Drilling pipe trip flow and loss identification method and system based on unsupervised learning

Abstract

The present application relates to the technical field of drilling engineering, and particularly relates to a drilling tripping overflow and loss identification method and system based on unsupervised learning, which comprises the following steps: automatically identifying the drilling state of drilling based on real-time collected comprehensive logging data; dynamically dividing a plurality of tripping calculation windows; obtaining the tripping tank drilling fluid volume change amount and outlet flow integral in each tripping calculation window; obtaining the open drilling fluid volume corresponding to each common drill pipe column under the open discharge condition and the closed discharge condition, respectively; determining whether loss or overflow occurs under the drilling state based on the open drilling fluid volume corresponding to each common drill pipe column, the tripping tank drilling fluid volume change amount in each tripping calculation window and the outlet flow integral. The problem that it is difficult to obtain the time point of overflow and loss identification and judgment in a unified manner due to different grouting methods and time intervals during tripping is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of drilling engineering, and particularly relates to a drilling tripping overflow and loss circulation identification method and system based on unsupervised learning. BACKGROUND

[0002] Drilling is an important means of exploring and developing oil and gas. In the drilling operation, drilling complex conditions and accidents threaten the entire process of drilling from beginning to end, seriously affecting drilling speed, well construction quality and exploration and development efficiency. Among them, well leakage and overflow are the two most common downhole complex accidents affecting drilling safety. Well leakage and overflow not only cause serious reservoir damage, increase exploration and development costs, and cause low oil and gas development efficiency, but also, if not controlled, can induce major accidents such as sticking, well collapse, blowout, etc., causing losses and negative social impact. Therefore, real-time identification and judgment of overflow and loss during drilling is of great significance.

[0003] At present, more researches have been carried out on overflow and loss monitoring, and the formed technologies mainly include wellhead monitoring technology, downhole drilling monitoring technology and artificial intelligence monitoring technology, specifically:

[0004] 1) The wellhead monitoring technology mainly identifies the occurrence of overflow and loss by monitoring the changes of ground measurement parameters (mud tank volume, outlet flow, standpipe pressure, etc.); this method has low cost and mature technology, but it depends on human experience to a great extent.

[0005] 2) The downhole drilling monitoring technology can quickly and early monitor overflow and loss by real-time measurement of downhole temperature, pressure and other parameters, but the cost is high, and the drilling measurement instrument has the risk of failure.

[0006] 3) Artificial intelligence monitoring technology: with the rapid development of information technology and artificial intelligence technology, domestic and foreign researchers have applied artificial intelligence technology to the monitoring of overflow and loss, but these artificial intelligence methods generally have the problems of complex modeling and difficult popularization.

[0007] For overflow and loss identification during tripping, tripping operation will cause changes in the volume of drilling cuttings in the wellbore, thereby causing changes in the volume of drilling fluid in the tripping tank. The existing technology usually needs to input the drilling tool size information in combination with the real-time measured changes in the volume of drilling fluid in the tripping tank, and apply artificial judgment or mathematical methods to identify the occurrence of overflow and loss; since human input information is needed, there is inevitably a lag and subjectivity in the information, and it is difficult to realize automation of the entire process. Among them, the existing technology is as follows:

[0008] 1) The invention with application number CN201811625006.2 discloses a method for overflow loss early warning trend analysis during drilling tripping operation and during drilling running operation. Through the invention, a new method for analyzing overflow loss trend according to mud change amount is provided, that is, on the one hand, by comprehensively reviewing the relevant influencing factors of drilling construction process, on the other hand, by applying advanced computer technology and intelligent algorithm, and by using original fuzzy mathematical processing method and trend analysis algorithm, the mud overflow loss trend during drilling tripping can be quickly identified after the grouting amount or the liquid discharge amount of each column of drill pipe is newly obtained, the original cognitive limitations are broken, the problems of complex, slow and untimely calculation existing in current overflow monitoring and early warning are solved, so as to fill the industry gap, and provide a new monitoring method and means for existing drilling construction safety, and reach the international leading level.

[0009] 2) The invention with application number CN111749633B provides a continuous tripping overflow and loss monitoring method, which comprises the following steps: acquiring the real-time out / in well volume of the drill string; adjusting the real-time discharge amount of the drilling fluid into the well according to the real-time out / in well volume of the drill string under the condition of acquiring the real-time out / in well volume of the drill string; judging whether overflow and loss occur once every predetermined time, and calculating the overflow and loss amount in the tripping process within the predetermined time in the case of overflow and loss, and judging the overflow and loss level in the tripping process; issuing a graded alarm according to the overflow and loss level of the tripping process; and repeating the above process. The present invention has the advantages of realizing continuous tripping drilling overflow and loss automatic monitoring, improving overflow and loss judgment accuracy, etc.

[0010] In summary, the following technical problems exist at present:

[0011] Due to the different grouting methods and time intervals during tripping, it is difficult to obtain the time point of overflow and loss identification and judgment in a unified way. SUMMARY

[0012] The technical problem to be solved by the present invention is to provide a drilling tripping overflow and loss identification method and system based on unsupervised learning to solve the problems of the prior art.

[0013] The technical scheme of the drilling tripping overflow and loss identification method based on unsupervised learning of the present invention is as follows:

[0014] Based on the real-time acquisition of comprehensive logging data, the drilling state of the drilling is automatically identified;

[0015] A plurality of tripping calculation windows are dynamically divided;

[0016] The volume change amount of the drilling fluid in the tripping tank and the outlet flow integral in each tripping calculation window are obtained;

[0017] According to the specification parameters of any common drill rod, the open row and closed row conditions are calculated respectively, until the open row drilling fluid volume corresponding to the column composed of each common drill rod is obtained.

[0018] Based on the open row drilling fluid volume corresponding to the column composed of each common drill rod, the drilling fluid volume change amount in the tripping tank and the outlet flow integral in each tripping calculation window, it is determined whether loss or overflow occurs in the drilling state.

[0019] The technical scheme of the drilling tripping overflow and loss identification system based on unsupervised learning is as follows:

[0020] The automatic identification module, the dynamic division module, the acquisition module, the calculation module and the determination module are included.

[0021] The automatic identification module is used to automatically identify the drilling state of drilling based on the real-time collected comprehensive logging data.

[0022] The dynamic division module is used to dynamically divide a plurality of tripping calculation windows.

[0023] The acquisition module is used to acquire the drilling fluid volume change amount in the tripping tank and the outlet flow integral in each tripping calculation window.

[0024] The calculation module is used to calculate the open row drilling fluid volume corresponding to the column composed of the common drill rod according to the specification parameters of any common drill rod, respectively in the open row and closed row conditions, until the open row drilling fluid volume corresponding to the column composed of each common drill rod is obtained.

[0025] The determination module is used to determine whether loss or overflow occurs in the drilling state based on the open row drilling fluid volume corresponding to the column composed of each common drill rod, the drilling fluid volume change amount in the tripping tank and the outlet flow integral in each tripping calculation window.

[0026] The storage medium of the application, the storage medium stores instructions, when the computer reads the instructions, the computer executes the above-mentioned any one based on the unsupervised learning of the drilling tripping overflow and loss identification method.

[0027] The electronic device of the application includes a processor and the above-mentioned storage medium, and the processor executes the instructions in the storage medium.

[0028] The beneficial effects of the application are as follows:

[0029] Based on the real-time collected comprehensive logging data, the drilling state of the drilling is automatically identified, and a plurality of tripping calculation windows are dynamically divided, so that the problem that the time points for overflow and loss identification and judgment are difficult to obtain in a unified manner due to different grouting modes and time intervals during tripping is solved. BRIEF DESCRIPTION OF DRAWINGS

[0030] Other features, objects, and advantages of the application will become more apparent from the following detailed description of non-limiting embodiments, made with reference to the accompanying drawings:

[0031] Figure 1 A flowchart of a drilling tripping overflow and loss identification method based on unsupervised learning according to an embodiment of the application;

[0032] Figure 2 Comprehensive logging data is collected;

[0033] Figure 3 An identification result diagram for identifying the drilling state of the drilling;

[0034] Figure 4 The tripping calculation window is tripped into the drilling fluid volume change situation;

[0035] Figure 5 The diagram shows the drilling fluid volume of each column of different specifications of drill pipe vs. the actual historical well drilling fluid volume of each column;

[0036] Figure 6 A structure diagram of a drilling tripping overflow and loss identification system based on unsupervised learning according to an embodiment of the application. DETAILED DESCRIPTION

[0037] As shown in Figure 1 , a drilling tripping overflow and loss identification method based on unsupervised learning according to an embodiment of the application includes the following steps:

[0038] S1, based on the real-time collected comprehensive logging data, the drilling state of the drilling is automatically identified;

[0039] Among them, the real-time collected comprehensive logging data is a multi-dimensional drilling time series data indexed by time, including drill bit depth sequence, well depth sequence, hook load sequence, inlet flow sequence, outlet flow sequence, tripping tank drilling fluid volume sequence, etc. Specifically:

[0040] 1) Time sequence is denoted as T, T=T1,T2,T3,…,T n ;

[0041] 2) The drill bit depth sequence is denoted as BDEP, BDEP=BDEP1,BDEP2,BDEP3,…,BDEP n, BDEP1 represents the drill bit depth at T1, BDEP2 represents the drill bit depth at T2, …, BDEP n represents the drill bit depth at T n , unit: m;

[0042] 3) The well depth sequence is denoted as HDEP, HDEP = HDEP1, HDEP2, HDEP3, …, HDEP n , HDEP1 represents the well depth at T1, HDEP2 represents the well depth at T2, …, HDEP n represents the well depth at T n , unit: m;

[0043] 4) The hook load sequence is HKLD, HKLD = HKLD1, HKLD2, HKLD3, …, HKLD n , HKLD1 represents the hook load at T1, HKLD2 represents the hook load at T2, …, HKLD n represents the hook load at T n , unit: kN;

[0044] Wherein, the drilling state includes a tripping state and a drilling state, the tripping state includes at least one of a tripping operation and a pipe-off operation, the drilling state includes at least one of a drilling operation and a pipe-on operation; the process of automatically identifying the drilling state of the drilling is as follows:

[0045] 1) If BDEP i <BDEP i-1 &&(HDEP i -BDEP i >TH trip ||DS i-1 is the tripping operation ||DS i-1 is the pipe-off operation), then DS i is the tripping operation, since the tripping state includes at least one of the tripping operation and the pipe-off operation, it is determined that the drilling state at T i is the tripping state, BDEP i represents the drill bit depth at T i , i = 1, 2, 3…n, BDEP i-1 represents the drill bit depth at T i-1 , HDEP i represents the well depth at T i , TH trip represents the tripping depth threshold, which is set by human according to the actual situation, DS i-1 represents the drilling operation at T i-1 .

[0046] 2) If BDEP i=BDEP i-1 &&(HDEP i -BDEP i >TH trip ||DS i-1 is a running-in operation||DS i-1 is a running-out operation)&&HKLD i <TH HKLDmin , DS i is a running-out operation, since the running-in state includes at least one of the running-in operation and the running-out operation, it is determined that the drilling state at T i is the running-in state, HKLD i represents the hook load at T i , TH HKLDmin represents the hook load threshold value at the running-in and running-out, in units of KN, in the present application, TH HKLDmin is 20 KN, and can also be set according to actual conditions.

[0047] 3) If BDEP i >BDEP i-1 &&(HDEP i -BDEP i >TH trip ||DS i-1 is a running-in operation||DS i-1 is a running-in operation, DS i is a running-in operation, since the running-in state includes at least one of the running-in operation and the running-out operation, it is determined that the drilling state at T i is the running-in state.

[0048] 4) If BDEP i =BDEP i-1 &&(HDEP i -BDEP i >TH trip ||DS i-1 is a running-in operation||DS i-1 is a running-in operation)&&HKLD<TH HKLDmin , DS i is a running-in operation, since the running-in state includes at least one of the running-in operation and the running-out operation, it is determined that the drilling state at T i is the running-in state.

[0049] S2, dynamically dividing a plurality of running-in and running-out calculation windows:

[0050] 1) In the running-in state, the implementation process of dynamically dividing a plurality of running-in and running-out calculation windows is as follows:

[0051] ① When the grouting method is continuous grouting, the drill bit position change is calculated based on the real-time collected comprehensive logging data. Multiple tripping and running calculation windows are divided with the drill bit position change as the dividing standard of one column length.

[0052] ② When the grouting method is intermittent grouting, the time period between two adjacent grouting times is divided into a tripping calculation window based on the real-time collected comprehensive logging data.

[0053] The process of obtaining the pouring time is as follows:

[0054] Based on real-time collected comprehensive logging data, the volume of drilling fluid in the tripping and tripping hopper is tracked in real time. When the volume of drilling fluid in the tripping and tripping hopper drops rapidly at a certain moment, and the rate of change exceeds a preset rate of change threshold, and the change in drill bit position exceeds the length of a single drill pipe, that moment is determined as the mud-pouring moment. The preset rate of change threshold can be set according to actual conditions.

[0055] 2) The process of dynamically dividing multiple tripping and tripping calculation windows during the drilling process is as follows:

[0056] During the drilling process, the drill bit position change is calculated based on the real-time collected comprehensive logging data. Multiple tripping and tripping calculation windows are divided with the drill bit position change as the dividing standard of one column length.

[0057] S3. Obtain the volume change of drilling fluid in the tripping tank and the integral of the outlet flow rate within each tripping calculation window.

[0058] TOW calculation window based on the nth drilling start. n Taking the data within TOW as an example, n The calculation process for the volume change of drilling fluid in the tripping and tripping tank and the integral of the outlet flow rate is explained:

[0059] 1) Calculate TOW using the first formula n TripPVVar, the volume change of drilling fluid inside the tripping and pulling container. n The first formula is: TripPVVar n =E(TripPV′)-E(TripPV″).

[0060] in, TOW n Starting from the first moment in the tripping and dredging process, the drilling fluid volume in the tripping and dredging container is taken as the starting point. The drilling fluid volumes in the tripping and dredging container are then sequentially taken at m moments, resulting in TripPV′1, TripPV′2, ..., TripPV′. m, wherein TripPV'1 represents the volume of the drilling fluid in the tripping tank at the first time point, TripPV'2 represents the volume of the drilling fluid in the tripping tank at the mth time point, and TripPV' represents the volume of the drilling fluid in the tripping tank at the last time point within TOW m , wherein 2m is less than TOW n , and m is an integer greater than 1.

[0061] , wherein the total number of all data points of the volume of the drilling fluid in the tripping tank is represented by N(TripPV'), and the first expected value is represented by E(TripPV').

[0061] , wherein the total number of all data points of the volume of the drilling fluid in the tripping tank is represented by N(TripPV'), and the first expected value is represented by E(TripPV'). n , wherein the volume of the drilling fluid in the tripping tank at the last time point within TOW m is represented by TripPV"2, the volume of the drilling fluid in the tripping tank at the (m-1)th time point before the last time point within TOW m is represented by TripPV"1, and the volume of the drilling fluid in the tripping tank at the mth time point before the last time point within TOW n is represented by TripPV". m , wherein the volume of the drilling fluid in the tripping tank at the last time point within TOW m is represented by TripPV"2, the volume of the drilling fluid in the tripping tank at the (m-1)th time point before the last time point within TOW n is represented by TripPV"1, and the second expected value is represented by E(TripPV").

[0062] 2) the second formula is used to calculate the outlet flow integral FLOVarArea n within TOW n结束 , and the second formula is as follows:

[0063]

[0064] , wherein Index(TOW n ) represents the index of the end time of TOW n起始 in the comprehensive logging data, Index(TOW n ) represents the index of the start time of TOW p , FLO p represents the outlet flow value at the time point, and FLO base represents the outlet flow base value, which is calculated by selecting the outlet flow value at the pump stopping and stable period.

[0065] S4, according to the specification parameters of any commonly used drill pipe, the volume of the drilling fluid corresponding to the stand composed of the commonly used drill pipe is calculated under the open and closed conditions, respectively, until the volume of the drilling fluid corresponding to the stand composed of each commonly used drill pipe is obtained, and the specific calculation process is as follows:

[0066] 1) under the open condition, the formula V total-out=π×(Tool.OD 2 -Tool.ID 2 )×Tool.Length, calculates the volume of drilling fluid displaced (V) corresponding to any commonly used drill pipe column. total-out Tool.OD represents the outer diameter of the commonly used drill pipe in meters (m), Tool.ID represents the inner diameter of the commonly used drill pipe in meters (m), and Tool.Length represents the single length of the commonly used drill pipe in meters (m).

[0067] 2) In the case of a closed arrangement, using the formula: V total-out =π×(Tool.OD 2 -Tool.ID 2 )×Tool.Length, calculates the volume of drilling fluid displaced (V) corresponding to any commonly used drill pipe column. total-out .

[0068] Each set of three commonly used drill pipes forms a column for that set of drill pipes. Commonly used drill pipes include 5" drill pipes, 2 3 / 8" drill pipes, 4" drill pipes, and 3 1 / 2" drill pipes.

[0069] S5. Based on the volume of drilling fluid displaced by each commonly used drill pipe column, the change in drilling fluid volume in the tripping tank within each tripping calculation window, and the integral of the outlet flow rate, determine whether leakage or overflow occurs during drilling. Specifically:

[0070] 1) When the drilling status is in the running-in phase, the process of determining whether leakage or overflow has occurred includes:

[0071] S50. Using the volume of drilling fluid displaced by each commonly used drill pipe column as the initial centroid heuristic value, and based on historical integrated logging data, calculate the change in drilling fluid volume in the tripping container and the integral of the outlet flow rate when each column is lowered, and use the K-means algorithm to perform clustering to obtain the centroid and the first maximum radius of each first cluster.

[0072] S51. Using the centroid of each first cluster as the initial centroid, calculate the change in drilling fluid volume in the tripping tank within each tripping calculation window based on the real-time collected comprehensive logging data, and convert it into the change in drilling fluid volume in the tripping tank of each stand. Calculate the outlet flow rate integral within each tripping calculation window and convert it into the outlet flow rate integral of each stand.

[0073] S52. Calculate the initial distance between the change in drilling fluid volume in the drilling tank and the product of the outlet flow rate of each column and the centroid of each first cluster. If the maximum distance is greater than the first maximum radius, it is determined that an abnormality has occurred in the drilling. Combined with the drilling status, determine whether leakage or overflow has occurred.

[0074] 2) when the drilling state is a tripping-in state, the process of determining whether the leakage or overflow occurs comprises:

[0075] S53, taking the drilling fluid volume of each common drill pipe column as the initial centroid heuristic value, calculating the tripping-in tank drilling fluid volume change amount when each column is tripped out according to historical comprehensive logging data, and using the K-means algorithm for clustering to obtain the centroid and the second maximum radius of each second cluster;

[0076] It should be noted that the unsupervised learning mentioned in the present application is the K-means algorithm, which can directly cluster data without labeling the data, and the present application uses the characteristics of each cluster after clustering to identify overflow and leakage.

[0077] S54, taking the centroid of each second cluster as the initial centroid, calculating the tripping-in tank drilling fluid volume change amount in each tripping-in calculation window according to the real-time collected comprehensive logging data, and converting it into the tripping-in tank drilling fluid volume change amount of each column;

[0078] S55, calculating the initial distance between the tripping-in tank drilling fluid volume change amount of each column and the centroid of each second cluster, and if the maximum distance is greater than the second maximum radius, it is determined that the drilling is abnormal, and whether the leakage or overflow occurs is determined in combination with the drilling state.

[0079] The present application still has the following technical problems:

[0080] The drilling tool information of the tripped-in or tripped-out wellbore needs to be manually input, and the tripping-in overflow and leakage identification process is difficult to realize full automation. In the present application, the threshold value is automatically extracted for overflow and leakage judgment in the normal drilling by clustering analysis of the historical tripping-in data, avoiding manual input and realizing the automation of the identification process.

[0081] Optionally, in the above technical solution, when it is determined that the drilling is abnormal, and whether the leakage or overflow occurs is determined in combination with the drilling state, comprising:

[0082] 1) when the drilling state is a tripping-in state, determining whether the current tripping-in tank drilling fluid volume change amount is less than the expected value, if yes, it is determined that the leakage occurs, and if no, it is determined that the overflow occurs;

[0083] When the drilling state is a tripping-in state, the current tripping-in tank drilling fluid volume change amount is an increment.

[0084] 2) when the drilling state is a tripping-in state, determining whether the current tripping-in tank drilling fluid volume change amount is less than the expected value, if yes, it is determined that the leakage occurs, and if no, it is determined that the overflow occurs.

[0085] Wherein, when the drilling state is a tripping state, the current tripping tank drilling fluid volume change amount is a decrease amount.

[0086] Compared with the prior art, the present application has the following beneficial effects:

[0087] 1) The present application adopts automatic identification of tripping operation process and dynamic division of overflow and loss identification window, realizing automation of overflow and loss identification process of tripping operation.

[0088] 2) Based on the optimized K-means method, the change rule and useful information of tripping tank drilling fluid volume and outlet flow during historical well tripping operation are mined and extracted, and combined with comprehensive logging data of normal drilling, overflow and loss during tripping are identified in real time, effectively avoiding the tediousness and hysteresis of manual input of drilling tool information, and improving the intelligent degree of overflow and loss identification.

[0089] The beneficial effects of the present application will be described below through an embodiment:

[0090] A certain well in a certain oilfield drilled to a well depth of 3159m, and the circulation was short to a well depth of 2792m for static observation, and it was found that the outlet was continuously flowing, and the well was immediately closed for observation, and the casing pressure rose to determine that overflow occurred. The present application was used to timely and accurately identify the overflow. The following is a detailed description of the identification process, including:

[0091] S101, identifying the drilling state of the well:

[0092] The identification start time is August 13, 16:48, and the end time is August 13, 18:43; the real-time collected comprehensive logging data required for automatic identification of the drilling state are read: including drill bit depth sequence, well depth sequence, hook load sequence, etc.; multi-dimensional data curve display is as shown in Figure 2 .

[0093] Based on multi-dimensional time series data correlation analysis, the drilling state is identified, and the identification result is as shown in Figure 3 . Due to the length of the article, only the identification results of part of the time period are displayed.

[0094] S102, dynamically dividing a plurality of tripping calculation windows:

[0095] A total of 12 columns were tripped between 16:48 on August 13 and 18:43 on August 13; see the above for the specific division process, and the tripping calculation window for the division includes: [2022 / 8 / 1317:11:09, 2022 / 8 / 1317:16:35], [2022 / 8 / 1317:16:40, 2022 / 8 / 1317:21:00], [2022 / 8 / 1317:21:05, 2022 / 8 / 1317:25:19], [2022 / 8 / 1317:25:24, 2022 / 8 / 1317:29:19], [2022 / 8 / 1317:29:24, 2022 / 8 / 1317:33:50], [2022 / 8 / 1317:33:55, 2022 / 8 / 1317:38:26], [2022 / 8 / 1317:38:36, 2022 / 8 / 1317:43:19], [2022 / 8 / 1317:43:24, 2022 / 8 / 1317:47:55], [2022 / 8 / 1317:48:00, 2022 / 8 / 1317:51:51], [2022 / 8 / 1317:51:56, 2022 / 8 / 1317:55:44], [2022 / 8 / 1317:55:54, 2022 / 8 / 1317:59:45], [2022 / 8 / 1317:59:50, 2022 / 8 / 1318:06:31], as shown in Figure 4

[0096] From the tripping tank volume change in the tripping calculation window, it can be seen that 3 times of back slurry occurred in the tripping calculation window, and the combined window according to the back slurry interval time is as follows: [2022 / 8 / 1317:11:09, 2022 / 8 / 1317:29:19], [2022 / 8 / 1317:29:24, 2022 / 8 / 1317:47:55], [2022 / 8 / 1317:48:00, 2022 / 8 / 1318:06:31].

[0097] S103, obtaining the drilling fluid volume change in the tripping tank in each tripping calculation window;

[0098] The drilling fluid volume change in the tripping tank in the three tripping calculation windows is shown in Table 1.

[0099] Table 1:

[0100]

[0101] ​S104. Iterate through the commonly used drill pipe specifications and calculate the volume of displaced drilling fluid corresponding to a single drill pipe and a column under both open and closed conditions. In other words, based on the specification parameters of any commonly used drill pipe, calculate the volume of displaced drilling fluid corresponding to the column composed of that commonly used drill pipe under both open and closed conditions, until the volume of displaced drilling fluid corresponding to the column composed of each commonly used drill pipe is obtained.

[0102] Based on the information on the inner and outer diameters and lengths of commonly used drill pipes, and according to the formulas mentioned above, the volume of drilling fluid displaced by drill pipe columns of different specifications is calculated. Some of the calculation results are shown in Table 2 below.

[0103] Table 2:

[0104] Common Drill Pipe Names 1 Column Volume (m3) - Open Stand 2 3 / 8" Drill Pipe 0.04 3 1 / 2" Drill Pipe 0.08 4" Heavy Weight Drill Pipe 0.16 4 1 / 2" Heavy Weight Drill Pipe 0.17 5" Drill Pipe 0.12 5" Heavy Weight Drill Pipe 0.27 4 3 / 4" Drill Collar 0.25

[0105] S105. Overflow identification during drilling;

[0106] This step uses an improved K-means algorithm to identify overflows during tripping in. The specific identification process is as follows:

[0107] 1) Based on the changes in drilling fluid volume in the tripping and tripping hopper during each grouting operation, and taking the volume of drilling fluid displaced by the commonly used drill pipe column as the initial centroid, calculate the distance between the actual changes in drilling fluid volume in the tripping and tripping hopper and the centroid (i.e., the sum of squared errors). The results are shown in the table below. Based on the calculation results, select the centroid with the smallest distance sum, i.e., the column volume of the 5" drill pipe, as shown in Table 3.

[0108] Table 3:

[0109]

[0110]

[0111] 2) Obtain the column volume of the 5" drill pipe in historical wells, such as... Figure 5 As shown; calculate the volume of drilling fluid dispersed in a single drill pipe relative to its center of mass (0.12m). 3 The average distance (sum of squared errors / number of pillars erected) was calculated, and outliers were removed according to the rule that the average distance is greater than 3 times the average distance. The average distance was then recalculated, and the result was 0.001424, as shown in Table 4.

[0112] Table 4:

[0113]

[0114]

[0115] 3) The average distance calculated is 0.001424 m3, which is the threshold value. Combined with the distance between the grouting volume of each stand and the centroid calculated in step a), it can be identified that overflow occurs when the third stand is started (the distance value is 0.001444 m3> threshold value 0.001424 m3) when the third stand is started. That is, the occurrence of overflow is identified at 2022 / 8 / 13 18:06:31.

[0116] In the above embodiments, although the steps are numbered S1, S2, etc., it is only a specific embodiment given by the present application, and those skilled in the art can adjust the execution order of S1, S2, etc. according to the actual situation, which is within the protection scope of the present application. It can be understood that in some embodiments, some or all of the above embodiments can be included.

[0117] As shown in Figure 6 The drilling overflow loss identification system 200 based on unsupervised learning of the embodiment of the present application includes an automatic identification module 210, a dynamic division module 220, an acquisition module 230, a calculation module 240, and a determination module 250.

[0118] The automatic identification module 210 is configured to automatically identify the drilling state of the drilling based on the real-time collected comprehensive logging data.

[0119] The dynamic division module 220 is configured to dynamically divide a plurality of tripping calculation windows.

[0120] The acquisition module 230 is configured to acquire the tripping tank drilling fluid volume change amount and the outlet flow integral in each tripping calculation window.

[0121] The calculation module 240 is configured to calculate the open-drilling fluid volume corresponding to the stand composed of any commonly used drill pipe under the open-drilling condition and the closed-drilling condition, respectively, until the open-drilling fluid volume corresponding to the stand composed of each commonly used drill pipe is obtained.

[0122] The determination module 250 is configured to determine whether loss or overflow occurs in the drilling state based on the open-drilling fluid volume corresponding to the stand composed of each commonly used drill pipe, the tripping tank drilling fluid volume change amount in each tripping calculation window, and the outlet flow integral.

[0123] Optionally, in the above technical solution, the drilling state is a drilling-in state or a tripping-out state, and the determination module 250 is specifically configured to:

[0124] When the drilling state is the drilling-in state, the process of determining whether loss or overflow occurs includes:

[0125] The volume of the drilling fluid displaced by each commonly used drill rod is taken as an initial centroid heuristic value, and the volume of the drilling fluid in the tripping tank and the outlet flow integral are calculated according to historical comprehensive logging data when each column is lowered, and the K-means algorithm is used for clustering to obtain the centroid and the first maximum radius of each first cluster;

[0126] The centroid of each first cluster is taken as an initial centroid, and the volume of the drilling fluid in the tripping tank is calculated according to real-time comprehensive logging data in each tripping calculation window, and is converted into the volume of the drilling fluid in the tripping tank of each column, and the outlet flow integral is calculated in each tripping calculation window and is converted into the outlet flow integral of each column;

[0127] The initial distance between the volume of the drilling fluid in the tripping tank of each column and the centroid of each first cluster is calculated, and if the maximum distance is greater than the first maximum radius, it is determined that the drilling is abnormal, and whether loss or overflow occurs is determined in combination with the drilling state;

[0128] When the drilling state is tripping, the process of determining whether loss or overflow occurs includes:

[0129] The volume of the drilling fluid displaced by each commonly used drill rod is taken as an initial centroid heuristic value, and the volume of the drilling fluid in the tripping tank is calculated according to historical comprehensive logging data when each column is lowered, and the K-means algorithm is used for clustering to obtain the centroid and the second maximum radius of each second cluster;

[0130] The centroid of each second cluster is taken as an initial centroid, and the volume of the drilling fluid in the tripping tank is calculated according to real-time comprehensive logging data in each tripping calculation window, and is converted into the volume of the drilling fluid in the tripping tank of each column;

[0131] The initial distance between the volume of the drilling fluid in the tripping tank of each column and the centroid of each second cluster is calculated, and if the maximum distance is greater than the second maximum radius, it is determined that the drilling is abnormal, and whether loss or overflow occurs is determined in combination with the drilling state.

[0132] Optionally, in the above technical solution, the determination module 250 is further specifically used for:

[0133] When it is determined that the drilling is abnormal, whether loss or overflow occurs is determined in combination with the drilling state, including:

[0134] When the drilling state is tripping, it is determined whether the volume of the drilling fluid in the tripping tank is less than an expected value, and if yes, it is determined that loss occurs, and if no, it is determined that overflow occurs;

[0135] When the drilling state is a tripping state, it is judged whether the current tripping tank drilling fluid volume change amount is less than an expected value, if yes, overflow is determined to occur, and if no, loss is determined to occur.

[0136] Optionally, in the technical solution, the real-time comprehensive logging data is time-indexed multi-dimensional drilling time series data, and specifically includes a drill bit depth sequence, a well depth sequence, and a hook load sequence, an inlet flow sequence, an outlet flow sequence, and a tripping tank drilling fluid volume sequence.

[0137] The steps of implementing the respective functions of the parameters and the unit modules in the drilling tripping overflow and loss identification system 200 based on unsupervised learning can refer to the parameters and steps in the embodiments of the drilling tripping overflow and loss identification method based on unsupervised learning, and will not be repeated here.

[0138] The storage medium of the embodiment of the present application stores instructions, and when the computer reads the instructions, the computer executes any one of the drilling tripping overflow and loss identification methods based on unsupervised learning.

[0139] The electronic device of the embodiment of the present application includes a processor and the above-mentioned storage medium, and the processor executes the instructions in the storage medium, wherein the electronic device can be a computer, a mobile phone, etc.

[0140] Those skilled in the art know that the present application can be implemented as a system, a method or a computer program product.

[0141] Therefore, the present disclosure can be embodied in the form of a complete hardware, a complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software, which is generally referred to as "circuitry", "module" or "system" herein. In addition, in some embodiments, the present application can also be implemented in the form of a computer program product in one or more computer readable media, which includes computer readable program codes.

[0142] Any combination of one or more computer readable medium can be utilized. The computer readable medium can be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium include: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium can be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

[0143] Although the embodiments of the present application have been shown and described above, it should be understood by those skilled in the art that the above embodiments are exemplary, and cannot be construed as limiting the present application, and those skilled in the art can make changes, modifications, replacements, and variations to the above embodiments within the scope of the present application.

Claims

1. A method for identifying lost circulation during drilling tripping based on unsupervised learning, the method comprising: The method comprises the following steps: Based on real-time comprehensive logging data, automatically identify the drilling state of the well; Dynamic division of multiple tripping calculation windows; Obtain the tripping tank drilling fluid volume change and outlet flow integral in each tripping calculation window; Calculate the TOW using the first formula n The volume of the drilling fluid in the trip-out tank at the first time point n The first formula is: TripPVVar n = E(TripPV ′ ) - E(TripPV″), wherein, The volume of the drilling fluid in the trip-out tank at the first time point n The volume of the drilling fluid in the trip-out tank at the first time point m The volume of the drilling fluid in the trip-out tank at the first time point ′ The volume of the drilling fluid in the trip-out tank at the first time point ′ The volume of the drilling fluid in the trip-out tank at the first time point m The volume of the drilling fluid in the trip-out tank at the first time point ′ The volume of the drilling fluid in the trip-out tank at the first time point n The volume of the drilling fluid in the trip-out tank at the first time point ′ The volume of the drilling fluid in the trip-out tank at the first time point End of trip tank volume at last time in TOW n End of trip tank volume at last time in TOW m End of trip tank volume at last time in TOW m End of trip tank volume at last time in TOW n End of trip tank volume at last time in TOW m End of trip tank volume at last time in TOW m End of trip tank volume at last time in TOW According to the specification parameters of any commonly used drill pipe, calculate the open drilling fluid volume corresponding to the stand composed of the commonly used drill pipe under open and closed conditions until the open drilling fluid volume corresponding to the stand composed of each commonly used drill pipe is obtained; Based on the open drilling fluid volume corresponding to the stand composed of each commonly used drill pipe, the tripping tank drilling fluid volume change and outlet flow integral in each tripping calculation window, determine whether loss or overflow occurs under the drilling state; When the drilling state is tripping in, the process of determining whether loss or overflow occurs includes: Taking the open drilling fluid volume corresponding to the stand composed of each commonly used drill pipe as the initial centroid heuristic value, calculating the tripping tank drilling fluid volume change and outlet flow integral when each stand is tripped in according to historical comprehensive logging data, and using K-means algorithm for clustering to obtain the centroid and first maximum radius of each first cluster; Taking the centroid of each first cluster as the initial centroid, calculating the tripping tank drilling fluid volume change in each tripping calculation window according to real-time comprehensive logging data, and converting it into the tripping tank drilling fluid volume change of each stand, calculating the outlet flow integral in each tripping calculation window, and converting it into the outlet flow integral of each stand; Calculating the initial distance between the tripping tank drilling fluid volume change and outlet flow integral of each stand and the centroid of each first cluster, if the maximum distance is greater than the first maximum radius, it is determined that the drilling is abnormal, and combined with the drilling state, it is determined whether loss or overflow occurs.

2. The method of claim 1, wherein, When the drilling state is tripping out, the process of determining whether loss or overflow occurs includes: Taking the open drilling fluid volume corresponding to the stand composed of each commonly used drill pipe as the initial centroid heuristic value, calculating the tripping tank drilling fluid volume change when each stand is tripped out according to historical comprehensive logging data, and using K-means algorithm for clustering to obtain the centroid and second maximum radius of each second cluster; Taking the centroid of each second cluster as the initial centroid, calculating the tripping tank drilling fluid volume change in each tripping calculation window according to real-time comprehensive logging data, and converting it into the tripping tank drilling fluid volume change of each stand; Calculating the initial distance between the tripping tank drilling fluid volume change of each stand and the centroid of each second cluster, if the maximum distance is greater than the second maximum radius, it is determined that the drilling is abnormal, and combined with the drilling state, it is determined whether loss or overflow occurs.

3. The method of claim 2, wherein, When it is determined that the drilling is abnormal, and combined with the drilling state, it is determined whether loss or overflow occurs, including: When the drilling state is tripping in, it is determined whether the current tripping tank drilling fluid volume change is less than the expected value, if yes, it is determined that loss occurs, if not, it is determined that overflow occurs; When the drilling state is a tripping-in state, it is determined whether the current tripping-in tank drilling fluid volume change amount is less than an expected value, if yes, it is determined that overflow occurs, and if no, it is determined that loss occurs.

4. The method of claim 1 to 3, wherein, The real-time collected comprehensive logging data is time-indexed multi-dimensional drilling time sequence data, and specifically includes a drill bit depth sequence, a well depth sequence, and a hook load sequence, an inlet flow sequence, an outlet flow sequence, and a tripping-in tank drilling fluid volume sequence.

5. A system for identifying lost circulation during drilling tripping based on unsupervised learning, the system comprising: The method comprises an automatic identification module, a dynamic division module, an acquisition module, a calculation module, and a determination module. The automatic identification module is configured to automatically identify a drilling state of drilling based on the real-time collected comprehensive logging data. The dynamic division module is configured to dynamically divide a plurality of tripping-in calculation windows. The acquisition module is configured to acquire a tripping-in tank drilling fluid volume change amount and an outlet flow integral in each tripping-in calculation window. Calculate TOW using the first formula n TripPVVar, the volume change of drilling fluid inside the tripping and pulling container. n The first formula is: TripPVVar n =E(TripPV) ′ )-E(TripPV″), where, TOW n Starting from the volume of drilling fluid in the tripping and dredging container at the first moment within the tripping and dredging interval, the volumes of drilling fluid in the tripping and dredging container at m subsequent moments are obtained as TripPV′1, TripPV′2, ..., TripPV′. m Among them, TripPV1 ′ TripPV′2 represents the drilling fluid volume in the tripping container at the first moment, and TripPV′2 represents the drilling fluid volume in the tripping container at the second moment. m ′ This represents the volume of drilling fluid in the tripping container at time m, where 2m is less than TOW. n The total number of data points for the volume of drilling fluid in all tripping and pulling-out hoppers; E(TripPV) ′ () represents the first expected value; End of trip tank volume at the last time within TOW n End of trip tank volume at the last time within TOW m End of trip tank volume at the last time within TOW m End of trip tank volume at the last time within TOW n End of trip tank volume at the last time within TOW m End of trip tank volume at the last time within TOW m End of trip tank volume at the last time within TOW The calculation module is configured to calculate, according to a specification parameter of any common drill pipe, a stand composed of the common drill pipe corresponding to a tripped-out drilling fluid volume in an open-out condition and a closed-out condition, until a stand composed of each common drill pipe respectively corresponding to a tripped-out drilling fluid volume is obtained. The determination module is configured to determine, based on the stand composed of each common drill pipe respectively corresponding to the tripped-out drilling fluid volume, the tripping-in tank drilling fluid volume change amount in each tripping-in calculation window, and the outlet flow integral, whether loss or overflow occurs in the drilling state. The determination module is specifically configured to: When the drilling state is a tripping-in state, the process of determining whether loss or overflow occurs comprises: The stand composed of each common drill pipe respectively corresponding to the tripped-out drilling fluid volume is taken as an initial centroid heuristic value, the tripping-in tank drilling fluid volume change amount and the outlet flow integral when each stand is tripped in are calculated according to historical comprehensive logging data, and clustering is performed by using a K-means algorithm to acquire a centroid and a first maximum radius of each first cluster. The centroid of each first cluster is taken as an initial centroid, the tripping-in tank drilling fluid volume change amount in each tripping-in calculation window is calculated according to the real-time collected comprehensive logging data, and is converted into the tripping-in tank drilling fluid volume change amount of each stand, the outlet flow integral in each tripping-in calculation window is calculated, and is converted into the outlet flow integral of each stand. The initial distance between the tripping-in tank drilling fluid volume change amount and the outlet flow integral of each stand and the centroid of each first cluster is calculated, and if the maximum distance is greater than the first maximum radius, it is determined that drilling is abnormal, and whether loss or overflow occurs is determined in combination with the drilling state.

6. The un-supervised learning based drilling trip kick loss identification system of claim 5, wherein, The determination module is further specifically configured to: When the drilling state is a tripping-in state, the process of determining whether loss or overflow occurs comprises: The stand composed of each common drill pipe respectively corresponding to the tripped-out drilling fluid volume is taken as an initial centroid heuristic value, the tripping-in tank drilling fluid volume change amount when each stand is tripped out is calculated according to historical comprehensive logging data, and clustering is performed by using a K-means algorithm to acquire a centroid and a second maximum radius of each second cluster. The drilling fluid volume change amount in the tripping tank in each tripping calculation window is calculated according to the real-time comprehensive logging data, and is converted into the drilling fluid volume change amount in the tripping tank of each stand, taking the centroid of each second cluster as the initial centroid; The initial distance of the drilling fluid volume change amount in the tripping tank of each stand and the centroid of each second cluster is calculated, and if the maximum distance is greater than the second maximum radius, it is determined that the drilling is abnormal, and whether loss or overflow occurs is determined in combination with the drilling state.

7. The un-supervised learning based drilling trip kick loss identification system of claim 6, wherein, The determination module is further specifically used for: When it is determined that the drilling is abnormal, whether loss or overflow occurs is determined in combination with the drilling state, including: When the drilling state is the tripping-in state, whether the current drilling fluid volume change amount in the tripping tank is less than the expected value is judged, if yes, it is determined that loss occurs, and if no, it is determined that overflow occurs; When the drilling state is the tripping-out state, whether the current drilling fluid volume change amount in the tripping tank is less than the expected value is judged, if yes, it is determined that overflow occurs, and if no, it is determined that loss occurs.

8. The un-supervised learning based drilling trip kick loss identification system of any one of claims 5 to 7, wherein, The real-time comprehensive logging data is multi-dimensional drilling time series data indexed by time, specifically including a drill bit depth sequence, a well depth sequence, a hook load sequence, an inlet flow sequence, an outlet flow sequence, and a tripping tank drilling fluid volume sequence.

9. A storage medium, characterized by The storage medium has instructions stored therein, and when the computer reads the instructions, the computer executes the drilling tripping overflow and loss identification method based on unsupervised learning according to any one of claims 1 to 4.

10. An electronic device, comprising: The storage medium has instructions stored therein, and when the computer reads the instructions, the computer executes the drilling tripping overflow and loss identification method based on unsupervised learning according to any one of claims 1 to 4. The storage medium has instructions stored therein, and when the computer reads the instructions, the computer executes the drilling tripping overflow and loss identification method based on unsupervised learning according to any one of claims 1 to 4.

Images