A multi-model solution method for time-effectiveness analysis of zipper fracturing construction

Abstract

A multi-model solution method for analyzing the timeliness of zipper-type fracturing construction, firstly, the timeliness factors of fracturing construction are set; secondly, the grey correlation analysis is used to obtain the correlation ranking, the key factors affecting the timeliness of fracturing construction are calculated, and the optimal correlation conclusion is selected for sub-factor clustering to simplify the research system; again, the AHP model is constructed according to the clustering results, and hypothesis analysis and sensitivity analysis are performed to predict the future development law of the timeliness of zipper-type fracturing construction; finally, the average construction efficiency is set as the target data, the process capability analysis of the construction company is performed, and the construction company is evaluated in the category of quality engineering. The present invention monitors the whole process of fracturing construction through grey correlation analysis and hierarchical analysis conclusions, and dynamically pre-controls the factors affecting the efficiency of fracturing at different stages; through the conclusions of process capability analysis, the advantages and disadvantages of different companies under the premise of quality requirements are compared to guide bidding and workload distribution; the algorithm is simple and easy to program, and control optimization can be performed accordingly.

Description

Technical Field

[0001] The present invention relates to the technical field of shale gas fracturing, and in particular to a multi-model solution method for timeliness analysis details of zipper-type fracturing construction based on grey correlation analysis, hierarchical analysis method and process capability analysis. Background Art

[0002] At present, the factory-based fracturing of multiple horizontal wells is 50% more time-efficient and 30% more complex than single-well fracturing. The "factory-based" fracturing mode is divided into two types: zipper-type fracturing and synchronous fracturing. The principle is to use a limited well site to deploy multiple sets of fracturing units, and simultaneously perform segmented multi-cluster fracturing between two or more adjacent wells, or perform zipper-type alternating fracturing between adjacent wells, so that the shale formation can withstand higher pressure, enhance stress interference between adjacent wells, and produce a more complex fracture network. Zipper-type alternating fracturing has more advantages in economy, timeliness, and practicality. The implementation of zipper-type fracturing requires the coordinated cooperation of the fracturing team, the gas testing team, and the logging team, and can only work with the support of auxiliary factors such as water supply, power supply, and proppant supply. The various tasks are highly connected, and there are many external influencing factors. It is a typical fuzzy system. However, the screening of key processes and the impact of timeliness of the entire "factory-based" zipper-type fracturing construction process are still unclear. The business matching evaluation and bidding decision-making of different construction companies in the same area are relatively general. It is necessary to explore the regularity of key factors under the fuzzy system and fuzzy decision-making under different conditions, solve the selection of key evaluation points of the fracturing process in different work areas and different working conditions, and rank the advantages and disadvantages of different companies. Summary of the invention

[0003] The purpose of the present invention is to overcome the shortcomings of the above-mentioned background technology and provide a multi-model joint solution method suitable for zipper-type fracturing construction, which takes into account the sorting out of the key points of timeliness analysis and the construction company's optimal decision-making. The analysis method includes a grey relational analysis (GRA) model, an analytical hierarchy process (AHP) model and a process capability analysis model. The GRA model can intuitively analyze dynamic problems under current working conditions; the AHP model can describe the future development law of the problem based on the GRA analysis and achieve relative prediction; the process capability analysis model can serve as a demand supplement for engineering problems and make fuzzy decisions. This multi-task model collaborative working method can give full play to the strengths of different mathematical models under the premise of data sharing, reasonably use the conclusions of different models to substitute into the new model of the next stage, reduce analysis errors, and more clearly describe the essence of the problem.

[0004] To achieve the above object, the present invention provides a multi-model solution method for zipper-type fracturing construction timeliness analysis, comprising the following steps:

[0005] S1: Data preparation: Collect the average construction efficiency data, timeliness influencing factors data and average construction efficiency data of the best wells of all construction companies in the work area;

[0006] S2: Analysis of factors affecting timeliness:

[0007] S21: Taking the timeliness influencing factors as sub-factors and the average construction efficiency as the parent factor, the grey correlation analysis model calculation method is used to calculate the comprehensive correlation of the timeliness influencing factors of zipper fracturing construction with respect to the average construction efficiency, and the key factors of fracturing timeliness are analyzed based on the comprehensive correlation;

[0008] S22: According to the comprehensive correlation calculation result in step S21, select the grey clustering sub-factors, construct a triangular matrix table to calculate the comprehensive correlation between the grey clustering sub-factors, and perform cluster analysis;

[0009] S3: Prediction of factors affecting timeliness: Based on the results of cluster analysis, a hierarchical analysis model is constructed, and hypothesis analysis and sensitivity analysis are performed. Based on the analysis results, the future development law of the timeliness of zipper fracturing construction is predicted;

[0010] S4: Construction company quality engineering assessment: Use process capability analysis calculation method, set average construction efficiency as target data, use the average construction efficiency data of the optimal well to conduct process capability analysis of the construction company, and assess the construction company in the scope of quality engineering.

[0011] Furthermore, the factors affecting the timeliness of the zipper-type fracturing construction include: fracturing construction efficiency, joint construction efficiency, average failure rate of key and critical parts, average efficiency of solving key and critical faults, efficiency of solving sand blockage, conventional process connection time, first-stage perforation timeliness, water supply timeliness, sand supply timeliness, high-pressure parts leakage treatment efficiency, and average skill score of team members.

[0012] Furthermore, the calculation method of the comprehensive correlation degree in the above step S2 is:

[0013] S2101: Calculate the gray absolute correlation degree, set X i =(x i (1),x i (2),…,x i (n)) is the system behavior sequence, let: k=1,2,…,n, n represents the number of elements in the sequence; then For X i The zeroing image of the initial point, D is called the zeroing operator; let X i ,X j are all system behavior sequences, then define: Where ε ijrepresents the absolute correlation; i, j = 0, 1, 2, ... m; m represents the system behavior sequence number; s represents the correlation coefficient;

[0014] S2102: Calculate the relative grey correlation, let X i =(x i (1),x i (2),…,x i (n)) is the system behavior sequence, let X i (k)c=x i (k) / x i (1); k = 1, 2, ..., n; then X i C=X' i =(x i (1)c,x i (2)c,…,x i (n)c) is X i The zero image of the starting point, C is the initialization operator; let X i ,X j They are all system behavior sequences, and their initial value images are X' i ,X' j ; The zero images of the starting points are calculated as follows: Then define: Where r ij Represents relative correlation; s' represents correlation coefficient

[0015] S2103: Calculate the grey comprehensive correlation degree, set X i ,X j are all system behavior sequences, and ε ij ,r ij are the absolute correlation and relative correlation of the two sequences respectively, then we define: ij =θε ij +(1-θ)r ij , where ρ ij Represents the comprehensive correlation, θ is the adjustment coefficient, θ∈[0,1].

[0016] Furthermore, in the above step S2103, the parameter θ is used to adjust the influence of the absolute correlation and the relative correlation on the comprehensive correlation, and preferably θ=0.5.

[0017] Furthermore, in the above step S2, when the correlation degree in the calculation result is greater than 0.5, it is determined to be the main influencing factor; if the correlation degrees of all factors are greater than 0.5, the threshold is set to 0.85 and used as the basis for determining gray clustering.

[0018] Furthermore, in the above step S3, the construction of the hierarchical analysis model includes the following steps:

[0019] S301: First, identify the problem, build a physical model and collect data; second, build a hierarchical model of the highest layer, i.e., the target layer, the middle layer, i.e., the criterion layer, and the lowest layer, i.e., the solution layer; perform pairwise comparison and scoring based on the engineering evaluation results to determine the score of the lower layer to the upper layer; finally, perform hierarchical synthesis calculation and consistency test, using the formula The formula for calculating the consistency index CI is: max is the maximum eigenvalue of the judgment matrix; n is the order of the judgment matrix;

[0020] S302: Adopting Calculate the consistency ratio CR. When CR < 0.10, it means that it has passed the one-time test. Otherwise, it does not have satisfactory consistency. In the formula, CI is the consistency index; RI is the random consistency index;

[0021] S303: Calculate the weight vector, use the geometric mean method, take the square root of the continuous product of each variable value, and use the geometric mean to calculate the development speed of the predicted target: Where W represents the development speed; i = 1, 2, 3..., n; a ij is the element in the i-th row and j-th column of the judgment matrix.

[0022] Furthermore, in the above step S4, the process capability index C is used pk Conduct process capability determination:

[0023]

[0024] Where: C pk : Process capability index without offset; USL: upper limit of product quality requirement; LSL: lower limit of product quality requirement; σ: standard deviation of normal distribution of process characteristics; μ: center of distribution of process quality characteristics; C pu : Process capability index with only upper limit requirement; C pl : Process capability index with only lower limit requirements.

[0025] Furthermore, according to the process capability index C pk , the process capability of the construction company is determined according to the ISO / TS16949 process capability level evaluation standard, and finally the engineering guidelines are obtained.

[0026] Compared with the prior art, the present invention has the following advantages:

[0027] Firstly, the workflow of grey correlation analysis → cluster analysis → hierarchical analysis is adopted in the present invention to obtain the key points that urgently need to be promoted and improved under the current boundary conditions, and to make certain predictions on the future changes of the working conditions; the grey correlation analysis parent factor is selected as the target data for process capability analysis, and the results of grey correlation analysis and hierarchical analysis can be combined to evaluate the construction company, so as to guide the bidding work under the preset boundary conditions.

[0028] Secondly, the present invention defines explicit causes, implicit causes and indirect causes respectively according to the results of on-site analysis and in combination with general fracturing conditions. Explicit causes: can directly affect the overall construction efficiency of zipper fracturing, and even lead to shutdown in severe cases. Implicit causes: key and critical processes that affect fracturing construction, but the probability of abnormalities occurring on site is accidental. Indirect causes: indirect influencing causes that will occur with a certain probability, and if abnormalities occur, they can be rectified in a short time. However, if the abnormalities are not handled in time, they will induce implicit causes or explicit causes after a certain period of accumulation. The ideas are clear, the logic is clear, and it can be connected with the on-site data system to dynamically analyze the working conditions.

[0029] Thirdly, the invention has low cost, and the GRA model and the AHP model have a good data coupling relationship in problem analysis, and can perform data combing, simplification and calculation under the fuzzy system. The workflow can effectively use all the data on site for analysis, without specific data, and without installing special instruments or sampling analysis instruments. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a process logic diagram of a multi-model solution method for time-effectiveness analysis of zipper-type fracturing construction.

[0031] Figure 2 It is a configuration diagram of the result of the correlation calculation in the embodiment of the present invention.

[0032] Figure 3 This is an intuitive diagram of observed sample clustering in an embodiment of the present invention.

[0033] Figure 4 It is a structural model diagram in an embodiment of the present invention.

[0034] Figure 5 It is a bar chart of the sample weight calculation results of the criterion layer and the scheme layer in the embodiment of the present invention.

[0035] Figure 6 This is a bar chart showing the calculation results of the solution layer sample weights assuming that the criterion layer weights are the same in an embodiment of the present invention.

[0036] Figure 7 This is a sensitivity analysis calculation result diagram under the B1 assumption in an embodiment of the present invention.

[0037] Figure 8This is a sensitivity analysis calculation result diagram under the B2 assumption in an embodiment of the present invention.

[0038] Fig. 9 This is a sensitivity analysis calculation result diagram under the B3 assumption in an embodiment of the present invention.

[0039] Fig.10 It is a normal test chart of the average construction efficiency of Company A in the embodiment of the present invention.

[0040] Fig.11 It is a normal test chart of the average construction efficiency of Company B in the embodiment of the present invention.

[0041] Fig.12 It is a normal test chart of the average construction efficiency of Company C in the embodiment of the present invention.

[0042] Fig.13 It is a normal test chart of the average construction efficiency of Company D in the embodiment of the present invention.

[0043] Fig.14 It is a normal test chart of the average construction efficiency of Company E in the embodiment of the present invention.

[0044] Fig.15 It is a normal test chart of the average construction efficiency of F Company in the embodiment of the present invention.

[0045] Fig.16 This is the process capability conclusion diagram of Company A in an embodiment of the present invention.

[0046] Fig.17 This is the process capability conclusion diagram of Company B in the embodiment of the present invention.

[0047] Fig.18 This is the process capability conclusion diagram of Company C in an embodiment of the present invention.

[0048] Fig.19 This is the process capability conclusion diagram of Company D in the embodiment of the present invention.

[0049] Fig. 20 This is the process capability conclusion diagram of Company E in the embodiment of the present invention.

[0050] Fig.21 This is the process capability conclusion diagram of Company F in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] The following embodiments are used to describe the implementation of the present invention in detail, but they do not limit the present invention and are only used as examples. At the same time, the advantages of the present invention will become clearer and easier to understand.

[0052] A multi-model solution method for timeliness analysis of zipper fracturing construction, characterized by comprising the following steps:

[0053] S1: Data preparation: Collect the average construction efficiency data, timeliness influencing factors data and average construction efficiency data of the optimal wells of all construction companies in the work area for zipper fracturing construction.

[0054] The timeliness analysis of zipper fracturing construction is divided into two tasks: analysis and prediction of timeliness influencing factors and quality engineering assessment of construction companies.

[0055] S2: Analysis of factors affecting timeliness:

[0056] S21: Taking the timeliness influencing factors as sub-factors and the average construction efficiency as the parent factor, the grey correlation analysis model calculation method is used to calculate the comprehensive correlation between the timeliness influencing factors of zipper fracturing construction and the average construction efficiency, and the key factors of fracturing timeliness are analyzed based on the comprehensive correlation.

[0057] S22: According to the comprehensive correlation calculation result in step S21, select the grey clustering sub-factors, construct a triangular matrix table to calculate the comprehensive correlation between the grey clustering sub-factors, and perform cluster analysis.

[0058] The zipper-type fracturing construction is divided into construction details and influencing factors (average construction efficiency, fracturing construction efficiency, joint construction efficiency, average failure rate of key and critical parts, average efficiency of solving key and critical faults, efficiency of solving sand blockage, connection time of conventional processes, time efficiency of first-stage perforation, time efficiency of water supply, time efficiency of sand supply, efficiency of handling puncture leakage of high-pressure parts, and average skill score of team members), and the average construction efficiency is set as the objective function.

[0059] The calculation method of comprehensive correlation is:

[0060] S2101: Calculate the gray absolute correlation degree, set X i =(x i (1),x i (2),…,x i (n)) is the system behavior sequence, let: k=1,2,…,n, n represents the number of elements in the sequence; then For X i The zeroing image of the initial point, D is called the zeroing operator; let X i ,X j are all system behavior sequences, then define: Where ε ij represents the absolute correlation; i, j = 0, 1, 2, ... m; m represents the system behavior sequence number; s represents the correlation coefficient;

[0061] S2102: Calculate the relative grey correlation, let X i =(x i (1),x i(2),…,x i (n)) is the system behavior sequence, let X i (k)c=x i (k) / x i (1); k = 1, 2, ..., n; then X i C=X' i =(x i (1)c,x i (2)c,…,x i (n)c) is X i The zero image of the starting point, C is the initialization operator; let X i ,X j They are all system behavior sequences, and their initial value images are X' i ,X' j ; The zero images of the starting points are calculated as follows: Then define: Where r ij Represents relative correlation; s' represents correlation coefficient

[0062] S2103: Calculate the grey comprehensive correlation degree, set X i ,X j are all system behavior sequences, and ε ij ,r ij are the absolute correlation and relative correlation of the two sequences respectively, then we define: ij =θε ij +(1-θ)r ij , where ρ ij Represents the comprehensive correlation, θ is the adjustment coefficient, θ∈[0,1].

[0063] Furthermore, in step S2103, θ∈[0,1], this parameter is used to adjust the influence of the absolute correlation and the relative correlation on the comprehensive correlation, and generally θ=0.5 can be taken.

[0064] Furthermore, the comprehensive correlation combines the advantages of absolute correlation and relative correlation, which not only reflects the similarity between the two sequence images, but also reflects the closeness of the change rates of the two sequences relative to their respective starting points. It is a quantitative indicator that accurately describes the closeness of the connection between the sequences. Therefore, the final result of the grey correlation modeling part is based on the final result of step S2103.

[0065] Furthermore, the grey correlation modeling part sets the zipper fracturing construction details factor as a sub-factor according to the field conditions, and the average construction efficiency as the parent factor. In the calculation results, when the correlation degree is greater than 0.5, it is determined to be the main influencing factor. If the correlation degree of all factors is greater than 0.5, the threshold is set to 0.85 and used as the basis for grey clustering.

[0066] According to engineering experience, the clustering results usually divide all factors into three to four categories to facilitate the next step of AHP modeling. Dividing into three categories can comprehensively consider the calculation of engineering quantities and engineering reality. In the analysis of the timeliness of fracturing construction, it is best to classify all factors into three categories: explicit, implicit, and indirect causes.

[0067] S3: Prediction of factors affecting timeliness: Based on the results of cluster analysis, a hierarchical analysis model is constructed, and hypothesis analysis and sensitivity analysis are performed. Based on the analysis results, the future development law of the timeliness of zipper fracturing construction is predicted.

[0068] The analytic hierarchy process is to decompose the abstract complex problem into various components through decision making, and then set the factors and problem logic into the target layer, criterion layer, and solution layer with a grouped hierarchical structure. Finally, through matrix construction, the mutual influence relationship between the factors is compared two by two, so as to quantify the components of the problem and calculate the weight of each factor in the system. Finally, the relative importance of each factor is determined to achieve the purpose of analysis or decision-making.

[0069] The construction of hierarchical analysis model includes the following steps:

[0070] S301: First, identify the problem, build a physical model and collect data; then, construct the highest layer (target layer), middle layer (criteria layer), and lowest layer (solution layer) hierarchical structure model in turn; perform pairwise comparison and scoring based on the engineering evaluation results to determine the score of the lower layer to the upper layer; finally, perform hierarchical synthesis calculation and consistency test, using the formula The formula for calculating the consistency index CI is: max is the maximum eigenvalue of the judgment matrix; n is the order of the judgment matrix.

[0071] S302: Adopting Calculate the consistency ratio CR. When CR < 0.10, it means that it has passed the one-time test, otherwise it does not have satisfactory consistency; CI is the consistency index; RI is the random consistency index.

[0072] S303: Calculate the weight vector, use the geometric mean method, take the square root of the continuous product of each variable value, and use the geometric mean to calculate the development speed of the predicted target: Where W represents the development speed; i = 1, 2, 3..., n; a ij is the element in the i-th row and j-th column of the judgment matrix.

[0073] S4: Construction company quality engineering assessment: Use process capability analysis calculation method, set average construction efficiency as target data, use the average construction efficiency data of the optimal well to conduct process capability analysis of the construction company, and assess the construction company in the scope of quality engineering.

[0074] Process capability analysis is an activity that studies the quality status of a process. It refers to the ability of the process to meet the quality of the process. It measures the internal consistency of the process and the minimum fluctuation in the most stable state. The process capability index C is usually used. pk Make a judgment:

[0075] C pk : Process capability index without offset; USL: upper limit of product quality requirement; LSL: lower limit of product quality requirement; σ: standard deviation of normal distribution of process characteristics; μ: center of distribution of process quality characteristics; C pu : Process capability index with only upper limit requirement; C pl : Process capability index with only lower limit requirement;

[0076] Finally, the process capability is determined according to the ISO / TS16949 process capability level evaluation standard (Table 1), and the engineering guidance is finally obtained.

[0077] Table 1 Evaluation indicators

[0078]

[0079]

[0080] like Figure 1 As shown, the present invention proposes a multi-model solution method for analyzing the timeliness of zipper-type fracturing construction, and evaluates the timeliness of zipper-type fracturing construction under a fuzzy system. It solves the selection of key evaluation points of the fracturing process in different work areas and different working conditions, and ranks the advantages and disadvantages of different companies. The specific steps are as follows:

[0081] The construction efficiency of the six companies in the work area in 2020 (Table 2) was collected as research data for grey correlation analysis and hierarchical analysis; the average construction efficiency of the best five wells of the six companies each year from 2015 to 2020 (Tables 3 to 8) was collected as research data for process capability analysis.

[0082] Table 2 Construction time efficiency of 6 companies in the work area in 2020

[0083]

[0084] Table 3 Average construction efficiency of the best five wells of Company A from 2015 to 2020

[0085]

[0086] Table 4 Average construction efficiency of the best five wells of Company B from 2015 to 2020

[0087]

[0088] Table 5 Average construction efficiency of the best five wells of Company C from 2015 to 2020

[0089]

[0090]

[0091] Table 6 Average construction efficiency of the best five wells of Company D from 2015 to 2020

[0092]

[0093] Table 7 Average construction efficiency of the best five wells of Company E from 2015 to 2020

[0094]

[0095]

[0096] Table 8 Average construction efficiency of the best five wells of Company F from 2015 to 2020

[0097]

[0098] The first step is to use the grey correlation analysis model calculation method, assuming that the average construction efficiency is x1~x 11 Prepare data for modeling (Table 9).

[0099] Table 9 Grey correlation analysis data preparation

[0100]

[0101]

[0102] Here we take the calculation of the correlation between x0 and x1 as an example:

[0103] Absolute correlation calculation:

[0104] (1) Zeroing image of the starting point of the sequence:

[0105] x0 sequence: 0.0000,-0.0400,0.3700,-0.8700,-0.1300,0.4100,

[0106] x1 sequence: 0.0000, 0.0800, -0.0700, 0.4000, 0.1100, -0.1300,

[0107] (2) Calculate |s0|, |s1|, |s1-s0|

[0108] |s0|=0.465; |s1|=0.455; |s1-s0|=0.010

[0109] ε ij =0.9948

[0110] Relative correlation calculation:

[0111] (1) Initial value image of the sequence:

[0112] x0 sequence: 1.0000, 0.9886, 1.1054, 0.7521, 0.9630, 1.1168,

[0113] x1 sequence: 1.0000, 1.0285, 0.9751, 1.1423, 1.0391, 0.9537,

[0114] (2) Zeroing image of the starting point of the sequence:

[0115] x1 sequence: 0.0000,-0.0114,0.1054,-0.2479,-0.0370,0.1168,

[0116] x2 sequence: 0.0000, 0.0285, -0.0249, 0.1423, 0.0391, -0.0463,

[0117] (3) Calculate |s'0|, |s'1|, |s'1-s'0|

[0118] |s′0|=0.1325; |s′1|=0.16185; |s′1-s′0|=0.02935

[0119] r ij =0.9778

[0120] Comprehensive correlation calculation:

[0121] ρ ij =θε ij +(1-θ)r ij =0.5×0.9948+(1-0.5)×0.9778=0.9863

[0122] It can be concluded that x1~x 11 ε of the sequence with respect to the x0 sequence ij ,r ij ,ρ ij result( Figure 2 ).

[0123] pass Figure 2It was concluded that the absolute correlation and relative correlation curves have the same changing trend, but there is an intersection of development trends between sequences 3 to 5 in the curve, which is judged as inconsistency in the correlation data. Therefore, simply considering the similarity between the two sequence images and the rate of change of the starting point cannot accurately describe the system rules. The use of comprehensive correlation results can analyze the key factors of fracturing effectiveness under the premise of strong model robustness. The calculation shows that the comprehensive correlation results of each sub-factor are greater than 0.5. According to the grey correlation analysis standard, all factors are judged to be important. Therefore, all sub-factors must be included in the subsequent grey clustering calculation. The comprehensive correlation of the 11 groups of sequences is ranked as follows: x1>x4>x6>x2>x 11 >x7>x5>x 10 >x3>x5>x9>x8, among which the correlation degree of factors x1, x4 and x6 is greater than 0.85, which can be used as key factors for quality assessment in the fracturing construction process.

[0124] In the second step, according to the calculation results of the comprehensive correlation, the grey clustering sub-factors are selected and a triangular matrix table is constructed to continue the calculation of the comprehensive correlation between the factors. Here, all sub-factors are selected to participate in the cluster analysis and expressed in the form of a triangular matrix (Table 10).

[0125] Table 10 x1~x 11 Comprehensive correlation matrix of sequence observation samples

[0126]

[0127]

[0128] Grey correlation clustering is used to define similar indicators. It reduces the number of indicators on the basis of dimensionality reduction processing data and simplifies research problems. The grey correlation clustering threshold needs to be determined according to the needs of the actual problem. Usually, the closer the threshold is to 1, the finer the classification is, and the fewer attribute indicators are in each component; the smaller the threshold is, the coarser the classification is, and the more attribute indicators are in each component. According to data characteristics and engineering experience, the correlation of the main factors x1, x4, and x6 obtained from the comprehensive correlation calculation results is greater than 0.85. Therefore, the clustering threshold is set to 0.85, and the comprehensive correlation values ​​that exceed the threshold are marked in Table 10. Search in the row and column directions of the matrix, and classify the samples represented by the target column and the samples represented by its row into one category. After sorting the data, ( Figure 3 ): x1, x4, x6 are one type; x5, x7 are one type; x2, x3, x8, x9, x 10 、x 11 For a category.

[0129] The third step is to divide the sub-factors x1~x into 11The sequences are divided into three categories and brought into the hierarchical analysis model for further calculation. According to engineering experience, the factors affecting the timeliness of fracturing construction are x1~x 12 It runs through the entire construction process and appears at different construction nodes with a certain probability, and forms a relationship of mutual influence. Therefore, x1, x4, and x6 are classified as explicit causes; x5 and x7 are classified as implicit causes; x2, x3, x8, x9, and x10 are classified as implicit causes. 10 、x 11 Classified as indirect causes. The explicit causes can directly affect the overall construction efficiency of zipper fracturing, and even cause shutdowns in serious cases. Hidden causes affect the key processes of fracturing construction, but the probability of abnormalities on site is accidental. Indirect causes are indirect influencing causes that occur with a certain probability. If abnormalities occur, they can be rectified in a short time. However, if the abnormalities are not handled in time, they will accumulate over a period of time and induce implicit or explicit causes.

[0130] Draw the structural model diagram of the timeliness analysis of fracturing construction ( Figure 4 ). To facilitate matrix operations, the target layer and criterion layer are labeled with A and B respectively. The process is as follows:

[0131] According to the comprehensive correlation analysis results, each judgment matrix is ​​constructed and the weight vector W is calculated.

[0132] A B3 B1 B2

[0133]

[0134] Weight vector W = (0.5936, 0.1571, 0.2493); CR = 0.0516

[0135] B1 x1 x4 x6

[0136]

[0137] Weight vector W = (0.6738, 0.2255, 0.1007); CR = 0.0825

[0138] B2 x5 x7

[0139]

[0140] Weight vector W = (0.3333, 0.6667); CR = 0

[0141] B3 x8 x9 x10 x11 x3 x2

[0142]

[0143] Weight vector W = (0.0619, 0.0693, 0.1394, 0.3598, 0.0944, 0.2752); CR = 0.0317.

[0144] The above CR calculation results are all less than 0.1 and pass the one-time test. The weight matrix W of each criterion element of different schemes is multiplied by the relative weight matrix (B1-B3) W of each criterion element to obtain the sample weights ( Figure 5 ).

[0145] pass Figure 5 The result is that the weight ranking of the scheme layer is: x1>x4>x7>x11>x2>x6>x5>x10>x3>x9>x8, which is consistent with Figure 3 There is a certain inconsistency in the ranking results of the comprehensive correlation calculation, but the general conclusions are consistent. Compared with the comprehensive correlation conclusion, the sample weights of the solution layer obtained by the hierarchical analysis calculation are smoother and have smaller fluctuations except for the x1 weight, which indicates that the calculation results are less disturbed by the outside world and the model is more robust. The weight ranking of the criterion layer is: B1>B3>B2, which is in line with the actual engineering. However, in order to avoid the lack of basic data statistics, it cannot fully reflect the defects of the fuzzy system. Here, the functional expansion of unsupervised mathematical operations is carried out to further perform hypothesis analysis and sensitivity prediction.

[0146] Data interaction is performed under the premise of model stability and one-time compliance inspection. Assuming that the weights of B1, B2, and B3 in the criterion layer are the same, the engineering explanation is that different types of clustering reasons occur with the same probability during the fracturing process. The results show that ( Figure 6 ): The weight ranking of the scheme layer is: x1>x7>x11>x5>x2>x4>x10>x6>x3>x9>x8.

[0147] The purpose of sensitivity analysis is to determine the impact of changes in system parameters or surrounding conditions on the current model state. In layman's terms, it is the relationship between changes in a factor in the criterion layer and changes in the factors in the solution layer. The engineering significance is to predict the future change trend of each sample in the solution layer by assuming the weight development of the criterion layer factors under the premise of ensuring the stability of the model. Here, based on the worst engineering practice, it is assumed that the weights of the B1, B2, and B3 categories in the criterion layer all show an increasing trend, and the change trend of each sample in the solution layer is calculated ( Figure 7-9 ).

[0148] According to engineering experience, 0.15 is set as the judgment threshold. Under the premise of model stability, when the maximum value of the observed sample in the future is greater than 0.15, the slope of the development curve exceeds the sensitivity limit, and the sample is judged to be sensitive. The engineering significance is: the sample judged to be sensitive will have an important impact on the timeliness of zipper fracturing construction in the future. Prevention and control are needed in advance. In order to simplify the calculation process, only Figure 7-9The curve of the observed sample data determined to be sensitive is shown in Figure 2. The sensitivity analysis results show that under the B1 assumption, x1, x2, x4, x7, and x11 are determined to be sensitive, and the slope ranking is: x1>x7>x11>x4>x2; under the B2 assumption, x1, x4, x5, and x7 are determined to be sensitive, and the slope ranking is: x7>x1>x5>x4; under the B3 assumption, x1, x2, x4, and x11 are determined to be sensitive. The slope ranking is: x1>x11>x2>x4.

[0149] The fourth step is to use the process capability analysis calculation method, set the average construction efficiency as the target data, and substitute the data in Table 2-7 into the algorithm to perform process capability analysis of the construction company.

[0150] First, the Anderson-Darling method was used to test the normality of all data ( Figure 10-15 ).

[0151] The normality test shows that the empirical cumulative distribution function of the sample data has a high degree of fit with the normal distribution simulation function of the assumed data, and the P value is greater than 0.005. It is determined that all company data obey the normal distribution. There is no need to convert the data in the next step, and capability analysis can be carried out directly.

[0152] Then set USL = 4, LSL = 3. Execute: Statistics → Quality Tools → Capability Analysis → Normal operation, and get the result (such as Figure 16-21 ).

[0153] Finally, the process capability was determined according to the ISO / TS16949 process capability level evaluation standard (Table 1), and the engineering guidance was finally obtained. Combined with the evaluation indicators, the process capability of the six companies in the work area was determined (Table 11)

[0154] Table 11 Process capability determination results of 6 companies in the work area

[0155]

[0156] The invention discloses a multi-model solution method for zipper-type fracturing construction timeliness analysis, and the analysis method performs unsupervised mathematical analysis of construction timeliness for zipper-type fracturing construction. First, the fracturing construction timeliness factors are set: average construction efficiency, fracturing construction efficiency, joint construction efficiency, average key and critical part failure rate, average key and critical part failure resolution efficiency, sand plugging resolution efficiency, conventional process connection time, first-stage perforation timeliness, water supply timeliness, sand supply timeliness, high-pressure parts puncture leakage processing efficiency, and team personnel quality. According to industry standards, the average construction efficiency is selected as the objective function; secondly, the grey correlation analysis is used to obtain the correlation ranking, and the key factors affecting the fracturing construction timeliness are calculated. The absolute correlation, relative correlation, and comprehensive correlation result configurations are compared, and the optimal correlation conclusion is selected for sub-factor clustering to simplify the research system; again, the AHP model is constructed according to the clustering results, and hypothesis analysis and sensitivity analysis are performed. Predict the future development law of zipper-type fracturing construction timeliness; finally, the average construction efficiency is set as the target data, and the process capability analysis of the construction company is performed, and the construction company is evaluated in the scope of quality engineering.

[0157] The engineering significance of this method is that the whole process of fracturing construction is monitored through the conclusions of grey correlation analysis and hierarchical analysis, and the factors affecting fracturing efficiency at different stages are dynamically pre-controlled. Through the conclusions of process capability analysis, the advantages and disadvantages of different companies under the premise of quality requirements are compared to guide bidding and workload allocation. The algorithm of this method is simple and easy to program, and it can ultimately be used for control optimization.

[0158] The contents not described in detail in this specification belong to the prior art known to professional and technical personnel in this field.

Claims

1. A multi-model solution method for timeliness analysis of zipper fracturing construction, characterized in that: The steps include: S1: Data preparation: Collect the average construction efficiency data, timeliness influencing factors data and average construction efficiency data of the best wells of all construction companies in the work area; S2: Analysis of factors affecting timeliness: S21: Taking the timeliness influencing factors as sub-factors and the average construction efficiency as the parent factor, the grey correlation analysis model calculation method is used to calculate the comprehensive correlation of the timeliness influencing factors of zipper fracturing construction with respect to the average construction efficiency, and the key factors of fracturing timeliness are analyzed based on the comprehensive correlation; S22: According to the comprehensive correlation calculation result in step S21, select the grey clustering sub-factors, construct a triangular matrix table to calculate the comprehensive correlation between the grey clustering sub-factors, and perform cluster analysis; S3: Based on the cluster analysis results, a hierarchical analysis model is constructed, and hypothesis analysis and sensitivity analysis are performed to predict the future development law of the zipper fracturing construction time efficiency based on the analysis results; S4: Construction company quality engineering assessment: Use process capability analysis calculation method, set average construction efficiency as target data, use the average construction efficiency data of the optimal well to conduct process capability analysis of the construction company, and conduct construction company assessment in the scope of quality engineering; The calculation method of the comprehensive correlation degree in the above step S2 is: S2101: Calculate the gray absolute correlation degree, set X i =(x i (1),x i (2),…,x i (n)) is the system behavior sequence, let: n represents the number of elements in the sequence; then For X i The zeroing image of the initial point, D is called the zeroing operator; let X i ,X j are all system behavior sequences, then define: Where ε ij represents the absolute correlation; i, j = 0, 1, 2, ... m; m represents the system behavior sequence number; s represents the correlation coefficient; S2102: Calculate the gray relative correlation, set X i =(x i (1),x i (2),…,x i (n)) is the system behavior sequence, let X i (k)c=x i (k) / x i (1); k = 1, 2, ..., n; then X i C=X′ i =(x i (1)c,x i (2)c,…,x i (n)c) is X i The zero image of the starting point, C is the initialization operator; let X i ,X j They are all system behavior sequences, and their initial value images are X′ i ,X′ j ; The zero images of the starting points are calculated as follows: Then define: Where r ij Indicates relative correlation; s' represents the correlation coefficient S2103: Calculate the grey comprehensive correlation degree, let X i ,X j are all system behavior sequences, and ε ij ,r ij are the absolute correlation and relative correlation of the two sequences respectively, then we define: ij =θε ij +(1-θ)r ij , where ρ ij represents the comprehensive correlation, θ is the adjustment coefficient, θ∈[0,1]; In the above step S3, the construction of the hierarchical analysis model includes the following steps: S301: First, identify the problem, build a physical model and collect data; second, build a hierarchical model of the highest layer, i.e., the target layer, the middle layer, i.e., the criterion layer, and the lowest layer, i.e., the solution layer; perform pairwise comparison and scoring based on the engineering evaluation results to determine the score of the lower layer to the upper layer; finally, perform hierarchical synthesis calculation and consistency test, using the formula The formula for calculating the consistency index CI is: max is the maximum eigenvalue of the judgment matrix; n is the order of the judgment matrix; S302: Adopting Calculate the consistency ratio CR. When CR < 0.10, it means that it has passed the one-time test. Otherwise, it does not have satisfactory consistency. In the formula, CI is the consistency index; RI is the random consistency index; S303: Calculate the weight vector, use the geometric mean method, take the square root of the continuous product of each variable value, and use the geometric mean to calculate the development speed of the predicted target: Where W represents the development speed; i = 1, 2, 3..., n; a ij is the element in the i-th row and j-th column of the judgment matrix.

2. A multi-model solution method for time-effectiveness analysis of zipper fracturing construction according to claim 1, characterized in that: The factors affecting the timeliness of zipper-type fracturing construction include: fracturing construction efficiency, joint construction efficiency, average failure rate of key and critical parts, average efficiency of solving key and critical faults, efficiency of solving sand blockage, connection time of conventional processes, timeliness of first-stage perforation, timeliness of water supply, timeliness of sand supply, efficiency of handling puncture leakage of high-pressure parts, and average skill score of team members.

3. A multi-model solution method for time-effectiveness analysis of zipper fracturing construction according to claim 1, characterized in that: In the above step S2103, the parameter θ is used to adjust the influence of the absolute correlation and the relative correlation on the comprehensive correlation, θ=0.

5.

4. A multi-model solution method for time-effectiveness analysis of zipper fracturing construction according to claim 1, characterized in that: In the above step S2, when the correlation degree in the calculation result is greater than 0.5, it is determined to be the main influencing factor; if the correlation degree of all factors is greater than 0.5, the threshold is set to 0.85 and used as the basis for determining gray clustering.

5. A multi-model solution method for time-effectiveness analysis of zipper fracturing construction according to claim 1, characterized in that: In the above step S4, the process capability index C is used pk Conduct process capability determination: Where: C pk : Process capability index without offset; USL: upper limit of product quality requirement; LSL: lower limit of product quality requirement; σ: standard deviation of normal distribution of process characteristics; μ: center of distribution of process quality characteristics; C pu : Process capability index with only upper limit requirement; C pl : Process capability index with only lower limit requirements.

6. A multi-model solution method for time-effectiveness analysis of zipper fracturing construction according to claim 5, characterized in that: According to the process capability index C pk , the process capability of the construction company is determined according to the ISO / TS16949 process capability level evaluation standard, and finally the engineering guidelines are obtained.

Images