A method for regional description of drilling engineering risks

Abstract

The present invention discloses a method for regional description of drilling engineering risks, including: step one, obtaining the formation pressure with uncertainty of the drilled well; step two, using the formation pressure to construct a drilling risk quantitative assessment model; step three, using geostatistical methods to perform a two-dimensional regional description of the risk assessment results of the drilling risk quantitative assessment model. It can be seen from the above technical scheme that this scheme has the following beneficial effects: 1. Using geostatistical methods, the risk assessment results of the drilled wells in the block are described in a two-dimensional regional manner within the stratigraphic layer group, so as to achieve the horizontal distribution of the vertical risk assessment results according to the stratigraphic layer group; 2. The method of the present invention realizes the regional description of drilling risks, so that when drilling a new well in the block, the risk situations that may occur in each layer group at the new well location point can be judged, which improves the comprehensiveness of the considerations of designers when designing drilling projects to a certain extent, and maximizes the drilling safety in the design stage.

Description

Technical Field

[0001] The invention relates to the technical field of deep well complex stratum drilling, and in particular to a method for regionalizing the description of drilling engineering risks. Background Art

[0002] As oil and gas exploration and development gradually move towards deep wells, complex formations and deep ocean waters, the geological environment encountered in drilling is becoming more and more complex, such as strong fold and fault zones, alternating soft and hard formations, large formation dips, and strong formation anisotropy. Due to the complexity of structural conditions, it is difficult to understand the spatial distribution of formation pressure in the block, resulting in frequent underground complex conditions and accidents such as blowouts, leaks, collapses, stuck, and broken drill tools during drilling.

[0003] Most of the existing risk assessment methods are post-drilling risk analysis and summary of drilled wells.

[0004] When planning to drill a new well in a block, we are still unable to judge the possible risks of each layer group at the location of the new well, which to some extent affects the comprehensiveness of the designers' considerations when designing the drilling project. Summary of the invention

[0005] In view of this, the present invention provides a method for regionalizing the description of drilling engineering risks, which can make the drilling engineering design plan more reasonable and the plan implementation process safer.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for regionalizing description of drilling engineering risks, comprising:

[0008] Step 1: Obtain the formation pressure of the drilled well with uncertainty;

[0009] Step 2: Use formation pressure to construct a drilling risk quantitative assessment model;

[0010] Step 3: Use geostatistical methods to perform a two-dimensional regional description of the risk assessment results of the drilling risk quantitative assessment model.

[0011] Preferably, in step 1, obtaining the formation pressure with uncertainty in the drilled well includes:

[0012] The logging data of key wells drilled in the area are collected, and the formation porosity, formation collapse and formation fracture pressure with uncertainty are calculated using the Monte Carlo simulation method.

[0013] Preferably, in step 2, a drilling risk quantitative assessment model is constructed using formation pressure, including:

[0014] Combined with the obtained formation pores, formation collapse and formation fracture pressure interval profiles with uncertainties in the drilled wells, a safe drilling fluid density window with uncertainties is established;

[0015] The drilling risk probability calculation model is obtained based on reliability theory and probability statistical analysis theory.

[0016] Preferably, the step of establishing a safe drilling fluid density window with uncertainty by combining the obtained formation pores, formation collapse and formation fracture pressure interval profiles with uncertainty in the wells includes:

[0017] (1) Probability distribution of lower limit of drilling fluid density for preventing well kick

[0018]

[0019] (2) Probability distribution of the lower limit of drilling fluid density for preventing wellbore collapse

[0020]

[0021] (3) Probability distribution of the upper limit of drilling fluid density to prevent differential pressure sticking

[0022]

[0023] (4) Probability distribution of the upper limit of the density of the anti-lost drilling fluid

[0024]

[0025] In the formula, S b is the suction pressure coefficient, g / cm 3 ; S g is the excitation pressure coefficient, g / cm 3 ; Δρ is the additional drilling fluid density, g / cm 3 ; S f The upper limit safety factor for preventing leakage, g / cm 3 ; S c is the cycle pressure loss coefficient, g / cm 3 ; ΔP is the allowable value of differential pressure for drilling, MPa; P p is the pore pressure equivalent drilling fluid density, g / cm 3 ;P c is the collapse pressure equivalent drilling fluid density, g / cm 3 ;P f is the drilling fluid density equivalent to the fracture pressure, g / cm 3 ;

[0026] According to the distribution of the upper and lower limits of the drilling fluid density, the safe drilling fluid density window with a cumulative probability of j0 is:

[0027]

[0028] It can be seen that the depth h i The upper and lower limits of safe drilling fluid density when the cumulative probability j = j0 are as follows:

[0029]

[0030]

[0031] When the cumulative probability is j0, the upper and lower limit curve expressions of safe drilling fluid density are shown in equations (8) and (9):

[0032]

[0033]

[0034] In the formula, are the lower and upper limits of safe drilling fluid density with cumulative probability j0; similarly, the lower limit of safe mud density with cumulative probability j1 is obtained and upper limit In this way, the |j1-j0|×100% safe mud density interval profile is obtained;

[0035] The drilling risk probability calculation model obtained according to the reliability theory and probability statistical analysis theory includes:

[0036] Kick risk probability calculation model:

[0037] R k,h =P(ρ d,h <ρ k,h )=1-F k,h (ρ d,h ); (17)

[0038] Well leakage risk probability calculation model:

[0039] R l,h =P(ρ d,h >ρ l,h )=F l,h (ρ d,h ); (18)

[0040] Well wall collapse risk probability calculation model:

[0041] R c,h =P(ρ d,h <ρ c,h )=1-F c,h (ρd,h ); (19)

[0042] Calculation model of the risk probability of differential pressure sticking:

[0043] R s,k =P(ρ d,h >ρ sk,h )=F sk,h (ρ d,h ). (20)

[0044] Preferably, in step 3, a geostatistical method is used to perform a two-dimensional regional description of the risk assessment results of the drilling risk quantitative assessment model, including:

[0045] The support vector machine algorithm was used to fit the experimental variogram to obtain the GA-SVM-Kriging interpolation algorithm;

[0046] The risk assessment results of the drilling risk quantitative assessment model are described in two dimensions based on the GA-SVM-Kriging interpolation algorithm.

[0047] Preferably, before the support vector machine algorithm is used to fit the experimental variation function to obtain the GA-SVM-Kriging interpolation algorithm, the method further includes:

[0048] A genetic algorithm is used to optimize and determine three important parameters in the support vector machine algorithm; the three important parameters include: insensitive loss parameter ε, penalty parameter C and kernel parameter σ of Gaussian RBF kernel function.

[0049] Preferably, the genetic algorithm is used to optimize and determine three important parameters in the support vector machine algorithm, including:

[0050] ① Initialize the population: Use binary coding to randomly generate M initial populations, and preset the number of iterations to N;

[0051] ② Calculate the fitness value: train the SVM with the parameters corresponding to each individual in the group and calculate the fitness value of each individual;

[0052] ③Choice: Determine the initial population G according to the roulette method (0) The number of times an individual is selected and the probability of individual selection are considered, and the excellent individuals with high probability are selected, and then their information is inherited to the next generation of new population;

[0053] ④Crossover: Let the crossover probability P c , randomly pair the parent population to generate new individuals;

[0054] ⑤ Mutation: For a given mutation probability P m , first randomly select M×P mindividuals, let the kth character undergo mutation operation, replace the information at the corresponding position of the selected chromosome, and generate a new individual;

[0055] ⑥ Termination condition judgment: if the termination condition is met, stop calculating and outputting the results; otherwise, go to step ② to continue calculating;

[0056] ⑦Decoding: Decode the optimal individual and finally obtain the optimal solution of the three important parameters in the SVM model.

[0057] Preferably, the risk assessment results of the drilling risk quantitative assessment model are described in two dimensions based on the GA-SVM-Kriging interpolation algorithm, including:

[0058] (1) Calculate the experimental coefficient of variation γ * (h);

[0059]

[0060] (2) The GA-SVM-Kriging algorithm is used to calculate the experimental coefficient of variation γ * (h) Perform regression fitting to obtain the theoretical coefficient of variation γ(h);

[0061] (3) For the position point x to be estimated o , find the corresponding weight vector [λ1,λ2,...λ n ], and then substitute it into the Kriging interpolation formula to get o The estimated value p at the location * (x0);

[0062] Assume that there are N wells drilled in this area, and the coordinates of the wells are (X i ,Y i ), i = 1, 2, ..., N; in a layer group j, the i-th well has a total of n discrete points of drilling risk probability data, which can be expressed as follows:

[0063]

[0064] If the target well location is (X o ,Y o ), based on the GA-SVM-Kriging interpolation algorithm, in formation j, the risk probability data of the i-th well that has been drilled is transplanted to the target well as follows:

[0065]

[0066] Where: i is the weight coefficient of the i-th well;

[0067] Transplant all N drilled well risk parameters to the target well according to the above method, and finally obtain the drilling risk probability of the target well in layer group j:

[0068]

[0069] Preferably, in step 3, after the risk assessment results of the drilling risk quantitative assessment model are described in two dimensions using a geostatistical method, the method further includes:

[0070] Step 4: Construct and visualize the regionalization model of drilling engineering risks.

[0071] Preferably, in step 4, constructing and visualizing a regionalized model of drilling engineering risks includes:

[0072] The drilling risk probability values ​​of multiple wells in the target area are calculated according to the GA-SVM-Kriging interpolation algorithm. Then, based on the calculation results, the lateral distribution cloud map of drilling engineering risks in the target stratum group is obtained by spatial interpolation, and the model is visualized using programming language.

[0073] It can be seen from the above technical solutions that the method for regionalizing the description of drilling engineering risks provided by the present invention has the following beneficial effects:

[0074] 1. The present invention uses geostatistical methods to perform two-dimensional regional description of the risk assessment results of the wells drilled in the block within the stratigraphic group, and realizes the horizontal distribution of the vertical risk assessment results according to the stratigraphic group, so as to make the drilling engineering design plan more reasonable and the implementation process of the plan safer;

[0075] 2. The method proposed in the present invention realizes the regional description of drilling risks. When drilling a new well in a block, it can judge the possible risks of each layer group at the location of the new well. To a certain extent, it improves the comprehensiveness of the considerations of designers when designing drilling projects, and maximizes the drilling safety in the design stage. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0077] Figure 1 A flowchart for describing the regionalization of drilling risks provided by an embodiment of the present invention;

[0078] Figure 2A flow chart for obtaining formation pressure with uncertainty provided by an embodiment of the present invention;

[0079] Figure 3 A schematic diagram of a lower limit value for preventing kick-outs and a circulating equivalent drilling fluid density distribution interference model provided in an embodiment of the present invention;

[0080] Figure 4a A risk assessment result diagram of a well X1 drilled in Block XX provided in an embodiment of the present invention;

[0081] Figure 4b A risk assessment result diagram of the drilling project in Block X2 of the XX area provided in an embodiment of the present invention;

[0082] Figure 4c A risk assessment result diagram of a well X3 in the XX block provided in an embodiment of the present invention;

[0083] Figure 4d A risk assessment result diagram of the drilling project X4 in the XX block provided in an embodiment of the present invention;

[0084] Figure 4e A risk assessment result diagram of the drilling project X5 in the XX block provided in an embodiment of the present invention;

[0085] Figure 4f A risk assessment result diagram of the drilling project X6 in Block XX provided in an embodiment of the present invention;

[0086] Figure 4g A risk assessment result diagram of the drilling project X7 in the XX block provided in an embodiment of the present invention;

[0087] Figure 4h A risk assessment result diagram of the drilling project X8 in the XX block provided in an embodiment of the present invention;

[0088] Figure 5a A cloud map of drilling risk areas (well kick) of the LLH group in the XX block provided in an embodiment of the present invention;

[0089] Figure 5b A cloud map of drilling risk areas (lost circulation) of the LLH group in the XX block provided in an embodiment of the present invention;

[0090] Figure 5c A cloud map of drilling risk areas (well wall collapse) for the LLH group in the XX block provided in an embodiment of the present invention;

[0091] Figure 5d A cloud map of drilling risk areas of the LLH group in the XX block (pressure differential stuck drilling) provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0092] In response to the technical problems in the background technology, the present invention discloses a regionalized description method for drilling engineering risks, which fully utilizes the good continuity and high resolution of logging data, and aims to establish a regionalized distribution cloud map of drilling risks in areas with complex geological structures, so as to more clearly understand the probability distribution and change laws of drilling risks in the block, and provide the necessary technical basis for reducing the occurrence of complex underground accidents in the block.

[0093] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0094] The method for regionalizing the risk of drilling engineering provided by the embodiment of the present invention is as follows: Figure 1 As shown, including:

[0095] Step 1: Obtain the formation pressure of the drilled well with uncertainty;

[0096] Step 2: Use formation pressure to construct a drilling risk quantitative assessment model;

[0097] Step 3: Use geostatistical methods to perform a two-dimensional regional description of the risk assessment results of the drilling risk quantitative assessment model.

[0098] It can be seen from the above technical solutions that the method for regionalizing the description of drilling engineering risks provided by the embodiment of the present invention has the following beneficial effects:

[0099] 1. The present invention uses geostatistical methods to perform two-dimensional regional description of the risk assessment results of the wells drilled in the block within the stratigraphic group, and realizes the horizontal distribution of the vertical risk assessment results according to the stratigraphic group, so as to make the drilling engineering design plan more reasonable and the implementation process of the plan safer;

[0100] 2. The method proposed in the present invention realizes the regional description of drilling risks. When drilling a new well in a block, it can judge the possible risks of each layer group at the location of the new well. To a certain extent, it improves the comprehensiveness of the considerations of designers when designing drilling projects, and maximizes the drilling safety in the design stage.

[0101] Furthermore, in step 1 of this scheme, the logging data of key wells drilled in the area are collected, and the formation porosity, formation collapse and formation fracture pressure with uncertainty are obtained according to the Monte Carlo simulation method. The design of step 1 of this scheme can make full use of the good continuity and high resolution characteristics of logging data, while also considering the uncertainty of the geological environment of deep well complex formation drilling, so that the formation pressure is more consistent with the actual situation downhole, which helps to make the drilling engineering design plan more reasonable. Specifically:

[0102] Step 1: Obtaining the formation pressure with uncertainty in the drilled well:

[0103] According to the formation pressure uncertainty analysis method proposed by Sheng Yanan of Petroleum University [Sheng Yanan, Guan Zhichuan, Zhang Guohui, Ye Lang, Liu Shujie, Feng Huanying. Wellbore structure optimization method based on pre-drilling risk prediction [J]. Petroleum Drilling and Production Technology, 2016, 38(04): 415-421.], the formation pores, formation collapse and formation fracture pressure with uncertainty are obtained. The process of obtaining the pressure is as follows: Figure 2 As shown, no further description is given here.

[0104] Furthermore, in step 2 of this solution, the uncertainty-containing formation pores, formation collapse and formation fracture pressure interval profiles of the drilled wells are combined to establish a safe drilling fluid density window with uncertainty; on this basis, the drilling risk probability calculation model is obtained according to the reliability theory and probability statistical analysis theory. Step 2 of this solution is designed in this way to obtain the probability values ​​of well kick, well wall collapse, well leakage and pipe sticking risks within the entire well depth range. Specifically:

[0105] Step 2: Quantitative risk assessment of drilling projects:

[0106] Combined with the constructed formation pores, formation collapse and fracture pressure interval profiles with uncertainty, a safe drilling fluid density window with uncertainty for risk assessment is established:

[0107] (1) Probability distribution of lower limit of drilling fluid density for preventing well kick

[0108]

[0109] (2) Probability distribution of the lower limit of drilling fluid density for preventing wellbore collapse

[0110]

[0111] (3) Probability distribution of the upper limit of drilling fluid density to prevent differential pressure sticking

[0112]

[0113] (4) Probability distribution of the upper limit of the density of the anti-lost drilling fluid

[0114]

[0115] In the formula, S b is the suction pressure coefficient, g / cm 3 ; S g is the excitation pressure coefficient, g / cm 3 ; Δρ is the additional drilling fluid density, g / cm 3 ; S f The upper limit safety factor for preventing leakage, g / cm 3 ; S c is the cycle pressure loss coefficient, g / cm 3 ; ΔP is the allowable value of differential pressure for drilling, MPa; P p is the pore pressure equivalent drilling fluid density, g / cm 3 ;P c is the collapse pressure equivalent drilling fluid density, g / cm 3 ;P f is the drilling fluid density equivalent to the fracture pressure, g / cm 3 ;

[0116] According to the distribution of the upper and lower limits of the drilling fluid density, the safe drilling fluid density window with a cumulative probability of j0 is:

[0117]

[0118] It can be seen that the depth h i The upper and lower limits of safe drilling fluid density when the cumulative probability j = j0 are as follows:

[0119]

[0120]

[0121] When the cumulative probability is j0, the upper and lower limit curve expressions of safe drilling fluid density are shown in equations (8) and (9):

[0122]

[0123]

[0124] In the formula, are the lower and upper limits of safe drilling fluid density with cumulative probability j0; similarly, the lower limit of safe mud density L(h) with cumulative probability j1 is obtained j＝j1 and upper limit H(h) j＝j1, thus obtaining the |j1-j0|×100% safe mud density interval profile; that is, constructing the uncertainty-containing safe drilling fluid density window required for risk assessment;

[0125] The present invention establishes a drilling risk prediction method based on the reliability theory of generalized stress and strength interference. According to the risk-causing mechanism of drilling risks, generalized stress is defined as a risk factor, that is, the pressure of the drilling fluid column in the wellbore; in addition, strength interference is defined as a safety factor, that is, the upper and lower limits of the safe drilling fluid density window; the functional function is defined as a risk function of downhole complexity and accidents. In order to better understand the process of analysis and derivation of drilling risk probabilities, the analysis and derivation process is now specifically described using the well kick risk as an example. The reason for the well kick risk is that the pressure of the drilling fluid column in the wellbore cannot balance the formation pressure, causing the fluid in the formation to enter the wellbore, causing the wellbore safety barrier to fail. The parameter describing the function of the drilling fluid is a random variable of reliability, and its reliability refers to the probability that the drilling fluid column can balance the formation pressure, that is, the probability R that the drilling fluid equivalent density is greater than the lower limit of the well kick prevention, and the calculation formula is:

[0126] R=P(Q>S)=P(QS>0)=P(Q / S>1) (10)

[0127] Where Q is the random variable of drilling fluid equivalent density, and S is the random variable of the lower limit of well kick prevention.

[0128] Taking well kick risk as an example, a specific analysis of the risk probability quantitative calculation model is carried out: the well kick risk value at a certain depth h is the drilling fluid density ρ during drilling d Less than the lower limit of the anti-kickout drilling fluid density at this depth ρ k(h) The probability value p k(h) (ρ d < k(h) ), p k(h) (ρ) is the probability density distribution function of the lower limit of the anti-blowout drilling fluid density. Therefore, the probability P(ρ) that the drilling fluid density is less than the lower limit of the anti-blowout drilling fluid density is d < k(h) ) is the area of ​​the shaded part in the figure. Figure 3 As shown; according to probability analysis theory, its value is 1-F ρk(h) (ρ d ), where F ρk(h) (ρ) is the cumulative probability distribution function of the lower limit of the anti-kickback drilling fluid density, F ρk(h) (ρ d ) is the lower limit of the density of the anti-kickback drilling fluid ρ k(h) Equal to the density of the drilling fluid during drilling ρ d Similarly, the probability basic theory of other risk models (lost circulation, well wall collapse, and stuck pipe) is the same as that of well kick risk, which will not be repeated here.

[0129] According to the reliability theory and probability statistical analysis theory, the probability calculation formula of pre-drilling engineering design risks (well kick, well wall collapse, well leakage, and differential pressure sticking) is obtained, as shown in Table 1.

[0130] Table 1 Drilling engineering risk probability calculation model

[0131] Risk Type Risk Probability Calculation Model Well <![CDATA[R k,h =P(ρ d,h <r k,h )=1-F k,h (r d,h ) (17)]]> Well leakage <![CDATA[R l,h =P(ρ d,h >r l,h )=F l,h (r d,h ) (18)]]> Well wall collapse <![CDATA[R c,h =P(ρ d,h <r c,h )=1-F c,h (r d,h ) (19)]]> Differential pressure drill <![CDATA[R s,k =P(ρ d,h >r sk,h )=F sk,h (r d,h ) (20)]]>

[0132] Furthermore, in step three of this scheme, the support vector machine algorithm is used to fit the experimental variation function to obtain the GA-SVM-Kriging interpolation algorithm; based on the GA-SVM-Kriging interpolation algorithm, the risk assessment results of the drilling risk quantitative assessment model are described in two dimensions. Step three of this scheme is designed in this way to facilitate the two-dimensional regional description of the risk assessment results of the wells that have been drilled in the block within the stratigraphic group, so as to realize the horizontal distribution of the vertical risk assessment results according to the stratigraphic group. In this way, the drilling engineering design scheme can be more reasonable and the implementation process of the scheme can be safer. Specifically:

[0133] Step 3: Regional description of drilling risks based on GA-SVM-Kriging algorithm:

[0134] Geostatistics takes spatial regionalized variables as research objects, and studies spatial parameters with randomness, structure, correlation and dependence. Through the evaluation process of drilling engineering risks, we can know that the risk evaluation results are related to factors such as geological structure, formation fluid and formation pressure. The risk evaluation results can be regarded as a function of geological structure, formation fluid and formation pressure. These influencing factors have regional characteristics. Therefore, the risk evaluation results can be discussed and analyzed as regionalized variables. The risk evaluation results have the following regional characteristics: spatial limitations, different degrees of continuity, and different types of anisotropy. Therefore, they can be regarded as regionalized variables.

[0135] The present invention proposes a regional description algorithm for drilling risk based on the principle of geostatistics, which is mainly composed of two parts, namely, a variation function that can perform regional formation pressure correlation and variability analysis, and a Kriging interpolation algorithm suitable for formation pressure spatial interpolation. In view of the defects of conventional spatial interpolation algorithms that are complex to calculate and require subjective judgment of spatial difference models, a support vector machine algorithm is used to fit the experimental variation function. This method does not require the optimal fitting of the theoretical coefficient of variation, thus making up for the shortcomings of the ordinary Kriging interpolation method. The Gaussian RBF kernel function is selected as the kernel function of the support vector machine, and the regression function form is:

[0136]

[0137] In the SVM regression model, there are three parameters that need to be determined: the insensitive loss parameter ε, the penalty parameter C, and the kernel parameter σ of the Gaussian RBF kernel function. In order to obtain a better fitting effect, this scheme uses a genetic algorithm to optimize the three important parameters in the SVM model. The parameters of the support vector machine model are optimized based on the genetic algorithm. The steps are as follows:

[0138] ① Initialize the population: Use binary coding to randomly generate M initial populations, and preset the number of iterations to N;

[0139] ② Calculate the fitness value: train the SVM with the parameters corresponding to each individual in the group and calculate the fitness value of each individual;

[0140] ③Choice: Determine the initial population G according to the roulette method (0) The number of times an individual is selected and the probability of individual selection are considered, and the excellent individuals with high probability are selected, and then their information is inherited to the next generation of new population;

[0141] ④Crossover: Let the crossover probability P c , randomly pair the parent population to generate new individuals;

[0142] ⑤ Mutation: For a given mutation probability P m , first randomly select M×P m individuals, let the kth character undergo mutation operation, replace the information at the corresponding position of the selected chromosome, and generate a new individual;

[0143] ⑥ Termination condition judgment: if the termination condition is met, stop calculating and outputting the results; otherwise, go to step ② to continue calculating;

[0144] ⑦Decoding: Decode the optimal individual and finally obtain the optimal solution of the three important parameters in the SVM model.

[0145] Select a function that reflects the SVM regression performance as the fitness function:

[0146]

[0147] Where: F(x i ) is the fitness value of the ith individual; y j ,q j They represent the expected output value and training sample output value of the SVM composed of the i-th individual respectively; m is the number of training samples.

[0148] Through the above research, a GA-SVM-Kriging interpolation algorithm was established. Based on this algorithm, the regional description of drilling risks is carried out. The process is as follows: Figure 1 As shown, it is mainly divided into the following three steps:

[0149] (1) Calculate the experimental coefficient of variation γ * (h);

[0150]

[0151] (2) The GA-SVM-Kriging algorithm is used to calculate the experimental coefficient of variation γ * (h) Perform regression fitting to obtain the theoretical coefficient of variation γ(h);

[0152] (3) For the position point x to be estimated o , find the corresponding weight vector [λ1,λ2,...λ n ], and then substitute it into the Kriging interpolation formula to get o The estimated value p at the location * (x0);

[0153] Assume that there are N wells drilled in this area, and the coordinates of the wells are (X i ,Y i ), i = 1, 2, ..., N; in a layer group j, the i-th well has a total of n discrete points of drilling risk probability data, which can be expressed as follows:

[0154]

[0155] If the target well location is (X o ,Y o ), based on the GA-SVM-Kriging interpolation algorithm, in formation j, the risk probability data of the i-th well that has been drilled is transplanted to the target well as follows:

[0156]

[0157] Where: i is the weight coefficient of the i-th well;

[0158] Transplant all N drilled well risk parameters to the target well according to the above method, and finally obtain the drilling risk probability of the target well in layer group j:

[0159]

[0160] Furthermore, after step 3 of this solution, the following steps are also included:

[0161] Step 4: Construction and visualization of regionalized model of drilling engineering risks. Specifically:

[0162] According to the GA-SVM-Kriging interpolation algorithm in step three above, the drilling risk probability values ​​of multiple wells in the target area can be calculated, and then the lateral distribution cloud map of drilling engineering risks in the target stratum group can be obtained based on the spatial interpolation of the calculation results, and the model can be visualized using programming language. That is, in step four, the establishment of the drilling risk area cloud map can grasp the regional rules as a whole, understand the general direction and possibility of the occurrence of formation instability, and provide a certain analytical basis and theoretical basis for the safe and smooth development of subsequent oil and gas reservoirs.

[0163] The technical solution of the present invention is further introduced below:

[0164] First, we collected logging data from key wells in the region and calculated the formation porosity, collapse and fracture pressures with uncertainties using the Monte Carlo simulation method. The calculation process is as follows: Figure 2 As shown; then, using the established formation porosity, collapse and fracture pressure with uncertainty, based on the reliability theory of generalized stress and strength interference, a drilling risk quantitative assessment model is constructed, which can obtain the probability values ​​of well kick, well wall collapse, well leakage and pipe sticking risks within the entire well depth range; then, using the geostatistical method, the risk assessment results of the drilled wells in the block are described in two dimensions within the formation group, so as to achieve the horizontal distribution of the vertical risk assessment results according to the formation group. The method proposed in the present invention takes the target formation group in the region as the basic unit, considers the geological characteristics of the target formation, and establishes a regional description method for drilling engineering risks in the target area. The method is mainly divided into the following steps: (1) Obtaining the formation pressure with uncertainty of the drilled well; (2) Quantitative assessment of the engineering risk of the drilled well; (3) Regional description of drilling risk based on GA-SVM-Kriging algorithm; (4) Construction and visualization of the regional model of drilling engineering risk.

[0165] In order to explain the implementation mode of the present invention in more detail, the present invention is further introduced in conjunction with specific embodiments below:

[0166] The geological research results of Block XX show that various underground complex situations and accidents such as blowout, leakage, collapse, stuck and broken drill tools frequently occur during drilling. For the 8 wells drilled in Block XX, the drilling engineering risk assessment was carried out according to the above process. The risk assessment results are as follows: Figure 4a to Figure 4h Then, the risk assessment results are extended in the LLH group using geostatistical methods to form a cloud map of drilling risk areas. The results are shown in Figure 5a to Figure 5d According to the drilling risk regionalization cloud map, the possible risk conditions of a new well M were predicted before drilling, and the prediction results are shown in Table 2:

[0167] Table 2 Drilling risk prediction results of Well M

[0168]

[0169] After the drilling of Well M was completed, the actual risk occurrence during the construction process can be seen from the well history data as shown in Table 3. According to Tables 2 and 3, the actual risk situations that occurred during the drilling project can be predicted by the regional cloud map, which shows that the establishment of the drilling risk regional cloud map has a great reference value for the prediction of complex underground accidents.

[0170] Table 3 Actual drilling risk of Well M

[0171] Occurrence well section / m Occurrence Layer Group Risk Category 2998～3015 LLH group Well leakage

[0172] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0173] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for regionalizing the description of drilling engineering risks, characterized in that: include: Step 1: Obtain the formation pressure of the drilled well with uncertainty; Step 2: Use formation pressure to construct a drilling risk quantitative assessment model; Step 3: Use geostatistical methods to perform two-dimensional regional description of the risk assessment results of the drilling risk quantitative assessment model; In the second step, a drilling risk quantitative assessment model is constructed using formation pressure, including: Combined with the obtained formation pores, formation collapse and formation fracture pressure interval profiles with uncertainties in the drilled wells, a safe drilling fluid density window with uncertainties is established; The drilling risk probability calculation model is obtained based on reliability theory and probability statistical analysis theory; The step of establishing a safe drilling fluid density window with uncertainty by combining the obtained formation pores, formation collapse and formation fracture pressure interval profiles with uncertainty in the drilled wells includes: (1) Probability distribution of lower limit of drilling fluid density for preventing well kick (2) Probability distribution of the lower limit of drilling fluid density for preventing wellbore collapse (3) Probability distribution of the upper limit of drilling fluid density to prevent differential pressure sticking (4) Probability distribution of the upper limit of the density of the anti-lost drilling fluid In the formula, S b is the suction pressure coefficient, g / cm 3 ; S g is the excitation pressure coefficient, g / cm 3 ; Δρ is the additional drilling fluid density, g / cm 3 ; S f The upper limit safety factor for preventing leakage, g / cm 3 ; S c is the cycle pressure loss coefficient, g / cm 3 ; ΔP is the allowable value of differential pressure for drilling, MPa; P p is the pore pressure equivalent drilling fluid density, g / cm 3 ;P c is the collapse pressure equivalent drilling fluid density, g / cm 3 ;P f is the drilling fluid density equivalent to the fracture pressure, g / cm 3 ; According to the distribution of the upper and lower limits of the drilling fluid density, the safe drilling fluid density window with a cumulative probability of j0 is: It can be seen that the depth h i The upper and lower limits of safe drilling fluid density when the cumulative probability j = j0 are as follows: When the cumulative probability is j0, the upper and lower limit curve expressions of safe drilling fluid density are shown in equations (8) and (9): In the formula, are the lower and upper limits of safe drilling fluid density with cumulative probability j0; similarly, the lower limit of safe mud density with cumulative probability j1 is obtained and upper limit In this way, the |j1-j0|×100% safe mud density interval profile is obtained; The drilling risk probability calculation model obtained according to the reliability theory and probability statistical analysis theory includes: Kick risk probability calculation model: R k,h =P(ρ d,h <p k,h )=1-F k,h (r d,h ); (17) Well leakage risk probability calculation model: R l,h =P(ρ d,h >r l,h )=F l,h (r d,h ); (18) Well collapse risk probability calculation model: R c,h =P(ρ d,h <p c,h )=1-F c,h (r d,h ); (19) Calculation model of the risk probability of differential pressure sticking: R s,k =P(ρ d,h >r sk,h )=F sk,h (r d,h ) (20).

2. The method for regionalizing the description of drilling engineering risks according to claim 1, characterized in that: In the step 1, obtaining the formation pressure with uncertainty in the drilled well includes: The logging data of key wells drilled in the area are collected, and the formation porosity, formation collapse and formation fracture pressure with uncertainty are calculated using the Monte Carlo simulation method.

3. The method for regionalizing the description of drilling engineering risks according to claim 1, characterized in that: In the step 3, a geostatistical method is used to perform a two-dimensional regional description of the risk assessment results of the drilling risk quantitative assessment model, including: The support vector machine algorithm was used to fit the experimental variogram to obtain the GA-SVM-Kriging interpolation algorithm; The risk assessment results of the drilling risk quantitative assessment model are described in two dimensions based on the GA-SVM-Kriging interpolation algorithm.

4. The method for regionalizing the description of drilling engineering risks according to claim 3, characterized in that: Before the support vector machine algorithm is used to fit the experimental variation function to obtain the GA-SVM-Kriging interpolation algorithm, the method further includes: A genetic algorithm is used to optimize and determine three important parameters in the support vector machine algorithm; the three important parameters include: insensitive loss parameter ε, penalty parameter C and kernel parameter σ of Gaussian RBF kernel function.

5. The method for regionalizing the description of drilling engineering risks according to claim 4, characterized in that: The genetic algorithm is used to optimize and determine three important parameters in the support vector machine algorithm, including: ① Initialize the population: Use binary coding to randomly generate M initial populations, and preset the number of iterations to N; ② Calculate the fitness value: train the SVM with the parameters corresponding to each individual in the group and calculate the fitness value of each individual; ③Choice: Determine the initial population G according to the roulette method (0) The number of times an individual is selected and the probability of individual selection are considered, and the excellent individuals with high probability are selected, and then their information is inherited to the next generation of new population; ④Crossover: Let the crossover probability P c , randomly pair the parent population to generate new individuals; ⑤ Mutation: For a given mutation probability P m , first randomly select M×P m individuals, let the kth character undergo mutation operation, replace the information at the corresponding position of the selected chromosome, and generate a new individual; ⑥ Termination condition judgment: if the termination condition is met, stop calculating and outputting the results; otherwise, go to step ② to continue calculating; ⑦Decoding: Decode the optimal individual and finally obtain the optimal solution of the three important parameters in the SVM model.

6. The method for regionalizing the description of drilling engineering risks according to claim 3, characterized in that: The risk assessment results of the drilling risk quantitative assessment model are described in two dimensions based on the GA-SVM-Kriging interpolation algorithm, including: (1) Calculate the experimental coefficient of variation γ * (h); (2) The GA-SVM-Kriging algorithm is used to calculate the experimental coefficient of variation γ * (h) Perform regression fitting to obtain the theoretical coefficient of variation γ(h); (3) For the position point x to be estimated o , find the corresponding weight vector [λ1,λ2,...λ n ], and then substitute it into the Kriging interpolation formula to calculate the estimated value p at the xo position * (x0); Assume that there are N wells drilled in this area, and the coordinates of the wells are (X i ,Y i ), i = 1, 2, ..., N; in a layer group j, the i-th well has a total of n discrete points of drilling risk probability data, which can be expressed as follows: If the target well location is (X o ,Y o ), based on the GA-SVM-Kriging interpolation algorithm, in formation j, the risk probability data of the i-th well that has been drilled is transplanted to the target well as follows: Where: i is the weight coefficient of the i-th well; Transplant all N drilled well risk parameters to the target well according to the above method, and finally obtain the drilling risk probability of the target well in layer group j:

7. The method for regionalizing the description of drilling engineering risks according to claim 1, characterized in that: In the step 3, after the risk assessment results of the drilling risk quantitative assessment model are described in two dimensions using a geostatistical method, the method further includes: Step 4: Construct and visualize the regionalization model of drilling engineering risks.

8. The method for regionalizing the description of drilling engineering risks according to claim 7, characterized in that: In step 4, a drilling engineering risk regionalization model is constructed and visualized, including: The drilling risk probability values ​​of multiple wells in the target area are calculated according to the GA-SVM-Kriging interpolation algorithm. Then, based on the calculation results, the lateral distribution cloud map of drilling engineering risks in the target stratum group is obtained by spatial interpolation, and the model is visualized using programming language.

Images