CN103967477B - A kind of horizontal well parameter detection method based on conducting probe array and information fusion technology - Google Patents

Abstract

The present invention relates to a kind of horizontal well parameter detection method based on conducting probe array and information fusion technology, the multiple parameters that can be used in centering low production liquid horizontal well are measured simultaneously. The method is demarcated the sample set of setting up conducting probe array output vector by experiment; Utilize sample set to determine that electricity leads threshold value, the data of conducting probe array response vector are carried out to pretreatment; The data of pretreatment vector are carried out to standardization; In conjunction with the information fusion method based on degree of belief function, the data of normalized response vector are carried out to similarity matching, obtain matched sample; Utilize matched sample to obtain the estimated value of oil-water interfaces height, conducting probe array orientation angle and water electrical conductivity. The invention has the beneficial effects as follows oil-water interfaces height, conducting probe array orientation angle and the water electrical conductivity on-line measurement simultaneously that can utilize in conducting probe array centering low production liquid horizontal well, made up the deficiency that conventional logging method is applied in horizontal well, certainty of measurement is high, good reliability.

Description

A kind of horizontal well parameter detection method based on conducting probe array and information fusion technology

Technical field

The present invention relates to a kind of horizontal well parameter detection method based on conducting probe array and information fusion technology, can be used forMultiple parameters in centering low production liquid horizontal well are measured simultaneously.

Background technology

Horizontal well technology is the new technology that China develops rapidly the late 20th century, owing to improving output and exploitation effectBenefit aspect has shown huge potentiality and advantage, has obtained the generally attention in oil-gas field development field. In theory,In horizontal well, may there is oil-water stratified flow (ST), oil-water stratified flow (ST&MI), water with interface fluctuationBag oil and water layer (Do/w&w), oil-in-water (o/w), Water-In-Oil and oil-in-water (Dw/o&o/w) and Water-In-Oil (w/o)Etc. multiple flow pattern. But requiring of every kind of flow pattern flow operating mode is different, stratified flow (ST) to the requirement of production fluid amountLow, and Water-In-Oil flow pattern (w/o) is all very harsh to the requirement of production fluid amount and moisture content. Recent research result shows, asIn the horizontal pipe that it is 125mm that fruit is wanted at internal diameter, obtain above-mentioned whole flow patterns, the adjustable range of total flow need cover 0～500m³/ d, the adjustable range of moisture content need cover 0～90%. At present, most domestic oil field entered low production fluid,High moisture mining period, the oil well of some oil field, old liberated area ultra-high water-containing, special low production fluid is also very common. Each large oil at homeField, total fluid production is lower than 30m³The low-yield liquid oil well of/d occupies certain ratio, and total fluid production is lower than 60m³In/dThe share that low liquid producing well has overwhelming superiority. Therefore, above-mentioned flow pattern is not general in the horizontal well of each elephant at homeAll over exist, and only have stratified flow (ST) and stratified-wavy flow (ST&MI) more common.

Due to the metering system of production profile logging employing intrusive mood, generally, the area on apparatus measures cross section is heldActual specific retention in water rate and well is not etc. In addition, obtain the separate phase flow rate of horizontal well profit two-phase, also need to understandThe VELOCITY DISTRIBUTION of each phase fluid. Under normal circumstances, specific retention measurement mechanism and flow rate measuring device are installed on loggerDiverse location, the specific retention that they are measured on cross section separately may not wait, but oil-water interfaces height is consistent. Therefore,Solve the measurement problem of low production liquid horizontal well specific retention, be applicable to adopting indirect measurement strategies, first measure oil water surfaceHeight, then calculate specific retention according to geometrical correspondence between oil-water interfaces height and specific retention.

Summary of the invention

In recent years, horizontal well drilling technology is day by day perfect, and most oil fields all have the horizontal well of higher proportion to come into operation. LongThough the exploitation of phase has obtained certain economic benefit, the production status of horizontal well also becomes more unstable. Part oilEven there is moisture content fast rise in well, the phenomenon that oil production significantly declines. Under the condition of this complexity, traditional productGoing out profile measurement method is difficult to be applied directly in horizontal well. For profit two phase stratification stream, if oil-water interfaces height hKnown, just can utilize the geometrical correspondence between oil-water interfaces height h and specific retention to calculate specific retention value. For realizingThis purpose, this patent is for the flow pattern feature of middle low production liquid horizontal well, in conjunction with conducting probe array and information fusion technologyPropose a kind of horizontal well parameter detection method, can realize oil-water interfaces height h, conducting probe array orientation angle θ and waterWhen conductivityσ, measure. This patent is characterised in that and comprises the following steps:

A. utilize calibration experiment to set up sample set S;

In the calibration process of sample set, keep water conductivityσ=σ_cConstant, adjust oil-water interfaces height h and electricity and lead spyPin array orientation angle θ; Under different parameters combination (h, θ), by the output response record of conducting probe array to sampleIn this collection S, and be labeled as sample vector C^s；

B. utilize sample set S to determine that electricity leads threshold value c_t；

Data arrangement in sample set S is become to an increasing sequence { c_n, the sequence of calculation { c_nDifference sequence { d_n, computing formula is:

d_n＝c_n+1-c_n，

In formula, c_nAnd c_n+1Represent respectively sequence { c_nIn sequence number be the element of n and n+1; Determine sequence { d_nIn the sequence number of greatest memberK, finds out sequence { c_nIn sequence number be respectively the element c of k and k+1_kAnd c_k+1, and calculate electricity and lead threshold value c_t, computing formula is:

$c_t=\frac{c_k+c_{k+1}}{2};$

C. the data of conducting probe array response vector C are carried out to pretreatment, obtain pretreatment vector T;

Pretreated computing formula is:

$t_i=\left\{\begin{array}{cc}{c}_i& c_i>c_t\\ 0& c_i\≤c_t\end{array}\right.,$

In formula, t_iAnd c_iBe respectively i element in vector T and vectorial C, c_tFor predefined electricity is led threshold value; Conducting probeThe response vector C of array can be sample vector C^sOr in-site measurement vector C^m, corresponding pretreatment vector is respectively T^sAnd T^m；

Generally, the relative dielectric constant of crude oil is 3～5, even if conducting probe is immersed in oil completely, and its output valveAlso non-vanishing, but be slightly larger than zero. Because this output valve is very little, be difficult to accurately measure. The response of conducting probe array toIn amount, calculate if such component is more and all participate in follow-up matching degree, will inevitably disturb to a certain extent sampleThe accuracy of coupling, and then affect final parameter Estimation result. In order to eliminate the impact of these components, mate at executing dataNeed sample data and measurement data to carry out threshold process before. The Main Function of threshold process is by concrete threshold decisionThe state of conducting probe, probe still contacts with water in oil phase, and the probe response in oil is replaced with " 0 ",Real response and its Mathematical Modeling of conducting probe array are consistent.

D. the data of conducting probe array pretreatment vector T are carried out to standardization, obtain normalized response vector R;

Standardized computing formula is:

$r_i=\frac{t_i}{t_{\max }},$

In formula, r_iAnd t_iBe respectively i element in vector T and vectorial R, t_maxFor the greatest member in vector T; VectorT can be by sample vector C^sOr in-site measurement vector C^mCalculate, corresponding standardized vector is respectively R^sWithR^m；

In formation water, be rich in mineral matter, its electrical conductivity is not constant conventionally, but along with well depth, the gentle well of well are pressed constantlyChange. In the time that the salinity of formation water raises, in conducting probe array response vector, the absolute value of each component also can becomeRatio increases. Because the response vector of conducting probe array is directly proportional to water electrical conductivity, in calibration process, can keepWater electrical conductivity is constant. While measurement at the scene, can obtain by data normalization the normalized response of conducting probe array,The impact of the variation of elimination water electrical conductivity on data similarity matching. Normalized response vector R only with oil-water interfaces heightH is relevant with azimuth angle theta; Data normalization has changed the absolute value of each component in vector T, but keeps the ratio between each componentExample relation is constant.

E. in conjunction with the information fusion method based on degree of belief function, to the number of conducting probe array normalized response vector RAccording to carrying out similarity matching, obtain matched sample M;

In sample set, finding out the sample consistent with the corresponding oil-water interfaces height of measurement data and azimuth, is oneData similarity matching problem. The similarity of data, refers to two similarity degrees between data object, and Data MatchingTo carry out registration according to certain inner link between data. Carry out data similarity coupling, need to select suitable similarDegree measure function. In the time carrying out horizontal well parameter detecting, also need to consider the reliability of measuring method. Due to underground surveyComplicated condition, harshness, the fault rate of logger is conventionally very high. In the process of going into the well or in measuring process, if portionDivide probe to break down, the authenticity of measurement data will be affected so. Like this, conventional similarity measure functionCan not meet the demands, but need to define a new similarity measure function:

$d_{\mathrm{fus}}=\underset{i=1}{\overset{24}{\mathrm{\Σ}}}{w}_i{x}_i,$

In formula, x_i＝|r_i ^s-r_i ^m|，w_iFor x_iWeight coefficient, r_i ^sAnd r_i ^mBe respectively normalized response vector R^sAnd R^mIIndividual element; Matched sample M makes similarity measure function d in sample set S_fusGet the sample of minimum of a value;

Weight coefficient w_iDeterministic process as follows:

First, definition x_iAnd x_jBetween degree of belief function be:

$b_{\mathrm{ij}}=\left\{\begin{array}{cc}1& |x_i-x_j|<M_1\\ \frac{e^{-|x_i-x_j|}-e^{-M_2}}{e^{-M_1}-e^{-M_2}}& M_1\&le;|x_i-x_j|<M_2\\ 0& |x_i-x_j|\&GreaterEqual;M_2\end{array}\right.,$

In formula, M₂>M₁>0；

Then, degree of breaking the wall of mistrust matrix B

$B=\left[\begin{array}{cccc}{b}_{11}& b_{12}& \&CenterDot;\&CenterDot;\&CenterDot;& b_{1n}\\ b_{21}& b_{22}& \&CenterDot;\&CenterDot;\&CenterDot;& b_{2n}\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ b_{n1}& b_{n2}& \&CenterDot;\&CenterDot;\&CenterDot;& b_{\mathrm{nn}}\end{array}\right];$

Finally, determine the corresponding characteristic vector A of eigenvalue of maximum of matrix B, utilize vectorial A to calculate w_i, calculate publicFormula is:

$w_i=\frac{a_i}{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{a}_j},$

In formula, a_iAnd a_jBe respectively i and j element in vectorial A;

F. utilize matched sample M to obtain the estimation of oil-water interfaces height, conducting probe array orientation angle and water electrical conductivityValue h, θ and σ;

It is right that the estimated values theta at the estimated value h of oil-water interfaces height and conducting probe array orientation angle is respectively matched sample M instituteThe oil-water interfaces height of answering and conducting probe array orientation angle; The estimation values sigma of water electrical conductivity is:

$\mathrm{\&sigma;}=\frac{\underset{i=1}{\overset{24}{\mathrm{\&Sigma;}}}{t}_i^m}{\underset{i=1}{\overset{24}{\mathrm{\&Sigma;}}}{m}_i}\&CenterDot;{\mathrm{\&sigma;}}^c,$

In formula, σ^cFor the water electrical conductivity in calibration experiment, m_iFor i the component of matched sample M,For toAmount T^mI element.

The invention has the beneficial effects as follows the oil-water interfaces height that can utilize in conducting probe array centering low production liquid horizontal well,The on-line measurement simultaneously of conducting probe array orientation angle and water electrical conductivity. Made up conventional logging method applies in horizontal wellDeficiency, certainty of measurement is high, good reliability.

Brief description of the drawings

Fig. 1 is flow chart of the present invention;

Fig. 2 is the measurement sectional view of conducting probe array in detailed description of the invention, in figure: oil water surface (201), electricityLead probe (202), metal sleeve (203), apparatus wall (204), support arm (205);

Fig. 3 is the structure chart of experimental provision in detailed description of the invention, in figure: water injection hole (301), measurement electrode (302),Metallic cylinder (303), scale (304), switch switching circuit (305), hard ware measure circuit (306), upperMachine program (307).

Detailed description of the invention

Fig. 1 is flow chart of the present invention, and Fig. 2 is the measurement sectional view of conducting probe array in detailed description of the invention, and electricity is ledProbe array is made up of 24 conducting probes (202) that are distributed in inside loop and outside loop, is arranged on respectively 12 equal angles and dividesThe support arm (205) of cloth is upper, and apparatus wall (204) is coaxial with horizontal well, for fixing and installation conducting probe; LevelWell internal diameter is 125mm, and the external diameter of apparatus wall (204) is 50mm; The azimuth of conducting probe array is that it is around axisThe angle turning over counterclockwise, and specify the initial orientation angle for conducting probe array showing in Fig. 2; In the time measuring,Sleeve pipe (203) and apparatus wall (204) current potential equate, are considered to the ground of whole measuring system. Fig. 3 is concrete enforcementThe structure chart of experimental provision in mode, part of detecting is made up of metallic cylinder (303) and built-in concentric metal shaft, respectivelyBe used for simulation oil well sleeve pipe and measuring instrument outer wall, length is 300mm. The reserved water injection hole (301) in top of metal cylinderBe used for adjusting water level. One side of metal cylinder is provided with scale (304) and is used for reading water surface elevation; At opposite side, be provided with24 measurement electrode (302), are used for analog electrical to lead probe array. Measurement electrode is complied with by switch switching circuit (305)Inferior gating is connected to hard ware measure circuit (306), and host computer procedure (307) is used for synchro switch commutation circuit and hardwareMeasuring circuit, realizes collection and the storage of data. Existing accompanying drawings the specific embodiment of the present invention.

A. utilize calibration experiment to set up sample set S;

Because empty G&O has similar electrology characteristic, for ease of carrying out calibration experiment, in experiment, replace oil with air,Utilize the oil water surface under gas-water interface face mould pseudo level well stratified flow condition. Utilize conducting probe array at geometryOn symmetry simplify calibration process, azimuth angle theta only regulates within the scope of 0 °～15 °, regulate step-length be 3 °; Solid at certainUnder fixed azimuth, regulate the height of gas-water interface face in test pipeline section, increment of adjustment is 2% of test pipeline section diameter.In calibration process, use normal domestic running water, electrical conductivity is 0.078S/m, under different parameters combination (h, θ),The output response record of conducting probe array, in sample set S, and is labeled as to sample vector C^s. Because the number of electrodeAmount is 24, so the response vector of conducting probe array comprises 24 elements.

B. utilize sample set S to determine that electricity leads threshold value c_t；

To carry out threshold process to test data or sample data, first need definite electricity to lead threshold value c_t. By the number in sample set SAccording to being arranged in an increasing sequence { c_n, the sequence of calculation { c_nDifference sequence { d_n}，d_nComputing formula be:

d_n＝c_n+1-c_n，

In formula, c_nAnd c_n+1Represent respectively sequence { c_nIn sequence number be the element of n and n+1; Determine sequence { d_nIn the sequence number of greatest memberK, finds out sequence { c_nIn sequence number be respectively the element c of k and k+1_kAnd c_k+1, and calculate electricity and lead threshold value c_t, computing formula is:

$c_t=\frac{c_k+c_{k+1}}{2};$

According to calculating, finally determine c_t＝1.32×10^-5S。

C. the response vector C of conducting probe array is carried out to data pretreatment, obtain pretreatment vector T;

The pretreated computing formula of data is:

$t_i=\left\{\begin{array}{cc}{c}_i& c_i>c_t\\ 0& c_i\&le;c_t\end{array}\right.,$

In formula, t_iAnd c_iBe respectively i element in vector T and vectorial C, c_tFor predefined electricity is led threshold value; Conducting probeThe response vector C of array can be sample vector C^sOr in-site measurement vector C^m, corresponding pretreatment vector is respectively T^sAnd T^m。

D. the pretreatment vector T of conducting probe array is carried out to data normalization, obtain normalized response vector R;

The computing formula of data normalization is:

$r_i=\frac{t_i}{t_{\max }},$

In formula, r_iAnd t_iBe respectively i element in vector T and vectorial R, t_maxFor the greatest member in vector T; VectorT can be by sample vector C^sOr in-site measurement vector C^mCalculate, corresponding standardized vector is respectivelyR^sAnd R^m；

E. the normalized response vector R of conducting probe array is carried out to data similarity coupling, obtain matched sample M;

Definition similarity measure function is:

$d_{\mathrm{fus}}=\underset{i=1}{\overset{24}{\mathrm{\&Sigma;}}}{w}_i{x}_i,$

In formula, x_i＝|r_i ^s-r_i ^m|，w_iFor x_iWeight coefficient, r_i ^sAnd r_i ^mBe respectively normalized response vector R^sAnd R^mIIndividual element; Matched sample M makes similarity measure function d in sample set S_fusGet the sample of minimum of a value;

Weight coefficient w_iDeterministic process as follows:

First, definition x_iAnd x_jBetween degree of belief be:

$b_{\mathrm{ij}}=\left\{\begin{array}{cc}1& |x_i-x_j|<M_1\\ \frac{e^{-|x_i-x_j|}-e^{-M_2}}{e^{-M_1}-e^{-M_2}}& M_1\&le;|x_i-x_j|<M_2\\ 0& |x_i-x_j|\&GreaterEqual;M_2\end{array}\right.,$

In formula, M₂>M₁>0；

Then, degree of breaking the wall of mistrust matrix B

$B=\left[\begin{array}{cccc}{b}_{11}& b_{12}& \&CenterDot;\&CenterDot;\&CenterDot;& b_{1n}\\ b_{21}& b_{22}& \&CenterDot;\&CenterDot;\&CenterDot;& b_{2n}\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ b_{n1}& b_{n2}& \&CenterDot;\&CenterDot;\&CenterDot;& b_{\mathrm{nn}}\end{array}\right];$

Finally, determine the corresponding characteristic vector A of eigenvalue of maximum of matrix B, utilize vectorial A to calculate w_i, calculate publicFormula is:

$w_i=\frac{a_i}{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{a}_j},$

In formula, a_iAnd a_jBe respectively i and j element in vectorial A;

F. utilize matched sample M to obtain the estimation of oil-water interfaces height, conducting probe array orientation angle and water electrical conductivityValue h, θ and σ;

It is right that the estimated values theta at the estimated value h of oil-water interfaces height and conducting probe array orientation angle is respectively matched sample M instituteThe oil-water interfaces height of answering and conducting probe array orientation angle; The estimation values sigma of water electrical conductivity is:

$\mathrm{\&sigma;}=\frac{\underset{i=1}{\overset{24}{\mathrm{\&Sigma;}}}{t}_i^m}{\underset{i=1}{\overset{24}{\mathrm{\&Sigma;}}}{m}_i}\&CenterDot;{\mathrm{\&sigma;}}^c,$

In formula, σ^cFor the water electrical conductivity in calibration experiment, m_iFor i the component of matched sample M,ForVector T^mI element. Test result show the error of h, θ and σ be respectively 2.00%, 3.00 °, 1.89%,Illustrate that the method is feasible, has higher certainty of measurement and reliability.

Description to the present invention and embodiment thereof, is not limited to this above, is only enforcement of the present invention shown in accompanying drawingOne of mode. In the situation that not departing from the invention aim, without create design with this technical scheme similarStructure or embodiment, all belong to protection domain of the present invention.

Claims (1)

1. the horizontal well parameter detection method based on conducting probe array and information fusion technology, is characterized in that comprisingThe following step:

A. utilize calibration experiment to set up sample set S;

In the calibration process of sample set, keep water conductivityσ=σ_cConstant, adjust oil-water interfaces height h and electricity and lead spyPin array orientation angle θ; Under different parameters combination (h, θ), by the output response record of conducting probe array to sampleIn this collection S, and be labeled as sample vector C^s；

B. utilize sample set S to determine that electricity leads threshold value c_t；

Data arrangement in sample set S is become to an increasing sequence { c_n, the sequence of calculation { c_nDifference sequence { d_n, computing formula is:

d_n＝c_n+1-c_n，

In formula, c_nAnd c_n+1Represent respectively sequence { c_nIn sequence number be the element of n and n+1; Determine sequence { d_nIn the sequence number of greatest memberK, finds out sequence { c_nIn sequence number be respectively the element c of k and k+1_kAnd c_k+1, and calculate electricity and lead threshold value c_t, computing formula is:

$c_t=\frac{c_k+c_{k+1}}{2};$

C. the data of conducting probe array response vector C are carried out to pretreatment, obtain pretreatment vector T;

Pretreated computing formula is:

$t_i=\left\{\begin{array}{cc}{c}_i& c_i>c_t\\ 0& c_i\&le;c_t\end{array}\right.,$

In formula, t_iAnd c_iBe respectively i element in vector T and vectorial C, c_tFor predefined electricity is led threshold value; Conducting probeThe response vector C of array is sample vector C^sOr in-site measurement vector C^m, corresponding pretreatment vector is respectively T^sWithT^m；

D. the data of conducting probe array pretreatment vector T are carried out to standardization, obtain normalized response vector R;

Standardized computing formula is:

$r_i=\frac{t_i}{t_{max}},$ (formula (1))

In formula, r_iAnd t_iBe respectively i element in vector T and vectorial R, t_maxFor the greatest member in vector T; VectorT is by sample vector C^sOr in-site measurement vector C^mCalculate, corresponding standardized vector is respectively R^sAnd R^m；

E. in conjunction with the information fusion method based on degree of belief function, to the number of conducting probe array normalized response vector RAccording to carrying out similarity matching, obtain matched sample M;

Definition similarity measure function is:

$d_{fus}=\underset{i=1}{\overset{24}{\&Sigma;}}{w}_i{x}_i,$

In formula,w_iFor x_iWeight coefficient, r_i ^sAnd r_i ^mBe respectively normalized response vector R^sAnd R^mIIndividual element; Matched sample M makes similarity measure function d in sample set S_fusGet the sample of minimum of a value;

Weight coefficient w_iDeterministic process as follows:

First, definition x_iAnd x_jBetween degree of belief function be:

$b_{ij}=\left\{\begin{array}{cc}1& |x_i-x_j|<M_1\\ \frac{e^{-|x_i-x_j|}-e^{-M_2}}{e^{-M_1}-e^{-M_2}}& M_1\&le;|x_i-x_j|<M_2\\ 0& |x_i-x_j|\&GreaterEqual;M_2\end{array}\right.,$

In formula, M₂>M₁>0；

Then, degree of breaking the wall of mistrust matrix B is as follows:

$B=\left[\begin{array}{cccc}{b}_{11}& b_{12}& \mathrm{...}& b_{1n}\\ b_{21}& b_{22}& \mathrm{...}& b_{2n}\\ .& .& & .\\ .& .& & .\\ .& .& & .\\ b_{n1}& b_{n2}& \mathrm{...}& b_{nn}\end{array}\right];$

Finally, determine the corresponding characteristic vector A of eigenvalue of maximum of matrix B, utilize vectorial A to calculate w_i, calculate publicFormula is:

$w_i=\frac{a_i}{\underset{j=1}{\overset{n}{\&Sigma;}}{a}_j},$

In formula, a_iAnd a_jBe respectively i and j element in vectorial A;

F. utilize matched sample M to obtain the estimation of oil-water interfaces height, conducting probe array orientation angle and water electrical conductivityValueWith

The estimated value of oil-water interfaces heightEstimated value with conducting probe array orientation angleBe respectively matched sample M institute rightThe oil-water interfaces height of answering and conducting probe array orientation angle; The estimated value of water electrical conductivityFor:

$\widehat{\mathrm{\&sigma;}}=\frac{\underset{i=1}{\overset{24}{\&Sigma;}}{t}_i^m}{\underset{i=1}{\overset{24}{\&Sigma;}}{m}_i}\&CenterDot;{\mathrm{\&sigma;}}^c,$

In formula, σ^cFor the water electrical conductivity in calibration experiment, m_iFor i the component of matched sample M,For toAmount T^mI element.