A spline membership function-based drilling condition fuzzy logic identification method

Abstract

This invention relates to a fuzzy logic identification method for drilling conditions based on spline membership functions. The method includes the following steps: establishing a set of drilling characteristic parameters; selecting drilling characteristic parameters and constructing a representation of their variation with drilling conditions; constructing spline membership functions; constructing a fuzzy logic matrix equation; and comparing the eigenvalues ​​with the spline membership function values ​​to determine the drilling condition type. This method combines spline membership functions with fuzzy logic matrix equations, determines the spline membership function curve that approximates the parameter variation trend, substitutes actual logging data to obtain membership function values, and compares the membership function values ​​with the eigenvalues ​​of the matrix equation to determine the drilling condition. This method improves the accuracy of drilling condition identification.

Description

Technical Field

[0001] This invention belongs to the field of drilling engineering technology and relates to a fuzzy logic identification method for drilling conditions based on spline membership functions. Background Technology

[0002] Automatic drilling condition identification technology has been widely used in the drilling industry, but it is still mainly done manually. In addition, most existing intelligent drilling condition identification systems use threshold methods to judge drilling conditions, which has the defects of incomplete drilling condition identification or single identification method, and cannot guarantee the efficiency of automatic drilling condition identification.

[0003] Currently, drilling parameter characterization relies on membership functions, including typical bell-shaped membership functions, typical trapezoidal membership functions, and typical triangular membership functions. The drawback of these functions is that the drilling parameter data points and shapes are fixed, and the trend description is relatively simple, which affects the accuracy of identification. Conventional membership functions are limited by the number and shape of drilling parameter data points, and the functions cannot infinitely approximate the trend of drilling parameter changes.

[0004] Patent No. CN201811293635.X discloses a method and system for identifying drilling conditions while drilling. The method includes: Step 1, constructing a condition identification model at the surface using historical data of acquired measurement parameters and corresponding drilling conditions. The condition identification model can determine the corresponding drilling conditions based on the measurement data. Step 2, writing the condition identification model into the drilling instrument so that the drilling instrument can determine the real-time drilling conditions during drilling using the condition identification model based on the real-time data of acquired measurement parameters.

[0005] Patent No. CN201910308999.9 discloses a method and apparatus for predicting drilling overflow and leakage conditions. The method includes: acquiring drilling data of the condition to be predicted; and using the drilling data of the condition to be predicted and a pre-established depth belief network model to predict drilling overflow and leakage conditions. It can establish a method for timely detection of overflow and leakage drilling conditions, thereby enabling effective measures to be taken in advance to address and handle complex overflow and leakage conditions in actual production, reducing drilling losses caused by complex conditions. Summary of the Invention

[0006] To address the problems existing in the prior art, this invention discloses a fuzzy logic identification method for drilling conditions based on spline membership functions. This method combines spline membership functions with fuzzy logic matrix equations. By determining the spline membership function curve that approximates the trend of parameter changes, and substituting actual logging data to obtain the membership function value, the method compares the membership function value with the eigenvalue of the matrix equation to determine the drilling conditions. This method improves the accuracy of drilling condition identification.

[0007] This invention specifically discloses a fuzzy logic recognition method for drilling conditions based on spline membership functions, which includes the following steps:

[0008] S1. Establish a set of drilling characteristic parameters: Select key drilling characteristic parameters to form a set of drilling characteristic parameters X0 based on the usage of drilling parameters;

[0009] S2. Select drilling characteristic parameters and construct a representation of their variation with drilling conditions: By analyzing drilling data and logging data from several sample wells, construct a table showing the variation of different key drilling characteristic parameters under different drilling conditions;

[0010] S3. Construct spline membership functions: The key drilling characteristic parameters in step S1 are summarized and unified, and spline membership functions of the corresponding drilling characteristic parameters are constructed using recursive functions, and the membership function values ​​of the corresponding drilling characteristic parameters are calculated.

[0011] S4. Constructing a fuzzy logic matrix equation: Based on the historical data of drilling conditions and corresponding logging data statistically obtained in step S2, different drilling conditions are taken as the anomaly domain of drilling parameters, and the logging data corresponding to different drilling conditions are taken as the symptom domain of drilling parameters. A fuzzy relation matrix equation for drilling condition identification is established. The feature parameters of the well to be identified are input into the fuzzy relation matrix equation for drilling condition identification, and the feature values ​​of the fuzzy relation matrix equation corresponding to the well to be identified are calculated.

[0012] S5. Drilling Condition Identification: The drilling condition type is determined by comparing the feature values ​​obtained in step S4 with the spline membership function values ​​obtained in step S3.

[0013] Furthermore, the key drilling characteristic parameters selected in step S1 include drill bit position, suspended weight, drilling pressure, rotational speed, torque, pump pressure, outlet discharge, mud pit volume, and gas measurement value.

[0014] Furthermore, the specific steps for constructing the spline membership function in step S3 are as follows:

[0015] S31. First, based on the drilling characteristic parameter set X0 in step S1, set each parameter in the drilling characteristic parameter set as x1, x2, x3, x4…x n , where 1, 2, 3, 4, ..., n is the number of drilling characteristic parameters selected;

[0016] S32. Define vector X as a set of m+1 non-decreasing numbers, x0≤x1≤x2≤...≤x, based on the acquisition time of different drilling characteristic parameters. m , where x m Let X be a node, and let X be a node vector. Then, node x... m -x m-1For each fixed constant, 0, 1, 2, 3, ..., m represent time points respectively;

[0017] S33. Use a recursive function to determine the spline basis function for the corresponding drilling characteristic parameters, where the recursive function is defined as:

[0018]

[0019] Where α j (x), β j (x) represents the characteristic parameters of the interpolation, and j is the number of drilling parameters selected in the process of constructing the spline membership function;

[0020] S34. Based on drilling characteristic parameters and the corresponding time points for different drilling characteristic parameters, determine the interval [x] j x j+1 Membership function on ]:

[0021]

[0022] Among them, h j x is the interpolation step size; j For node variables; m j The boundary thresholds set for corresponding drilling conditions are obtained using boundary conditions.

[0023] Furthermore, the specific method for constructing the fuzzy logic matrix equation in step S4 is as follows:

[0024] S41. Using vector X as the set of drilling characteristic parameters and vector Y as the set of drilling condition types, construct a fuzzy relation matrix for drilling conditions:

[0025]

[0026] Where X = {x1, x2, x3, ..., x} i}={x p 1,2,3,...,i},Y={y1,y2,y3,...,y j}={y q 1,2,3,...,j},0≤r pq ≤1, 0≤p≤i, 0≤q≤j, i represents different drilling characteristic parameters, j represents different drilling condition types; matrix R is an order matrix, the row elements of matrix R represent drilling characteristic parameters, the column elements represent drilling condition types, and rij represents the weight of drilling characteristic parameter xi in drilling condition type yj.

[0027] S42. Based on the actual logging data of the well to be identified, obtain the feature vector X of the well to be identified:

[0028] X = {x1, x2, x3, ..., x} i};

[0029] S43. Normalize the feature vector X of the well to be identified to obtain the normalized vector corresponding to the feature vector.

[0030]

[0031] Where x ij Represents the eigenvector element x i Membership degree to condition j;

[0032] S44. Solve the fuzzy relation matrix equation based on the principles of fuzzy mathematics. Right now:

[0033]

[0034] S45. The eigenvalues ​​of the corresponding fuzzy relation matrix equation are obtained by solving the fuzzy relation matrix equation.

[0035] Furthermore, in step S41, the drilling characteristic parameter x i In drilling operation type y j The weights of the parameters are obtained from the table of changes in drilling conditions and key drilling characteristic parameters constructed in step S2.

[0036] Furthermore, the specific steps in S5 for determining the drilling condition type by comparing eigenvalues ​​with spline membership function values ​​are as follows:

[0037] S51. First, the membership function values ​​of the drilling characteristic parameters calculated based on the spline membership functions constructed in step S3 are associated with the corresponding drilling conditions. Each drilling characteristic parameter corresponds to one drilling condition.

[0038] S52. Compare the eigenvalues ​​calculated in step S4 with the membership function values ​​in S3, and select the highly correlated membership function values. The drilling condition type corresponding to the highly correlated membership function value is the drilling condition type determined.

[0039] 1) The fuzzy logic automatic identification method for drilling conditions in this invention first uses spline membership functions to characterize drilling parameters. Given arbitrary drilling parameter data from the well to be logged, the spline membership function can be used to determine the spline curve that best approximates the parameter variation trend, making it easier for computers to quickly and accurately identify drilling conditions. Furthermore, by analyzing the relationship between the characteristics of drilling parameter variations in sample wells and drilling conditions, nine characteristic parameters—suspended weight, drilling pressure, rotational speed, torque, pump pressure, outlet discharge, mud pit volume, gas measurement values, drill bit position, and well depth—are selected to characterize the patterns and construct relevant membership function curves. By calculating the eigenvalues ​​of the fuzzy relation matrix equation of the well to be identified and comparing them with the membership function values ​​constructed from the sample wells, the accuracy of drilling condition identification is improved. Attached Figure Description

[0040] Figure 1 This is a flowchart illustrating a fuzzy logic identification method for drilling conditions based on spline membership functions in this embodiment.

[0041] Figure 2 This is a schematic diagram of the membership function of the multi-point spline in this embodiment;

[0042] Figure 3 This is a schematic diagram of the membership function of the three-point spline in this embodiment;

[0043] Figure 4 This is a comparison chart of the drilling condition identification of the method in this embodiment and the threshold model under the reaming drilling condition;

[0044] Figure 5 A comparison diagram of drilling condition identification using the method of this embodiment and the threshold model under downhole drilling conditions;

[0045] Figure 6 This is a comparison chart showing the drilling condition identification of the method in this embodiment and the threshold model under the sitting drilling condition. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0048] Example 1:

[0049] Combination Figure 1 As shown in the figure, this embodiment specifically discloses a fuzzy logic recognition method for drilling conditions based on spline membership functions, which includes the following steps:

[0050] S1. Establish a set of drilling characteristic parameters: Based on the usage of drilling parameters, select key drilling characteristic parameters to form a set of drilling characteristic parameters X0; In this embodiment, the key drilling characteristic parameters selected include drill bit position, suspended weight, drilling pressure, rotation speed, torque, pump pressure, outlet displacement, mud pit volume, and gas measurement value, i.e., X={x1,x2,x3,...,x9}, where 1-9 represent drill bit position, suspended weight, drilling pressure, rotation speed, torque, pump pressure, outlet displacement, mud pit volume, and gas measurement value, respectively.

[0051] S2. Select drilling characteristic parameters and construct a variation representation law with drilling conditions: Through drilling data and logging data in several sample wells, construct a table of variation law of different key drilling characteristic parameters under different drilling conditions, as shown in Table 1 below;

[0052] Table 1

[0053]

[0054] S3. Construct spline membership functions: The key drilling characteristic parameters in step S1 are summarized and unified, and spline membership functions of the corresponding drilling characteristic parameters are constructed using recursive functions, and the membership function values ​​of the corresponding drilling characteristic parameters are calculated.

[0055] like Figure 2 and 3 As shown, the specific steps for constructing the spline membership function of the well to be identified using the traditional spline membership function are as follows:

[0056] S31. First, based on the drilling characteristic parameter set X0 in step S1, set each parameter in the drilling characteristic parameter set as x1, x2, x3, x4…x n Where 1, 2, 3, 4, ..., n is the number of drilling characteristic parameters selected. For example, when describing the working condition as overflow, the parameters selected are pump pressure, outlet flow rate, inlet flow rate, mud pit volume, and gas measurement value, and n = 5.

[0057] S32. Define vector X as a set of m+1 non-decreasing numbers, x0≤x1≤x2≤...≤x, based on the acquisition time of different drilling characteristic parameters. m , where x m Let X be a node, and let X be a node vector. Then, node x... m -x m-1 For each fixed constant, 0, 1, 2, 3, ..., m represent time points respectively;

[0058] S33. Use a recursive function to determine the spline basis function for the corresponding drilling characteristic parameters, where the recursive function is defined as:

[0059]

[0060] Where α j (x), β j (x) represents the characteristic parameters of the interpolation, and j is the number of drilling parameters selected in the process of constructing the spline membership function;

[0061] S34. Based on drilling characteristic parameters and the corresponding time points for different drilling characteristic parameters, determine the interval [x] j x j+1 Membership function on ]:

[0062]

[0063] Among them, h j x is the interpolation step size; j For node variables; m j The boundary thresholds are set for corresponding drilling conditions and obtained using boundary conditions. For example, when the drilling condition is set to overflow or leakage, m j This is expressed as the threshold value for the difference between inlet and outlet flow rates; when the drilling condition is set to stuck pipe, m j It is expressed as drill string torque.

[0064] S4. Constructing a fuzzy logic matrix equation: Based on the historical data of drilling conditions and corresponding logging data statistically obtained in step S2, different drilling conditions are taken as the anomaly domain of drilling parameters, and the logging data corresponding to different drilling conditions are taken as the symptom domain of drilling parameters. A fuzzy relation matrix equation for drilling condition identification is established. The feature parameters of the well to be identified are input into the fuzzy relation matrix equation for drilling condition identification, and the feature values ​​of the fuzzy relation matrix equation corresponding to the well to be identified are calculated.

[0065] The specific method for constructing the fuzzy logic matrix equation in step S4 is as follows:

[0066] S41. Using vector X as the set of drilling characteristic parameters and vector Y as the set of drilling condition types, construct a fuzzy relation matrix for drilling conditions:

[0067]

[0068] Where X = {x1, x2, x3, ..., x} i}={x p 1,2,3,...,i},Y={y1,y2,y3,...,y j}={y q 1,2,3,...,j},0≤r pq ≤1, 0≤p≤i, 0≤q≤j, where i represents different drilling characteristic parameters and j represents different drilling condition types; matrix R is an dimensional matrix, where the row elements of matrix R represent drilling characteristic parameters and the column elements represent drilling condition types. ij Represents drilling characteristic parameter x i In drilling operation type y j The weight of drilling characteristic parameter xi in drilling condition type yj is obtained by constructing the table of changes in drilling conditions and key drilling characteristic parameters in step S2.

[0069] S42. Based on the actual logging data of the well to be identified, obtain the feature vector X of the well to be identified:

[0070] X = {x1, x2, x3, ..., x} i};

[0071] S43. Normalize the feature vector X of the well to be identified to obtain the normalized vector corresponding to the feature vector.

[0072]

[0073] Where x ij Represents the eigenvector element x i Membership degree to condition j;

[0074] S44. Solve the fuzzy relation matrix equation based on the principles of fuzzy mathematics. Right now:

[0075]

[0076] S45. The eigenvalues ​​of the corresponding fuzzy relation matrix equation are obtained by solving the fuzzy relation matrix equation.

[0077] S5. Drilling Condition Identification: The drilling condition type is determined by comparing the eigenvalues ​​obtained in step S4 with the spline membership function values ​​obtained in step S3. The specific steps are as follows:

[0078] S51. First, the membership function values ​​of the drilling characteristic parameters calculated based on the spline membership functions constructed in step S3 are associated with the corresponding drilling conditions. Each drilling characteristic parameter corresponds to one drilling condition.

[0079] S52. Compare the eigenvalues ​​calculated in step S4 with the membership function values ​​in S3, and select the highly correlated membership function values. The drilling condition type corresponding to these highly correlated membership function values ​​is the determined drilling condition type. For example... Figure 4-6 As shown, through field logging data survey, taking well N25 as an example, 168,200 sets of data were transmitted through logging data. The traditional threshold model found that 35,509 sets of data did not identify the working conditions, with an unidentification rate as high as 21.112%. However, the model in this embodiment identified all 168,200 sets of data from well N25, with an identification rate of 100%.

[0080] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A fuzzy logic identification method for drilling conditions based on spline membership functions, characterized in that, Includes the following steps: S1. Establish a set of drilling characteristic parameters: Based on the usage of drilling parameters, select key drilling characteristic parameters to form a set of drilling characteristic parameters X0; S2. Select drilling characteristic parameters and construct a representation of their variation with drilling conditions: By analyzing drilling data and logging data from several sample wells, construct a table showing the variation of different key drilling characteristic parameters under different drilling conditions; S3. Construct spline membership functions: The key drilling characteristic parameters in step S1 are summarized and unified, and spline membership functions of the corresponding drilling characteristic parameters are constructed using recursive functions, and the membership function values ​​of the corresponding drilling characteristic parameters are calculated. S4. Constructing a fuzzy logic matrix equation: Based on the historical data of drilling conditions and corresponding logging data statistically obtained in step S2, different drilling conditions are taken as the anomaly domain of drilling parameters, and the logging data corresponding to different drilling conditions are taken as the symptom domain of drilling parameters. A fuzzy relation matrix equation for drilling condition identification is established. The feature parameters of the well to be identified are input into the fuzzy relation matrix equation for drilling condition identification, and the feature values ​​of the fuzzy relation matrix equation corresponding to the well to be identified are calculated. S5. Drilling Condition Identification: The drilling condition type is determined by comparing the eigenvalues ​​obtained in step S4 with the spline membership function values ​​obtained in step S3. The specific steps for constructing the spline membership function in step S3 are as follows: S31. First, based on the drilling characteristic parameter set X0 in step S1, set each parameter in the drilling characteristic parameter set as x1, x2, x3, x4…x n , where 1, 2, 3, 4, ..., n is the number of drilling characteristic parameters selected; S32. Define vector X as a set of m+1 non-decreasing numbers, x0≤x1≤x2≤...≤x, based on the acquisition time of each drilling characteristic parameter. m , where x m Let X be a node, and let X be a node vector. Then, node x... m -x m-1 For each fixed constant, 0, 1, 2, 3, ..., m represent time points respectively; S33. Use a recursive function to determine the spline basis function for the corresponding drilling characteristic parameters, where the recursive function is defined as: ， Where α j (x), β j (x) represents the characteristic parameters of the interpolation, j is the number of drilling parameters selected in the process of constructing the spline membership function, and y j This refers to the drilling operation type; S34. Based on drilling characteristic parameters and the corresponding time points for different drilling characteristic parameters, determine the interval [x] j x j+1 Membership function on ]: , Among them, h j x is the step size; j For node variables; m j The boundary thresholds set for the corresponding drilling conditions are obtained using boundary conditions.

2. The fuzzy logic identification method for drilling conditions according to claim 1, characterized in that: The key drilling characteristic parameters selected in step S1 include drill bit position, suspended weight, drilling pressure, rotational speed, torque, pump pressure, outlet displacement, inlet flow rate, mud pit volume, and gas measurement value.

3. The fuzzy logic identification method for drilling conditions according to claim 1, characterized in that: The specific method for constructing the fuzzy logic matrix equation in step S4 is as follows: S41. Using vector X as the set of drilling characteristic parameters and vector Y as the set of drilling condition types, construct a fuzzy relation matrix for drilling conditions: ， in , ,0≤ ≤1, 0≤p≤i, 0≤q≤j, where i represents different drilling characteristic parameters and j represents different drilling condition types; matrix R is an dimensional matrix, where the row elements of matrix R represent drilling characteristic parameters and the column elements represent drilling condition types. ij Represents drilling characteristic parameter x i In drilling operation type y j The weight it occupies in the middle; S42. Based on the actual logging data of the well to be identified, obtain the feature vector X of the well to be identified: ； S43. Normalize the feature vector X of the well to be identified to obtain the normalized vector corresponding to the feature vector. : , Where x ij Represents the eigenvector element x i For working conditions j Membership degree; S44. Solve the fuzzy relation matrix equation based on the principles of fuzzy mathematics. ,Right now: ， S45. The eigenvalues ​​of the corresponding fuzzy relation matrix equation are obtained by solving the fuzzy relation matrix equation.

4. The fuzzy logic identification method for drilling conditions according to claim 1, characterized in that: In step S41, the drilling characteristic parameter x i In drilling operation type y j The weights in the data are obtained through the table of changes in drilling conditions and key drilling characteristic parameters constructed in step S2.

5. The fuzzy logic identification method for drilling conditions according to claim 1, characterized in that: The specific steps in S5 for determining the drilling condition type by comparing eigenvalues ​​with spline membership function values ​​are as follows: S51. First, the membership function values ​​of the drilling characteristic parameters calculated based on the spline membership functions constructed in step S3 are associated with the corresponding drilling conditions. Each drilling characteristic parameter corresponds to one drilling condition. S52. Compare the eigenvalues ​​calculated in step S4 with the membership function values ​​in S3, and select the highly correlated membership function values. The drilling condition type corresponding to the highly correlated membership function value is the drilling condition type determined.

Images