision tree algorithm | |
| CN106847293A (en) | Facility cultivation sheep stress behavior acoustical signal monitoring method | |
| CN104751229A (en) | Bearing fault diagnosis method capable of recovering missing data of back propagation neural network estimation values | |
| CN106874950A (en) | A kind of method for identifying and classifying of transient power quality recorder data | |
| CN109061426B (en) | Transformer partial discharge fault diagnosis method and online monitoring device | |
| CN112183643B (en) | Hard rock tension-shear fracture identification method and device based on acoustic emission | |
| CN112149230B (en) | Method for predicting comfort deterioration of wind-induced train of strong wind railway | |
| CN110880369A (en) | Gas marker detection method based on radial basis function neural network and application | |
| CN102661759A (en) | Method for identifying and predicting nonlinear multivariable key parameters of circulating fluidized bed boiler | |
| CN112597702B (en) | Pneumatic modeling generation type confrontation network model training method based on radial basis function | |
| CN107657274A (en) | A kind of y-bend SVM tree unbalanced data industry Fault Classifications based on k means | |
| CN105139025B (en) | Gas-solid fluidized bed flow pattern online intelligent recognition method based on nonlinear analysis method | |
| CN108509601A (en) | A kind of flavour of food products assessment method based on big data analysis | |
| CN114429152A (en) | Rolling bearing fault diagnosis method based on dynamic index antagonism self-adaption | |
| CN110991689B (en) | Distributed photovoltaic power generation system short-term prediction method based on LSTM-Morlet model | |
| CN109376903B (en) | PM2.5 concentration value prediction method based on game neural network | |
| CN114510974A (en) | Intelligent identification method for gas-liquid two-phase flow pattern in porous medium | |
| CN106955097A (en) | A kind of fetal heart frequency state classification method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |