CN107013812B - A kind of THM coupling line leakage method - Google Patents
A kind of THM coupling line leakage method Download PDFInfo
- Publication number
- CN107013812B CN107013812B CN201710310608.8A CN201710310608A CN107013812B CN 107013812 B CN107013812 B CN 107013812B CN 201710310608 A CN201710310608 A CN 201710310608A CN 107013812 B CN107013812 B CN 107013812B
- Authority
- CN
- China
- Prior art keywords
- leakage
- pipeline
- field
- optical fiber
- parameter
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Examining Or Testing Airtightness (AREA)
Abstract
The invention discloses a kind of THM coupling line leakage method, include the following steps: to construct pipeline THM coupling sensor-based system, the simulation of pipeline THM coupling sensor-based system monitoring tested pipeline light condition, the simulation of pipeline THM coupling sensor-based system monitoring tested pipeline nominal situation, the simulation of pipeline THM coupling sensor-based system incident of leakage, Monitoring Pinpelines neural net model establishing and study, line leakage.The method of the object of the present invention is to provide the monitorings of pipe leakage report alert, positioning and leak sizes judgement, it is connected each other by acquiring three field parameters of pipeline and establishing between detection parameters, purpose is that wrong report can be efficiently reduced, avoids failing to report, leakage point is accurately positioned, and the size of leakage is provided by neural network algorithm, reliable basis is provided to formulate maintenance program.
Description
Technical field
The invention belongs to Monitoring Pinpelines technical fields, relate particularly to a kind of THM coupling line leakage method.
Background technique
The underground of modern city is based on various demands, has buried various metallic conduits, wherein comprising being largely used to heat supply
Large-scale pipeline, as a pipe network of heating power supply, for being made to enterprise and domestic heat since urban size is wide, demand is big
Complicated at the pipe-line of provided underground, conveying distance is long, and length of embedment is deeper and deeper, once leakage will make society and property
At heavy losses and influence.
Since underground environment is complicated and changeable, the heat distribution pipeline damaged by largely observing leakage is found due to expanding with heat and contract with cold,
Hydraulic pressure impact, the factors such as soil corrosion eventually lead to the leakage of metallic conduit fatigue cracking, substantially reduce pipeline and use the longevity
Life.In current pipeline leakage detection method, pressure sensor sectional monitoring pipeline morning water inlet is generallyd use, with water outlet
Pressure change, by extract characteristic value, and combine neural network model simulation method detection leakage and positioning, but due to pass
Sensor distributed quantity is limited, and monitoring data are relatively simple, limited sample size, and the model error trained is excessive, often results in
Wrong report is failed to report, and positioning accuracy error is excessive, causes maintenance difficult, and is difficult estimation leakage size, it is difficult to maintenance project is formulated,
It is coped with according to the actual situation after can only excavating, substantially prolong maintenance time and increases maintenance cost.
Summary of the invention
The method of the object of the present invention is to provide the monitorings of pipe leakage report alert, positioning and leak sizes judgement, passes through acquisition
Three field parameters of pipeline and establishing between detection parameters connect each other, it is therefore intended that can efficiently reduce wrong report, avoid leaking
Report, be accurately positioned leakage point, and by neural network algorithm provide leakage size, for formulation maintenance program provide reliably according to
According to.
To achieve the above object, the technical scheme is that
A kind of THM coupling line leakage method, includes the following steps:
Step 1, building pipeline THM coupling sensor-based system, including temperature field detection fiber system, moisture field detection fiber
System, stress field detection fiber system, and determine relative mounting positions parameter；
Step 2, the monitoring tested pipeline light condition simulation of pipeline THM coupling sensor-based system, and output parameter one is recorded,
As sample data one；
Step 3, the monitoring tested pipeline nominal situation simulation of pipeline THM coupling sensor-based system, and output parameter two is recorded,
As sample data two；
Step 4, the simulation of pipeline THM coupling sensor-based system incident of leakage, and record simulation input parameter and output parameter
Three, as sample data three；
Step 5, Monitoring Pinpelines neural net model establishing and study；
Step 6, line leakage: the real-time output parameter in pipeline THM coupling sensor-based system monitoring process is inputted
Into trained Monitoring Pinpelines neural network, Monitoring Pinpelines neural network voluntarily judges and exports monitoring result；
Wherein, neural network structure is modeled in step 5, the simulation dynamic annealing process of single neuron uses formula (1-
2) it indicates,
V (t+1)=kv (t)-E+T (t) (u (t)-I0) (2)
In formula:
U is the transient state connection weight for inputting neuron and output neuron；
V is the internal state connection weight for inputting neuron and output neuron；
I0For neuron input deviation；
K is neu damping factor (0≤k≤1)；
E is the objective function of neural network；
ε is the decay factor of output function；
T (t) is self-feedback connection weights weight；
Annealing strategy is based on equation (3-7) and is applied in trained network, this network model is n defeated in input layer
Ingress, c output node and n x c connection weight, the mathematic(al) representation of the model are as follows:
vx；j(t+1)=kvx；j(t)+E-T(t)(ux；j(t)-I0) (5)
Δwj=η (zx-wj)ux；j (6)
In formula, E is the energy function of n input and c output node network, ux；jAnd vx；jRespectively transient state connection weight and interior
Portion's state connection weight.
Further, step 1 specifically comprises the following steps:
Step 101: pipeline THM coupling sensing system hardware parameter determines, in which:
Temperature field detection fiber system, moisture field detection fiber system, stress field detection fiber system are all made of distribution
Fibre optical sensor is respectively used to temperature field, the stress field, moisture field delta data of measurement tested pipeline；
THM coupling sensor-based system includes temperature detection optical fiber, Humidity Detection optical fiber, stress mornitoring optical fiber, photoelectric conversion dress
It sets, signal processing apparatus and central processing unit；
According to tested pipeline length, diameter and monitoring accuracy temperature detection fiber, Humidity Detection optical fiber, stress inspection
Survey the length of optical fiber；Coiling length determines the accuracy of detection accuracy and leakage positioning.
Step 102: the installation of pipeline THM coupling sensor-based system, in which:
Temperature detection optical fiber, Humidity Detection optical fiber and stress mornitoring light are laid in a manner of spiral winding outside tested pipeline
Fibre, and the length of temperature detection optical fiber, Humidity Detection optical fiber, the length of stress mornitoring optical fiber and tested pipeline is recorded respectively；
The input terminal of temperature detection optical fiber, Humidity Detection optical fiber, stress mornitoring optical fiber is connected with photoelectric conversion device respectively
It connects, output end is connect with signal processing apparatus respectively, and the output end of signal processing apparatus is connect with central processing unit.
Further, temperature detection optical fiber, Humidity Detection optical fiber and stress mornitoring optical fiber are with 120 ° of interval spiral windings
In tested pipeline outer wall.
Further, in step 2 specifically includes the following steps:
Step 201: tested pipeline processing: excluding internal flow and stand to arrive working temperature；
Step 202: starting and debug THM coupling sensor-based system；
Step 203: after system output parameter is stablized, testing and record output parameter one, wherein output parameter one includes
Temperature field initial signal, moisture field initial signal and stress field initial signal.
Further, step 3 specifically comprises the following steps:
Step 301: tested pipeline water filling reaches normal discharge and pressure；
Step 302: measurement output parameter two simultaneously records；One group of output parameter two is recorded at interval of the set time, it is continuous to survey
Amount record 24-48 hours；Output parameter two be same time point under record temperature field working condition signal, moisture field working condition signal with
And stress field working condition signal；
Step 303: by collected multiple groups temperature field working condition signal, moisture field working condition signal and stress field working condition signal
Tabulation record, it is spare as reference data.
Further, step 4 specifically comprises the following steps:
Step 401: determine that leakiness simulates grade: use simulates pipe with liquid identical in tested pipeline on pipeline
Road leakage, fluid flow are divided into multiple grades from small to large, respectively correspond different leakage class；
Step 402: determining leakage simulation dot density: leakage simulation point, tunnelling ray are set in pipeline external surface uniform array
Quasi- dot density is arranged according to detection accuracy；
Step 403: incident of leakage simulation: successively carrying out same levels fluid flow drippage behaviour on each leakage simulation point
Make, and record corresponds to the output parameter three of each leakage simulation point respectively；
After first grade fluid flow simulation, tested pipeline is handled, until carrying out after three output parameters are stablized
Second wheel is different from first grade fluid flow drippage simulation, and record corresponds to the output of each leakage simulation point respectively
Parameter three；
It repeats the above process, until completing the simulation of all grade fluid floies；
Wherein output parameter three includes temperature field leakage signal, moisture field leakage signal and stress field leakage signal, mould
Quasi- input parameter includes leakage simulation point position parameter and fluid flow class parameter；
Each group of temperature field leakage signal, moisture field leakage signal and stress field leakage signal correspond to one group of tunnelling ray
Quasi- point location parameter and fluid flow class parameter, record above-mentioned corresponding data as sample data three respectively.
Further, step 5 specifically comprises the following steps:
Step 501: according to the length and tested pipeline of temperature detection optical fiber, Humidity Detection optical fiber and stress mornitoring optical fiber
Length, establish temperature field detection fiber system, moisture field detection fiber system, stress field detection fiber system output parameter with
Positioning corresponding relationship between tested pipeline physical location；
Step 502: according to temperature field detection fiber system, moisture field detection fiber system, stress field detection fiber system
Output parameter and tested pipeline physical location between positioning corresponding relationship design neural network structure, determine target component and
Neuron parameter；
Step 503: initialization neural network parameter；
Step 504: sample data one, sample data two and sample data three input training data: being input to design
In good neural network structure；
Step 505: gradient declines undated parameter；
Step 506: test: entire neural network being tested；
Step 507: whether measuring accuracy reaches requirement: difference between the predicted value and actual value of neural network model with
Threshold value compares, if difference is greater than threshold value, repeatedly step 505, until difference is less than threshold value；
Step 508: obtain trained neural network: after training for several times, difference is less than threshold value, is trained at this time
Good neural network.
Further, it includes each of collected three field data and tested pipeline that training data is inputted in step 504
Position, wherein input layer number is equal with the number of actually detected sampled point.
Further, step 6 specifically comprises the following steps:
The temperature field, moisture field and the stress field data that are collected around tested pipeline in real time are input to trained mind
Through in network, outgoing event when by the incident of leakage and simulated leakage of the reality output obtained is compared, and is judged and is divided
Leak sizes intensity grade, and the maximum position of Fiber-optic Sensors with Data change rate occurred according to incident of leakage, with practical pipe
The position in road is corresponding as a result, obtain the position of leakage point, and the space according to temperature field, moisture field and stress field after leakage
With Annual distribution variation relation, the diffusion velocity of leakage is obtained, correct the leak sizes intensity grade obtained by neural network, it is comprehensive
It closes and obtains final leak position and leak sizes class information after judging.
Further, leak sizes intensity grade includes three: normal, general leakage, serious leak.
Compared with prior art, the invention has the advantages that
1, because detection fiber is mounted on the outside of pipeline, there is no relative displacement when detection with pipeline, the temperature of fiber laser arrays,
The measurement data of humidity and stress, but it is corresponding with the length of pipeline and position, when pipe leakage, pass through difference portion on optical fiber
The Parameters variation comparison of position can leak development trend with locating leaks in pipes position, leak sizes；When not leaking, Ke Yitong
Stress distribution size at piping different location judges stress concentration region, or the pipeline location easily leaked.
Since three field parameters are there are connected effect, whether comprehensive three field parameters can be leaked with real-time judge, if leakage can
With at once determine leak position, and according to temperature field surrounding, moisture field, stress field real-time change situation, judge leak sizes
And development trend；It can find the pipeline location of stress abnormality in time by pipe stress distribution field, replacement avoids leaking in time.
So the present invention can prevent the hair of pipe leakage by analyzing three field datas by setting THM coupling optical fiber sensing system
It is raw, it can also grasp position and tendency information when leaking and occurring, the linkage performance of temperature and humidity effectively avoids the generation of wrong report.
2, artificial neural network is formed by connecting by numerous adjustable connection weights of neuron, is had at large-scale parallel
Reason, distributed information storage, good self-organizing adaptability, and there is very strong learning ability.
The present invention monitors tested pipeline light condition analogue data, normal work by record pipeline THM coupling sensor-based system
Condition analogue data, incident of leakage analogue data, and three groups of data are input in established neural network, to the nerve net
Network is trained, by obtaining trained neural network after repeated multiple times training.
Real-time output parameter in pipeline THM coupling sensor-based system monitoring process is input to trained Monitoring Pinpelines
In neural network, Monitoring Pinpelines neural network voluntarily judges and exports monitoring result.
Method raising of the invention efficiently reduces wrong report in actual use, avoids failing to report, leakage point is accurately positioned,
And the size of leakage is provided by neural network algorithm, reliable basis is provided to formulate maintenance program.Meanwhile not destroying primary tube
Road structure is laid with simple, long service life.
3, the present invention is laid on tested winding three kinds of distributed fiberoptic sensors in a manner of the spiral of 120 ° of interval
On the outside of pipeline, this kind of canoe can form the solid package to tested pipeline, mainly can increase the laying length of optical fiber, mention
The resolution ratio and sensitivity of high detection.
4, temperature detection fiber, Humidity Detection optical fiber by fiber mounting bracket are fixedly mounted on tested pipeline in the present invention
On outer wall, such temperature detection optical fiber and Humidity Detection optical fiber will avoid wave length shift caused by pipe stress variation.Stress inspection
It surveys optical fiber to be adhered directly on the outside of pipeline, making stress mornitoring optical fiber and pipeline, deformation occurs together, and the more preferable stress for detecting pipeline becomes
Change.
Detailed description of the invention
Illustrate the embodiment of the present invention or technical solution in the prior art in order to clearer, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it is clear that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
Other attached drawings are obtained according to these attached drawings.
Fig. 1 is the method flow diagram of THM coupling line leakage method of the present invention；
Fig. 2 is step 5 Monitoring Pinpelines neural net model establishing and study in THM coupling line leakage method of the present invention
Flow chart；
Fig. 3 is that the chaos annealing Competitive ANN in Competitive ANN is added in chaotic annealing in the present embodiment
Structure chart.
Specific embodiment
Below in conjunction with the attached drawing in the present invention, technical solution in the embodiment of the present invention carry out it is clear, completely retouch
It states, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Based on the present invention
In embodiment, those skilled in the art's all other reality obtained without making creative work
Example is applied, protection scope of the present invention is belonged to.
As shown in Figure 1, being the method flow diagram of THM coupling line leakage method of the present invention, THM coupling pipeline is let out
The method of leakage monitoring method includes the following steps:
Step 1, building pipeline THM coupling sensor-based system, including temperature field detection fiber system, moisture field detection fiber
System, stress field detection fiber system, and determine relative mounting positions parameter；
Step 2, the monitoring tested pipeline light condition simulation of pipeline THM coupling sensor-based system, and output parameter one is recorded,
As sample data one；
Step 3, the monitoring tested pipeline nominal situation simulation of pipeline THM coupling sensor-based system, and output parameter two is recorded,
As sample data two；
Step 4, the simulation of pipeline THM coupling sensor-based system incident of leakage, and record simulation input parameter and output parameter
Three, as sample data three；
Step 5, Monitoring Pinpelines neural net model establishing and study；
Step 6, line leakage: the real-time output parameter in pipeline THM coupling sensor-based system monitoring process is inputted
Into trained Monitoring Pinpelines neural network, Monitoring Pinpelines neural network voluntarily judges and exports monitoring result.
Because detection fiber is mounted on the outside of pipeline, there is no relative displacement when detection with pipeline, it is the temperature of fiber laser arrays, wet
The measurement data of degree and stress, but it is corresponding with the length of pipeline and position, when pipe leakage, pass through different parts on optical fiber
Parameters variation comparison can be with locating leaks in pipes position, leak sizes leak development trend；When not leaking, it can pass through
Stress distribution size at pipeline different location judges stress concentration region, or the pipeline location easily leaked.
Since three field parameters are there are connected effect, whether comprehensive three field parameters can be leaked with real-time judge, if leakage can
With at once determine leak position, and according to temperature field surrounding, moisture field, stress field real-time change situation, judge leak sizes
And development trend；It can find the pipeline location of stress abnormality in time by pipe stress distribution field, replacement avoids leaking in time.
So the present invention can prevent the hair of pipe leakage by analyzing three field datas by setting THM coupling optical fiber sensing system
It is raw, it can also grasp position and tendency information when leaking and occurring, the linkage performance of temperature and humidity effectively avoids the generation of wrong report.
Artificial neural network is formed by connecting by numerous adjustable connection weights of neuron, is had at large-scale parallel
Reason, distributed information storage, good self-organizing adaptability, and there is very strong learning ability.
The present invention monitors tested pipeline light condition analogue data, normal work by record pipeline THM coupling sensor-based system
Condition analogue data, incident of leakage analogue data, and three groups of data are input in established neural network, to the nerve net
Network is trained, by obtaining trained neural network after repeated multiple times training.
Real-time output parameter in pipeline THM coupling sensor-based system monitoring process is input to trained Monitoring Pinpelines
In neural network, Monitoring Pinpelines neural network voluntarily judges and exports monitoring result.
Method of the invention efficiently reduces wrong report in actual use, avoids failing to report, being accurately positioned leakage point, and
The size of leakage is provided by neural network algorithm, provides reliable basis to formulate maintenance program.Meanwhile not destroying original pipeline
Structure is laid with simple, long service life.
Specifically, step 1 specifically comprises the following steps:
Step 101: pipeline THM coupling sensing system hardware parameter determines, in which: temperature field detection fiber system, humidity
Field detecting fibre system, stress field detection fiber system are all made of distributed fiberoptic sensor, are respectively used to measurement tested pipeline
Temperature field, stress field, moisture field delta data；THM coupling sensor-based system include temperature detection optical fiber, Humidity Detection optical fiber,
Stress mornitoring optical fiber, photoelectric conversion device, signal processing apparatus and central processing unit；According to tested pipeline length, diameter with
And monitoring accuracy temperature detection fiber, the length of Humidity Detection optical fiber, stress mornitoring optical fiber；Coiling length determines detection essence
The accuracy of degree and leakage positioning.
Step 102: pipeline THM coupling sensor-based system installation, in which: spread in a manner of spiral winding outside tested pipeline
If temperature detection optical fiber, Humidity Detection optical fiber and stress mornitoring optical fiber, and temperature detection optical fiber, Humidity Detection light are recorded respectively
The length of the fine, length of stress mornitoring optical fiber and tested pipeline；By temperature detection optical fiber, Humidity Detection optical fiber, stress mornitoring
The input terminal of optical fiber is connect with photoelectric conversion device respectively, and output end is connect with signal processing apparatus respectively, signal processing apparatus
Output end connect with central processing unit.
Temperature detection fiber, Humidity Detection optical fiber are fixedly mounted on outside tested pipeline by fiber mounting bracket in the present invention
On wall, such temperature detection optical fiber and Humidity Detection optical fiber will avoid wave length shift caused by pipe stress variation.Stress mornitoring
Optical fiber is adhered directly on the outside of pipeline, and making stress mornitoring optical fiber and pipeline, deformation occurs together, the more preferable stress variation for detecting pipeline.
Specifically, temperature detection optical fiber, Humidity Detection optical fiber and stress mornitoring optical fiber with 120 ° of intervals be spirally wound on by
Test tube pipeline outer wall.
The present invention is laid on measured tube winding three kinds of distributed fiberoptic sensors in a manner of the spiral of 120 ° of interval
On the outside of road, this kind of canoe can form the solid package to tested pipeline, mainly can increase the laying length of optical fiber, improve
The resolution ratio and sensitivity of detection.
Specifically, in step 2 specifically includes the following steps:
Step 201: tested pipeline processing: excluding internal flow and stand to arrive working temperature；
Step 202: starting and debug THM coupling sensor-based system；
Step 203: after system output parameter is stablized, testing and record output parameter one, wherein output parameter one includes
Temperature field initial signal, moisture field initial signal and stress field initial signal.
Pipeline THM coupling sensor-based system monitors the simulation of tested pipeline light condition, gives fibre optical sensor established standards, makes
At the middle part of fibre optical sensor detection interval, and in actual work, temperature, humidity, the size of stress are greater than a reference value certainly
The a reference value then illustrates that fibre optical sensor breaks down when finding that actually detected data are less than a reference value, needs to pass optical fiber
Sensor repairs.
It specifically walks, step 3 specifically comprises the following steps:
Step 301: tested pipeline water filling reaches normal discharge and pressure；
Step 302: measurement output parameter two simultaneously records；One group of output parameter two is recorded at interval of the set time, it is continuous to survey
Amount record 24-48 hours；Output parameter two be same time point under record temperature field working condition signal, moisture field working condition signal with
And stress field working condition signal；
Step 303: by collected multiple groups temperature field working condition signal, moisture field working condition signal and stress field working condition signal
Tabulation record, it is spare as reference data.
Due to using spiral winding mode, it is possible to show that the temperature field of pipeline, moisture field, stress field three dimensional change become
Gesture, the trend can be used for analyzing leak sizes and development trend as time goes by.
Specifically, step 4 specifically comprises the following steps:
Step 401: determine that leakiness simulates grade: use simulates pipe with liquid identical in tested pipeline on pipeline
Road leakage, fluid flow are divided into multiple grades from small to large, respectively correspond different leakage class；
Step 402: determining leakage simulation dot density: leakage simulation point, tunnelling ray are set in pipeline external surface uniform array
Quasi- dot density is arranged according to detection accuracy；
Step 403: incident of leakage simulation: successively carrying out same levels fluid flow drippage behaviour on each leakage simulation point
Make, and record corresponds to the output parameter three of each leakage simulation point respectively；After first grade fluid flow simulation,
Tested pipeline is handled, until the second wheel of progress is different from first grade fluid flow and drips mould after three output parameters are stablized
It is quasi-, and record corresponds to the output parameter three of each leakage simulation point respectively；It repeats the above process, until completing all grades
The simulation of fluid flow；Wherein output parameter three includes temperature field leakage signal, moisture field leakage signal and stress field leakage
Signal, simulation input parameter include leakage simulation point position parameter and fluid flow class parameter；Each group of temperature field leakage
Signal, moisture field leakage signal and stress field leakage signal correspond to one group of leakage simulation point position parameter and fluid flow
Class parameter records above-mentioned corresponding data as sample data three respectively.
Simulated leakage process in the present embodiment specifically: at a certain position of pipeline (usually in the connector of two knot pipelines
Place, is easiest to leak), artificial destruction or produce a leakage point, record leakage point position and aperture size (if
It is circular hole, then records Circularhole diameter；If it is square hole or other irregular holes, then area is recorded), and covered with sandy soil, it covers
Depth is depending on the practical depth of burying of pipeline, and after water flowing to normal pressure, the change of three field parameters is detected using fibre optical sensor
Change.
It can change the size (such as 3-5 grade, such as 1cm, 2cm, 3cm, 4cm, 5cm) of leak, water when simulation
The temperature (according to a range of water temperature when practical pipeline work, for example being equally divided into 3-5 grade) of stream, the pressure of water flow
(according to a range of hydraulic pressure when practical pipeline work, for example being equally divided into 3-5 grade) then grouping mea-sure and records number
According to as sample data.
As shown in Fig. 2, step 5 specifically comprises the following steps:
Step 501: according to the length and tested pipeline of temperature detection optical fiber, Humidity Detection optical fiber and stress mornitoring optical fiber
Length, establish temperature field detection fiber system, moisture field detection fiber system, stress field detection fiber system output parameter with
Positioning corresponding relationship between tested pipeline physical location.
Step 502: according to temperature field detection fiber system, moisture field detection fiber system, stress field detection fiber system
Output parameter and tested pipeline physical location between positioning corresponding relationship design neural network structure, determine target component and
Neuron parameter.
As shown in figure 3, for the chaos annealing that chaotic annealing is added in Competitive ANN is competed in the present embodiment
Neural network structure figure, designs neural network structure, and the simulation dynamic annealing process of single neuron uses formula (1-2) table
Show,
V (t+1)=kv (t)-E+T (t) (u (t)-I0) (2)
In formula:
U is the transient state connection weight for inputting neuron and output neuron；
V is the internal state connection weight for inputting neuron and output neuron；
I0For neuron input deviation；
K is neu damping factor (0≤k≤1)；
E is the objective function of neural network；
ε is the decay factor of output function；
T (t) is self-feedback connection weights weight.
Step 503: initialization neural network parameter.
Step 504: sample data one, sample data two and sample data three input training data: being input to design
In good neural network structure.
Step 505: gradient declines undated parameter.
The transient state u of input layer and output layer in the training processx；jWith internal weight state vx；jBy being embedded in simulated annealing
Function progressivelyes reach steady-state process.Output state reaches the renewal process gradually to decay with a small Studying factors.Dividing
During shape, these parameters are obtained in real time by parallel computation.Neuron state passes through function vx；jReal-time update.Simulated annealing
Strategy is applied in trained network based on equation (3-7).This network model is n input node, c output section in input layer
Point and n x c connection weight.The mathematic(al) representation of the model is as follows:
vx；j(t+1)=kvx；j(t)+E-T(t)(ux；j(t)-I0) (5)
Δwj=η (zx-wj)ux；j (6)
In formula, E is the energy function of n input and c output node network.ux；jAnd vx；jRespectively transient state connection weight and interior
Portion's state connection weight.
Step 506: test: entire neural network being tested.
Step 507: whether measuring accuracy reaches requirement: difference between the predicted value and actual value of neural network model with
Threshold value compares, if difference is greater than threshold value, repeatedly step 505, until difference is less than threshold value.
Step 508: obtain trained neural network: after training for several times, difference is less than threshold value, is trained at this time
Good neural network.
Chaos annealing Competitive ANN setting: 54 input neurons, 3 output neurons.
Input vector=[temperature, humidity, stress value]
Export classification=[severe leakage, slight leakage are normal]
Initial value: ε=0.004, k=0.9, E=0, T (t)=0.033, I0=0.65, T0=0.09, y0=0.5, β=
500,3000 step of iterative step surveys trained network using 1000 groups of data using 300 groups of training data training networks
Examination, test result are shown: 27 groups of data test mistakes, then precision is 97.3%.
Specifically, each position that training data includes collected three field data and tested pipeline is inputted in step 504,
Wherein, input layer number is equal with the number of actually detected sampled point.
Specifically, step 6 specifically comprises the following steps:
The temperature field, moisture field and the stress field data that are collected around tested pipeline in real time are input to trained mind
Through in network, outgoing event when by the incident of leakage and simulated leakage of the reality output obtained is compared, and is judged and is divided
Leak sizes intensity grade, and the maximum position of Fiber-optic Sensors with Data change rate occurred according to incident of leakage, with practical pipe
The position in road is corresponding as a result, obtain the position of leakage point, and the space according to temperature field, moisture field and stress field after leakage
With Annual distribution variation relation, the diffusion velocity of leakage is obtained, correct the leak sizes intensity grade obtained by neural network, it is comprehensive
It closes and obtains final leak position and leak sizes class information after judging.
Specifically, leak sizes intensity grade includes three: normal, general leakage, serious leak.
When actual monitoring, near the leakage point before and after pipe leakage, temperature field, moisture field and stress field can be with leakages
Size and time change, and the parameter size of different location is related with the distance of leakage point, and the length and position when optical fiber is laid with are believed
Breath, has one-to-one relationship with the physical length and laying density of pipeline and position, will all inspections related with location information
Input of the data of survey as neural network.I.e. input sample is the numerical value for being respectively temperature, humidity and stress, number of samples
It is determined by tested pipeline length and measurement accuracy, since temperature and stress fibre optical sensor are distributed data acquisition, is actually adopted
Sampling point is tested pipeline total length L (rice) divided by 0.1 (rice), such as pipeline is 100 meters long, then number of sampling points is 1000
Sample.Output is the size degree three grades of leakage: normal (A), general leakage (B), severe leakage (C), exporting is 000,
010,100 represents A, B, C.Since the laying length and duct length of distribution type fiber-optic have one-to-one relationship, leakage is big
The small maximum point of variation is the position of leakage point.
The above description is only an embodiment of the present invention, is not intended to limit the scope of the invention, all to utilize this hair
Equivalent structure or equivalent flow shift made by bright specification and accompanying drawing content is applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of THM coupling line leakage method, which comprises the steps of:
Step 1, building pipeline THM coupling sensor-based system, including temperature field detection fiber system, moisture field detection fiber system,
Stress field detection fiber system, and determine relative mounting positions parameter；
Step 2, the monitoring tested pipeline light condition simulation of pipeline THM coupling sensor-based system, and output parameter one is recorded, as
Sample data one；
Step 3, the monitoring tested pipeline nominal situation simulation of pipeline THM coupling sensor-based system, and output parameter two is recorded, as
Sample data two；
Step 4, the simulation of pipeline THM coupling sensor-based system incident of leakage, and simulation input parameter and output parameter three are recorded,
As sample data three；
Step 5, Monitoring Pinpelines neural net model establishing and study；
Step 6, line leakage: the real-time output parameter in pipeline THM coupling sensor-based system monitoring process is input to instruction
In the Monitoring Pinpelines neural network perfected, Monitoring Pinpelines neural network voluntarily judges and exports monitoring result；
Wherein, neural network structure is modeled in step 5, the simulation dynamic annealing process of single neuron uses formula (1-2) table
Show,
V (t+1)=kv (t)-E+T (t) (u (t)-I0) (2)
In formula:
U is the transient state connection weight for inputting neuron and output neuron；
V is the internal state connection weight for inputting neuron and output neuron；
I0For neuron input deviation；
K is neu damping factor (0≤k≤1)；
E is the objective function of neural network；
ε is the decay factor of output function；
T (t) is self-feedback connection weights weight；
Annealing strategy is based on equation (3-7) and is applied in trained network, this network model is n input section in input layer
Point, c output node and nxc connection weight, the mathematic(al) representation of the model are as follows:
vx；j(t+1)=kvx；j(t)+E-T(t)(ux；j(t)-I0) (5)
Δwj=η (zx-wj)ux；j (6)
In formula, E is the energy function of n input and c output node network, ux；jAnd vx；jRespectively transient state connection weight and internal shape
State connection weight.
2. THM coupling line leakage method according to claim 1, which is characterized in that the step 1 is specifically wrapped
Include following steps:
Step 101: pipeline THM coupling sensing system hardware parameter determines, in which:
The temperature field detection fiber system, moisture field detection fiber system, stress field detection fiber system are all made of distribution
Fibre optical sensor is respectively used to temperature field, the stress field, moisture field delta data of measurement tested pipeline；
The THM coupling sensor-based system includes temperature detection optical fiber, Humidity Detection optical fiber, stress mornitoring optical fiber, photoelectric conversion dress
It sets, signal processing apparatus and central processing unit；
According to tested pipeline length, diameter and monitoring accuracy temperature detection fiber, Humidity Detection optical fiber, stress mornitoring light
Fine length；
Step 102: the installation of pipeline THM coupling sensor-based system, in which:
Temperature detection optical fiber, Humidity Detection optical fiber and stress mornitoring optical fiber are laid in a manner of spiral winding outside tested pipeline,
And the length of temperature detection optical fiber, Humidity Detection optical fiber, stress mornitoring optical fiber and tested pipeline is recorded respectively；
The input terminal of temperature detection optical fiber, Humidity Detection optical fiber, stress mornitoring optical fiber is connect with photoelectric conversion device respectively, it is defeated
Outlet is connect with signal processing apparatus respectively, and the output end of signal processing apparatus is connect with central processing unit.
3. THM coupling line leakage method according to claim 2, which is characterized in that the temperature detection light
Fibre, Humidity Detection optical fiber and stress mornitoring optical fiber are spirally wound on tested pipeline outer wall with 120 ° of intervals.
4. THM coupling line leakage method according to claim 1, which is characterized in that specific in the step 2
The following steps are included:
Step 201: tested pipeline processing: excluding internal flow and stand to arrive working temperature；
Step 202: starting and debug THM coupling sensor-based system；
Step 203: after system output parameter is stablized, testing and record output parameter one, wherein output parameter one includes temperature
Field initial signal, moisture field initial signal and stress field initial signal.
5. THM coupling line leakage method, feature described in any one of -4 claims exist according to claim 1
In the step 3 specifically comprises the following steps:
Step 301: tested pipeline water filling reaches normal discharge and pressure；
Step 302: measurement output parameter two simultaneously records；One group of output parameter two, continuous measurement note are recorded at interval of the set time
Record 24-48 hours；Output parameter two is the temperature field working condition signal recorded under same time point, moisture field working condition signal and answers
Field of force working condition signal；
Step 303: collected multiple groups temperature field working condition signal, moisture field working condition signal and stress field working condition signal are tabulated
Record, it is spare as reference data.
6. THM coupling line leakage method according to claim 5, which is characterized in that the step 4 is specifically wrapped
Include following steps:
Step 401: determining that leakiness simulates grade: being let out using pipeline is simulated on pipeline with liquid identical in tested pipeline
Leakage, fluid flow are divided into multiple grades from small to large, respectively correspond different leakage class；
Step 402: determining leakage simulation dot density: leakage simulation point, leakage simulation point are set in pipeline external surface uniform array
Density is arranged according to detection accuracy；
Step 403: incident of leakage simulation: successively carrying out the drippage operation of same levels fluid flow on each leakage simulation point, and
Record corresponds to the output parameter three of each leakage simulation point respectively；
After first grade fluid flow simulation, tested pipeline is handled, until carrying out second after three output parameters are stablized
Wheel is different from first grade fluid flow drippage simulation, and record corresponds to the output parameter of each leakage simulation point respectively
Three；
It repeats the above process, until completing the simulation of all grade fluid floies；
Wherein output parameter three includes temperature field leakage signal, moisture field leakage signal and stress field leakage signal, is simulated defeated
Entering parameter includes leakage simulation point position parameter and fluid flow class parameter；
Each group of temperature field leakage signal, moisture field leakage signal and stress field leakage signal correspond to one group of leakage and simulate point
Location parameter and fluid flow class parameter record above-mentioned corresponding data as sample data three respectively.
7. THM coupling line leakage method according to claim 6, which is characterized in that the step 5 is specifically wrapped
Include following steps:
Step 501: according to temperature detection optical fiber, the length of Humidity Detection optical fiber and stress mornitoring optical fiber and measured tube road length
Degree, establishes the output parameter and quilt of temperature field detection fiber system, moisture field detection fiber system, stress field detection fiber system
Positioning corresponding relationship between test tube road physical location；
Step 502: according to temperature field detection fiber system, moisture field detection fiber system, stress field detection fiber system it is defeated
The positioning corresponding relationship between parameter and tested pipeline physical location designs neural network structure out, determines target component and nerve
First parameter；
Step 503: initialization neural network parameter；
Step 504: input training data: sample data one, sample data two and sample data three being input to designed
In neural network structure；
Step 505: gradient declines undated parameter；
Step 506: test: entire neural network being tested；
Step 507: whether measuring accuracy reaches requirement: difference and threshold value between the predicted value and actual value of neural network model
It compares, if difference is greater than threshold value, repeatedly step 505, until difference is less than threshold value；
Step 508: obtain trained neural network: after training for several times, difference is less than threshold value, obtains at this time trained
Neural network.
8. THM coupling line leakage method according to claim 7, which is characterized in that defeated in the step 504
Enter each position that training data includes collected three field data and tested pipeline, wherein input layer number and practical inspection
The number for surveying sampled point is equal.
9. THM coupling line leakage method according to claim 1, which is characterized in that the step 6 is specifically wrapped
Include following steps:
The temperature field, moisture field and the stress field data that are collected around tested pipeline in real time are input to trained nerve net
In network, outgoing event when by the incident of leakage and simulated leakage of the reality output obtained is compared, and is judged and is divided leakage
Size intensity grade, and the maximum position of Fiber-optic Sensors with Data change rate occurred according to incident of leakage, with practical pipeline
Position is corresponding as a result, obtain the position of leakage point, and according to temperature field after leakage, moisture field and stress field space and when
Between changes in distribution relationship, obtain the diffusion velocity of leakage, correct the leak sizes intensity grade obtained by neural network, synthesis is commented
Final leak position and leak sizes class information are obtained after sentencing.
10. THM coupling line leakage method according to claim 9, which is characterized in that the leak sizes journey
Spending grade includes three: normal, general leakage, serious leak.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710310608.8A CN107013812B (en) | 2017-05-05 | 2017-05-05 | A kind of THM coupling line leakage method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710310608.8A CN107013812B (en) | 2017-05-05 | 2017-05-05 | A kind of THM coupling line leakage method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107013812A CN107013812A (en) | 2017-08-04 |
| CN107013812B true CN107013812B (en) | 2019-07-12 |
Family
ID=59448821
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710310608.8A Expired - Fee Related CN107013812B (en) | 2017-05-05 | 2017-05-05 | A kind of THM coupling line leakage method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107013812B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108051035B (en) * | 2017-10-24 | 2019-08-09 | 清华大学 | The pipe network model recognition methods of neural network model based on gating cycle unit |
| CN109027704B (en) * | 2018-05-30 | 2020-07-28 | 华中科技大学 | Pipeline monitoring system and monitoring method based on microstructure optical fiber distributed sensing |
| CN109506848B (en) * | 2018-12-29 | 2024-05-31 | 汉威科技集团股份有限公司 | Novel online scanning ultrasonic gas leakage detection system |
| EP3942271A1 (en) * | 2019-03-12 | 2022-01-26 | Siemens Aktiengesellschaft | Method and device for operating a distribution network |
| CN112212224A (en) * | 2020-10-15 | 2021-01-12 | 海盐于通管件制造有限公司 | Pipeline water leakage detection method |
| CN116698285B (en) * | 2023-07-31 | 2023-10-24 | 合肥航谱时代科技有限公司 | Dangerous chemical gas pipeline leakage monitoring system based on infrared imager |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101183899A (en) * | 2007-12-19 | 2008-05-21 | 天津大学 | BP network based pipeline security identifying method for optical fiber pipeline leakage monitoring device |
| CN201464144U (en) * | 2009-07-07 | 2010-05-12 | 紫光捷通科技股份有限公司 | Pipeline leakage and blockage detection system |
| KR100973662B1 (en) * | 2010-01-20 | 2010-08-03 | 주식회사 유앤유소프트 | Water leakage position estimating system |
| CN104654024A (en) * | 2015-02-12 | 2015-05-27 | 常州大学 | Method for locating and analyzing leakage of city gas pipeline based on GRNN (Generalized Regression Neural Network) |
| KR101613396B1 (en) * | 2015-05-28 | 2016-04-18 | 주식회사 삼천리 | Apparatus for detecting fluid leakage |
| CN105805556A (en) * | 2015-09-30 | 2016-07-27 | 李儒峰 | Distributed optical fiber leakage monitoring system |
-
2017
- 2017-05-05 CN CN201710310608.8A patent/CN107013812B/en not_active Expired - Fee Related
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101183899A (en) * | 2007-12-19 | 2008-05-21 | 天津大学 | BP network based pipeline security identifying method for optical fiber pipeline leakage monitoring device |
| CN201464144U (en) * | 2009-07-07 | 2010-05-12 | 紫光捷通科技股份有限公司 | Pipeline leakage and blockage detection system |
| KR100973662B1 (en) * | 2010-01-20 | 2010-08-03 | 주식회사 유앤유소프트 | Water leakage position estimating system |
| CN104654024A (en) * | 2015-02-12 | 2015-05-27 | 常州大学 | Method for locating and analyzing leakage of city gas pipeline based on GRNN (Generalized Regression Neural Network) |
| KR101613396B1 (en) * | 2015-05-28 | 2016-04-18 | 주식회사 삼천리 | Apparatus for detecting fluid leakage |
| CN105805556A (en) * | 2015-09-30 | 2016-07-27 | 李儒峰 | Distributed optical fiber leakage monitoring system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107013812A (en) | 2017-08-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107013812B (en) | A kind of THM coupling line leakage method | |
| CN106015951B (en) | A kind of gas pipeline leak detection system and method adapting to various states variation | |
| CN108984873B (en) | Water supply network real-time leakage detection method, device, system and storage medium | |
| US10242414B2 (en) | Method for locating a leak in a fluid network | |
| CN104930355B (en) | Suitable for the online lossless detection method and its device of city supply water pipeline system | |
| CN101255952B (en) | Test system for monitoring pipe leakage and early warning safety | |
| CN201014212Y (en) | Pipeline leakage monitoring and safety early warning test system | |
| CN105221933A (en) | A kind of pipeline network leak detecting method in conjunction with resistance identification | |
| CN107869653A (en) | A kind of pipeline flow sensitivity matrix soil's rigidity method | |
| CN109838696A (en) | Pipeline fault diagnostic method based on convolutional neural networks | |
| CN115451347B (en) | Intelligent monitoring system and method for petroleum pipeline safety | |
| CN108050394A (en) | Fuel gas pipeline leakage based on sound pressure signal identification detects positioning experiment platform | |
| CN106844814A (en) | A kind of large complicated gas distributing system system leak detection method | |
| CN111024821A (en) | Composite material storage box health monitoring system and method | |
| CN109738014A (en) | The intelligent diagnosing method and system of city integrated piping lane equipment fault | |
| CN113424114A (en) | Pressure control in a supply network | |
| CN115127036B (en) | Municipal gas pipe network leakage positioning method and system | |
| JPH0896039A (en) | Water pipeline information management device | |
| CN108506740A (en) | A kind of fluid pipeline leakage region based on current meter determines method and system | |
| Kutyłowska | Prediction of failure frequency of water-pipe network in the selected city | |
| Ravula et al. | Experimental validation of leak and water-ingression detection in low-pressure gas pipeline using pressure and flow measurements | |
| Wu et al. | Leveraging optical communication fiber and AI for distributed water pipe leak detection | |
| CN201184848Y (en) | Instrument for measuring shallow layer ground-temperature energy parameter | |
| CN116164241A (en) | Intelligent detection method for leakage faults of gas extraction pipe network | |
| RU2767263C1 (en) | Method for integrated assessment of indicators determining the technical condition of pipeline systems, and a monitoring system for its implementation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190712 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |