CN103870670B - A kind of tube corrosion degree Forecasting Methodology and device - Google Patents
A kind of tube corrosion degree Forecasting Methodology and device Download PDFInfo
- Publication number
- CN103870670B CN103870670B CN201210548324.XA CN201210548324A CN103870670B CN 103870670 B CN103870670 B CN 103870670B CN 201210548324 A CN201210548324 A CN 201210548324A CN 103870670 B CN103870670 B CN 103870670B
- Authority
- CN
- China
- Prior art keywords
- data
- oil pipe
- corrosion
- etch rate
- data sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Abstract
The invention discloses a kind of tube corrosion degree Forecasting Methodology and device, belong to oil field well technical field.This method includes:S1, chooses at least two groups underground working data and its actual etch rate data of corresponding oil pipe, and carries out normalization to the data of selection, and the data after normalization are divided into training data sample and inspection data sample;S2, with reference to Fuzzy Linear Regression method and least square method, is modeled to training data sample, obtains predicting etch rate data as the tube corrosion forecast model of variable as dependent variable, oil pipe by underground working data;S3, the correctness of the tube corrosion forecast model is verified using the inspection data sample, if upchecked, S4 is performed, otherwise, transposition S1, is chosen new Data duplication and is performed flow;S4, using obtained tube corrosion forecast model, extent of corrosion prediction is carried out to the in-service oil pipe under designated environment.Using the present invention, the degree of accuracy of down-hole oil tube extent of corrosion prediction can be improved.
Description
Technical field
The present invention relates to oil field well technical field, more particularly to a kind of tube corrosion degree Forecasting Methodology and device.
Background technology
With the exploitation of oil gas field, hydrogen sulfide, carbon dioxide, chlorine that the HTHP multiphase flow environment under Oil/gas Well contains
The media such as ion, water and microorganism can occur as attendants, make down-hole oil tube too corroded, so not only have impact on
The normal production of Oil/gas Well, and bring many difficulties to examination workover treatment.
The factor of influence oil pipe steel corrosion can be divided into environmental factor, material factor and mechanics factor etc..Wherein, environmental factor
Including temperature, partial pressure, corrosion products film, pH value, flow velocity, flow pattern, solution degree of supersaturation, bacterium etc..The above factor
The corrosion of oil pipe is affected to some extent, therefore prediction tube corrosion degree is mainly set about from this several respects factor.Actual work
The corrosion influence factors of down-hole oil tube are extremely complex in journey, for example:The change of well depth, the difference of operating condition and construction cycle
Process difference cause complicated and changeable, the tube corrosion journey such as underground diverse location, the concentration of medium of different times and working condition
Degree differs.In addition, oil film corrosion Inhibition Mechanism in the gap of many such as oil connection positions presence and galvanic corrosion problem, pipe
And current not yet thoroughly research in the art the problems such as down-hole string compound stress.Due to existing corrosion tests
With equipment still can not simulating down-hole actual behavior, therefore only by field working conditions Data Detection tube corrosion situation completely, or
Only by existing corrosion prediction theoretical model and simulation operating mode corrosion test all can not well to in-service oil pipe corrosion failure
Carry out early warning.
In the prior art, by means of classical theory formula and laboratory simulation data, the corrosion condition of oil pipe is sentenced
Disconnected, conventional prediction extent of corrosion criterion has:(1) CO is passed through2Partial pressure judges:WhenInterval scale oil pipe is serious
Corrosion;And work asWhen show oil pipe generality corrosion;WhenInterval scale oil pipe is without corruption
Erosion.(2) basisRatio in judgement:WhenWhen be CO2Corrosion;WhenFor H2S
Corrosion.(3) according to Ca2+/HCO3Ratio in judgement:WhenAnd in the presence of having stratum water, Ca2+/HCO3<
0.5 shows that tube corrosion speed is relatively low, Ca2+/HCO3> 1000 shows that tube corrosion speed is medium, Ca2+/HCO3=0.5~
1000 represent oil pipe and occur heavy corrosion.
During the present invention is realized, inventor has found that prior art at least has problems with:
Judgement early warning is carried out to tube corrosion according to some classical theory formula or laboratory simulation data in the prior art,
Often the single factors only for influence tube corrosion are analyzed to judge the extent of corrosion of oil pipe.And in fact, due to oil
The complexity of pipe subsurface environment, the method for prior art can not be accurately judged the corrosion condition of oil pipe.
The content of the invention
In order to solve the problems of the prior art, the invention provides a kind of tube corrosion degree Forecasting Methodology and device,
To improve the degree of accuracy of down-hole oil tube extent of corrosion prediction.The technical scheme that the present invention is provided is as follows:
On the one hand there is provided a kind of tube corrosion degree Forecasting Methodology, multigroup varying environment is prestored in database
Middle underground working data and the actual etch rate data of corresponding oil pipe, the underground working data include multiple influence factors
Data, methods described includes:
S1, chooses at least two groups underground working data and its actual etch rate data of corresponding oil pipe of storage, and right
The data of selection are normalized, and the data after normalization are divided into training data sample and inspection data sample;
S2, with reference to Fuzzy Linear Regression method and least square method, is modeled to the training data sample, obtains by well
Lower floor data predicts etch rate data as the tube corrosion forecast model of dependent variable as independent variable, oil pipe;
S3, the correctness of the tube corrosion forecast model is verified using the inspection data sample, if upchecked,
S4 is then performed, otherwise, S1 is gone to, new Data duplication is chosen and performs flow;
S4, using resulting tube corrosion forecast model, carries out extent of corrosion pre- to the in-service oil pipe under designated environment
Survey.
It is preferred that, the group number of the training data sample is not less than the number of the independent variable.
It is preferred that, the data of described pair of selection are normalized, and are specially:
It is determined that the span z ∈ [z for the data chosenmin,zmax], and determine normalized range scaleWherein z is the data chosen,For the data after normalization;
According toWith the data of the selection, the data after normalization are drawn.
It is preferred that, the underground working data include CO2Concentration, Cl-Concentration, H2S concentration, temperature and pressure data.
It is preferred that, the combination Fuzzy Linear Regression method and least square method are modeled to the training data sample,
Obtain predicting etch rate data as the tube corrosion prediction mould of dependent variable as independent variable, oil pipe by underground working data
Type, be specially:
To typical linear regression model yi=a1x1i+a2x2i+…+amxmiUsing least square method, and based on training data
Sample, determines aj, wherein, xjiThe data of j-th of the influence factor substituted into i-th group of training data sample, yiSubstitute into i-th group of instruction
Practice the actual etch rate data of oil pipe in data sample, j=1,2 ..., m;
According toWith the training data sample and a determinedj, determine cj, wherein, H
For default confidence value, and 0≤H≤1;
According to the membership function of Triangular Fuzzy NumberWith a determinedj、cj,
Draw Triangular Fuzzy Number Aj(aj,cj) value, be used as Estimates of Fuzzy Linear Regression Model y=A1x1+A2x2+…+AmxmIn recurrence system
Number Aj, it regard the Estimates of Fuzzy Linear Regression Model as the tube corrosion forecast model.
It is preferred that, the correctness that the tube corrosion forecast model is verified using the inspection data sample, specifically
For:
Underground working data in the inspection data sample are substituted into the tube corrosion forecast model, oil pipe are obtained pre-
Survey etch rate data, by the actual etch rate data of the underground working data corresponding oil pipe in the inspection data sample and
The difference of the oil pipe prediction etch rate data is compared with the ratio of the actual etch rate data of the oil pipe with predetermined threshold value, to test
The correctness of the tube corrosion forecast model is demonstrate,proved, if the ratio is no more than predetermined threshold value, is verified, otherwise, checking
Failure.
On the other hand there is provided a kind of tube corrosion degree Forecasting Methodology, multigroup different rings are prestored in database
Underground working data and the actual etch rate data of corresponding oil pipe in border, the underground working data include multiple influence factors
Data, methods described includes:
S1, chooses at least two groups underground working data and its actual etch rate data of corresponding oil pipe of storage, and right
The data of selection are normalized, and the data after normalization are divided into training data sample and inspection data sample;
S2, with reference to the Fuzzy Nonlinear Return Law and least square method, is modeled to the training data sample, obtain by
Underground working data predict etch rate data as the tube corrosion forecast model of dependent variable as independent variable, oil pipe;
S3, the correctness of the tube corrosion forecast model is verified using the inspection data sample, if upchecked,
S4 is then performed, otherwise, S1 is gone to, new Data duplication is chosen and performs flow;
S4, using resulting tube corrosion forecast model, carries out extent of corrosion pre- to the in-service oil pipe under designated environment
Survey.
On the other hand there is provided a kind of tube corrosion degree prediction meanss, described device includes:
Memory module, for prestoring underground working data and the actual corrosion speed of corresponding oil pipe in multigroup varying environment
Rate data, the underground working data include the data of multiple influence factors;
Modeling module, for performing:S1, at least two groups underground working data and its corresponding oil pipe for choosing storage are actual
Etch rate data, and the data of selection are normalized, the data after normalization are divided into training data sample and inspection
Data sample;
S2, with reference to Fuzzy Linear Regression method and least square method, is modeled to the training data sample, obtains by well
Lower floor data predicts etch rate data as the tube corrosion forecast model of dependent variable as independent variable, oil pipe;
S3, the correctness of the tube corrosion forecast model is verified using the inspection data sample, if upchecked,
Then notify prediction module to perform S4, otherwise, go to S1, choose new Data duplication and perform flow;
Prediction module, for performing:S4, using resulting tube corrosion forecast model, to in-service under designated environment
Oil pipe carries out extent of corrosion prediction.
It is preferred that, the group number of the training data sample is not less than the number of the independent variable.
It is preferred that, the modeling module, specifically for:
It is determined that the span z ∈ [z for the data chosenmin,zmax], and determine normalized range scaleWherein z is the data chosen,For the data after normalization;
According toWith the data of the selection, the data after normalization are drawn.
It is preferred that, the underground working data include CO2Concentration, Cl-Concentration, H2S concentration, temperature and pressure data.
It is preferred that, the modeling module, specifically for:
To typical linear regression model yi=a1x1i+a2x2i+…+amxmiUsing least square method, and based on training data
Sample, determines aj, wherein, xjiThe data of j-th of the influence factor substituted into i-th group of training data sample, yiSubstitute into i-th group of instruction
Practice the actual etch rate data of oil pipe in data sample, j=1,2 ..., m;
According toWith the training data sample and a determinedj, determine cj, wherein, H
For default confidence value, and 0≤H≤1;
According to the membership function of Triangular Fuzzy NumberWith a determinedj、cj,
Draw Triangular Fuzzy Number Aj(aj,cj) value, be used as Estimates of Fuzzy Linear Regression Model y=A1x1+A2x2+…+AmxmIn recurrence system
Number Aj, it regard the Estimates of Fuzzy Linear Regression Model as the tube corrosion forecast model.
It is preferred that, the modeling module, specifically for:
Underground working data in the inspection data sample are substituted into the tube corrosion forecast model, oil pipe are obtained pre-
Survey etch rate data, by the actual etch rate data of the underground working data corresponding oil pipe in the inspection data sample and
The difference of the oil pipe prediction etch rate data is compared with the ratio of the actual etch rate data of the oil pipe with predetermined threshold value, to test
The correctness of the tube corrosion forecast model is demonstrate,proved, if the ratio is no more than predetermined threshold value, is verified, otherwise, checking
Failure.
On the other hand there is provided a kind of tube corrosion degree prediction meanss, described device includes:
Memory module, for prestoring underground working data and the actual corrosion speed of corresponding oil pipe in multigroup varying environment
Rate data, the underground working data include the data of multiple influence factors;
Modeling module, for performing:S1, at least two groups underground working data and its corresponding oil pipe for choosing storage are actual
Etch rate data, and the data of selection are normalized, the data after normalization are divided into training data sample and inspection
Data sample;
S2, with reference to the Fuzzy Nonlinear Return Law and least square method, is modeled to the training data sample, obtain by
Underground working data predict etch rate data as the tube corrosion forecast model of dependent variable as independent variable, oil pipe;
S3, the correctness of the tube corrosion forecast model is verified using the inspection data sample, if upchecked,
Then notify prediction module to perform S4, otherwise, go to S1, choose new Data duplication and perform flow;
Prediction module, for performing:S4, using resulting tube corrosion forecast model, to in-service under designated environment
Oil pipe carries out extent of corrosion prediction.
The beneficial effect that the technical scheme that the present invention is provided is brought is:
Based on actually detected underground working data and the actual etch rate data of corresponding oil pipe, by combining unsharp line
Property the Return Law and least square method set up by underground working data as independent variable, oil pipe predict etch rate data as because become
The tube corrosion forecast model of amount, and according to the model, extent of corrosion prediction is carried out to oil pipe, so as to improve oil field well
The degree of accuracy of tube corrosion degree prediction.
Brief description of the drawings
Technical scheme in order to illustrate the embodiments of the present invention more clearly, makes required in being described below to embodiment
Accompanying drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the present invention, for
For those of ordinary skill in the art, on the premise of not paying creative work, other can also be obtained according to these accompanying drawings
Accompanying drawing.
Fig. 1 is tube corrosion degree Forecasting Methodology flow chart provided in an embodiment of the present invention;
Fig. 2 is tube corrosion degree Forecasting Methodology flow chart provided in an embodiment of the present invention;
Fig. 3 is tube corrosion degree Forecasting Methodology flow chart provided in an embodiment of the present invention;
Fig. 4 is tube corrosion degree prediction meanss structural representation provided in an embodiment of the present invention;
Fig. 5 is tube corrosion degree prediction meanss structural representation provided in an embodiment of the present invention.
Embodiment
To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with accompanying drawing to embodiment party of the present invention
Formula is described in further detail.
Embodiment one
The embodiments of the invention provide a kind of tube corrosion degree Forecasting Methodology, this method is according to collection, arrangement in engineering
Underground working data and tube corrosion situation data, first build variable (actual corrosion rate) with independent variable (corrosion impact
Factor) database, then set up tube corrosion forecast model using a kind of algorithm of new solution Fuzzy Linear Regression, finally
The tube corrosion result under particular cases is predicted using the tube corrosion forecast model of foundation.
Fig. 1 show a kind of tube corrosion degree Forecasting Methodology flow chart provided in an embodiment of the present invention, and this method includes
Following steps:
S0:Underground working data and the actual etch rate data of corresponding oil pipe in multigroup varying environment are obtained, number is stored in
According in storehouse.Wherein, underground working data include the data of multiple influence factors, for example, can include:Temperature, pressure, corrosion are situated between
Matter concentration (such as CO2、Cl-、H2S concentration) etc..The actual corrosion rate of oil pipe defines method and is:To the corrosion failure feelings of on-the-spot record
Condition is classified, and records corrosion pit depth, and the corrosion class drafted has 4 grades:1. without any corrosion;2. slight erosion (box cupling,
Tube wall pit shape etc.);3. heavy corrosion (box cupling, tube wall hole shape, sheet etc.);4. corrosion failure (box cupling or body
Corrosion failure etc.), then, with reference to oil pipe failure level, according to corrosion pit depth and oil pipe service life, define oil pipe actual rotten
Lose speed.It is preferred that obtaining the CO under multigroup different situations in this step2、Cl-、H2S, temperature and pressure data and corresponding
In tube corrosion level data deposit database.
S1:In database, the actual corrosion speed of at least two groups underground working data and its corresponding oil pipe of storage is chosen
Rate data, and the data of selection are normalized, the data after normalization are divided into training data sample and inspection data sample
This.It is preferred that, in this step, CO is chosen in database2Concentration, Cl-Concentration, H2S concentration, temperature and pressure data (i.e. underground
Floor data includes CO2Concentration, Cl-Concentration, H2S concentration, temperature and pressure data) and its actual corrosion rate of corresponding oil pipe
Data, and these data are normalized.
It is preferred that, in this step, the group number of the data of selection is more than the number of the independent variable described in step S2, wherein,
Inspection data sample can be one group, and the group number of training data sample is not less than the number of the independent variable.
S2:With reference to Fuzzy Linear Regression method and least square method, training data sample is modeled, obtained by underground work
Condition data as independent variable, the actual etch rate data of oil pipe as variable tube corrosion forecast model.
S3:The correctness of the tube corrosion forecast model is verified using inspection data sample, if upchecked, is performed
S4, otherwise, goes to S1, chooses new Data duplication and performs flow.
S4:Using resulting tube corrosion forecast model, extent of corrosion is carried out to the in-service oil pipe under designated environment pre-
Survey.
In above-mentioned steps S1, normalized uses linear-scale algorithm, it is possible to achieve initial data (in database
The measured data of storage) change to the sized data of required range scale.Built using the data after dimensional variation
Mould, model is inaccurate caused by can eliminating the data using different scale scope.
Specifically, the process that the data to selection are carried out can be as follows:
First, it is determined that the span for the data chosen, and determines normalized range scale.
If z represents the data chosen, i.e., without sized data sample, its span is z ∈ [zmin,zmax], warp
Cross the use of the data after the normalization obtained after linear-scale algorithm processRepresent, its span is
Then, according to below equation and the data chosen, the data after normalization are drawn.
Z andTherebetween conversion formula can be:
Above formula (13) realizes the sized data for initial data being changed to required range scale, and formula (14) is real
Show and sized data are changed to initial data again.(13) formula of utilization can selected underground working data and its
The corresponding actual etch rate data of oil pipe is transformed in new range scale (e.g., [- 1,1]).
With reference to Fuzzy Linear Regression method, the implementation of S2 in flow shown in Fig. 1 is illustrated.
Fuzzy Linear Regression method is first introduced, it is similar with classical linear regression analysis, in Fuzzy Linear Regression, if
Variable y independent variable xs related to it1、x2、…、xmJust like offline sexual intercourse:
Y=A1x1+A2x2+…+Amxm (1)
Regression analysis is to utilize known n groups observation data yi、x1i、x2i、…、xmi(wherein, i=1,2 ..., n), go to estimate
Count regression coefficient Aj.But in fuzzy linear regression analysis, it is believed that model has ambiguity, i.e. regression coefficient AjIt is fuzzy number, in
It is the match value of modelWith observation yiBetween deviation be as caused by this ambiguity.The one of typical linear regression model
As representation be:
yi=a1x1i+a2x2i+…+amxmi (8)
Generally take AjFor Triangular Fuzzy Number A (a, c), its membership function is:
The membership function for drawing y in formula (1) by above formula (2) is:
To make fitting function (1) observe data y to known n groupsi、x1i、x2i、…、xmi(wherein, i=1,2 ..., n) intend
Close best, in linear regression analysis, it is necessary to while meeting following 2 criterions:
(a) the fuzzy amplitude sum of each regression coefficient must be made minimum (i.e. precision is maximum), i.e.,:
(b) according to certain confidence level H, it is necessary to all observation data y can be coveredi, i.e.,:
hy(yi)≥H,0≤H≤1 (5)
The criterion ensure that no degree of membership<H yi。
It can be obtained according to formula (3) and formula (5):
In Fuzzy Linear Regression method, convolution (4) and formula (6) can will be converted into solution the problem of solving Fuzzy Linear Regression
Linear programming shown in equation group (7):
A is can obtain after solving equation group (7)j(aj,cj) value, so as to obtain yiWith independent variable x1、x2、…、xmBetween
Function expression.
The embodiment of the present invention use above-mentioned Fuzzy Linear Regression idea about modeling, but do not use above-mentioned equation group (7) come
Fuzzy coefficient is determined, but fuzzy coefficient is calculated with reference to Fuzzy Linear Regression method and least square method, Fig. 1 is illustrated in figure 2
Middle step S2 specific implementation method flow, may include steps of:
S21:Least square method is used to the general type (such as following (formula 15)) of typical linear regression model, and based on instruction
Practice data sample, determine aj, wherein, xjiThe data of j-th of the influence factor substituted into i-th group of training data sample, yiSubstitute into
The actual etch rate data of oil pipe in i-th group of training data sample, j=1,2 ..., m.
yi=a1x1i+a2x2i+…+amxmi (15)
If the match value of above-mentioned model isSolved using least square method:
It can be seen from mathematical knowledge, whenWhen, the Q in formula (16) obtains extreme value, i.e. need full
Foot:
The Blur center value a of the Estimates of Fuzzy Linear Regression Model shown in formula (15) can be obtained as equation group (17)1、a2、
a3、...、am, the center simulation model y of typical linear recurrence can be obtainedi=a1x1i+a2x2i+…+amxmi。
S22:According to following formula (10) and training data sample and a determinedj, determine cj。
Specifically, solving fuzzy amplitude c according to above-mentioned formula (6)jDuring, it can obtain with following formula (10), according to
Formula (10) and training data sample and a determinedjC can be calculatedj.Finally, the unsharp line shown in formula (11) can be obtained
Property regression model.
S23:According to the membership function (shown in formula (2)) of Triangular Fuzzy Number and a determinedj、cj, draw Triangular Fuzzy Number
Aj(aj,cj) value, be used as Estimates of Fuzzy Linear Regression Model y=A1x1+A2x2+…+AmxmIn regression coefficient Aj, this is obscured
Linear regression model (LRM) is used as the tube corrosion forecast model.The Estimates of Fuzzy Linear Regression Model is also denoted as such as following formula
(11)。
Wherein, Aj(aj,cj) it is Triangular Fuzzy Number, its value can change according to the difference of z values within the scope of one.
CO2Concentration, Cl-Concentration, H2S concentration, temperature and pressure are larger for the corrosion rate influence of oil pipe, for underground
Floor data includes CO2Concentration, Cl-Concentration, H2The situation of S concentration, temperature and pressure data, CO2Concentration, Cl-Concentration, H2S is dense
The model of degree, temperature and pressure data and oil pipe prediction etch rate data can be as follows:
ΔVcorr=f (Δ CO2,ΔCl-,ΔH2S,ΔF,ΔT) (9)
Wherein, f () is Estimates of Fuzzy Linear Regression Model, and formula (11) is substituted into formula (9), then can obtain tube corrosion prediction
Model:
Vcorr=(a1,c1)*CO2+(a2,c2)*Cl-+(a3,c3)*ΔH2S+(a4,c4)*ΔF+(a5,c5)*ΔT (12)
Here, due to Aj(aj,cj) it is Triangular Fuzzy Number, and non-determined numerical value, so, in actual applications, can basis
Concrete condition sets z values, to determine each AjNumerical value, for example, can be to all AjThe value for setting z is aj+cj。
In step s3, the correctness of tube corrosion forecast model is verified using inspection data sample, can be with specific as follows:
Underground working data in inspection data sample are substituted into the tube corrosion forecast model, oil pipe prediction corrosion is obtained
Speed data, the actual etch rate data of the corresponding oil pipe of the underground working data and the oil pipe in inspection data sample is pre-
The difference for surveying etch rate data is compared with the ratio of the actual etch rate data of the oil pipe with predetermined threshold value, to verify the oil pipe
The correctness of corrosion prediction model, if the ratio is no more than predetermined threshold value, is verified, otherwise, authentication failed.Specifically may be used
With equation below:
In formula, as long as Ω value is within 30% (predetermined threshold value), then it is assumed that be verified.If authentication failed, say
The precision of bright model is unsatisfactory for use requirement, can re-execute step S1, S2, carries out data-optimized selection, or collect foot
After enough data samples, then it is modeled.
In the embodiment of the present invention, based on actually detected underground working data and the actual corrosion rate number of corresponding oil pipe
According to by being used as independent variable, oil pipe prediction corruption by underground working data with reference to Fuzzy Linear Regression method and least square method foundation
Speed data is lost as the tube corrosion forecast model of dependent variable, and according to the model, extent of corrosion prediction is carried out to oil pipe, from
And the degree of accuracy of oil field well tube corrosion degree prediction can be improved.
Embodiment two
The embodiments of the invention provide a kind of tube corrosion degree Forecasting Methodology, multigroup difference is prestored in database
Underground working data and the actual etch rate data of corresponding oil pipe in environment, the underground working data include multiple influences because
The data of element, as shown in figure 3, this method may include steps of:
S1, chooses at least two groups underground working data and its actual etch rate data of corresponding oil pipe of storage, and right
The data of selection are normalized, and the data after normalization are divided into training data sample and inspection data sample;
S2, with reference to the Fuzzy Nonlinear Return Law and least square method, is modeled to the training data sample, obtain by
Underground working data predict etch rate data as the tube corrosion forecast model of dependent variable as independent variable, oil pipe;
S3, the correctness of the tube corrosion forecast model is verified using the inspection data sample, if upchecked,
S4 is then performed, otherwise, S1 is gone to, new Data duplication is chosen and performs flow;
S4, using resulting tube corrosion forecast model, carries out extent of corrosion pre- to the in-service oil pipe under designated environment
Survey.
In the embodiment of the present invention, based on actually detected underground working data and the actual corrosion rate number of corresponding oil pipe
According to by being used as independent variable, oil pipe prediction by underground working data with reference to the Fuzzy Nonlinear Return Law and least square method foundation
Etch rate data and according to the model, extent of corrosion prediction is carried out to oil pipe as the tube corrosion forecast model of dependent variable,
So as to improve the degree of accuracy of oil field well tube corrosion degree prediction.
Embodiment three
Based on identical technical concept, the embodiments of the invention provide a kind of tube corrosion degree prediction meanss, such as Fig. 4 institutes
Show, described device includes:
Memory module 410, it is actual rotten for prestoring underground working data and corresponding oil pipe in multigroup varying environment
Speed data is lost, the underground working data include the data of multiple influence factors;
Modeling module 420, for performing:S1, chooses at least two groups underground working data and its corresponding oil pipe of storage
Actual etch rate data, and the data of selection are normalized, by the data after normalization be divided into training data sample and
Inspection data sample;
S2, with reference to Fuzzy Linear Regression method and least square method, is modeled to the training data sample, obtains by well
Lower floor data predicts etch rate data as the tube corrosion forecast model of dependent variable as independent variable, oil pipe;
S3, the correctness of the tube corrosion forecast model is verified using the inspection data sample, if upchecked,
Then notify prediction module to perform S4, otherwise, go to S1, choose new Data duplication and perform flow;
Prediction module 430, for performing:S4, using resulting tube corrosion forecast model, under designated environment
Use as a servant oil pipe and carry out extent of corrosion prediction.
It is preferred that, the group number of the training data sample is not less than the number of the independent variable.
It is preferred that, the modeling module 420, specifically for:
It is determined that the span z ∈ [z for the data chosenmin,zmax], and determine normalized range scaleWherein z is the data chosen,For the data after normalization;
According toWith the data of the selection, the data after normalization are drawn.
It is preferred that, the underground working data include CO2Concentration, Cl-Concentration, H2S concentration, temperature and pressure data.
It is preferred that, the modeling module 420, specifically for:
To typical linear regression model yi=a1x1i+a2x2i+…+amxmiUsing least square method, and based on training data
Sample, determines aj, wherein, xjiThe data of j-th of the influence factor substituted into i-th group of training data sample, yiSubstitute into i-th group of instruction
Practice the actual etch rate data of oil pipe in data sample, j=1,2 ..., m;
According toWith the training data sample and a determinedj, determine cj, wherein, H
For default confidence value, and 0≤H≤1;
According to the membership function of Triangular Fuzzy NumberWith a determinedj、cj,
Draw Triangular Fuzzy Number Aj(aj,cj) value, be used as Estimates of Fuzzy Linear Regression Model y=A1x1+A2x2+…+AmxmIn recurrence system
Number Aj, it regard the Estimates of Fuzzy Linear Regression Model as the tube corrosion forecast model.
It is preferred that, the modeling module 420, specifically for:
Underground working data in the inspection data sample are substituted into the tube corrosion forecast model, oil pipe are obtained pre-
Survey etch rate data, by the actual etch rate data of the underground working data corresponding oil pipe in the inspection data sample and
The difference of the oil pipe prediction etch rate data is compared with the ratio of the actual etch rate data of the oil pipe with predetermined threshold value, to test
The correctness of the tube corrosion forecast model is demonstrate,proved, if the ratio is no more than predetermined threshold value, is verified, otherwise, checking
Failure.
In the embodiment of the present invention, based on actually detected underground working data and the actual corrosion rate number of corresponding oil pipe
According to by being used as independent variable, oil pipe prediction corruption by underground working data with reference to Fuzzy Linear Regression method and least square method foundation
Speed data is lost as the tube corrosion forecast model of dependent variable, and according to the model, extent of corrosion prediction is carried out to oil pipe, from
And the degree of accuracy of oil field well tube corrosion degree prediction can be improved.
Example IV
Based on identical technical concept, the embodiments of the invention provide a kind of tube corrosion degree prediction meanss, such as Fig. 5 institutes
Show, described device includes:
Memory module 510, it is actual rotten for prestoring underground working data and corresponding oil pipe in multigroup varying environment
Speed data is lost, the underground working data include the data of multiple influence factors;
Modeling module 520, for performing:S1, chooses at least two groups underground working data and its corresponding oil pipe of storage
Actual etch rate data, and the data of selection are normalized, by the data after normalization be divided into training data sample and
Inspection data sample;
S2, with reference to the Fuzzy Nonlinear Return Law and least square method, is modeled to the training data sample, obtain by
Underground working data predict etch rate data as the tube corrosion forecast model of dependent variable as independent variable, oil pipe;
S3, the correctness of the tube corrosion forecast model is verified using the inspection data sample, if upchecked,
Then notify prediction module to perform S4, otherwise, go to S1, choose new Data duplication and perform flow;
Prediction module 530, for performing:S4, using resulting tube corrosion forecast model, under designated environment
Use as a servant oil pipe and carry out extent of corrosion prediction.
In the embodiment of the present invention, based on actually detected underground working data and the actual corrosion rate number of corresponding oil pipe
According to by being used as independent variable, oil pipe prediction by underground working data with reference to the Fuzzy Nonlinear Return Law and least square method foundation
Etch rate data and according to the model, extent of corrosion prediction is carried out to oil pipe as the tube corrosion forecast model of dependent variable,
So as to improve the degree of accuracy of oil field well tube corrosion degree prediction.
It should be noted that:The tube corrosion degree prediction meanss that above-described embodiment is provided are pre- in progress tube corrosion degree
, can be as needed and by above-mentioned functions only with the division progress of above-mentioned each functional module for example, in practical application during survey
Distribute and completed by different functional modules, i.e., the internal structure of device is divided into different functional modules, retouched with completing the above
The all or part of function of stating.In addition, tube corrosion degree prediction meanss and tube corrosion degree that above-described embodiment is provided
Forecasting Methodology embodiment belongs to same design, and it implements process and refers to embodiment of the method, repeats no more here.
The embodiments of the present invention are for illustration only, and the quality of embodiment is not represented.
One of ordinary skill in the art will appreciate that realizing that all or part of step of above-described embodiment can be by hardware
To complete, the hardware of correlation can also be instructed to complete by program, described program can be stored in a kind of computer-readable
In storage medium, storage medium mentioned above can be read-only storage, disk or CD etc..
The foregoing is only presently preferred embodiments of the present invention, be not intended to limit the invention, it is all the present invention spirit and
Within principle, any modification, equivalent substitution and improvements made etc. should be included in the scope of the protection.
Claims (8)
1. a kind of tube corrosion degree Forecasting Methodology, it is characterised in that well in multigroup varying environment is prestored in database
The actual etch rate data of lower floor data and corresponding oil pipe, the underground working data include CO2Concentration, Cl-Concentration, H2S
Concentration, temperature and pressure data, methods described include:
S1, chooses at least two groups underground working data and its actual etch rate data of corresponding oil pipe of storage, and to choosing
Data be normalized, the data after normalization are divided into training data sample and inspection data sample;
S2, to typical linear regression model yi=a1x1i+a2x2i+…+amxmiUsing least square method, and based on training data sample
This, determines aj, wherein, xjiThe data of j-th of the influence factor substituted into i-th group of training data sample, yiSubstitute into i-th group of training
The actual etch rate data of oil pipe in data sample, j=1,2 ..., m;
According toWith the training data sample and a determinedj, determine cj, wherein, H is pre-
If confidence value, and 0≤H≤1;
According to the membership function of Triangular Fuzzy NumberWith a determinedj、cj, draw
Triangular Fuzzy Number Aj(aj, cj) value, a and the c are constant, are used as Estimates of Fuzzy Linear Regression Model y=A1x1+A2x2
+…+AmxmIn regression coefficient Aj, using the Estimates of Fuzzy Linear Regression Model as the tube corrosion forecast model, the z is choosing
The floor data and its corresponding etch rate data taken, the x1, x2..., xmFor the related independent variable of the y, the y is
The actual corrosion rate of oil pipe, the x1, x2..., xmFor the underground working data;
S3, the correctness of the tube corrosion forecast model is verified using the inspection data sample, if upchecked, is held
Row S4, otherwise, goes to S1, chooses new Data duplication and performs flow;
S4, using resulting tube corrosion forecast model, extent of corrosion prediction is carried out to the in-service oil pipe under designated environment.
2. according to the method described in claim 1, it is characterised in that the group number of the training data sample is not less than described from change
The number of amount.
3. according to the method described in claim 1, it is characterised in that the data of described pair of selection are normalized, it is specially:
It is determined that the span z ∈ [z for the data chosenmin, zmax], and determine normalized range scaleIts
Middle z is the data chosen,For the data after normalization;
According toWith the data of the selection, the data after normalization are drawn.
4. according to the method described in claim 1, it is characterised in that described to verify the oil pipe using the inspection data sample
The correctness of corrosion prediction model, be specially:
Underground working data in the inspection data sample are substituted into the tube corrosion forecast model, oil pipe prediction are obtained rotten
Lose speed data, by the actual etch rate data of the corresponding oil pipe of the underground working data in the inspection data sample with it is described
The difference of oil pipe prediction etch rate data is compared with the ratio of the actual etch rate data of the oil pipe with predetermined threshold value, to verify
The correctness of tube corrosion forecast model is stated, if the ratio is no more than predetermined threshold value, is verified, otherwise, authentication failed.
5. a kind of tube corrosion degree prediction meanss, it is characterised in that described device includes:
Memory module, for prestoring underground working data and the actual corrosion rate number of corresponding oil pipe in multigroup varying environment
According to the underground working data include CO2Concentration, Cl-Concentration, H2S concentration, temperature and pressure data;
Modeling module, for performing:S1, chooses at least two groups underground working data and its actual corrosion of corresponding oil pipe of storage
Speed data, and the data of selection are normalized, the data after normalization are divided into training data sample and inspection data
Sample;
S2, to typical linear regression model yi=a1x1i+a2x2i+…+amxmiUsing least square method, and based on training data sample
This, determines aj, wherein, xjiThe data of j-th of the influence factor substituted into i-th group of training data sample, yiSubstitute into i-th group of training
The actual etch rate data of oil pipe in data sample, j=1,2 ..., m;
According toWith the training data sample and a determinedj, determine cj, wherein, H is pre-
If confidence value, and 0≤H≤1;
According to the membership function of Triangular Fuzzy NumberWith a determinedj、cj, draw
Triangular Fuzzy Number Aj(aj, cj) value, be used as Estimates of Fuzzy Linear Regression Model y=A1x1+A2x2+…+AmxmIn regression coefficient
Aj, using the Estimates of Fuzzy Linear Regression Model as the tube corrosion forecast model, the z is for the floor data chosen and its correspondingly
Etch rate data, the x1, x2..., xmFor the related independent variable of the y, the y is the actual corrosion rate of oil pipe, described
x1, x2..., xmFor the underground working data;
S3, the correctness of the tube corrosion forecast model is verified using the inspection data sample, if upchecked, is led to
Know that prediction module performs S4, otherwise, go to S1, choose new Data duplication and perform flow;
Prediction module, for performing:S4, using resulting tube corrosion forecast model, to the in-service oil pipe under designated environment
Carry out extent of corrosion prediction.
6. device according to claim 5, it is characterised in that the group number of the training data sample is not less than described from change
The number of amount.
7. device according to claim 5, it is characterised in that the modeling module, specifically for:
It is determined that the span z ∈ [z for the data chosenmin, zmax], and determine normalized range scaleIts
Middle z is the data chosen,For the data after normalization;
According toWith the data of the selection, the data after normalization are drawn.
8. device according to claim 5, it is characterised in that the modeling module, specifically for:
Underground working data in the inspection data sample are substituted into the tube corrosion forecast model, oil pipe prediction are obtained rotten
Lose speed data, by the actual etch rate data of the corresponding oil pipe of the underground working data in the inspection data sample with it is described
The difference of oil pipe prediction etch rate data is compared with the ratio of the actual etch rate data of the oil pipe with predetermined threshold value, to verify
The correctness of tube corrosion forecast model is stated, if the ratio is no more than predetermined threshold value, is verified, otherwise, authentication failed.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210548324.XA CN103870670B (en) | 2012-12-17 | 2012-12-17 | A kind of tube corrosion degree Forecasting Methodology and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210548324.XA CN103870670B (en) | 2012-12-17 | 2012-12-17 | A kind of tube corrosion degree Forecasting Methodology and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103870670A CN103870670A (en) | 2014-06-18 |
CN103870670B true CN103870670B (en) | 2017-10-17 |
Family
ID=50909197
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201210548324.XA Active CN103870670B (en) | 2012-12-17 | 2012-12-17 | A kind of tube corrosion degree Forecasting Methodology and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103870670B (en) |
Families Citing this family (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105678053A (en) * | 2015-12-29 | 2016-06-15 | 中国石油天然气集团公司 | Method for predicting corrosion and invalidity of in-service tubing strings based on data mining |
CN105653791A (en) * | 2015-12-29 | 2016-06-08 | 中国石油天然气集团公司 | Data mining based corrosion failure prediction system for on-service oil pipe column |
BR112019009500A2 (en) * | 2016-12-21 | 2019-08-06 | Gen Electric | corrosion protection for air-cooled capacitors |
CN106951616B (en) * | 2017-03-10 | 2020-01-14 | 西安交通大学 | Carbon steel pipeline CO based on computational fluid mechanics2Method for predicting corrosion rate of solution |
CN109710410B (en) * | 2018-12-24 | 2021-08-13 | 微梦创科网络科技(中国)有限公司 | Internet information resource allocation method and device |
CN110751339A (en) * | 2019-10-24 | 2020-02-04 | 北京化工大学 | Method and device for predicting corrosion rate of pipeline and computer equipment |
CN111177947B (en) * | 2020-01-12 | 2022-02-22 | 西南石油大学 | Multi-factor considered CO2Corrosion prediction plate establishing method |
CN111798930B (en) * | 2020-07-17 | 2022-02-25 | 西南石油大学 | CO considering influence of corrosion product film2Corrosion rate prediction method |
CN112668206B (en) * | 2021-01-20 | 2022-03-15 | 西南石油大学 | Multi-factor-considered acid gas field corrosion prediction model and parameter determination method |
CN113806964B (en) * | 2021-09-30 | 2022-08-23 | 西南石油大学 | Corrosion and scaling rate prediction method considering multi-factor coupling effect |
CN115506777B (en) * | 2022-10-08 | 2023-09-08 | 中国石油大学(北京) | Method and device for determining safety coefficient of sleeve |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102052076A (en) * | 2009-10-30 | 2011-05-11 | 中国石油化工股份有限公司 | System for monitoring components of shaft fluid of H2S/CO2-containing gas field and analysis method thereof |
CN102282411A (en) * | 2009-01-19 | 2011-12-14 | Bp北美公司 | Method and system for predicting corrosion rates using mechanistic models |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012202792A (en) * | 2011-03-25 | 2012-10-22 | Tokyo Electric Power Co Inc:The | Life prediction method for optical fiber compound overhead ground line |
-
2012
- 2012-12-17 CN CN201210548324.XA patent/CN103870670B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102282411A (en) * | 2009-01-19 | 2011-12-14 | Bp北美公司 | Method and system for predicting corrosion rates using mechanistic models |
CN102052076A (en) * | 2009-10-30 | 2011-05-11 | 中国石油化工股份有限公司 | System for monitoring components of shaft fluid of H2S/CO2-containing gas field and analysis method thereof |
Non-Patent Citations (1)
Title |
---|
松辽盆地北部深层气井CO2腐蚀预测方法;刚振宝等;《天然气勘探与开发》;20080630;第64-68页 * |
Also Published As
Publication number | Publication date |
---|---|
CN103870670A (en) | 2014-06-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103870670B (en) | A kind of tube corrosion degree Forecasting Methodology and device | |
Yan et al. | Harnessing the power of machine learning for carbon capture, utilisation, and storage (CCUS)–a state-of-the-art review | |
Ayvaz | A linked simulation–optimization model for solving the unknown groundwater pollution source identification problems | |
CN104063588B (en) | Multi-source data fusion-based method for predicting pipeline corrosion defect size | |
Kim et al. | Global and local parameters for characterizing and modeling external corrosion in underground coated steel pipelines: A review of critical factors | |
Wang et al. | Study of electrochemical corrosion on Q235A steel under stray current excitation using combined analysis by electrochemical impedance spectroscopy and artificial neural network | |
Muther et al. | Socio-inspired multi-cohort intelligence and teaching-learning-based optimization for hydraulic fracturing parameters design in tight formations | |
Ahmadi et al. | Fuzzy modeling and experimental investigation of minimum miscible pressure in gas injection process | |
Wolodko et al. | Modeling of microbiologically influenced corrosion (MIC) in the oil and gas industry-past, present and future | |
Ponomareva et al. | A new approach to predict the formation pressure using multiple regression analysis: Case study from Sukharev oil field reservoir–Russia | |
CN114818363A (en) | Urban deep drainage tunnel silting risk assessment and prediction early warning method | |
Krechowicz et al. | Risk assessment in energy infrastructure installations by horizontal directional drilling using machine learning | |
RU2459136C2 (en) | Method to monitor pipeline corrosion and device for its realisation | |
Brodny et al. | The use of the neuro-fuzzy model to predict the methane hazard during the underground coal mining production process | |
Sambo et al. | Application of adaptive neuro-fuzzy inference system and optimization algorithms for predicting methane gas viscosity at high pressures and high temperatures conditions | |
WO2010065343A2 (en) | Method of predicting the location of microbiologically influenced corrosion of underground items | |
Du et al. | Deeppipe: Theory-guided prediction method based automatic machine learning for maximum pitting corrosion depth of oil and gas pipeline | |
Zhang et al. | SVM model for estimating the parameters of the probability-integral method of predicting mining subsidence | |
Hilgefort | Big data analysis using Bayesian network modeling: a case study with WG-ICDA of a gas storage field | |
CN107860795A (en) | Multifunctional corrosion real-time monitoring device and method in a kind of gun barrel | |
Guan et al. | Application of probabilistic model in pipeline direct assessment | |
Zhang et al. | Hybrid BN approach to analyzing risk in tunnel-induced bridge damage | |
CN104463362A (en) | Reservoir fluid yield prediction method and device | |
Katashov et al. | Digital platform as a tool for efficient reservoir management | |
Zhao et al. | Application of internal corrosion direct assessment in CO2 slug flow submarine pipelines |
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 |