CN106015951B - A kind of gas pipeline leak detection system and method adapting to various states variation - Google Patents
A kind of gas pipeline leak detection system and method adapting to various states variation Download PDFInfo
- Publication number
- CN106015951B CN106015951B CN201610530135.8A CN201610530135A CN106015951B CN 106015951 B CN106015951 B CN 106015951B CN 201610530135 A CN201610530135 A CN 201610530135A CN 106015951 B CN106015951 B CN 106015951B
- Authority
- CN
- China
- Prior art keywords
- gas pipeline
- model
- gas
- pipeline
- neural network
- 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.)
- Expired - Fee Related
Links
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F17—STORING OR DISTRIBUTING GASES OR LIQUIDS
- F17D—PIPE-LINE SYSTEMS; PIPE-LINES
- F17D5/00—Protection or supervision of installations
- F17D5/02—Preventing, monitoring, or locating loss
- F17D5/06—Preventing, monitoring, or locating loss using electric or acoustic means
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Examining Or Testing Airtightness (AREA)
Abstract
It is a kind of adapt to various states variation gas pipeline leak detection system and method include it is a kind of adaptation various states variation gas pipeline leak detection system (abbreviation this system) and it is a kind of adapt to various states variation gas pipeline leakage detection method (abbreviation this method)；This system includes input module, gas pipeline transporting mechanism model, neural network compensation model, pipe-line system signal acquisition module, corrects output processing module and error acquisition module；This method is to establish gas pipeline mathematical model, the flow signal for obtaining gas pipeline state space equation, handling gas pipeline state space equation, establish RBF neural and verification detection model.In conjunction with gas pipeline flow mathematical model and Knowledge based engineering detection method, research end load, caliber, pipe bending, environment temperature, gas velocity of wave and the compressed coefficient, the pipeline hydraulic coefficient of friction resistance etc. simultaneously monitor gas pipeline leakage.Training pipeline model, obtains the accurate model for adapting to the variation of pipeline operating mode.
Description
Technical field
The present invention relates to a kind of the gas pipeline leak detection systems and method of the variation of adaptation various states, belong to automatic control
System and leak detection technology and gas pipe leakage detection field.
Background technology
After decades of development, domestic and international researcher proposes pipelines leak detection and localization method and applies
In engineering reality.Hardware based method is let out using the sensor of the various different mechanism carried or on pipeline to detect
It leaks and is positioned, this method is the method for the leakage detection and localization mainly used early stage.Currently, based on software, it is soft, hard
The leakage detection and localization technology acquired with data based on monitoring control that part is combined widely is paid close attention to, and is increasingly becoming
The mainstream of Discussion on Pipe Leakage Detection Technology.
Detection method based on software is measurement, quality and volume conservation, the pipeline stream by flow and pressure in pipeline
The principles such as dynamic kinetic model realize leak detection by computer programming.Method based on software can be divided into based at signal
Method, Knowledge based engineering method and the method based on pipeline mathematical model of reason.Method based on signal processing have negative pressure wave method,
Mass/volume balancing method, pressure spot analytic approach, acoustic wave detection etc.；Knowledge based engineering method have method based on expert system,
Method based on pattern-recognition, the method based on neural network；Pipeline leakage detection and location method based on mathematical model can
It is divided into transient model method, filter method, identification method etc..
It is entitled " a kind of pipeline gas leak detecting device " application No. is 201210563120.3；Application No. is
200920293127.1 entitled " a kind of pressure pipeline leakage testing device "；It is entitled application No. is 200620127803.4
" a kind of underground pipeline leak detection device "；Application No. is 200810104233.0, a kind of entitled " pipeline leakage testing system
System "；It is entitled " a kind of gas pipe leakage detection device " application No. is 201010002941.0；Application number
201220078512.6, entitled " leak source for gas pipeline detects positioning system " etc., these above-mentioned patents all stress
In design using the device of various kinds of sensors and instrument detection leakage, belong to hardware based Discussion on Pipe Leakage Detection Technology.Hardware
Detection technique has many advantages, such as to detect accurate, high sensitivity, and can determine the size of leak position and leakage rate, but hard
Part detection technique have the shortcomings that it is apparent, this is because hardware detection needs arrange that the hardware such as a large amount of sensor are set on pipeline
Standby, cost of installation and maintenance is very huge, and for the complexity of the detection hardware detection of buried pipeline by bigger, and current one
As not separately as pipe detection means, be generally mated software use.
Application No. is 201310169679.2, entitled " the gas pipeline leak detecting device based on sound wave signals and inspection
The patent of survey method " proposes, using computational fluid dynamics pressure signal in sound wave sensor acquisition pipeline, feature to be carried out to sound wave signals
The method extracted and carry out leakage judgement.But the collected dynamic pressure signal of sound wave sensor is fainter, easily by extraneous ring
Border is interfered, unsuitable teletransmission.
Application No. is 201310523173.7, the patent of entitled " a kind of pipe leakage experimental provision and experimental method ",
It proposes using instruments such as thermometer, volume flowmeter and pressure gauges, by comparing fluid steady state leakage rate calculations value and measurement
It is worth to fluid steady state leakage rate correction model, research fluid leakage feature and rule under different operating modes, are pipe-line maintenance
Foundation is provided with leak detection.
Application No. is 200610172271.0, entitled " detection method of the pipe leakage based on artificial neural network "
Patent, it is proposed that leak detection is carried out using neural network, using the stronger adaptability of neural network and predictive ability, but is lacked
Weary physical basis, to the data other than training sample, prediction effect is bad, sometimes even with actual industrial process contradiction.
Application No. is 201010004704.8, entitled " gas pipeline leakage detecting and positioning device and its detection positioning side
The patent of method " is proposed to establish the transient state mathematical model of gas flowing in pipe and is solved using the method for characteristic curves, using in model
Pressure, flow rate calculation value and pressure, the error of flow sensor measured value judge to leak, since mechanism model is in solution procedure
In need to carry out some and simplify and approximate, therefore error will necessarily be brought.
In conclusion above is referred to gas pipe leakage detection methods and techniques be mostly by pressure, flow, temperature,
The device and method that the various kinds of sensors such as sound wave detect signal and carry out signature analysis extraction leakage information, various leakage inspections
Survey method respectively has an advantage and disadvantage, and a kind of single leakage detection method is difficult to reach good effect in each performance indicator, such as
Fruit organically combines a variety of methods, learns from other's strong points to offset one's weaknesses, and will can improve the performance of leakage detection and localization.The application is based on god
Method through network and the method based on pipeline mathematical model are combined, and solve the problems, such as following two aspects：On the one hand, this Shen
Pressure signal please is acquired only with high-precision pressure sensor, reduces equipment initial investment, it is cost-effective；On the other hand, will have
There is the modelling of preferable physical basis to be combined with stronger adaptability and predictive ability neural network, improves gas pipeline
Leak detection precision.
The application is dedicated to " method based on gas pipeline flow mathematical model " together " Knowledge based engineering detection method "
Be combined, have complementary advantages, research gas pipeline end load variation, caliber changes, pipeline bends and environment
Detection model is suitable under the various states of Parameters variations such as temperature, gas velocity of wave, gas compressibility factor, the pipeline hydraulic coefficient of friction resistance
Ying Xing, the monitoring for gas pipeline and leak detection propose a kind of modeling of gas pipeline and leakage detection method.
Invention content
The application aim to overcome that hardware based gas pipeline leak detection technology is of high cost, operation and maintenance it is difficult with
And the present situation poor based on traditional gas pipeline transporting mechanism Model suitability, precision is low, propose that a kind of adaptation various states become
The gas pipeline leak detection system and method for change.
The application proposes a kind of the gas pipeline leak detection system and method for the variation of adaptation various states, including a kind of suitable
Answer the gas pipeline leak detection system (abbreviation this system) and a kind of gas transmission adapting to various states variation that various states change
Pipeline leakage detection method (abbreviation this method)；
Wherein, this system includes：Input module, gas pipeline, gas pipeline transporting mechanism model, neural network compensate mould
Type, corrects output processing module and error acquisition module at pipe-line system signal acquisition module；
Wherein, the input module includes real system input, gas pipeline transporting mechanism mode input and nerve net
Network compensation model inputs；
Wherein, system input, gas pipeline transporting mechanism mode input and neural network compensation model input three can be with
It is identical, it can also be different；Real system input is the parameters of the primary condition and boundary condition and gas pipeline of pipeline；
Wherein, the gas pipeline transporting mechanism model (follow-up abbreviation " mechanism model "), the process of foundation is：To gas
The continuity equation and the equation of motion that body transmits in gas pipeline are derived and are simplified, and gas pipeline stable state and dynamic are established
Mathematical model, and the mathematical model of gas pipeline non-linear partial differential form is solved, establish gas pipeline state space
Equation, and flow signal in model is handled；
Wherein, the neural network compensation model (follow-up abbreviation " neural network "), the process of foundation is：Due to RBF
(Radial Basis Function, RBF) existing biological context of neural network can approach non-linear letter with arbitrary accuracy again
Number, as long as and central point selection it is appropriate, it is only necessary to seldom neuron is achieved with good Approximation effect, also has unique
The advantages of best approximate point, the application choose RBF neural, determine network topology structure, network inputs vector sum target to
Amount establishes neural network model compensation；And network training and test are carried out in the case where loading different operating modes when changing, compensation is defeated
Feed channel transporting mechanism model is because of error caused by various states variation；
Wherein, the pipe-line system signal acquisition module refers to using the high-precision being mounted on along pipeline on each node
Pressure sensor, in gas pipeline transmission direction each node pressure signal be acquired, that is, obtain this system in pressure
Signal measured value；
Wherein, the amendment output processing module refers to by the calculating output valve of gas pipeline transporting mechanism model and god
Output through network building out model is summed：Wherein, mechanism model is master cast, and neural network is compensation model, for mending
Error existing for gas pipeline transporting mechanism model is repaid, master cast and compensation model combine, and establish the mixed model of gas pipeline,
Gas pipeline operating status can preferably be monitored and judge the generation of leakage, to effectively improve accuracy of detection；
Wherein, the error acquisition module calculates each node pressure sensor measured value of gas pipeline and mechanism model meter
Calculate the error between output valve, output vector of the error as neural metwork training；Error acquisition module also calculates pressure letter
Error after number measured value and compensation between the mixed model output valve of gas pipeline, the error sum of squares minimum is as network instruction
Experienced desired value, the value is close to 0 or the smaller the better；
A kind of connection relation for each module of gas pipeline leak detection system adapting to various states variation is as follows：
Input module and gas pipeline transporting mechanism model, neural network compensation model and pipe-line system signal acquisition module
It is connected, corrects output processing module and gas pipeline transporting mechanism model, neural network compensation model, pipe signal acquisition module
It is connected with error acquisition module, error acquisition module is connected with pipe signal acquisition module and amendment output processing module.
A kind of gas pipeline leakage detection method adapting to various states variation, i.e. this method are more based on a kind of adaptation
The gas pipeline leak detection system of kind state change carries out gas pipeline leak detection and compensation, is as follows：
Step 1: in view of installation flowmeter can destroy former flow field and influence measuring accuracy, loss can be also caused stress,
When building a kind of gas pipeline leak detection system that adaptation various states change, replaced using the higher pressure sensor of precision
Flowmeter in conventional pipeline leakage checking system, the unified pressure signal acquired on each node of gas pipeline；
Step 2: using the pressure value on the collected each node of initial time gas pipeline of step 1 as primary condition,
Gas pipeline first and last side pressure force snesor is in collected pressure value of each moment as boundary condition, primary condition and perimeter strip
The parameters of part and pipeline are as input condition known to this system；
Step 3: establishing gas pipeline transmission based on gas pipeline transmission continuity equation, the equation of motion and state equation
Mechanism model is as follows：
Step 3.1 derives and simplifies the continuity equation and the equation of motion of processing gas pipeline transporting mechanism model, and ties
The equation of gas state is closed, the simplification mathematical model of the dynamic partial differential equation form of gas pipeline is established；
Wherein, the continuity equation is：
Wherein, the equation of motion is：
Wherein, p --- gas pressure in gas pipeline, Pa；
qm--- gas pipeline mass flow, kg/s；
T --- time variable, s；
X --- along the length variable of gas pipeline, m；
A --- gas pipeline cross-sectional area, m2；
A --- gas velocity of wave, m/s；
G --- acceleration of gravity, m/s2；
θ --- angle between gas pipeline axis direction and horizontal plane, rad；
λ --- gas pipeline hydraulic simulation experiment, dimensionless；
D --- gas pipeline internal diameter, m；
Step 3.2 solves the simplification mathematical model that step 3.1 is established, specially：Based on center implicit difference method
Differencing is carried out to gas pipeline dynamical equation, obtains gas pipeline state space equation；
Wherein, the gas pipeline state space equation is：
X (j)=[p in formula1, j... pN, j, qM1, j... qMn, j]T,For constant coefficient matrix；With this
The function of system mode vector；It is item related with boundary condition, it is assumed herein that it is known that head end perimeter strip
Part X0；
Flow signal flows formula using gas in the gas pipeline state space equation that step 3.3 obtains step 3.2
Calculating method, 2 pressure differential methods and approximate model method are handled；
Step 4: establishing neural network compensation model, and compensate because of the approximate and simplified error brought to mechanism model,
It is as follows：
The advantage and disadvantage of the more various artificial neural networks of step 4.1 simultaneously select RBF neural to establish gas pipeline
Neural network compensation model；
Step 4.2 establishes the neural network compensation model of gas pipeline, concrete structure on the basis of step 4.1：Specifically
Using three layers of RBF neural containing input layer, hidden layer and output layer, and determine the input vector of neural network compensation model
For the calculation of pressure value of n node on j moment gas pipelines：P i.e. in step 3.21, j.…pN, j, output vector is the j+1 moment
Pressure divergence on gas pipeline on each node between actual measured value and mechanism model calculated value, hidden layer use Gaussian kernel letter
Number is used as radial basis function；
Step 4.3 is more prone to for the training and study that make neural network compensation model, to the god determined in step 4.2
Input vector and output vector through network building out model carry out data normalization processing, are transformed to [- 1,1] or [0,1]
In the range of, difficulty when RBF neural training is mitigated with this, and avoid because of inputoutput data order of magnitude difference
It is larger and cause neural network forecast error larger；
Step 4.4 is further designed RBF neural on the basis of step 4.2 and 4.3, is embodied in：?
RBF neural is trained and is tested under the changed a variety of different operating modes of end gas consumption of gas pipeline；
Specifically, writing .m file routines using MATLAB development environments, a kind of adaptation various states of the present invention are changed
Gas pipeline leak detection system carry out specific implementation and experimental verification, verification under different operating modes this detecting system it is accurate
Property, and the pressure data of abnormal conditions is tested, analyze basis for estimation when there is leakage in gas pipeline.
It is to complete a kind of gas pipeline leakage detection method of adaptation various states variation by aforementioned four step.
Advantageous effect
A kind of gas pipeline leak detection system and method adapting to various states variation proposed by the present invention, has as follows
Advantageous effect：
1. the gas pipeline leak detection system and method for a kind of adaptation various states variation of the present invention, only with pressure
Signal establishes the gas pipeline leak detection system and method for adapting to multimode variation, therefore only needs to install high-precision pressure sensing
Device acquires pressure signal, further saves initial investment and operation maintenance cost；
2. the gas pipeline leak detection system and method for a kind of adaptation various states variation of the present invention, by gas transmission
Pipeline continuity equation and the equation of motion carry out theory analysis, obtain a pair of of nonlinear partial differential equation, implicitly poor using center
Point-score carries out numerical solution to the simplified model, establishes gas pipeline No leakage state space equation, can be calculated based on this equation
The pressure for going out upper each node all moment along pipeline, reference frame is provided for leak detection, can be found in time and accurate
Positioning leak position, estimation leakage rate are simultaneously alarmed, to more efficiently monitor gas pipeline operation conditions；
3. the gas pipeline leak detection system and method for a kind of adaptation various states variation of the present invention, are managed by determining
Network topology structure and network inputs vector sum object vector of the road in No leakage compensate pipe machine using neural network
The variation of various states in model is managed, and network is trained and is tested under the changed various working of end load,
Gas pipeline RBF neural compensation model is established, it can error existing for effective compensation mechanism model；
4. a kind of gas pipeline leakage detection method of adaptation various states variation of the present invention, uses in gas pipeline
Mechanism model and neural network are combined, and are established mixed model using MATLAB environment and carried out specific implementation and theoretical simulation,
The variation that system proposed by the invention is suitable for pipeline various states and load is demonstrated, there is higher precision, it can be accurate
The operating status of gas pipeline is monitored, and can be judged rapidly when pipeline leaks, to realize that efficient gas pipeline is let out
Leakage positioning and detection.
Description of the drawings
Fig. 1 is a kind of gas pipeline leak detection system adapting to various states variation of the present invention and the system composition of method
Schematic diagram；
Fig. 2 is in a kind of gas pipeline leak detection system adapting to various states variation of the present invention and embodiment of the method
Gas pipeline RBF neural network structure figure；
Fig. 3 is real in a kind of gas pipeline leak detection system adapting to various states variation of the present invention and embodiment of the method
Test pipeline pressure sampled point and simulated leakage position schematic diagram；
Fig. 4 is in a kind of gas pipeline leak detection system adapting to various states variation of the present invention and embodiment of the method
It is compared between mixed model output valve, mechanism model calculated value and measured value when 2 each node test RBF neural of operating mode
Result figure.
Specific implementation mode
The present invention will be further described with reference to the accompanying drawings and examples and detailed description：
Fig. 1 is that a kind of system composition adapting to gas pipeline leak detection system and method that multimode changes of the present invention shows
It is intended to；RBF neural compensation model structure chart when Fig. 2 is gas pipeline No leakage proposed by the present invention, from Figure 2 it can be seen that should
RBF neural uses Three Tiered Network Architecture, is input layer, hidden layer and output layer, wherein network model input vector respectively
For the calculation of pressure value of n node on j moment pipelines, output vector be on j+1 moment pipelines on each node actual measured value with
Pressure divergence between mechanism model calculated value, hidden layer is using gaussian kernel function as radial basis function.
In a kind of gas pipeline leak detection system adapting to multimode variation proposed by the present invention, mechanism model is to pass through
What the continuity equation and the equation of motion that gas transmits in gas pipeline were set up, it is master cast, serves main；
RBF neural rubs to the angle of temperature, caliber, pipe bending, pipeline and horizontal direction, waterpower as model of error estimate
The dynamic change of the various states parameters such as resistance coefficient, gas compressibility factor is compensated to the error that mechanism model is brought, to carry
High mixed model precision plays fine tuning, and is trained to pipeline model under different operating modes, and acquisition is being adapted to pipeline just
The accurate model of normal operating mode variation.
Embodiment
Gas pipeline in a kind of gas pipeline leak detection system of adaptation multimode variation of the present embodiment is internal diameter
The steel pipe of 10mm, length 75m, wherein air source are air compressor, can generate the pressure that pressure limit is 0~0.7Mpa (gauge pressure)
Contracting gas, many places installation high-precision pressure sensor acquires pressure signal on gas pipeline, and is simulated along pipeline different location
The different leak of inner diameter size, gas pipeline end, which passes through, installs throttle valve adjustment flow simulation difference operating mode, throttle valve
It is followed by muffler and leads to air.A kind of gas pipeline leak detection system of adaptation multimode variation of the present embodiment simulates altogether
Five kinds of operating modes, end flow are 343.3,347.5,355.1,368.4,372.3, and unit is g/min (gram/minute), respectively
Operating mode 1, operating mode 2, operating mode 3, operating mode 4 and operating mode 5.When establishing RBF neural, using the data of operating mode 1,3,5 as training
Data in algorithm are used for Training RBF Neural Network, using 4 test network state of operating mode 2 and operating mode.Specific pressure acquisition position
It is as shown in Figure 3 to set (S1-S7) and simulated leakage position (L1-L3).
Operating mode 2 tests RBF neural mixed model output valve (correction value), mechanism model output valve (calculated value) and pressure
Comparing result between force snesor observed pressure data (measured value) on each force samples point is as shown in Figure 4.1st section
Point is head end boundary condition, shows the 2nd node to the 7th node pressure data in Fig. 4.
As can be seen from Figure 4 in addition on the 2nd point of gas pipeline individual sampled point mixed model errors missed than mechanism model
Difference is slightly bigger outer, and mixed model error will be less than mechanism model error on all sampled points on other each points, and explanation is built
Vertical compensation model plays good adjustment effect.Mechanism model error (except head end) on each force samples point on pipeline
(Error of Mechanism model, abbreviation EM) and mixed model error (Error of Hybrid model, abbreviation EH)
As shown in table 1 below, unit is Pa for the position (i) of maximum value and minimum value and corresponding sampled point (moment).It can from table 1
To find out, there is certain errors between actual measured value on node 3,4,5,6 for mechanism model calculation of pressure value, and mix
Error very little between molding type pressure prediction value and measured value, even less than pressure sensor measurement error is (in experimental system
Pressure sensor maximum error of measuring is in 1220Pa or so), precision of prediction is very high.
1 operating mode of table, 2 mechanism model and the comparison of mixed model error
Embodiment described above be the present invention it is a kind of adapt to various states variation gas pipeline leak detection system and
The preferred embodiment of method, the present invention should not be limited to the embodiment and attached drawing disclosure of that.It is every not depart from
The lower equivalent or modification completed of spirit disclosed in this invention, both falls within the scope of protection of the invention.
Claims (4)
1. a kind of gas pipeline leak detection system adapting to various states variation, it is characterised in that：Including：It is input module, defeated
Feed channel, neural network compensation model, pipe-line system signal acquisition module, is corrected at output gas pipeline transporting mechanism model
Manage module and error acquisition module；
The input module includes real system input, gas pipeline transporting mechanism mode input and neural network compensation model
The information of three kinds of inputs of input；
Wherein, real system input, gas pipeline transporting mechanism mode input and neural network compensation model input three can be with
It is identical, it can also be different；Real system input is the parameters of the primary condition and boundary condition and gas pipeline of pipeline；
The gas pipeline transporting mechanism model, follow-up abbreviation " mechanism model ", the process of foundation are：To gas in appendix
The continuity equation and the equation of motion transmitted in road is derived and is simplified, and gas pipeline stable state and dynamic mathematical models are established,
And the mathematical model of gas pipeline non-linear partial differential form is solved, gas pipeline state space equation is established, and right
Flow signal is handled in model；
The neural network compensation model, follow-up abbreviation " neural network ", the process of foundation are：RBF neural is chosen, really
Determine network topology structure, network inputs vector sum object vector, establishes neural network model compensation；And it is loading when changing
Network training and test are carried out under different operating modes, compensation gas pipeline transporting mechanism model is because caused by various states variation
Error；
The pipe-line system signal acquisition module uses the high-precision pressure sensor being mounted on along pipeline on each node, right
Each node pressure signal is acquired in gas pipeline transmission direction, obtains pressure signal measured value in this system；
The calculating output valve of gas pipeline transporting mechanism model and neural network are compensated mould by the amendment output processing module
The output of type is summed：
Wherein, mechanism model is master cast, and neural network is compensation model, is existed for compensating gas pipeline transporting mechanism model
Error, master cast and compensation model combine, establish the mixed model of gas pipeline, can preferably monitor gas pipeline operation
State and the generation for judging leakage, to effectively improve accuracy of detection；
Wherein, the error acquisition module calculates each node pressure sensor measured value of gas pipeline and mechanism model calculating is defeated
Go out the error between value, output vector of the error as neural metwork training；Error acquisition module also calculates pressure signal reality
Error after measured value and compensation between the mixed model output valve of gas pipeline, the error sum of squares minimum is as network training
Desired value, the value is close to 0 or the smaller the better；
The connection relation of each module is as follows：
Input module and gas pipeline transporting mechanism model, neural network compensation model and pipe-line system signal acquisition module phase
Even, correct output processing module and gas pipeline transporting mechanism model, neural network compensation model, pipe signal acquisition module and
Error acquisition module is connected, and error acquisition module is connected with pipe signal acquisition module and amendment output processing module.
2. a kind of gas pipeline leakage detection method adapting to various states variation, it is characterised in that：
It is as follows：
Step 1: when building a kind of gas pipeline leak detection system that adaptation various states change, it is higher using precision
Flowmeter in the conventional pipeline leakage checking system of pressure sensor substitution, the unified pressure acquired on each node of gas pipeline
Force signal；
Step 2: using the pressure value on the collected each node of initial time gas pipeline of step 1 as primary condition, gas transmission
Pipeline first and last side pressure force snesor collected pressure value of each moment as boundary condition, primary condition and boundary condition with
And the parameters of pipeline are as input condition known to this system；
Step 3: establishing gas pipeline transporting mechanism based on gas pipeline transmission continuity equation, the equation of motion and state equation
Model；
Step 4: establishing neural network compensation model, and compensate because of the approximate and simplified error brought to mechanism model,
It is as follows：
The advantage and disadvantage of the more various artificial neural networks of step 4.1 simultaneously select RBF neural to establish the nerve of gas pipeline
Network building out model；
Step 4.2 establishes the neural network compensation model of gas pipeline, concrete structure on the basis of step 4.1：It is specific to use
Three layers of RBF neural containing input layer, hidden layer and output layer, and determine that the input vector of neural network compensation model is j
The calculation of pressure value of n node on moment gas pipeline：P i.e. in step 3.21, j... pN, j, output vector is that the j+1 moment is defeated
Pressure divergence in feed channel on each node between actual measured value and mechanism model calculated value, hidden layer use gaussian kernel function
As radial basis function；
Step 4.3 is more prone to for the training and study that make neural network compensation model, to the nerve net determined in step 4.2
The input vector and output vector of network compensation model carry out data normalization processing, are transformed to the model of [- 1,1] or [0,1]
In enclosing, difficulty when RBF neural training is mitigated with this, and avoid because inputoutput data order of magnitude difference is larger
And cause neural network forecast error larger；
Step 4.4 is further designed RBF neural on the basis of step 4.2 and 4.3, is embodied in：In gas transmission
RBF neural is trained and is tested under the changed a variety of different operating modes of end gas consumption of pipeline.
3. a kind of gas pipeline leakage detection method adapting to various states variation as claimed in claim 2, it is characterised in that：
The foundation of step 3 gas pipeline transporting mechanism model, is as follows：
Step 3.1 derives and simplifies the continuity equation and the equation of motion of processing gas pipeline transporting mechanism model, and combines gas
Body state equation establishes the simplification mathematical model of the dynamic partial differential equation form of gas pipeline；
Wherein, the continuity equation is：
Wherein, the equation of motion is：
Wherein, p --- gas pressure in gas pipeline, Pa；
qm--- gas pipeline mass flow, kg/s；
T --- time variable, s；
X --- along the length variable of gas pipeline, m；
A --- gas pipeline cross-sectional area, m2；
A --- gas velocity of wave, m/s；
G --- acceleration of gravity, m/s2；
θ --- angle between gas pipeline axis direction and horizontal plane, rad；
λ --- gas pipeline hydraulic simulation experiment, dimensionless；
D --- gas pipeline internal diameter, m；
Step 3.2 solves the simplification mathematical model that step 3.1 is established, specially：Based on center implicit difference method to defeated
Feed channel dynamical equation carries out differencing, obtains gas pipeline state space equation；
Wherein, the gas pipeline state space equation is：
X (j)=[p in formula1, j... pN, j, qM1, j... qMn, j]T,For constant coefficient matrix；It is with this
The function for state vector of uniting；It is item related with boundary condition, it is assumed herein that it is known that head end boundary condition
X0；
Flow signal is calculated using gas flowing formula in the gas pipeline state space equation that step 3.3 obtains step 3.2
Method, 2 pressure differential methods and approximate model method are handled.
4. a kind of gas pipeline leakage detection method adapting to various states variation as claimed in claim 2, feature also exist
In：Step 4.4 is designed and realizes to RBF neural on the basis of step 4.2 and 4.3, further, specifically, using
MATLAB development environments write .m file routines, the gas pipeline leak detection system for adapting to various states variation a kind of to the present invention
System carries out specific implementation and experimental verification, verifies the accuracy of this detecting system under different operating modes, and to the pressure of abnormal conditions
Force data is tested, and basis for estimation when there is leakage in gas pipeline is analyzed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610530135.8A CN106015951B (en) | 2016-07-07 | 2016-07-07 | A kind of gas pipeline leak detection system and method adapting to various states variation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610530135.8A CN106015951B (en) | 2016-07-07 | 2016-07-07 | A kind of gas pipeline leak detection system and method adapting to various states variation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106015951A CN106015951A (en) | 2016-10-12 |
| CN106015951B true CN106015951B (en) | 2018-11-09 |
Family
ID=57107490
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610530135.8A Expired - Fee Related CN106015951B (en) | 2016-07-07 | 2016-07-07 | A kind of gas pipeline leak detection system and method adapting to various states variation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106015951B (en) |
Families Citing this family (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108758357B (en) * | 2018-06-01 | 2019-06-11 | 浙江大学 | A kind of localization of bursted pipe method based on the analysis of water supply network change value of pressure |
| CN109443456B (en) * | 2018-10-31 | 2021-01-12 | 华北电力大学（保定） | Flow measuring method and device |
| BE1026849B1 (en) * | 2018-12-07 | 2020-07-09 | Atlas Copco Airpower Nv | Gas network and method for simultaneously detecting leaks and obstructions in a gas network under pressure or under vacuum |
| CN109783972B (en) * | 2019-01-30 | 2022-11-08 | 河北工业大学 | Method for monitoring leakage flow in check valve based on fluid-solid coupling analysis and calculation |
| US11704458B2 (en) * | 2019-10-10 | 2023-07-18 | Voda, Inc. | System and method for optimizing utility pipe sensors placement using artificial intelligence technology |
| CN111022941B (en) * | 2019-12-27 | 2021-05-28 | 重庆科技学院 | Natural gas pipeline leakage detection method based on LSTM recurrent neural network |
| CN111237646B (en) * | 2020-02-17 | 2021-03-02 | 清华大学 | Automatic identification and positioning method for leakage of water supply pipe network |
| CN112212222A (en) * | 2020-09-09 | 2021-01-12 | 黄腾飞 | Method and system for detecting pipeline leakage |
| CN114508704B (en) * | 2020-11-16 | 2024-04-30 | 中国石油天然气股份有限公司 | Pipeline leakage detection method and device and storage medium |
| CN112665790B (en) * | 2020-12-21 | 2023-03-14 | 广东美的暖通设备有限公司 | Refrigerant leakage detection method, device and equipment |
| CN112861425A (en) * | 2021-01-13 | 2021-05-28 | 上海交通大学 | Method for detecting performance state of double-shaft gas turbine by combining mechanism and neural network |
| CN113608517A (en) * | 2021-07-20 | 2021-11-05 | 大唐互联科技(武汉)有限公司 | Mechanism model diagnosis system based on Internet of things |
| CN113551744A (en) * | 2021-08-10 | 2021-10-26 | 中国石油大学(北京) | Ultrasonic flowmeter performance online monitoring method and system |
| CN113803647B (en) * | 2021-08-25 | 2023-07-04 | 浙江工业大学 | Pipeline leakage detection method based on fusion of knowledge features and hybrid model |
| CN115111537B (en) * | 2022-08-24 | 2022-11-18 | 北京云庐科技有限公司 | Method, device and medium for determining the position of a leak in a gas pipeline network |
| CN116382100B (en) * | 2023-06-05 | 2023-08-01 | 中特检管道工程（北京）有限公司 | Oil and gas pipeline detection control system and control method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101761780A (en) * | 2010-01-11 | 2010-06-30 | 中国石油大学(华东) | Gas pipeline leakage detecting and positioning device and method thereof |
| CN101943324A (en) * | 2010-06-28 | 2011-01-12 | 东北大学 | Weak signal detection device and method based on wavelets and RBF neural network |
| CN103441499A (en) * | 2013-07-24 | 2013-12-11 | 河海大学常州校区 | Linearization feedback neural sliding-mode control method for three-phase parallel-connection active power filter |
| CN204153470U (en) * | 2014-09-09 | 2015-02-11 | 青岛厚科信息工程有限公司 | Based on the gas pipeline supervisory system of infrasonic sensor and pressure transducer |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014142825A1 (en) * | 2013-03-13 | 2014-09-18 | Bp Corporation North America Inc. | Virtual in-line inspection of wall loss due to corrosion in a pipeline |
-
2016
- 2016-07-07 CN CN201610530135.8A patent/CN106015951B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101761780A (en) * | 2010-01-11 | 2010-06-30 | 中国石油大学(华东) | Gas pipeline leakage detecting and positioning device and method thereof |
| CN101943324A (en) * | 2010-06-28 | 2011-01-12 | 东北大学 | Weak signal detection device and method based on wavelets and RBF neural network |
| CN103441499A (en) * | 2013-07-24 | 2013-12-11 | 河海大学常州校区 | Linearization feedback neural sliding-mode control method for three-phase parallel-connection active power filter |
| CN204153470U (en) * | 2014-09-09 | 2015-02-11 | 青岛厚科信息工程有限公司 | Based on the gas pipeline supervisory system of infrasonic sensor and pressure transducer |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106015951A (en) | 2016-10-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106015951B (en) | A kind of gas pipeline leak detection system and method adapting to various states variation | |
| CN206130547U (en) | Gas transmission pipeline leak testing system under multiplex condition | |
| US20130066568A1 (en) | Integrated system with acoustic technology, mass imbalance and neural network for detecting, locating and quantifying leaks in ducts | |
| Ostapkowicz | Leak detection in liquid transmission pipelines using simplified pressure analysis techniques employing a minimum of standard and non-standard measuring devices | |
| CN110197049B (en) | Non-metal pipeline leakage positioning method based on transient inverse problem | |
| CN105840987B (en) | It is a kind of that localization method and device are weighted based on the pipe leakage of pressure wave and sound wave | |
| Wang et al. | Experimental study on water pipeline leak using In-Pipe acoustic signal analysis and artificial neural network prediction | |
| CN108050394B (en) | Gas pipeline leakage detection positioning experiment platform based on sound pressure signal identification | |
| Lang et al. | A multiple leaks’ localization method in a pipeline based on change in the sound velocity | |
| CN103775832B (en) | Based on the device that the petroleum pipeline leakage of transient flow Inverse Problem Method detects | |
| CN101761780A (en) | Gas pipeline leakage detecting and positioning device and method thereof | |
| CN112628613A (en) | Method and system for monitoring pipeline leakage, positioning leakage and calculating leakage amount | |
| CN107013812A (en) | A kind of THM coupling line leakage method | |
| Ravula et al. | Experimental validation of leak and water-ingression detection in low-pressure gas pipeline using pressure and flow measurements | |
| Jung et al. | Systematic exploration of pipeline network calibration using transients | |
| Fu et al. | Identifying two-point leakages in parallel pipelines based on flow parameter analysis | |
| JP7374191B2 (en) | Gas networks and methods for simultaneous detection of leaks and faults in gas networks under pressure or vacuum | |
| Ostapkowicz et al. | Leak detection in liquid transmission pipelines during transient state related to a change of operating point | |
| CN113227642B (en) | Method for detecting leaks in a gas network under pressure or vacuum and gas network | |
| CN201100852Y (en) | A portable hydraulic failure diagnosis instrument | |
| CN114811448B (en) | Method for pipeline leakage detection, leakage flow velocity estimation and leakage positioning under flowing condition | |
| CN112912663B (en) | Gas pipe network and method for detecting blockage in gas pipe network under pressure or vacuum | |
| CN112393125B (en) | Gas pipe network leakage detection system and method based on inverse problem analysis | |
| CN110750756B (en) | Real-time on-line instrument checksum diagnosis method through optimal support vector machine algorithm | |
| Günther et al. | Experimental setup to examine leakage outflow in a scaled water distribution network |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20181109 Termination date: 20190707 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |