RU2446419C1 - Hydrocarbon deposit exploration method - Google Patents

Abstract

FIELD: oil and gas industry.

SUBSTANCE: geochemical investigation of geological site is performed, training sample of geochemical and geophysical data is created on "oil - empty" training wells, geochemical anomaly is recorded, oil-drainage boundary is determined as per parameters of natural electric field EF and magnetic field MF, oil-drainage boundary is simulated by means of discriminant analysis method. Training sample of geochemical and geophysical data is created on "oil - residual oil - empty" training wells, and geochemical probability GP parameter is obtained. In case geochemical anomaly is not available at geological site (GP<50%), it is assumed as non-prospective and excluded from geophysical investigations; at sites with GP>50% there obtained is sampling of four parameters, such as EF, MF, GP and AO, which are coordinate located at common investigation points, complex probability parameter CPP is obtained, and CPP map determining the oil-drainage boundary of geological site is built.

EFFECT: increasing prediction reliability of hydrocarbon deposits.

6 dwg

Description

The invention relates to methods for prospecting and exploration of hydrocarbon deposits and can be used to detect oil and gas promising objects.

A known method of exploration of oil and gas deposits, based on the use of artificially neural networks and algorithms. The method involves seismic profiling using the common depth point method, obtaining time sections, obtaining a training sample within the analyzed window in the image of a section of seismic terraces of time sections, including images of tectonic-facies zones in the form of seismic geological bodies using a training system and mapping as an algorithm (patent RF №2094828, publ. 10/27/1997).

The disadvantage of this method is to obtain secondary signs of oil, the ambiguity of the interpretation of seismic time sections.

A known method of exploration - biogeochemical testing (BHCT). The method has been widely used in structural drilling (RF patent No. 2200334, publ. 03/10/2003). Prediction of the oil content of the Paleozoic cover section is carried out using test wells drilled up to 550 m. The essence of the method is to increase the absolute values and gradients of the biogeochemical signal along the section in the well. As you approach the accumulation of hydrocarbons, an increase in the BGC signal is observed. Advantage - the method is based on the activation of the vital processes of microorganisms on the path of hydrocarbon migration and causes biogeochemical anomalies throughout the section, up to the surface horizons.

The disadvantages of the method:

- organisms are equally dispersed both in degraded, residual oil, and in its industrial reserves, therefore, their presence in the formations is not a prerequisite for the industrial oil content of the formation;

- the absence of the areal contour of the oil content (the method gives a point forecast);

- It is an expensive method, as it implies the cost of drilling a well.

A known method of geochemical research using GORE-SORBER technology. GORE-SORBER geochemical surveys specialize in "passive" collection of hydrocarbon gases from the soil using an adsorbent embedded in the soil (US patent No. 5235863, publ. 08/17/1993). The GORE-SORBER research method is the statistical processing and modeling of chemical compounds (in the C1-C18 series) obtained from each sample. The method includes hierarchical cluster analysis and distinctive analysis. Geochemical maps are computer-generated surface probabilities interpolated from sample values.

The disadvantages of the method: the detected anomalies must be consistent with structural factors: uplift, tectonic disturbance, stratigraphic trap; distortion of the contour of oil deposits due to the influence of lateral processes of gas migration on upward migration.

Closest to the proposed invention in technical essence is a method for predicting oil and gas deposits by a complex of geophysical and geochemical methods (GHHM) (RF patent No. 2298817, publ. 05/10/2007 - prototype), including geophysical integration (ground based survey of natural electric potential (EP), ground magnetic field survey (MP) of methods and geochemical examination (surface gas-geochemical survey). The disadvantage of this method is the distortion of the forecast contour of oil content due to the lack of a fundamental structural factor of the selected seismic object in the GHM complex model.

The technical problem to which this invention is directed is to simplify the production of a complex of geophysical and geochemical studies and increase the reliability of the forecast of hydrocarbon deposits by developing a new complex probability parameter (CPV), which is used to construct a model of the oil contour.

The problem is solved as follows: register a geochemical anomaly according to the technology of surface gas-geochemical studies, for example, according to the GORE-SORBER technology (or any other); identify the basis for further production of geophysical studies of EP (natural electric field) and magnetic field (magnetic field) at points of geochemical testing; establish correlation dependences of the data of EP, MP, geochemical probability (GV) and absolute elevation (AO) of the studied reservoir, as a result of which a new complex parameter of probability (CPV) of geophysical and geochemical research methods, including EP, MP, GV and AO, is derived, build a CPV model that allows you to select the predicted oil profile.

The works include: the creation of a training sample of geochemical and geophysical data, the production of geochemical studies at the studied object, statistical processing and modeling of chemical compounds (in series C1-C13 to C18) obtained from each sample, the construction of a geochemical model that confirms the influence of a deep hydrocarbon object , survey of geophysical fields of EP and MP, analysis of statistical relationships between values of geochemical probability (GV), values of fields of EP and MP and a lutely marks (JSC) producing horizon, the creation of the "oil-empty" model. The presence of a well with residual oil content allows ranking data according to the condition “oil - residual oil - empty”, a multivariate conversion of HV, EP, MP and AO is found, which classified these variables and allowed to derive the KPV parameter and create a model of KPV oil content. For the model, the conditions are accepted: the training oil well has an oil recovery factor of -100%, a well with a residual oil content of 75-50%, an empty well has an oil recovery factor of 0%.

The CPV model of oil content displays the projection onto the surface of the underlying hydrocarbon accumulation. A value of 75% of the CPV score is called an abnormal threshold. The obtained values above 75% are considered to be the characteristics of a hydrocarbon deposit. Areas where the obtained KPV values tend to zero are considered insufficient to characterize the hydrocarbon reservoir. The value of the CPV from 75-50% is taken as a characteristic of the condition of residual oil.

The features of the invention are:

1. creation of a training sample of geochemical and geophysical data on the training wells “empty” - “oil” and, as additional information, on a well with residual oil content;

2. conducting a geochemical study using a gas-geochemical method, for example, using the GORE-SORBER method (or any other method), of a seismic object to confirm the influence of a deep hydrocarbon object;

3. determination of the effect of a hydrocarbon deposit is carried out by applying discriminant analysis for the training samples “oil - empty” or “oil - residual oil - empty”;

4. building a geochemical model with the determination of hydrocarbon anomalies, deriving the parameter — geochemical probability (GV) by direct signs of the state of organic hydrocarbon compounds in the region from C1-C13 to C18, confirming the effect of the deep hydrocarbon object on the surface;

5. in the absence of a geochemical anomaly, with a GW <50%, a conclusion is made about the futility of the studied geological object, further studies at this site are stopped, which will significantly reduce financial costs and get an economic effect;

6. determination of the oil-bearing profile, identified by geochemical research (at HS> 50%), is carried out by surveying using geophysical methods of a natural electric field (EP) and high-precision magnetic exploration (MP);

7. obtaining a sample of four parameters (GW, EP, MP, AO), coordinate-located at common points of the study;

8. modeling of the oil profile by discriminant analysis. An image of a well with known oil content is the standard for identifying sections of oil reservoirs with the same or similar features, and, on the contrary, a well with no oil content displays a “geophysical-geochemical image” of the background of the study area. A well with signs of residual oil content in the reservoir will be the standard for a “geophysical-geochemical image” for identifying areas of residual oil content in the reservoir;

9. building a map of the complex probability parameter (CPV), which determines the oil profile of the geological object. The contour of hydrocarbon deposits is allocated by the value of the CPV, equal to and exceeding the value of 75%. KPV values of 75-50% correspond to the conditions of residual oil. Values of KPV below 50% correspond to the conditions of the absence of signs of oil content.

Signs 3, 5, 7, 8, 9 are significant distinguishing features. In the proposed method, in contrast to the prototype, first of all, geochemical studies are carried out, and according to their results, a decision is made whether or not to conduct geophysical research at individual abnormal seismic objects. Thus, the problem of optimizing the production volume of a complex of geophysical surveys and improving the accuracy of determining the contour of the reservoir is solved.

The proposed method is based on the following provisions:

1. the migration of hydrocarbons from the reservoir is accompanied both by ascending paths to the day surface and laterally, extending to several hundred meters;

2. hydrocarbon migration leads to the formation of stable geochemical anomalies in the surface layers of the sedimentary cover, depending on the geological conditions, a shift of the geochemical anomaly from the actual contour of the deposit is possible;

3. "fuel cells" are formed above the hydrocarbon deposits, where the hydrocarbon reservoir and secondary epigenetic mineral formations (for example, sulfides) act as electrodes, creating conditions for the presence of two media with different values of the pH value ph;

4. studies of geochemical halos, magnetometry and changes in the redox potentials of the “fuel cell” of the reservoir, as well as seismic data, allows you to coordinate the contour of the oil reservoir.

The main methodological principles for interpreting the results of experimental research are as follows.

CPV recorded in surface sediments is an integral quantitative characteristic of hydrocarbon migration from the reservoir and secondary epigenetic changes that overlap the rock reservoir. The distribution of this parameter over the area above the hydrocarbon reservoir is subject to certain laws.

Based on the experience and the studies described, the functional dependence of the CPV is established, which characterizes the spatial distribution of the contour of the hydrocarbon pool.

The technical result is achieved by the fact that at the stage preceding the setting of deep drilling, the predicted oil content is calculated using the complex probability parameter (CPV).

Concrete example

The proposed method on the example of the Shadkin structure is as follows. The presented figures explain the essence of the invention.

A geochemical survey is carried out over a seismic local object to determine the hydrocarbon potential, the sampling step is 750 m - 250 m to 100 m, depending on the size of the seismic object. To model the geochemical probability (GV) of the oil potential, wells with a set of distinctive characteristics of the oil and empty type are used as standards. According to the state of organic hydrocarbon compounds in the region from C1-C13 to C18, the presence of a deep hydrocarbon potential from the reservoir was determined.

Figure 1 presents a fragment of the results of a geochemical study using GORE-SORBER technology on the Shadkinskaya structure in the form of an extensive geochemical anomaly. A GW of more than 50% was established, which confirms the presence of oil-saturated formations. Exploratory well No. 635 discovered the oil-saturated formation Do in the sediments of the Kynovsky horizon of the Upper Devonian, resulting in 11 m ³ / s of oil.

Drilled well No. 690, also located in an anomalous field of geochemical probability, did not confirm the forecast for the presence of industrial oil content, according to the core data, residual oil was found in it, oil flow was not obtained. Thus, the geochemical anomaly cannot characterize the exact contour of the oil potential, but only determines the presence or absence of hydrocarbons at depth. The contours in figure 1 reflect the absolute elevation of the roof of the Sargaev horizon. The triangle in figure 1 marks the design wells recommended by seismic data.

After drilling unproductive well No. 690 at the Shadkinskoye site, geophysical work is carried out using the EP and MP method. According to the results of processing field survey materials, maps of the distribution of the physical fields of EP and MP were obtained (Figs. 2 and 3).

The analysis and selection of statistical criteria are carried out - the values of EP, MP, the probability of the geochemical field of the GW and the absolute elevation (AO) at the points along the steps of sampling at a seismic object (Table 1).

For the training sample, the values of the nodes of the constructed parameter grids are selected in the area of well No. 635 - oil class, No. 690 - class “stop. oil ”, area of point 512907 - class“ empty ”. The grouping of the values of the criteria of the variable is carried out, where the rows of the training sample are assigned to one or another class, in this case 1 - “oil”, 2 - “stop. oil ”, 3 -“ empty ”(table 2).

They conclude that the established differences between the given samples are not random and reflect the actual differences between the populations (“oil” - “remaining oil” - “empty”).

Based on the data of the training sample, the discriminant functions are calculated and analyzed according to the following parameters:

Wilks - Lambda Wilks - parameter to indicate the power of discrimination. Its value varies from 1 (there is no discrimination) to 0 (complete discrimination). The smaller the variable, the greater the contribution it makes to discrimination. As can be seen from the table, all variables have values close to zero, i.e. in the training sample, classes 1 - “oil”, 2 - “stop. oil ”, 3 -“ empty ”well separated from each other. Partial Lambda - Wilks private lambda - this is the Wilks lambda parameter for the single contribution of the corresponding variable to the discrimination between the oil - stop classes oil is empty. ” The smaller its value, the greater the contribution the variable makes to discrimination. The table shows that the variables GV and MP make the largest contribution to discrimination.

F - Fisher's criterion - this criterion acts as a measure of the information content of a variable, the higher it is, the greater the information content. The table shows that the most informative are the variables GV and MP.

Toler - tolerance - a measure of the redundancy of a variable relative to the totality of other variables. The lower the tolerance, the higher the level of redundancy of the variable in this model. The level of tolerance = 0.01 - is the threshold for the inclusion of a variable in this model. As can be seen from the table, it is much more than 0.01, so all the variables in the model are not redundant.

p-level - significance level - this is an indicator that is in decreasing dependence on the reliability of the result. A higher p-level corresponds to a lower level of confidence in the dependence between the variables found in the sample. It is the p-level that represents the probability of an error related to the distribution of the observed result to the entire population. For example, p-level = 0.05 (i.e. 1/20), shows that there is a 5% chance that the relationship between the variables found in the sample is only a random feature of this sample. Results with a p-level <0.001 are considered statistically highly significant. In this case, the p-level is much less than 0.001, so all variables are highly significant.

Table 3 Results of the analysis of parameters of discriminant functions Wilks'lambda Partial lambda F p-level Toler GP 0.002215 0.21 193.63 0.0000001 0.84 MP 0.001793 0.26 146.89 0.0000001 0.71 EP 0.000765 0.61 33.17 0.0000001 0.37 AO 0.000755 0.62 06/32 0.0000001 0.44

Table 3 shows that the sampling criteria are nonrandom, and the data from the training sample discriminate well against the model for the classes “oil - stop. oil is empty. ”

All values in the training sample are divided into classes “oil - stop. oil is empty ”(table 4), i.e. for example, all 29 lines in the training set, a priori assigned to the class “oil”, in the model also belong to the class “oil”, etc.

Table 4 Sensitivity assessment of the decision rules of the training sample Classes The percentage of correct values,% "oil" "ost. oil" empty "oil" one hundred 29th 0 0 "ost. oil" one hundred 0 40 0 empty one hundred 0 0 40 Total one hundred 29th 40 40 Lines: observable classes Columns: Predicted Classes

In the model “oil” - “stop. oil "-" empty "the discriminant function is used: The root of the equation (X (Y) value) = С + k ₁ * p ₁ + k ₂ * p ₂ ... + k _n * p _n

Where k ₁ , k ₂ , ... k _n are the coefficients of the variables, p ₁ , p ₂ , ... p _n are the values of the variables,

X (Root of equation-1) = C ₁ + k _{1 GP} * p _{1 GP} + k _1MP * p _1MP + k _1EP * p _1EP + k _1AO * p _1AO

Y (Root of Equation-2) = C ₂ + k _2GP * p _2GP + k _2MP * p _2MP + k _2EP * p _2EP + k _2AO * p _2AO

for the variables of the model “oil” - “stop. oil "-" empty ", p ₁ , p ₂ , ... p _n - values of variables, C - eigenvalue of the function.

The coefficients of the variables EP, MP, GP, and AO were calculated for each function X and Y (Table 5).

Table 5 The values of the coefficients of the discriminant functions 1 and 2 TO Function 1 Function 2 GP -0.9661 -0.10961 MP -0.1168 1.05781 EP 0.9961 0.29252 AO -0.9186 -0.15988 Eigenvalue 174.7431 11.22740

For the training sample, the values of the roots of the discriminant functions 1 and 2 were obtained (Table 6), as well as the values of the roots of the functions for the centroids of the classes (Table 7), which determine the position of the classes.

Figure 4 is a distribution chart of the values of the training sample with centroids of classes (Table 6), all three sets are well separated from each other and have no intersections.

Based on the equations of the discriminant functions of the oil model - “stop. oil "-" empty "the values of the roots of the equations for each investigated point are obtained (table 8).

Figure 5 represents the distribution of the studied points relative to the sets "oil", "stop. oil "and" empty ". As can be seen from the figure, basically unclassified points by their location gravitate to one or another aggregate. This arrangement determines their belonging to one or another aggregate with varying degrees of probability in a given two-dimensional space.

Based on the obtained roots, the Mahalanobis distances were calculated. The minimum distance of the studied point to the centroid of the aggregate (the distance of the Mahalanobis) determines its belonging to the classes “oil”, “stop. oil "and" empty ". Points are classified according to the given parameters “oil - residual oil - empty” (Fig. 5, Table 9).

Table 7 Class Centroid Coordinates Root 1 Root 2 "oil" -7.1873 -5.17693 "ost. oil" -11.7480 3.15699 empty 16.9588 0.59629

By solving the piecewise linear function, the Z value of the complex parameter of oil potential prospects or, in another way, the oil potential probability (CPV) is obtained (Table 9).

where Z is the probability of CPV, X is the min Mahalanobis distance, B is the max Mahalanobis distance.

After the classification of points, a map of the complex probability parameter (LPC) of the oil prospects of the Shadkinsky block is constructed (Fig. 6).

In our example, wells No. 635 and No. 690 are in the same conditions according to geochemical survey (GW), but in different conditions according to the structural factor (AO) and according to the results of electromagnetic survey survey (EP and MP).

Analyzing the CPV map presented in Fig. 5, it can be said that the contour of the predicted oil content of the Shadkinskoye site significantly decreased relative to the geochemical anomaly of GORE SORBER.

Two zones are distinguished on the site, contoured by the CPV contour with a maximum probability of 75% and higher. The first is located in the center of the study area in the area of well No. 635 with an absolute mark of ≈ -1495 m, the second is located in the north-west of the site in the uplift circuit in the area of project well No. 695, limited to iso-gypsum at points with an absolute mark of ≈ -1495 m.

The uplift located in the northwest of the studied area (in the area of the project well 692) turned out to be in the unpromising area of the CPV model. According to the results of modeling the integrated oil parameter parameter map, project well 692 is located outside the oil content circuit, in the zone of residual oil content, the CPV value is 75-50%. In the area of well 692, the KPV value was 55%.

Drilled well No. 690 with residual oil content lies in the KPV model value range of 65%, is located hypsometrically below well No. 635, in adverse conditions according to geophysical data (EP and MP), in addition, there is a deflection along the line of points 512796-513240 that separates the northern elevation .

The design wells No. 693 and 694 recommended by seismic exploration according to the KPV model data also found themselves in the unpromising zone of residual oil content, located in the KPV range below 75%. The location of the project well 693 according to the CPV model in the probability values of 67%, well 694 in the probability values of 55%.

According to the considered example, it was found that as a result of using the method of prospecting and exploration of a hydrocarbon deposit according to the proposed method, the predicted oil content contour was significantly reduced.

The economic effect is achieved by reducing geophysical exploration and preventing the drilling of "empty" wells.

As a result of these studies, it was recommended to abandon the drilling of wells No. 692, 693 and 694.

Project well 695 has all the features of the CPV model (100% probability), similar to the features of productive well 635.

Using the proposed method will identify new deposits; to map small oil deposits by increasing the accuracy of the oil forecast, to reduce the consumption of material and financial resources by saving money both in the production of geophysical and gas-geochemical studies, and by reducing the drilling of "empty" wells.

This method was also tested on Tanay and Alimovsky uplifts. According to the results of studies by the method of the complex probability parameter, a reduction in the contour of the oil pool was established.

Claims (1)

A method for search and exploration of a hydrocarbon deposit, including the creation of a training sample of geochemical and geophysical data on training wells “oil - residual oil - empty”, conducting a geochemical study of a geological object, characterized in that the parameter of the geochemical probability of the GW is derived, determining the effect of the hydrocarbon deposit by applying the discriminant analysis of training samples "oil - residual oil - empty", in the absence of a geochemical anomaly at a geological object (HB <50%) recognize it they are unpromising and excluded from geophysical surveys, at sites with a HV> 50%, a sample of the coordinates of the natural electric field of the electromagnetic field, magnetic field of the magnetic field, the geochemical probability of the magnetic field, the absolute elevation of the studied reservoir layer AO are obtained, the complex probability parameter ( KPV) of oil content, build a map of KPV of oil content, which determines the contour of the oil content of a geological object along the contour of 75% probability.

Images