Intrusion state correction method and device for stratum resistivity logging information interpretation

Abstract

The invention relates to the technical field of open hole well logging in petroleum and natural gas exploration and development, in particular to an intrusion state correction method and device for stratum resistivity logging information interpretation, and the method comprises the steps: setting stratum parameters corresponding to different layer thickness intervals and different intrusion states, carrying out forward modeling, calculating a theoretical K value and an Rd4-Rs18 value according to the measured parameters, and calculating a K value and an Rd4-Rs18 value according to the theoretical K value and the Rd4-Rs18 value; the established neural network model is trained to obtain a group of interpretation models, and the group of interpretation models can be used for interpreting three stratum parameters of a variable K value of each layer, a specific value of the stratum to the electrical resistivity of an intrusive zone and the radius of the intrusive zone; the method is used for solving the problem that quantitative interpretation of the true resistivity value of the stratum cannot be carried out due to the fact that an intrusion state correction chart of the corresponding layer thickness cannot be developed and intrusion and state influence correction cannot be carried out on the stratum due to the fact that the obtained apparent resistivity value of an ultra-thin layer is too low and the change range is too small in an original fixed K value interpretation method.

Description

Technical Field

[0001] The present invention relates to the technical field of open-hole well logging in oil and gas exploration and development, and particularly to an invasion state correction method and device for formation resistivity logging data interpretation. Background Art

[0002] The invasion state of the formation is generally divided into three cases: non-invasion, low invasion, and high invasion. The apparent resistivity of the formation is affected by different invasion states. As the ratio of the formation resistivity to the resistivity of the invaded zone R t / R xo gradually increases and the radius ri of the invaded zone gradually deepens, the apparent resistivity value gradually decreases in the low-invasion formation, and the two show a good inverse relationship. However, in the high-invasion formation, the apparent resistivity gradually increases, and the two show a good direct relationship. Therefore, to ensure the accuracy of the invasion state influence correction, before performing the invasion correction, it is necessary to first divide the invasion state of the formation as non-invasion, low-invasion, or high-invasion.

[0003] The traditional invasion state division methods are qualitative and quantitative. Among them, the qualitative division method uses the amplitude difference R d -R s value for qualitative distinction. A positive amplitude difference indicates low invasion, a negative amplitude difference indicates high invasion, and coincidence indicates non-invasion. However, since the amplitude difference value obtained by this method is low and the change range is small, the accuracy of dividing thick layers is low, and it cannot be used for thin layers and ultra-thin layers. The quantitative division method is to separately develop a group of resistivity invasion influence correction charts for low invasion and high invasion with different layer thicknesses, and distinguish by seeing which chart the intersection point of the interpreted layer falls on. The three-lateral low-invasion correction charts with layer thicknesses of 5.0 m and 0.8 m and the three-lateral high-invasion correction charts with layer thicknesses of 6.0 m and 0.8 m are as Figures 1-4 shown. As can be seen from Figures 1-4 this, for thick layers, as the layer thickness thins, the interpretation error gradually increases. For thin layers and ultra-thin layers, because the measured apparent resistivity value is too low and the change range is too small, it is impossible to develop corresponding high-invasion and low-invasion correction charts for the layer thickness, and the true resistivity of the formation cannot be quantitatively interpreted. Summary of the Invention

[0004] The present invention provides an invasion state correction method and device for formation resistivity logging data interpretation to solve the problem that in the original method, due to the too low measured apparent resistivity and small change range, it is impossible to develop corresponding high-invasion and low-invasion correction charts for the layer thickness, and it is impossible to correct the layer thickness and invasion state influence of the formation, and thus it is impossible to quantitatively interpret the true resistivity R t of the formation.

[0005] According to one aspect of the present invention, there is provided an invasion state correction method for formation resistivity logging data interpretation, including:

[0006] Set formation parameters corresponding to different layer thickness intervals and different invasion states, and perform forward modeling to obtain measured parameter values related to the formation invasion state. Determine the deep theoretical K value and the difference R between the apparent resistivities of the deep and shallow laterologs after resistivity correction based on the set formation parameters and measured parameter values d 4 -R s 18 ;

[0007] Optimize the set different layer thickness intervals to obtain the optimal layer thickness interval after optimization;

[0008] Use the optimal layer thickness interval set by the forward modeling, the formation parameters corresponding to different invasion states, the obtained measured parameter values, the deep theoretical K value, and R d 4 -R s 18 , train the established neural network model to obtain an interpretation model corresponding to different invasion states for each optimal layer thickness interval;

[0009] Obtain the measured parameter values related to the formation invasion state, the formation thickness, and R determined based on the measured parameter values corresponding to the target formation d 4 -R s 18 ;

[0010] Determine the difference R between the apparent resistivities of the deep and shallow laterologs after layer thickness and resistivity correction for the non-invaded formation d 13 -R s 13 , and based on the R d 13 -R s 13 determine the invasion state of the target formation;

[0011] Find the optimal layer thickness interval closest to the formation thickness of the target formation and the corresponding interpretation model of the invasion state, and input the measured parameter values related to the formation invasion state and R d 4 -R s 18 corresponding to the target formation into this interpretation model to run, and obtain the deep variable K value, invasion zone radius r i , formation specific invasion zone resistivity R t / R xo value interpretation results.

[0012] Preferably, the method for setting formation parameters corresponding to different layer thickness intervals and different invasion states and performing forward modeling to obtain measurement parameter values related to the formation invasion state includes:

[0013] Setting a group of formations with true formation resistivity R t , resistivity R xo of the invaded zone, and radius r i of the invaded zone, inputting them into the forward model for forward modeling, and obtaining measurement parameter values related to the formation invasion state corresponding to each formation.

[0014] Preferably, the measurement parameter values related to the formation invasion state at least include: the ratio of the supply current of the first shielding electrode to the main electrode of the deep triple lateral, i.e., deep I 1 / I 0 , the ratio of the sum of the supply currents of the first and second shielding electrodes of the deep triple lateral to the supply current of the main electrode, i.e., deep I 1 +I 2 / I 0 , the potential U d of the main electrode of the deep triple lateral, the ratio of the supply current of the shielding electrode to the main electrode of the shallow triple lateral, i.e., shallow I 1 / I 0 , and the potential U s of the main electrode of the shallow triple lateral.

[0015] Preferably, the method for determining the deep theoretical K value according to the set formation parameters and measurement parameter values includes:

[0016] The calculation formula for the deep theoretical K value is: deep theoretical K value = R t / U d ;

[0017] Where: R t is the true formation resistivity set during forward modeling, and U d is the potential of the main electrode of the deep triple lateral measured through forward modeling;

[0018] Preferably, the method for optimizing the set different layer thickness intervals to obtain the optimized best layer thickness interval includes:

[0019] Judging whether the average relative error of the deep theoretical K values of all formations between two adjacent layer thickness intervals is less than a predetermined percentage. If not, insert several layer thickness intervals between the two adjacent layer thickness intervals until the average relative error of the deep theoretical K values of all formations between every two adjacent inserted layer thickness intervals is less than the predetermined percentage, thereby obtaining the best layer thickness interval.

[0020] Preferably, the method for determining the difference R between the apparent resistivities of the deep and shallow laterologs after resistivity correction according to the measured parameter values includes: d 4 -R s 18 wherein the method for determining R after resistivity correction includes:

[0021] Determining the relationship between the deep I d / I 4

[0022] in the forward simulation result and the deep theoretical K value; 1 / I 0 Substituting the deep I

[0023] measured from the target formation into the relationship between the deep I 1 / I 0 and the deep theoretical K value, calculating the corresponding value, and multiplying the corresponding value by the U 1 / I 0 measured from the target formation to obtain R d d 4 ;

[0024]

[0025] wherein the method for determining R after resistivity correction includes: s 18 Determining the shallow theoretical K value according to the set formation parameters and measured parameter values, and the method includes:

[0026]

[0027] The calculation formula for the shallow theoretical K value is: shallow theoretical K value = R t / U s ;

[0028] In the formula: R t is the true resistivity of the formation set during forward simulation, and U s is the main electrode potential of the shallow laterolog measured through forward simulation.

[0028] Determining the relationship between the shallow I 1 / I 0 in the forward simulation result and the shallow theoretical K value;

[0029] Substituting the shallow I 1 / I 0 measured from the target formation into the relationship between the shallow I 1 / I 0 and the shallow theoretical K value, calculating the corresponding value, and multiplying the corresponding value by the U s measured from the target formation to obtain R s18 .

[0030] Preferably, the method for determining the invasion state of the target formation by using the apparent resistivity difference R after correcting the deep and shallow triple laterologs of the non-invaded formation for layer thickness and resistivity includes: d 13 -R s 13 Based on the R d 13 -R s 13 Determine the invasion state of the target formation, including:

[0031] According to the forward simulation results, establish the relationship between the deep I 1 +I 2 / I 0 of the non-invaded formation and the deep theoretical K value, and the relationship between the shallow I 1 / I 0 of the non-invaded formation and the shallow theoretical K value;

[0032] Substitute the measured deep I 1 +I 2 / I 0 value of the target formation into the relationship between the deep I 1 +I 2 / I 0 of the non-invaded formation and the deep theoretical K value, calculate the corresponding value, and multiply the corresponding value by the U d 0 measured from the target formation to obtain R d 13 ;

[0033] Substitute the measured shallow I 1 / I 0 value of the target formation into the relationship between the shallow I 1 / I 0 of the non-invaded formation and the shallow theoretical K value, calculate the corresponding value, and multiply the corresponding value by the U s 0 measured from the target formation to obtain R s 13 ;

[0034] If the difference between R d 13 and R s 13 is negative, the invasion state of the target formation is high invasion. If the difference between R d 13 and R s 13 is positive, the invasion state of the target formation is low invasion or non-invasion.

[0035] Preferably, it further includes: measuring the measured parameters during forward modeling, and determining R based on the measured parameters d 4 -R s 18 and inputting the formation thickness, the best layer thickness interval closest to the formation thickness, and the corresponding invasion state into the interpretation model, and running to obtain the corresponding deep-variable K value, r i , R t / R xo value interpretation result;

[0036] Comparing it with the deep theoretical K value of the forward modeling and the set r i , R t / R xo value to determine the relative error between the two. The relative error between the deep-variable K value and the deep theoretical K value is the relative error of the final formation true resistivity interpretation.

[0037] According to one aspect of the present invention, there is provided an invasion state correction device for interpreting formation resistivity logging data, including:

[0038] A forward modeling unit, configured to set formation parameters corresponding to different layer thickness intervals and different invasion states, perform forward modeling, obtain measured parameter values related to the formation invasion state, and determine the deep theoretical K value and the difference R between the apparent resistivities of the deep and shallow triple laterologs after resistivity correction according to the set formation parameters and measured parameter values d 4 -R s 18 ;

[0039] A layer thickness interval optimization unit, configured to optimize the set different layer thickness intervals to obtain the optimized best layer thickness interval;

[0040] An interpretation model establishment unit, configured to use the best layer thickness interval set by the forward modeling, the formation parameters corresponding to different invasion states, the obtained measured parameter values, the deep theoretical K value, and R d 4 -R s 18 to train the established neural network model to obtain an interpretation model corresponding to different invasion states of each best layer thickness interval;

[0041] An acquisition unit, configured to acquire the measured parameter values related to the formation invasion state, the formation thickness, and R determined according to the measured parameter values corresponding to the target formation d 4 -R s 18 ;

[0042] An invasion state determination unit for determining the apparent resistivity difference R after correcting the apparent resistivity of the deep and shallow laterologs through the formation thickness and resistivity for the non-invaded formation depth d 13 -R s 13 and determining the invasion state of the target formation according to the R d 13 -R s 13 ;

[0043] An interpretation result output unit for finding the best formation thickness interval closest to the formation thickness of the target formation and the corresponding interpretation model for the invasion state, and inputting the measurement parameters related to the formation invasion state and R d 4 -R s 18 corresponding to the target formation into the interpretation model to run, and obtaining the deep change K value, invasion zone radius r i , formation resistivity relative to invasion zone resistivity R t / R xo value interpretation result corresponding to the target formation.

[0044] The present invention has at least the following beneficial effects:

[0045] The present invention proposes an invasion state correction method and device for formation resistivity logging data interpretation. By performing forward simulation, parameters related to the formation invasion state are obtained, and then the neural network models with different invasion states established are trained through the relevant parameters to obtain the interpretation model; the formation invasion state is accurately determined, and finally the interpretation result is obtained by running the interpretation model corresponding to the formation invasion state, thereby realizing the correction of the influence of the true resistivity invasion state of the formation, and further being able to effectively improve the interpretation accuracy of the true resistivity value of the formation and solve the problem of the inability to quantitatively interpret the true resistivity value of the formation. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The drawings here are incorporated into the specification and constitute a part of this specification. These drawings show embodiments in accordance with the present invention and are used together with the specification to explain the technical solutions of the present invention.

[0047] Figure 1 Showing the triple lateral resistivity logging correction chart for a formation thickness of 5 meters according to an embodiment of the present invention;

[0048] Figure 2 Showing the triple lateral resistivity logging correction chart for a formation thickness of 0.8 meters according to an embodiment of the present invention;

[0049] Figure 3 Showing the triple lateral resistivity logging correction chart for a formation thickness of 6 meters with high invasion according to an embodiment of the present invention;

[0050] Figure 4 Shows the triple lateral high invasion correction chart with a layer thickness of 0.8 m according to an embodiment of the present invention;

[0051] Figure 5 Shows the main current line distribution diagram of a homogeneous formation according to an embodiment of the present invention;

[0052] Figure 6 Shows the main current line distribution diagram with a layer thickness of 4.8 m according to an embodiment of the present invention;

[0053] Figure 7 Shows the main current line distribution diagram with a layer thickness of 0.2 m according to an embodiment of the present invention;

[0054] Figure 8 Shows the R of the non-invaded formation according to an embodiment of the present invention t / R m And the relationship diagram of the deep theoretical K value;

[0055] Figure 9 Shows the deep I of the non-invaded formation according to an embodiment of the present invention 1 / I 0 And R t / R m Relationship diagram;

[0056] Figure 10 Shows the deep I of the fourth invasion state according to an embodiment of the present invention 1 / I 0 And the relationship diagram of the deep theoretical K value;

[0057] Figure 11 Shows the shallow I of the eighteenth invasion state according to an embodiment of the present invention 1 / I 0 And the relationship diagram of the shallow theoretical K value;

[0058] Figure 12 Shows the relationship diagram of the deep variable K value and the formation thickness according to an embodiment of the present invention;

[0059] Figure 13 Shows the relationship diagram of the deep theoretical K value of the non-invaded formation and the deep I according to an embodiment of the present invention 1 +I 2 / I 0 Relationship diagram;

[0060] Figure 14 Shows the relationship diagram of the shallow theoretical K value of the non-invaded formation and the shallow I according to an embodiment of the present invention 1 / I 0 Relationship diagram;

[0061] Figure 15 Shows the flow chart of the invasion state correction method for interpreting formation resistivity logging data according to an embodiment of the present invention. Detailed implementation mode

[0062] Various exemplary embodiments, features, and aspects of the present invention will be described in detail below with reference to the accompanying drawings. Identical reference numerals in the drawings denote elements having the same or similar functions. Although various aspects of the embodiments are shown in the drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0063] As used herein, the term "exemplary" means "serving as an example, embodiment, or illustration." Any embodiment described herein as "exemplary" is not necessarily to be construed as superior to or better than other embodiments.

[0064] As used herein, the term "and / or" merely describes an association relationship between associated objects and indicates that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the term "at least one" as used herein means any one of a plurality or any combination of at least two of a plurality. For example, including at least one of A, B, and C may represent including any one or more elements selected from the set composed of A, B, and C.

[0065] In addition, for a better description of the present invention, numerous specific details are given in the following detailed implementation manners. Those skilled in the art should understand that the present invention can also be implemented without certain specific details. In some instances, methods, means, elements, and circuits well known to those skilled in the art are not described in detail so as to highlight the gist of the present invention.

[0066] Figure 1 Showing the triple lateral resistivity logging low invasion correction chart for a formation thickness of 5 meters according to an embodiment of the present invention; Figure 2 Showing the triple lateral resistivity logging low invasion correction chart for a formation thickness of 0.8 meters according to an embodiment of the present invention; Figure 3 Showing the triple lateral resistivity logging high invasion correction chart for a formation thickness of 6 meters according to an embodiment of the present invention; Figure 4 Showing the triple lateral resistivity logging high invasion correction chart for a formation thickness of 0.8 meters according to an embodiment of the present invention; Figure 5 Showing the main current line distribution chart of a homogeneous formation according to an embodiment of the present invention;

[0067] Figure 6 Showing the main current line distribution chart for a formation thickness of 4.8 meters according to an embodiment of the present invention; Figure 7 Showing the main current line distribution chart for a formation thickness of 0.2 meters according to an embodiment of the present invention; Figure 8 Showing the R of an invasion-free formation according to an embodiment of the present invention t / R m Relationship diagram with the deep theoretical K value; Figure 9 Showing the deep I of an invasion-free formation according to an embodiment of the present invention 1 / I 0 versus R t / Rm Relationship diagram; Figure 10 Deep I / I showing the fourth invasion state according to an embodiment of the present invention 1 / I 0 Relationship diagram with deep theoretical K value; Figure 11 Shallow I / I showing the eighteenth invasion state according to an embodiment of the present invention 1 / I 0 Relationship diagram with shallow theoretical K value; Figure 12 Relationship diagram showing the relationship between the deep variable K value and the formation thickness according to an embodiment of the present invention; Figure 13 Showing the relationship between the deep theoretical K value of the non-invaded formation and deep I+I / I according to an embodiment of the present invention 1 +I 2 / I 0 Relationship diagram;

[0068] Figure 14 Showing the relationship between the shallow theoretical K value of the non-invaded formation and shallow I / I according to an embodiment of the present invention 1 / I 0 Relationship diagram; Figure 15 Flowchart showing the invasion state correction method for formation resistivity log data interpretation according to an embodiment of the present invention. As shown, a method for correcting the invasion state of formation resistivity log data interpretation includes: Step S01: Set formation parameters corresponding to different layer thickness intervals and different invasion states, and perform forward simulation to obtain measured parameter values related to the formation invasion state. Determine the deep theoretical K value and the difference in apparent resistivity between the deep and shallow triple lateral resistivity R-R after resistivity correction according to the set formation parameters and measured parameter values Figures 1-15 ; Step S02: Optimize the set different layer thickness intervals to obtain the optimized best layer thickness interval; Step S03: Use the best layer thickness interval set by the forward simulation, the formation parameters corresponding to different invasion states, and the obtained measured parameter values, deep theoretical K value, and R-R to train the established neural network model to obtain an interpretation model corresponding to different invasion states for each best layer thickness interval; Step S04: Obtain the measured parameter values related to the formation invasion state, formation thickness, and R-R determined according to the measured parameter values corresponding to the target formation d 4 -R s 18 ; Step S05: Determine the difference in apparent resistivity R-R between the deep and shallow triple lateral resistivity after layer thickness and resistivity correction for the non-invaded formation d 4 -R s 18 ; Step S06: Use the obtained R-R and the interpretation model corresponding to the best layer thickness interval to determine the invasion state of the target formation d 4 -R s 18 ; Step S07: Use the obtained R-R and the interpretation model corresponding to the best layer thickness interval to determine the invasion state of the target formation d 13 -R s13 , according to the said R d 13 -R s 13 Determine the invasion state of the target formation; Step S06: Find the best layer thickness interval closest to the formation thickness of the target formation and the corresponding interpretation model of the invasion state, and use the measurement parameters related to the formation invasion state corresponding to the target formation and R d 4 -R s 18 , input and run it in this interpretation model to obtain the deep change K value, invasion zone radius r i 、formation resistivity ratio invasion zone resistivity R t / R xo value interpretation result.

[0069] In the embodiment of the present invention, the original resistivity logging data is interpreted by using the fixed K value method. The basic formula in resistivity logging and interpretation is:

[0070] R a =K*U d / I 0 (1);

[0071] In the formula: U d is the main electrode potential, with the unit of millivolt (mv), I 0 is the main electrode supply current, with the unit of milliampere (mA), K is the electrode system constant, and R a is the apparent resistivity value, with the unit of ohm-meter (Ω·m). U d / I 0 can be measured, and R a and K are two variables. Since the resistivity logging and interpretation technology has been available, the traditional method has been to advance from the right end to the left end of the above formula, that is, multiply the measured U d / I 0 value by the fixed K value of the electrode system constant obtained from the homogeneous formation to obtain the apparent resistivity R a value, regard R a as a variable, and finally correct and interpret the true resistivity R t value of the formation through various influencing factors.

[0072] This set of interpretation methods with fixed K values has solved the problem of quantitative interpretation of the true resistivity R t of relatively thick formations well in the past few decades. It provides an important basis for accurately obtaining the oil saturation value of the formation and finally conducting reservoir evaluation, and has achieved obvious geological effects. However, this set of interpretation methods with fixed K values has the following two problems:

[0073] Question 1. For the two triple lateral resistivity logs with low invasion and invasion correction charts with layer thicknesses of 5 m and 0.8 m, as attached Figure 1 and Figure 2 shown, in Figure 1 and 2 , d is the well diameter, di is the diameter of the invaded zone, dn is the outer diameter of the tool, R LL3 d is the deep triple lateral apparent resistivity, R LL3 s is the shallow triple lateral apparent resistivity, R s is the resistivity of the surrounding rock, h is the layer thickness, R m is the resistivity of the mud. From the comparison between Figure 1 and Figure 2 , it can be seen that as the layer thickness gradually decreases from 5 m to 0.8 m, the opening degree of the correction chart gradually becomes narrower, resulting in a gradual increase in the interpretation error of the true resistivity R t value of the formation. Especially when the layer thickness reaches 0.6 m, due to the narrow chart and low values, some errors in the original data may cause the intersection point to fall outside the chart, making it impossible to quantitatively interpret the R t value.

[0074] Question 2. For ultra-thin layers less than 0.6 m, taking a 0.2 m ultra-thin layer as an example, due to being severely affected by the layer thickness, the apparent resistivity R a value is not only too low but also has a small variation range. It is impossible to develop a resistivity and invasion influence correction chart for ultra-thin layers, impossible to correct the invasion influence of ultra-thin layers, and impossible to quantitatively interpret the true resistivity R t / R m (R m = 1) value.

[0075] Ultra-thin layers generally refer to formations with a layer thickness of 0.2 m - 0.5 m. To solve the problem of quantitative interpretation of the true resistivity R t of ultra-thin layers, it is necessary to accurately find the main problems existing in the original fixed K-value interpretation method.

[0076] (1). For the first time, the main current line distribution diagram visually reveals the contradictory pair of layering ability and detection depth existing in the lateral logging method and its fundamental cause.

[0077] In the homogeneous formation with the K value calibrated, the main current distribution is as attached Figure 5 shown. From the attachment Figure 5 it can be seen that the main current can not only enter the formation entirely in a flat plate shape, but also the divergence width is only 0.8 m at a depth of 10 m, fully conforming to the basic principle of triple lateral logging. This shows that the shielding effect of the shielding electrode and the main electrode being equipotential (U 1 = U 0 ) is the best, the detection depth is the deepest, and the measured potential of the main electrode is the largest at 688.1 mv.

[0078] In the formation where the ratio of true resistivity of the formation to the resistivity of the mud R t / R m = 30 (R m = 1), with an invasion radius r i = 0.1 m in a non-invaded formation, the main current distribution in a formation with a thickness of 4.8 m is as shown in the appendix Figure 6 shown. From the appendix Figure 6 it can be seen that the main current initially enters the formation in a flat plate shape, but starts to diverge at a depth of 4 m, starts to shunt to the surrounding rock at a depth of 7 m, and 1 / 4 of the main current still flows through at a depth of 10 m. This shows that the equipotentiality of the shielding electrode and the main electrode (U 1 = U 0 ) has a good shielding effect, a relatively deep detection depth, and the measured main electrode potential of 516.1 mv is relatively large.

[0079] The main current distribution in an ultra-thin formation with a thickness of 0.2 m is as shown in the appendix Figure 7 shown. From the appendix Figure 7 it can be seen that most of the main current shunts to the surrounding rock along the wellbore, only a small part enters the formation and immediately shunts to the surrounding rock, and no main current flows through at a depth of 0.3 m. This shows that the equipotentiality of the shielding electrode and the main electrode (U 1 = U 0 ) has the worst shielding effect, the shallowest detection depth, and the measured main electrode potential of 114.3 is the smallest.

[0080] Therefore, it can be concluded that for lateral logging, if every heterogeneous formation is treated equally and the equipotentiality of the shielding electrode and the main electrode (U 1 = U 0 ) is used for measurement, it is the fundamental reason that the contradictory pair of layering ability and detection depth cannot be unified, making it impossible to interpret the true resistivity of ultra-thin formations.

[0081] (2). Deeply reveals the fundamental problems existing in the fixed K-value interpretation method.

[0082] In a non-invaded formation with R t / R m = 30 (R m = 1), the effects of different interpretation methods are shown in Table 1:

[0083] Table 1: Data table of results of different interpretation methods

[0084]

[0085] As can be seen from Table 1, as the formation gradually changes from a homogeneous layer to an ultra-thin layer with a thickness of 0.2 m, the measured main electrode potential U d drops from 688.1 mv to 114.3 mv, a decrease of 6.02 times. Multiplying by the fixed K-value of 0.0436 measured in the homogeneous formation, the calculated apparent resistivity Ra The value rapidly drops from 30 to 4.98, which is also reduced by 6.02 times. Thus, all the drawbacks existing in the lateral logging method are transferred to the apparent resistivity R by multiplying by the fixed K value. a In the determination of the value, finally, the apparent resistivity R of the 0.2-meter ultra-thin layer a not only has too low a value but also has too small a variation range, making it impossible to develop resistivity correction charts for low invasion and high invasion with corresponding layer thicknesses and to correct the invasion effect, and unable to quantitatively interpret the true resistivity R of the formation. t of the formation.

[0086] Thus, it is profoundly revealed that in the fixed K value interpretation method of the original resistivity logging data, for each heterogeneous formation, a one-size-fits-all approach is adopted, and the fixed K value measured in the homogeneous formation is used for calculating the apparent resistivity R a value. This transfers the drawback that the contradictory relationship between the layer separation ability and the detection depth existing in the logging method cannot be unified directly to the apparent resistivity interpretation. This is also the fundamental reason why the invasion state of the ultra-thin layer cannot be divided at present, resulting in the inability to quantitatively interpret the true resistivity R t value.

[0087] The invasion state correction method for interpreting formation resistivity logging data provided by the embodiments of the present invention specifically includes the following steps:

[0088] Step S01: Set a group of formation parameters corresponding to different layer thickness intervals and different invasion states, and perform forward simulation to obtain measurement parameter values related to the formation invasion state. Determine the deep theoretical K value and the difference R d 4 -R s 18 between the deep and shallow triple lateral apparent resistivities after resistivity correction according to the set formation parameters and measurement parameter values.

[0089] In the present invention, the method of setting formation parameters corresponding to different layer thickness intervals and different invasion states and performing forward simulation to obtain measurement parameter values related to the formation invasion state includes: setting the true resistivity R t of the formation, the resistivity R xo of the invaded zone, and the radius r i of the invaded zone for a group of formations, and inputting them into the forward model for forward simulation to obtain the measurement parameter values related to the formation invasion state corresponding to each formation.

[0090] In the present invention, the measurement parameter values related to the formation invasion state at least include: the ratio of the supply current of the first shielding electrode of the deep triple lateral to the supply current of the main electrode, i.e., deep I 1 / I 0 、the ratio of the sum of the supply currents of the first and second shielding electrodes of the deep triple lateral to the supply current of the main electrode, i.e., deep I1 +I 2 / I 0 and the potential U of the main electrode of the deep laterolog d and the ratio of the supply current of the shield electrode to the main electrode of the shallow laterolog, i.e., shallow I 1 / I 0 and the potential U of the main electrode of the shallow laterolog s .

[0091] In the embodiment of the present invention, among the measured parameters obtained by forward simulation, the U d and U s are two measured parameters in the original fixed K-value interpretation method, while the deep I 1 / I 0 , deep I 1 +I 2 / I 0 , shallow I 1 / I 0 have never been measured or utilized.

[0092] In the present invention, the method for determining the deep theoretical K-value according to the set formation parameters and measured parameter values includes: the calculation formula for the deep theoretical K-value is as follows

[0093] Deep theoretical K-value = R t / U d ;

[0094] Where: R t is the true formation resistivity set during forward simulation, and U d is the potential of the main electrode of the deep laterolog measured through forward simulation.

[0095] In the embodiment of the present invention, the variable K-value (deep variable K-value) interpretation method is based on reverse thinking on the basis of the original fixed K-value interpretation method. In the calculation formula of the original formation apparent resistivity R a , the K-value is regarded as a variable. If the potential of the main electrode measured for a thick layer is large, the K-value is taken a little smaller; if the potential of the main electrode measured for a thin layer is small, the K-value is taken a little larger. The deep variable K-value is used to offset the influence of layer thickness on measurement. The expected effect of the deep variable K-value interpretation method for the electric logging data of a 0.2-meter ultra-thin layer is shown in Table 1. As can be seen from Table 1, when gradually changing from a homogeneous layer to a 0.2-meter ultra-thin layer, although the value of the potential U d of the deep lateral main electrode rapidly drops from 688.1 mv to 114.3 mv, a decrease of 6.02 times, the calculated deep theoretical K-value rapidly increases from 0.0436 to 0.2625, which is an inverse increase of 6.02 times. Thus, the drawbacks existing in the lateral logging method are overcome, and the contradictory pair of layer separation ability and detection depth reaches a perfect unity in the deep variable K-value interpretation method. Finally, whether it is a 4.8-meter thick layer or a 0.2-meter ultra-thin layer, R a / Rm =R t / R m =30, that is, the true resistivity value of the formation, achieving the same effect as the homogeneous formation. It can be seen that as long as we try to change the depth of each layer to K value K d 0 Calculate it, correct various influencing factors, make it close to the deep theoretical K value, multiply it by the measured deep three-side main electrode potential U d , that is, the true resistivity R of the formation is obtained t value, thus the true resistivity R t The quantitative interpretation of the value is transformed into the deep variable K value K d 0 The exact solution of .

[0096] In the embodiment of the present invention, the resistivity effect is one of the most important factors in the interpretation of electrical logging data. The ratio of the true resistivity of the 0.2-meter ultra-thin layer without invasion to the mud resistivity is R t / R m and deep theoretical K value K d 0 The relationship is as follows Figure 8 As shown in the attached Figure 8 It can be seen that when R t / R m From 2 to 40, the corresponding K d 0 It increased from 0.063 to 0.351, an increase of 5.57 times. There is a good positive proportional relationship between the two, which shows that the deep theoretical K value K d 0 The effect of resistivity is very large and must be corrected for.

[0097] Because it is impossible to develop a corresponding chart to correct the influence of the true resistivity of the formation in the ultra-thin layer of 0.2 meters, and the true resistivity of the formation R t The value is unknown before interpretation, so if you want to change the K value K d 0 In order to correct the resistivity effect, innovative ideas and methods must be explored.

[0098] 0.2m ultra-thin layer without stratum depth I 1 / I 0 With R t / R m The relationship is as follows Figure 9 As shown, in Figure 9 In the case of R t / R m From 2 to 40, its depth I 1 / I 0The value also increases from 15.13 to 40.55, increasing by 2.64 times, and there is also a good direct proportional relationship between the two. In traditional lateral logging, deep I 1 / I 0 This parameter has not been measured and studied for application, but it can be easily obtained by additional measurement. Therefore, it is first proposed to use the ratio of the supply current of the first shielding electrode of the deep triple lateral to the supply current of the main electrode, deep I 1 / I 0 to replace the true resistivity of the formation R t / R m For the resistivity influence correction method of the deep variable K value K d 0 to solve the problem that the resistivity influence correction of the deep variable K value cannot be carried out for the 0.2-meter ultra-thin layer.

[0099] In the present invention, the method for determining the difference in apparent resistivity of the deep and shallow triple laterals after resistivity correction according to the measured parameter values, R d 4 -R s 18 includes: Among them, the method for determining R d 4 after resistivity correction includes: determining the relationship between deep I 1 / I 0 in the forward simulation result and the deep theoretical K value; substituting the deep I 1 / I 0 measured from the target formation into the relationship between deep I 1 / I 0 and the deep theoretical K value, calculating the corresponding value, and multiplying the corresponding value by the U d measured from the target formation to obtain R d 4 ;

[0100] Among them, the method for determining R s 18 after resistivity correction includes: determining the shallow theoretical K value according to the set formation parameters and measured parameter values, and the method includes: the calculation formula for the shallow theoretical K value is: shallow theoretical K value = R t / U s ;

[0101] In the formula: R t is the true resistivity of the formation set during forward simulation, and U s is the potential of the main electrode of the shallow triple lateral measured through forward simulation.

[0102] Determine the shallow I 1 / I 0The relational expression between the shallow I and the shallow theoretical K value; substitute the shallow I 1 / I 0 measured from the target formation into the relational expression between the shallow I 1 / I 0 and the shallow theoretical K value, calculate the corresponding value, and multiply the corresponding value by the U s measured from the target formation to obtain R s 18 .

[0103] In the embodiment of the present invention, taking the set formation thickness as an ultra-thin layer of 0.2 m and the deep triple lateral with R t / R xo = 2, r i = 0.3 m invasion condition as the criterion, the measured deep I 1 / I 0 and the deep theoretical K value are obtained through forward simulation, where the deep I 1 / I 0 and the deep theoretical K value K d 0 relation data table is shown in Table 2 below.

[0104] Table 2: Deep I 1 / I 0 and deep theoretical K value K d 0 relation data table

[0105] <![CDATA[R t / R m > 40 30 20 10 5 3 <![CDATA[R xo / R m > 20 15 10 5 2.5 1.5 <![CDATA[r i (m)]]> 0.30 0.30 0.30 0.30 0.30 0.30 <![CDATA[Shen I 1 / I 0 > 34.49 31.60 27.47 21.09 16.35 14.03 <![CDATA[Deep theory K value K d 0 > 0.421 0.349 0.275 0.191 0.138 0.107

[0106] Draw the cross plot of the deep I 1 / I 0 and the deep theoretical K value K d 0 , and the result is as Figure 10 shown. It can be seen that there is an obvious proportional relationship between the two, so as to establish the relationship between the deep I 1 / I 0 and the deep theoretical K value K d 0 as shown in the following formula (2).

[0107] Deep theoretical K value K d 0 = 0.0465e 0.0643x (2);

[0108] where x is the deep I 1 / I 0 .

[0109] Substitute the deep I 1 / I 0 value actually measured from the target formation into the relational expression (2) to calculate the corresponding value Kd 0 , since this type of invasion state is uniformly defined as the fourth invasion state, it is called K d 4 , multiply K d 4 by the U measured from the target formation d value, and then R can be obtained d 4 .

[0110] In the embodiment of the present invention, taking the set formation thickness as a 0.2-meter ultra-thin layer and the shallow triple lateral R t / R xo = 0.33, r i = 0.175 meters invasion condition as the standard, through forward simulation, the measured shallow I 1 / I 0 and the calculated shallow theoretical K value K s 0 are obtained. Among them, the relationship data table of the shallow I 1 / I 0 and the shallow theoretical K value K s 0 is shown in Table 3 below.

[0111] Table 3: Relationship data table of the shallow I 1 / I 0 and the shallow theoretical K value K s 0

[0112]

[0113]

[0114] Plot the cross plot of the shallow I 1 / I 0 and the shallow theoretical K value K s 0 . The result is as shown in Figure 11 . Thus, the relationship between the shallow I 1 / I 0 and the shallow theoretical K value K s 0 is shown in the following formula (3).

[0115] The shallow theoretical K value K s 0 = 0.0219e 0.057x (3);

[0116] Where x is the shallow I 1 / I 0 .

[0117] The shallow I actually measured from the target formation​1 / I 0 The corresponding K value calculated by substituting the value into relational expression (3). Since this intrusion state is uniformly defined as the eighteenth intrusion state, it is called K s 0 value. Multiply K s 18 by the U value measured from the target formation to obtain R s 18 value, that is, R is obtained s value, that is, R is obtained s 18 .

[0118] Among the input parameters, the parameter R d 4 -R s 18 is processed and calculated respectively from the deep and shallow triple lateral logging data. The maximum value of the difference between R d 4 -R s 18 is 57.99, which is very large. While the maximum value of the difference between the original deep and shallow apparent resistivity R d -R s is only 0.76, which is very small. The absolute value of the maximum difference between the two is 57.23 higher, and the relative value has increased by 76.3 times. The value range of R d 4 -R s 18 is between 57.99 - 1.94 at most, and the absolute value of the difference has changed nearly 30 times, with a very large change range; while the change range of the original R d -R s is between 0.76 - 0.04. Although the relative change of the difference has also reached 19 times, the absolute value of the difference has only changed by 0.72, with a really small change range.

[0119] Thus, it can be seen that using the processed difference between R d 4 -R s 18 as the input parameter, the effect of correcting the deep variable K value for the layer thickness, resistivity, intrusion and state by the neural network model is surely qualitatively improved compared with the original use of the difference between R d -R s .

[0120] Step S02: Optimize the set different layer thickness intervals to obtain the optimized best layer thickness interval.

[0121] In the embodiments of the present invention, the layer thickness effect is one of the most important influencing factors in the interpretation of resistivity logging data. The present invention scientifically reveals the objective law of the layer thickness effect of thin layers and ultra-thin layers for the first time. In a non-invaded formation where the ratio of the true resistivity of the formation to the resistivity of the mud is equal to 30, the deep-variable K value K d 0 has the relationship with the formation thickness h as shown in the appendix Figure 12 as follows. It can be seen from the appendix Figure 12 that as the layer thickness gradually thins, its variable K value gradually increases, and there is a good inverse relationship between the two. When the layer thickness is less than 0.5 meters, there is a sudden increase in the layer thickness effect. For example, when comparing a layer thickness of 0.5 meters with 4.8 meters, the layer thickness difference is 4.3 meters and the deep-variable K value only increases by 36%. However, when comparing a layer thickness of 0.2 meters with 0.3 meters, the layer thickness difference is only 0.1 meter, while the deep-variable K value increases by an average of 53% and the maximum increase is 92.3%, almost doubling. Thus, it can be seen that the variable K value is greatly affected by the layer thickness and must be corrected.

[0122] Since the existing methods for interpreting electrical logging data cannot correct the layer thickness effect of ultra-thin layers, new methods must be explored. After repeated research, it is finally determined that a neural network algorithm interpretation model with a set of different layer thickness intervals is used to correct the layer thickness.

[0123] In the present invention, the method for optimizing the set of different layer thickness intervals to obtain the optimized best layer thickness interval includes: determining whether the average relative error of all theoretical K values of the formations between two adjacent layer thickness intervals is less than a predetermined percentage. If not, several layer thickness intervals are inserted between the two adjacent layer thickness intervals until the average relative error of all theoretical K values of the formations between each two adjacent layer thickness intervals after insertion is less than the predetermined percentage, thereby obtaining the best layer thickness interval.

[0124] In the embodiments of the present invention, the predetermined percentage is 10%. The preferred standard for the best layer thickness interval is that the average relative error of all theoretical K values of the formations between two adjacent layer thickness intervals must be less than 10%, so as to ensure that the error of the final layer thickness correction is controlled within 5%. If the average relative error of the theoretical K values is greater than 10%, several layer thickness intervals are inserted between the two adjacent layer thickness intervals until the average relative error of all theoretical K values of the formations between each two adjacent layer thickness intervals after insertion is less than 10%. If the sum of the average relative errors of the theoretical K values between two adjacent layer thickness intervals among three adjacent layer thickness intervals is still less than 10%, one of the layer thickness intervals is deleted. The layer thickness intervals are optimized according to the principle that as the layer thickness gradually thins from thick to thin, the layer thickness intervals gradually become denser.

[0125] The initial set layer thickness is 4.8 m - 0.2 m, and there are 10 different layer thickness intervals, which are: 4.8, 3.6, 2.4, 1.2, 0.8, 0.6, 0.5, 0.4, 0.3, 0.2. Calculate the K value of the depth change K between each two layer thickness intervals d 0 The average relative error of K and the initial selection of layer thickness intervals are shown in Table 11 below.

[0126] Table 11: Initial Selection Statistical Table of Layer Thickness Intervals from 4.8 m to 0.2 m

[0127]

[0128]

[0129] As can be seen from Table 11, the average relative error of the theoretical K value of the depth between the two layer thickness intervals of 4.8 m and 3.6 m is 6.7%, so there is no need for sparsification or densification in the middle.

[0130] Although the average relative error of the theoretical K value of the depth between the six layer thickness intervals of 3.6 and 2.4, 2.4 and 1.2, 1.2 and 0.8, 0.8 and 0.6, 0.6 and 0.5 is between 7.4% and 9.8%, although it does not exceed 10%, when the formation resistivity is relatively large and the invasion is relatively deep, most layers exceed 10%. Therefore, five layer thickness intervals of 3.0, 1.8, 1.0, 0.7, and 0.55 must be densified in the middle.

[0131] The average relative error of the theoretical K value of the depth between the two layer thickness intervals of 0.5 and 0.4 m is 14.1%, so a layer thickness interval of 0.45 must be densified in the middle.

[0132] The average relative error of the theoretical K value of the depth between the two layer thickness intervals of 0.4 m and 0.3 m is as high as 26.4%, so 2 - 3 layer thickness intervals must be densified in the middle.

[0133] The average relative error of the theoretical K value of the depth between the two layer thickness intervals of 0.3 m and 0.2 m is as high as 53.7%, so 3 - 4 layer thickness intervals must be densified in the middle.

[0134] The error between the two layer thickness intervals is too large, exceeding 10%. Multiple layer thickness intervals must be densified in the middle. That is, several layer thickness intervals are inserted between the layer thickness intervals with an average relative error greater than 10% so that the relative error between the corresponding depth change K values of each adjacent two layer thickness intervals after insertion is less than or equal to 10%. In addition, for the two layer thickness intervals with an error less than 10% and greater than 7%, a layer thickness interval with an intermediate value can be inserted, so as to make the calibration accuracy higher.

[0135] To deeply understand the preferred situation of layer thickness intervals between 0.6 m and 0.2 m, 14 layer thickness intervals were measured, and the average relative error of the theoretical K values of all strata depths between every two layer thickness intervals and the preferred situation of layer thickness intervals are shown in Table 7.

[0136] Table 7: Preferred statistical table of layer thickness intervals from 0.2 m to 0.6 m

[0137]

[0138]

[0139] As can be seen from Table 7, the average relative error of the deep-varying K values between 12 layer thickness intervals from 0.5 m to 0.2 m is less than 10%, so there is no need for densification or sparsification. Among the 3 layer thickness intervals from 0.5 m to 0.6 m, the sum of the average relative errors of the deep theoretical K values of adjacent two layer thickness intervals is still less than 10%. Therefore, the layer thickness interval of 0.55 m in the middle can be deleted. The finally determined optimal layer thickness intervals are 4.8 m, 3.6 m, 3.0 m, 2.4 m, 1.8 m, 1.2 m, 1.0 m, 0.8 m, 0.7 m, 0.6 m, 0.5 m, 0.45 m, 0.4 m, 0.375 m, 0.35 m, 0.325 m, 0.3 m, 0.28 m, 0.26 m, 0.24 m, 0.22 m, 0.2 m, a total of 22 layer thickness intervals.

[0140] Step S03: Using the optimal layer thickness intervals set by the forward simulation, the formation parameters corresponding to different intrusion states, and the obtained measurement parameter values, deep theoretical K values, and R d 4 -R s 18 , train the established neural network model to obtain an interpretation model corresponding to different intrusion states for each optimal layer thickness interval.

[0141] In the embodiment of the present invention, the optimal layer thickness intervals set during forward simulation, different R corresponding to different intrusion states t 、R xo 、r i values, and the measurement parameters corresponding to the forward simulation, that is, deep I 1 / I 0 、deep I 1 +I 2 / I 0 、U d 、shallow I 1 / I 0 、U s and the deep theoretical K value K d 0 and R d 4 -Rs 18 As input data, it is input into the corresponding neural network model, and the neural network model is trained to obtain each optimal layer thickness interval, that is, 22 layer thickness intervals correspond to a total of 44 interpretation models for two invasion states of high invasion and low invasion. Among them, the neural network model is modeled using the Bayesian regularization backpropagation neural network algorithm.

[0142] In the embodiment of the present invention, six measured input parameters are used to model and solve the formation deep variation K value K d 0 , R t / R xo , r i For these three parameters, the preferred criteria for the input parameters are not only closely related to the layer thickness, resistivity, invasion and state, but also the values should be large, and the maximum variation range should be more than 2 times to ensure the calibration accuracy.

[0143] The maximum values and maximum variation multiples of the 8 parameters measured for layer thicknesses of 0.2 m and 1.0 m are shown in Table 4, where the maximum and minimum values are respectively in R t / R m = 40 non-invaded formation and R t / R m = 3, r i = 0.3 m invaded formation.

[0144] Table 4: Table of maximum values and maximum variation multiples of measured parameters for different layer thicknesses

[0145]

[0146] As can be seen from Table 4: 1. The deep I 1 / I 0 value is closely related to the layer thickness, true resistivity and invasion state parameters R t / R xo , r i . For a layer thickness of 0.2 m, its maximum value is 25.8, and the maximum variation range is 2.4 times. For a layer thickness of 1.0 m, the maximum value is 11, and the maximum variation range is 1 time, which meets the preferred criteria for input parameters. It is used for calibration of the influence of layer thickness, resistivity and invasion state on the deep variation K value, and calculation of the R d 4 value.

[0147] 2. The deep I 1 +I 2 / I 0 value is closely related to the layer thickness, true resistivity and invasion state parameters R t / R xo , r iClosely related, its value gradually increases as the layer thickness thickens. The maximum value is 1098.3 when the layer thickness is 0.2m, and the maximum change range is 4.7 times. The maximum value is 4737.6 when the layer thickness is 1m, and the maximum change range is 15.4 times, meeting the preferred standard of input parameters. Use it to correct the influence of layer thickness, resistivity, and invasion state on the deep-variable K value and calculate R d 4 As the layer thickness gradually thickens from 0.2 meters, its effect is better than that of deep I 1 / I 0 getting better and better.

[0148] 3. The U d value is closely related to the layer thickness, true resistivity, and invasion state parameter R t / R xo and r i Closely related. The maximum value is 173.7 when the layer thickness is 0.2m, and the maximum change range is 4.7 times. The maximum value is 877.5 when the layer thickness is 1m, and the maximum change range is 17.1 times. The effect is much better than that of the apparent resistivity R a calculated by the original method, meeting the preferred standard of input parameters. Use it to correct the influence of layer thickness, resistivity, and invasion state on the deep-variable K value and calculate the theoretical K value K d 0 of each layer and K d 4 calculation.

[0149] 4. The shallow I 1 / I 0 value is related to the layer thickness, true resistivity, and invasion state parameter R t / R xo and r i has a good relationship. The maximum value is 34.2 when the layer thickness is 0.2m, and the maximum change is 2.3 times. The maximum value is 39.1 when the layer thickness is 1m, and the maximum change range is 2.6 times, meeting the preferred standard of input parameters. Use it to correct the influence of layer thickness, resistivity, and invasion state on the deep-variable K value and calculate the theoretical K value K s 0 of each layer and the R s 18 value calculation.

[0150] 5. The U s value is related to the layer thickness, true resistivity, and invasion state parameter R t / R xo and r i has a good relationship. The maximum value is 42.3 when the layer thickness is 0.2m, and the maximum change is 3.3 times. The maximum value is 67.4 when the layer thickness is 1m, and the maximum change range is 5.0 times. The effect is much better than that of the original apparent resistivity of the shallow triple lateral R s meeting the preferred standard of input parameters. Use it to correct the influence of layer thickness, resistivity, and invasion state on the deep-variable K value and calculate the shallow theoretical K value Ks 0 and R s 18 calculation

[0151] 6. R d 4 -R s 18 value is closely related to the layer thickness, true resistivity and invasion state parameter R t / R xo and r i The layer thickness is 0.2m, the maximum value is 57.99, and the maximum change is 30 times, meeting the preferred standard of input parameters. It is used to correct the influence of layer thickness, resistivity and invasion state on the deep-variable K value

[0152] Step S04: Obtain the measurement parameter values related to the formation invasion state corresponding to the target formation, the formation thickness, and R determined according to the measurement parameter values d 4 -R s 18 .

[0153] In the embodiments of the present invention, the formation thickness of the target formation needs to be interpreted by the method of layer-by-layer value extraction from well logging data. For formations above 0.6 meters, conventional well logging data is used for division, and for ultra-thin layers of 0.2 - 0.5 meters, resistivity imaging logging data is used for division. The detailed requirements for the formation thickness division accuracy are clearly proposed for the first time: when the layer thickness is between 4.8 meters and 1.2 meters, the accuracy should reach within 0.6 meters; when the layer thickness is between 1.2 meters and 0.8 meters, the accuracy should reach within 0.2 meters; when the layer thickness is between 0.8 meters and 0.5 meters, the accuracy should reach within 0.1 meters; when the layer thickness is between 0.5 meters and 0.4 meters, the accuracy should reach within 0.05 meters; when the layer thickness is between 0.4 meters and 0.3 meters, the accuracy should reach within 0.025 meters; when the layer thickness is between 0.3 meters and 0.2 meters, the accuracy should reach within 0.02 meters

[0154] The measurement parameters corresponding to the target formation are deep I 1 / I 0 、deep I 1 +I 2 / I 0 、U d 、shallow I 1 / I 0 、U s , and then determine the corresponding R according to the measurement parameters d 4 -R s 18 .

[0155] Step S05: Determine the apparent resistivity difference R after correcting the deep and shallow triple lateral resistivity by layer thickness and resistivity for the non-invaded formationd 13 -R s 13 , according to the said R d 13 -R s 13 determine the invasion status of the target formation.

[0156] In the present invention, the apparent resistivity difference R after correcting the deep and shallow triple laterologs of the non-invaded formation by layer thickness and resistivity is determined d 13 -R s 13 , according to the said R d 13 -R s 13 A method for determining the invasion status of the target formation, comprising: according to the forward simulation results, establishing a relationship between the deep I of the non-invaded formation 1 +I 2 / I 0 and the deep theoretical K value, and a relationship between the shallow I 1 / I 0 and the shallow theoretical K value; substituting the deep I measured in the target formation 1 +I 2 / I 0 value into the relationship between the deep I 1 +I 2 / I 0 and the deep theoretical K value, calculating the corresponding value, and multiplying the corresponding value by the U measured in the target formation d 0 , to obtain R d 13 ; substituting the shallow I measured in the target formation 1 / I 0 value into the relationship between the shallow I 1 / I 0 and the shallow theoretical K value, calculating the corresponding value, and multiplying the corresponding value by the U measured in the target formation s 0 , to obtain R s 13 ; if the subtracted value of R s 13 and R d 13 is negative, the invasion status of the target formation is high invasion; if the subtracted value of R s 13 and R d 13 is positive, the invasion status of the target formation is low invasion or non-invasion.

[0157] In the embodiments of the present invention, to ensure the accuracy of the invasion effect correction for the deep resistivity variation factor K d 0 before the invasion correction, it is necessary to first divide the invasion state of the target formation, that is, whether it is low invasion (no invasion) or high invasion.

[0158] R d 13 is the apparent resistivity value of the formation solved by the variable K value method after the deep triple lateral logging data in the non-invaded formation is corrected for the influence of layer thickness and resistivity.

[0159] Taking the forward simulation measurement data of 20 non-invaded formations of the deep triple lateral with the 9th electrode array as an example, as shown in Table 8.

[0160] Table 8: Deep triple lateral measurement data table of non-invaded formation with the 9th electrode array

[0161]

[0162]

[0163] As can be seen from Table 8, as R t / R m gradually decreases, the deep I 1 +I 2 / I 0 and the K d 0 value also gradually decreases, and the two show a good proportional relationship. According to the data in Table 8, the relationship between the deep theoretical K value K d 0 and the deep I 1 +I 2 / I 0 is as shown in Figure 13 From Figure 13 the polynomial relationship between the two can be obtained as shown in the following formula (4):

[0164] Deep theoretical K value K d 0 = 6.8455755E-13 (deep I 1 + 2 / I 0 ) 4 - 1.4408700E-09 (deep I 1 + 2 / I 0 ) 3

[0165] + 1.1645956E-06 (deep I 1 + 2 / I 0 ) 2 - 2.9322345E-04 (deep I1 + 2 / I 0 )

[0166] +5.9341769E-02 (4);

[0167] Its correlation coefficient R 2 = 0.9999. The K calculated by formula (4) d 0 was obtained from 20 non-invaded formations, so it is collectively referred to as K d 13 . The K calculated according to equation (4) d 13 and the deep theoretical K value K d 0 have an average relative error of only 0.42%, with high precision.

[0168] Substitute the actually measured deep I of the target formation 1 +I 2 / I 0 value into the relational expression (4) to obtain the corresponding K d 13 , multiply it by the U d value actually measured from this target formation, and finally the corresponding R d 13 value can be obtained.

[0169] R s 13 is the apparent resistivity value of the formation solved by the variable K value method after correcting the shallow triple lateral logging data in the non-invaded formation for the influence of layer thickness and resistivity.

[0170] Taking the forward simulation measurement data of 20 non-invaded formations of the shallow triple lateral with the 9th electrode system as an example, as shown in Table 9:

[0171] Table 9: Shallow triple lateral measurement data table of non-invaded formations with the 9th electrode system

[0172]

[0173]

[0174] As can be seen from Table 9, as R t / R m gradually decreases, the shallow I 1 / I 0 and the K s 0 value also gradually decreases, and the two show a good proportional relationship. Based on the data in Table 9, establish the shallow theoretical K value K s 0 and the shallow I 1 / I0 The relationship is as shown in Figure 14 and the polynomial relationship between the two can be obtained from Figure 14 as shown in the following formula (5):

[0175] The shallow theoretical K value K s 0 = 2.0531988E-06 (shallow I 1 / I 0 ) 5 - 2.4165718E-04 (shallow I 1 / I 0 ) 4

[0176] + 1.1367868E-02 (shallow I 1 / I 0 ) 3 - 2.6584193E-01 (shallow I 1 / I 0 ) 2

[0177] + 3.0994132 (shallow I 1 / I 0 ) - 14.321412 (5);

[0178] The correlation coefficient R 2 = 0.9999. The K s 0 calculated by formula (5) was obtained from 20 non-invaded formations, so it is collectively referred to as K s 13 . The K s 13 calculated according to formula (5) and the shallow theoretical K value K s 0 have an average relative error of only 0.35%, indicating high accuracy.

[0179] Substitute the shallow I 1 / I 0 value actually measured from the target formation into the relationship formula (5) to obtain the corresponding K s 13 . Multiply it by the U s value actually measured from the target formation, and finally the corresponding R s 13 value can be obtained.

[0180] On this basis, subtract R d 13 from R s 13, the difference value can be obtained. To better divide the formation invasion state, it is best that the difference value is close to zero in the non-invaded formation. Now, the difference values of 20 layers are between +0.08 and -0.09, and the average difference value is 0.033, meeting the requirements of the assumption.

[0181] To understand the effect of the R value calculated by combining the deep and shallow triple lateral logging data of the No. 9 electrode system on dividing the formation invasion state, 351 formations of 0.2-meter ultra-thin layers are measured, including 12 non-invaded formations, 259 low-invaded formations, and 80 high-invaded formations. Statistics are carried out according to the standard that a positive difference value indicates low invasion or non-invasion, and a negative difference value indicates high invasion. Among the 12 non-invaded layers, 11 layers have a positive difference value, and the coincidence rate is 91%. Among the 259 low-invaded layers, all the difference values are positive, and the coincidence rate is 100%. Among the 80 high-invaded layers, 79 layers have a negative difference value, and the coincidence rate is 98.8%. In total, there are 351 layers, and 249 layers are in coincidence, with a total coincidence rate of 99.4%. d 13 -R s 13 This innovative method has preferably solved the world problem that the invasion states of all formation thicknesses including 0.2-meter ultra-thin layers cannot be accurately divided, providing an important guarantee for accurately correcting the invasion and state influence of variable K values.

[0182] This innovative method preferably solves the world problem that the invasion states of all formation thicknesses including 0.2-meter ultra-thin layers cannot be accurately divided, providing an important guarantee for accurately correcting the invasion and state influence of variable K values.

[0183] Step S06: Find the best layer thickness interval closest to the formation thickness of the target formation and the corresponding interpretation model for the invasion state, and input the measurement parameters related to the formation invasion state and R d 4 -R s 18 corresponding to the target formation into the interpretation model to run, and obtain the deep variable K value, invasion zone radius r i , formation specific invasion zone resistivity R t / R xo value interpretation results corresponding to the target formation.

[0184] In the embodiment of the present invention, if the thickness of the divided target formation is 0.36 meters and the invasion state is a low invasion state, then select the corresponding low invasion state neural network interpretation model with the best layer thickness interval of 0.35 meters closest to the target formation thickness, and input the five actually measured parameters of the target formation and R d 4 -R s 18 value calculated according to the measurement parameters of the target formation into the interpretation model to run, and obtain the interpretation results, that is, the deep variable K value, r i , R t / R xoFor the first time, the present invention uses the Bayesian regularization backpropagation neural network algorithm for modeling, and replaces the calibration chart in the original fixed-K value interpretation method through a neural network model to achieve accurate calibration of the influence of the invasion state on resistivity logging interpretation.

[0185] In the present invention, it further includes: inputting the measurement parameters obtained during forward simulation and the determined R d 4 -R s 18 and the formation thickness into the interpretation model with the best layer thickness interval closest to the formation thickness and the corresponding invasion state, and running to obtain the corresponding deep-variable K value, r i 、R t / R xo value interpretation results; comparing them with the deep theoretical K value of the forward simulation and the set r i 、R t / R xo value to determine the relative error between the two. The relative error between the deep-variable K value and the deep theoretical K value is the relative error of the final formation true resistivity interpretation.

[0186] In an embodiment of the present invention, after obtaining the interpretation model through training, if it is necessary to determine the error of the interpretation result of the interpretation model, it is necessary to find the formation with the set thickness during forward simulation corresponding to the best layer thickness interval of the interpretation model, and use the measurement parameters obtained from the forward simulation corresponding to this formation, that is, the deep theoretical K value K d 0 、deep I 1 / I 0 、deep I 1 +I 2 / I 0 、U d 、shallow I 1 / I 0 、U s and the calculated R d 4 -R s 18 input into the corresponding interpretation model, and after running, obtain the interpretation results, that is, r i 、R t / R xo 、deep-variable K value K d 0 , compare the interpretation results with the set r i 、R t / R xo and the calculated deep theoretical K value during the corresponding forward model respectively to determine the relative error between the two, which is the relative error of the final interpretation result obtained by running the final interpretation model. Among them, K d 0The relative error of value interpretation is the true resistivity R of the formation t The error of quantitative interpretation.

[0187] In the embodiment of the present invention, the No. 9 electrode system is selected to measure 360 formations with a thickness of 0.2 m ultra-thin layer and a layer thickness of 0.8 m (including 276 low-invasion layers and 84 high-invasion layers), and the variable K value method is used for interpretation. The interpretation results are shown in Tables 5 and 6.

[0188] Table 5: Data table of interpretation results for different invasion states of 0.2 m ultra-thin layer

[0189]

[0190] Table 6: Data table of interpretation results for different invasion states of formations with a layer thickness of 0.8 m

[0191]

[0192] As can be seen from Tables 5 and 6, 1. The average relative error of variable K value K d 0 interpretation is 9.59% for the 0.2 m ultra-thin layer in the low-invasion formation, and the interpretation accuracy is improved by 70.1 percentage points compared with the original fixed K value interpretation method; it is 7.77% for the formation with a layer thickness of 0.8 m, and the interpretation accuracy is improved by 45.1 percentage points compared with the original fixed K value interpretation method. In the high-invasion formation, it is 4.72% for the 0.2 m ultra-thin layer, and the interpretation accuracy is improved by 35.7 percentage points compared with the original fixed K value interpretation method; it is 9.54% for the formation with a layer thickness of 0.8 m, and the interpretation accuracy is improved by 19.6 percentage points compared with the original fixed K value interpretation method. The accuracy is very high. And the number of positive and negative error layers and the average relative error are very close. The relative error of all formations is less than 10%, and the effect is very good, which indicates that the result of formation invasion state division is good, and it can make the true resistivity R of low-invasion and high-invasion formations t Quantitative interpretation can achieve satisfactory results.

[0193] 2. The variable K value interpretation method can also provide R t / R xo and r i Two important parameters of formation invasion conditions. The average relative error in the low-invasion formation is 2.90% and 16.8% respectively for the 0.2 m ultra-thin layer, and 2.14% and 13.9% respectively for the formation with a layer thickness of 0.8 m. In the high-invasion formation, it is 0.81% and 3.33% respectively for the 0.2 m ultra-thin layer, and 2.90% and 6.48% respectively for the formation with a layer thickness of 0.8 m. The accuracy is very high. And the number of positive and negative error layers and the average relative error are very close. The relative error of most layers is less than 10%, and the effect is very good. It can be used for quantitative interpretation completely, so as to scientifically analyze the invasion conditions of each layer and provide important basic data for accurate influence correction.

[0194] In the embodiment of the present invention, 250 0.2-meter ultra-thin layers of the No. 5 electrode system (including 150 low-invasion layers and 100 high-invasion layers) are interpreted using the corresponding invasion state interpretation model and the opposite invasion state interpretation model (low-invasion formations are interpreted using the high-invasion model, and high-invasion formations are interpreted using the low-invasion model). The final interpretation results are shown in Table 10:

[0195] Table 10: Data table for interpretation of corresponding and interchangeable network models

[0196]

[0197] In Table 10, the upper left and lower right corners respectively represent the results interpreted using the corresponding and opposite invasion state models. As can be seen from Table 10: 1. When interpreted using the corresponding invasion state model, the interpretation accuracy of the three formation parameters is very high and the effect is very good; 2. When interpreted using the opposite invasion state model, the average relative errors of the variable K value K d 0 for low-invasion and high-invasion formations are 49.1% and 360% respectively, and the average relative errors of R t / R xo are 155.8% and 3257% respectively, and the average relative errors of r i are 756.9% and 243% respectively. Since the errors are too large, they cannot be used. This fully demonstrates how important the accurate division of the formation invasion state is for accurate invasion correction of the formation and quantitative interpretation of the true resistivity R t value of the formation.

[0198] The technical solution of the present invention is not only applicable to lateral logging, but also applicable to the interpretation of all resistivity logging data. It is not only applicable to 0.2 - 0.5-meter ultra-thin layers, but also applicable to formations with other layer thicknesses. The thinner the layer, the greater the improvement in accuracy compared to the original interpretation method. It has the advantages of high interpretation accuracy, low input cost, and high output benefit.

[0199] It can be understood that the above-mentioned method embodiments mentioned in the present invention can be combined with each other to form combined embodiments without violating the principle logic. Due to space limitations, the present invention will not elaborate further.

[0200] The execution entity of the invasion state correction method for formation resistivity log data interpretation can be a device for correcting the invasion state of formation resistivity log data interpretation. For example, the invasion state correction method for formation resistivity log data interpretation can be executed by a terminal device, a server, or other processing devices. Among them, the terminal device can be a user equipment (UE), a mobile device, a user terminal, a terminal, a cellular phone, a cordless phone, a personal digital assistant (PDA), a handheld device, a computing device, a vehicle-mounted device, a wearable device, etc. In some possible implementation manners, the invasion state correction method for formation resistivity log data interpretation can be implemented by a processor calling computer-readable instructions stored in a memory.

[0201] Those skilled in the art can understand that in the above method of the specific implementation manner, the writing order of each step does not mean a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined according to its function and possible internal logic.

[0202] The present invention also provides a device for correcting the invasion state of formation resistivity log data interpretation, including: a forward modeling unit, configured to set formation parameters corresponding to different layer thickness intervals and different invasion states, and perform forward modeling to obtain measurement parameter values related to the formation invasion state, and determine a deep theoretical K value and a difference R between the apparent resistivities of the deep and shallow triple lateral logs after resistivity correction according to the set formation parameters and measurement parameter values d 4 -R s 18 ; a layer thickness interval optimization unit, configured to optimize the set different layer thickness intervals to obtain an optimized best layer thickness interval; an interpretation model establishment unit, configured to use the best layer thickness interval set by the forward modeling, formation parameters corresponding to different invasion states, and the obtained measurement parameter values, deep theoretical K value, and R d 4 -R s 18 , train the established neural network model to obtain an interpretation model corresponding to different invasion states for each best layer thickness interval; an acquisition unit, configured to acquire the measurement parameter values related to the formation invasion state, formation thickness, and R determined according to the measurement parameter values corresponding to the target formation d 4 -R s 18 ; an invasion state determination unit, configured to determine a difference R between the apparent resistivities of the deep and shallow triple lateral logs after layer thickness and resistivity correction for a non-invaded formation d 13 -R s 13 , according to the Rd 13 -R s 13 Determine the invasion state of the target formation; an interpretation result output unit, configured to find the best layer thickness interval closest to the formation thickness of the target formation and the corresponding interpretation model corresponding to the invasion state, and measure the parameters related to the formation invasion state and R corresponding to the target formation d 4 -R s 18 , input and run in this interpretation model to obtain the deep-variable K value, invasion zone radius r corresponding to the target formation i , formation resistivity relative to invasion zone R t / R xo value interpretation result.

[0203] In some embodiments, the functions or modules and units included in the device provided by the embodiments of the present invention can be used to execute the methods described in the above method embodiments. The specific implementation can refer to the description of the above method embodiments. For the sake of brevity, it will not be repeated here.

[0204] The present invention replaces the original fixed K-value interpretation method with a variable K-value interpretation of invasion state influence correction method, and solves the problem that due to the R a value is not only too low but also the variation range is too small, resulting in the inability to develop low-invasion and high-invasion resistivity invasion influence correction charts for corresponding layer thicknesses, and the inability to accurately correct invasion influence and R t value quantitative interpretation problem. At the same time, the present invention first adopts the method of building a model with the Bayesian regularization backpropagation neural network algorithm to replace the correction chart in the original fixed K-value interpretation method to correct the resistivity and invasion and state influence of the variable K-value. By establishing interpretation models corresponding to different layer thickness intervals and invasion states, using the set formation parameters and six parameters obtained through forward simulation and the deep theoretical K value to train the neural network model corresponding to the optimal layer thickness interval and invasion state to obtain the interpretation model; through R d 13 -R s 13 difference to divide the formation invasion state, realizing the correction of the invasion state influence on the variable K-value; by using the best layer thickness interval closest to the actual formation thickness and the interpretation model corresponding to the actual formation invasion state to interpret the actual target formation, solving the problem that the invasion state influence correction could not be carried out originally, improving the invasion and state influence correction accuracy, making the formation variable K-value interpretation method a complete success, and preferably solving the world problem that the true resistivity R t value cannot be quantitatively interpreted.

[0205] The embodiments of the present invention have been described above. The above description is exemplary and not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The choice of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to technologies in the market, or to enable other ordinary skilled persons in the art to understand the embodiments disclosed herein.

Claims

1. A method for intrusion state correction in formation resistivity logging data interpretation, It is characterized in that include: Set the formation parameters corresponding to different layer thickness intervals and different invasion states, and perform forward simulation to obtain the measured parameter values ​​related to the formation invasion state. According to the set formation parameters and measured parameter values, determine the deep theoretical K value and the deep and shallow three-way apparent resistivity difference R after resistivity correction. d 4 -R s 18 ; Optimizing the different set layer thickness intervals to obtain the optimal layer thickness interval after optimization; The optimal layer thickness interval set by the forward simulation and the formation parameters corresponding to different invasion states and the obtained measured parameter values, deep theoretical K value and R d 4 -R s 18 , train the established neural network model to obtain the explanation model corresponding to different invasion states of each optimal layer thickness interval; Obtain the measured parameter value related to the formation invasion state, the formation thickness and the R value determined according to the measured parameter value corresponding to the target formation d 4 -R s 18 ; Determine the apparent resistivity difference R after layer thickness and resistivity correction in the three lateral directions of the non-invasion stratum depth and shallowness d 13 -R s 13 , according to the R d 13 -R s 13 determining an invasion state of the target formation; Find the best layer thickness interval closest to the layer thickness of the target layer and the interpretation model corresponding to the corresponding invasion state, and compare the measured parameter value related to the formation invasion state corresponding to the target layer and R d 4 -R s 18 , input into the interpretation model and run to obtain the deep-variable K value and invasion zone radius r corresponding to the target stratum i , formation resistivity R t / R xo Value explains the result.

2. The invasion state correction method for formation resistivity logging data interpretation according to claim 1, It is characterized in that The method of setting formation parameters corresponding to different layer thickness intervals and different invasion states, and performing forward simulation to obtain measurement parameter values ​​related to the formation invasion state includes: Set the true resistivity R of the formation corresponding to different layer thickness intervals and different invasion states t , Intrusion zone resistivity R xo , Intrusion zone radius r i A group of strata with different values ​​are input into the forward model for forward simulation to obtain the measured parameter values ​​related to the invasion state of each stratum.

3. The invasion state correction method for formation resistivity logging data interpretation according to claim 2, It is characterized in that The measured parameter values ​​related to the formation invasion state include at least: the ratio of the power supply current of the first shielding electrode in the deep three lateral directions to the main electrode, i.e., the deep I 1 / I 0 , the ratio of the sum of the supply currents of the first and second shielding electrodes to the supply current of the main electrode, i.e., the depth I 1 +I 2 / I 0 , deep three lateral main electrode potential U d , the ratio of the supply current of the shallow three lateral shielding electrodes to the main electrode, that is, shallow I 1 / I 0 , shallow three lateral main electrode potential U s .

4. The invasion state correction method for formation resistivity logging data interpretation according to claim 3, Features: The method for determining the deep theoretical K value according to the set formation parameters and measurement parameter values ​​comprises: The calculation formula of the deep theoretical K value is: deep theoretical K value = R t / U d ; Where: R t is the true resistivity of the formation set in the forward modeling, U d is the deep three lateral main electrode potential measured by forward modeling.

5. The invasion state correction method for formation resistivity logging data interpretation according to claim 2, It is characterized in that The method of optimizing the different set layer thickness intervals to obtain the optimal layer thickness interval after optimization includes: Determine whether the average relative error of all theoretical K values ​​of the formation depth between two adjacent layer thickness intervals is less than a predetermined percentage. If not, insert several layer thickness intervals between the two adjacent layer thickness intervals until the average relative error of all theoretical K values ​​of the formation depth between each two adjacent layer thickness intervals after insertion is less than a predetermined percentage, thereby obtaining the optimal layer thickness interval.

6. The invasion state correction method for formation resistivity logging data interpretation according to claim 2, It is characterized in that Determine the apparent resistivity difference R of the three lateral directions in the depth after resistivity correction according to the measured parameter value d 4 -R s 18 method, include: Among them, determine the R after resistivity correction d 4 methods, including: Determine the depth I in the forward simulation results 1 / I 0 The relationship between the K value of deep theory; The depth I measured by the target formation 1 / I 0 Substitute the depth I 1 / I 0 The corresponding value is calculated from the relationship between the theoretical K value and the deep K value, and the corresponding value is multiplied by the U measured in the target formation. d , and get R d 4 ; Among them, determine the R after resistivity correction s 18 methods, including: The method for determining the shallow theoretical K value according to the set formation parameters and the measured parameter values ​​comprises: The calculation formula of the shallow theoretical K value is: shallow theoretical K value = R t / U s ; Where: R t is the true resistivity of the formation set in the forward modeling, U s is the shallow three-lateral main electrode potential measured by forward modeling; Determine the shallow I in the forward simulation results 1 / I 0 The relationship between the shallow theoretical K value and the The shallow I 1 / I 0 Substitute the shallow I 1 / I 0 The corresponding value is calculated from the relationship between the theoretical K value and the target formation measured U s , and get R s 18 .

7. The invasion state correction method for formation resistivity logging data interpretation according to claim 6, It is characterized in that The determination of the apparent resistivity difference R after the layer thickness and resistivity correction in the three lateral directions of the non-invasion formation depth and shallowness d 13 -R s 13 , according to the R d 13 -R s 13 The method for determining the invasion state of the target formation comprises: According to the forward modeling results, the non-invasion formation depth I is established. 1 +I 2 / I 0 The relationship between the deep theoretical K value and the shallow I 1 / I 0 The relationship with the shallow theoretical K value; The depth I of the target formation is measured 1 +I 2 / I 0 Substitute the value into the depth I 1 +I 2 / I 0 The corresponding value is calculated in the relationship between the theoretical K value and the deep formation, and the corresponding value is multiplied by the U obtained by measuring the target formation. d 0 , and get R d 13 ; The shallow I 1 / I 0 Substitute the value into the shallow I 1 / I 0 The corresponding value is calculated in the relationship between the shallow theoretical K value and the target formation measured by U s 0 , and get R s 13 ; If R d 13 With R s 13 If the subtraction value is negative, the invasion state of the target formation is high invasion. d 13 With R s 13 If the subtraction value is positive, the invasion state of the target formation is low invasion or no invasion.

8. The invasion state correction method for formation resistivity logging data interpretation according to any one of claims 2 to 7, It is characterized in that Also includes: The measured parameters obtained during the forward simulation and the R determined based on the measured parameters d 4 -R s 18 The formation thickness is input into the best layer thickness interval closest to the formation thickness and the interpretation model of the corresponding invasion state, and the corresponding deep variable K value, r i , R t / R xo Values ​​explain the results; It is compared with the deep theoretical K value of the forward simulation and the set r i , R t / R xo The values ​​are compared to determine the relative error between the two. The relative error between the deep variable K value and the deep theoretical K value is the relative error of the final interpretation of the true resistivity of the formation.

9. An invasion state correction device for formation resistivity logging data interpretation, It is characterized in that include: The forward simulation unit is used to set the formation parameters corresponding to different layer thickness intervals and different invasion states, and perform forward simulation to obtain the measured parameter values ​​related to the formation invasion state, and determine the deep theoretical K value and the deep and shallow three-way apparent resistivity difference R after resistivity correction according to the set formation parameters and measured parameter values. d 4 -R s 18 ; The layer thickness interval optimization unit is used to optimize the different set layer thickness intervals to obtain the optimal layer thickness interval after optimization; The interpretation model building unit is used to use the optimal layer thickness interval set by the forward simulation and the formation parameters corresponding to different invasion states and the obtained measured parameter values, deep theoretical K value and R d 4 -R s 18 , train the established neural network model to obtain the explanation model corresponding to different invasion states of each optimal layer thickness interval; an acquisition unit, for acquiring the measured parameter value related to the formation invasion state, the formation thickness, and the R value determined according to the measured parameter value corresponding to the target formation; d 4 -R s 18 ; The invasion state determination unit is used to determine the apparent resistivity difference R after the layer thickness and resistivity correction of the deep and shallow three lateral directions of the non-invaded formation. d 13 -R s 13 , according to the R d 13 -R s 13 determining an invasion state of the target formation; The interpretation result output unit is used to find the best layer thickness interval closest to the layer thickness of the target layer and the interpretation model corresponding to the corresponding invasion state, and to output the measurement parameters related to the formation invasion state and R corresponding to the target layer. d 4 -R s 18 , input into the interpretation model and run to obtain the deep-variable K value and invasion zone radius r corresponding to the target stratum i , formation resistivity R t / R xo Value explains the result.

Images