CN110009151B - Petroleum drilling accident time marking method - Google Patents

Petroleum drilling accident time marking method Download PDF

Info

Publication number
CN110009151B
CN110009151B CN201910263364.1A CN201910263364A CN110009151B CN 110009151 B CN110009151 B CN 110009151B CN 201910263364 A CN201910263364 A CN 201910263364A CN 110009151 B CN110009151 B CN 110009151B
Authority
CN
China
Prior art keywords
accident
data
related parameter
time
parameter value
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201910263364.1A
Other languages
Chinese (zh)
Other versions
CN110009151A (en
Inventor
李鑫
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Mt Networks Co ltd
Original Assignee
Shanghai Mt Networks Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shanghai Mt Networks Co ltd filed Critical Shanghai Mt Networks Co ltd
Priority to CN201910263364.1A priority Critical patent/CN110009151B/en
Publication of CN110009151A publication Critical patent/CN110009151A/en
Application granted granted Critical
Publication of CN110009151B publication Critical patent/CN110009151B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/10—Text processing
    • G06F40/103—Formatting, i.e. changing of presentation of documents
    • G06F40/117—Tagging; Marking up; Designating a block; Setting of attributes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Abstract

The invention relates to the technical field of petroleum drilling, in particular to a method for marking occurrence and cut-off time of petroleum drilling accidents. In order to improve the prediction efficiency and the prediction accuracy of petroleum drilling accidents, the invention provides a labeling method for petroleum drilling accident time, which comprises the following steps: collecting historical data of petroleum drilling and sorting in groups; performing primary smoothing treatment on the historical data to obtain primary smoothed data, and adding a difference term of two adjacent primary smoothed data; extracting starting point data, end point data and accident data of accident occurrence according to the change trend of the primary smooth data; primary smoothing or secondary smoothing is respectively carried out on accident related parameter values according to the data length of accident data, and a processing result is saved; and determining final accident data, actual accident occurrence time and actual accident ending time according to the processing result. The actual accident occurrence time and the actual accident ending time marked by the marking method have high precision, and the prediction efficiency and the accuracy of the petroleum drilling accidents can be improved.

Description

Petroleum drilling accident time marking method
Technical Field
The invention relates to the technical field of petroleum drilling, in particular to a method for marking occurrence and cut-off time of petroleum drilling accidents.
Background
Petroleum drilling is a high-risk and costly system project. In the drilling construction process, engineering accidents such as blowout, lost circulation, drilling tool damage and the like can occur at any time, and once the engineering accidents occur, the engineering accidents can cause great losses of funds, time and even property at life.
In order to reduce or avoid the loss caused by engineering accidents in the petroleum drilling process, a petroleum drilling accident early warning system is designed by the personnel in the field so as to predict the occurrence time and the ending time of petroleum drilling accidents according to historical data in the petroleum drilling construction process. However, in the above-mentioned history data, the occurrence time and the end time of the oil drilling accident are manually marked by the driller after the drilling accident occurs. Therefore, the occurrence time and the ending time of the drilling accident marked by the manual operation are usually deviated from the actual occurrence time and the ending time, so that the essential cause of the drilling accident can not be accurately obtained in the later analysis, and the petroleum drilling accident early warning system can not accurately predict the occurrence of the petroleum drilling accident in the subsequent drilling construction process.
In addition, in the existing petroleum drilling accident early warning system, only a small part of processes are analyzed by a computer, and most of processes are completed by manual operation, so that the prediction efficiency and the prediction accuracy of the petroleum drilling accident early warning system are low.
Disclosure of Invention
In order to improve the prediction efficiency and the prediction accuracy of petroleum drilling accidents, the invention provides a labeling method for petroleum drilling accident time, which comprises the following steps:
step S1, collecting historical data in the petroleum drilling construction process, wherein the historical data are grouped according to corresponding drilling well numbers JH, accident-related parameter values P in each group of historical data are sequenced according to corresponding collection time t, the data sequence numbers are recorded as i, i=1, 2 and 3 Λ, the accident-related parameter values P refer to values of related parameters for judging drilling accidents, and when the change trend of the related parameter values for judging the drilling accidents is a descending trend, the accident-related parameter values P are positive values of the related parameters for judging the drilling accidents; when the change trend of the related parameter value for judging the drilling accident is an ascending trend when the accident happens, the accident related parameter value P is a negative value of the related parameter for judging the drilling accident;
step S2, performing primary smoothing processing on the accident-related parameter value P in each set of history data acquired in the step S1 to obtain primary smoothing data P1 of the accident-related parameter value P,
according to the accident occurrence time T marked by manpower Human mark Dividing the primary smoothing data P1 into an A1 part and an A2 part, wherein the A1 part represents the primary smoothing data P1 of the accident related parameter value P before accident occurrence, the A2 part represents the primary smoothing data P1 of the accident related parameter value P after accident occurrence, and a difference item DI of two adjacent primary smoothing data P1 is added in the A1 part and the A2 part, and DI i-1 ＝P1 i -P1 i-1 (i.gtoreq.2), wherein,
DI i-1 represents the added i-1 st difference term,
P1 i representing the value of the i-th primary smoothed data,
P1 i-1 a value representing the i-1 st primary smoothed data;
step S3, extracting starting point data SD, end point data CD and accident data AD of accident occurrence according to the change trend of primary smooth data P1 of the accident related parameter value P, wherein the starting point data SD is the primary smooth data of the last part of the A1, which corresponds to the difference item DI is less than or equal to 0, and the end point data CD is the primary smooth data of the first part of the A2, which corresponds to the difference item DI is more than or equal to 0; the accident data AD is data located between the start point data SD and the end point data CD;
s4, judging the data length of the accident data AD, when the data length of the accident data AD is greater than or equal to 10, intercepting an accident-related parameter value P 'corresponding to the accident data AD from the accident-related parameter value P, performing secondary smoothing on the accident-related parameter value P' to update the starting point data SD, the end point data CD and the accident data AD, and storing an update result and performing the next step;
when the data length of the accident data AD is smaller than 10, performing primary smoothing processing on the accident-related parameter value P again, and when the data length of the processed accident data AD is greater than or equal to 10, intercepting the accident-related parameter value P 'corresponding to the accident data AD from the accident-related parameter value P, performing secondary smoothing processing on the accident-related parameter value P', so as to update the starting point data SD, the end point data CD and the accident data AD, and storing an update result and performing the next step; when the data length of the processed accident data AD is smaller than 10, storing the processing result and carrying out the next step;
step S5, determining final accident data AD according to the updated or processed result in step S4 f Actual time T of accident Implementation And the actual end time T of the accident Real terminal 。
The marking method for the oil drilling accident time is adopted to mark the occurrence time and the ending time of the oil drilling accident, the condition of inaccurate manual marking can be effectively corrected, meanwhile, the actual accident occurrence time and the actual accident ending time which are accurate to seconds can be given, compared with manual marking, the marking precision is high, and the prediction efficiency and the prediction accuracy of the oil drilling accident can be further improved.
Preferably, in the step S2, a smoothing process is performed by using a local polynomial regression fitting method. Further, a weighted least square method is used for fitting.
Preferably, when smoothing the accident-related parameter value P by using a local polynomial regression fitting method, use is made of Σ i w i (P i -θi-b) 2 The value of (1) is the smallest, and primary smooth data P1 is obtained i =θi+b, wherein,
w i an accident-related parameter value P representing the ith data in the history data i Weights at fitting time, and w i ＝exp(i-i c 2 /2τ 2 )，i c Data sequence number of data representing the central position, τ represents the concentration degree of the history data, and τ=0.08 when the accident-related parameter value P is subjected to primary smoothing; τ=0.02 when the primary smoothing process is performed again on the accident-related parameter value P; when performing secondary smoothing processing on the accident-related parameter value P' corresponding to the accident data AD, τ=0.2;
θ and b represent fitting parameters.
Preferably, in the step S4, when the data length of the accident data AD is greater than or equal to 10, the starting point data SD and the end point data CD are redetermined according to the processing result of the secondary smoothing processing, and the starting point data SD is the secondary smoothing data corresponding to the case where the value range of the last difference term DI in the A1 part is [ -0.04, -0.03], and the end point data CD is the secondary smoothing data corresponding to the difference term di+.0 in the first A2 part; and when the data length of the accident data AD is smaller than 10, adjusting the concentration degree parameter tau of the historical data, and carrying out primary smoothing on the accident-related parameter value P again.
Preferably, the related parameter for determining the accident corresponding to the accident related parameter value P is the total pool volume of the well, the outlet/inlet flow or the riser pressure.
In addition, the invention also provides a petroleum drilling accident time marking system, which marks the petroleum drilling accident time by adopting any petroleum drilling accident time marking method.
Preferably, the petroleum drilling accident time marking system marks the petroleum drilling accident time and then obtains the final accident data AD f Summarizing and outputting the final accident data AD f Including the accident occurrence time T of well drilling well number and manual marking Human mark Time T of actual occurrence of accident Implementation And the actual end time T of the accident Real terminal 。
Drawings
FIG. 1 is a flow chart of the oil drilling accident time marking method of the invention;
FIG. 2 is a schematic diagram of a first experimental result of marking lost circulation accident time of petroleum drilling by adopting the petroleum drilling accident time marking method of the invention;
FIG. 3 is a schematic diagram of a second experimental result of marking lost circulation event time of oil drilling by using the oil drilling event time marking method of the invention.
Detailed Description
The petroleum drilling accident time marking method of the invention is described in detail below with reference to figures 1-3 and experience examples.
As shown in FIG. 1, the petroleum drilling accident time marking method of the invention comprises the following steps:
historical data in the petroleum drilling construction process is collected first, and the collected historical data are grouped according to corresponding drilling well numbers JH. When the historical data is collected, the collection time interval delta t is a fixed value, and preferably the value range of the delta t is 1s-10s. Then, sequencing each group of historical data and corresponding accident-related parameter values P in the acquisition time t according to the sequence of the acquisition time t, and marking the data sequence of each piece of data as i, wherein the accident-related parameter values P are values of related parameters for judging drilling accidents, and when the change trend of the related parameter values for judging the drilling accidents is a descending trend, the accident-related parameter values P are positive values of the related parameters for judging the drilling accidents; when the change trend of the related parameter value for judging the drilling accident is in an ascending trend when the accident occurs, the accident related parameter value is a negative value of the related parameter for judging the drilling accident. Preferably, the accident-related parameter value P corresponds to the relevant parameter for determining the accident, which may be the total well volume, outlet/inlet flow or riser pressure of the well. Therefore, when the petroleum drilling accident time marking method is adopted to mark the occurrence time and the ending time of the petroleum drilling accident, the person skilled in the art can mark the occurrence time and the ending time of different types of petroleum drilling accidents by adjusting specific parameters corresponding to the accident related parameter value P, and the application range and the practicability of the marking method are greatly improved.
And respectively carrying out primary smoothing treatment on the accident-related parameter value P in each group of collected historical data to obtain primary smoothing data P1 of the accident-related parameter value P. In the smoothing process, a local polynomial regression fitting method may be preferably employed. When smoothing accident related parameter values P in a group of historical data by adopting a local polynomial regression fitting method, the weighted least square method can be preferably adopted to fit the accident related parameter values P, and when smoothing, the sigma can be utilized i w i (P i -θi-b) 2 The value of (1) is the smallest, and primary smooth data P1 is obtained i =θi+b, wherein,
w i representing an accident-related parameter value P in an ith data in a set of historical data i Weights at fitting time, and w i ＝exp(i-i c 2 /2τ 2 )，i c Data sequence number of data in central position, tau represents concentration degree of historical data, and value of tau can be regulated according to smoothing requirement,
θ and b represent fitting parameters.
In performing the primary smoothing of the accident-related parameter value P, τ may be a relatively moderate value, such as τ=0.08. In this way, too little selectable data caused by too concentrated historical data can be avoided, and too large selectable data range caused by too scattered historical data can be avoided.
Then, according to the accident occurrence time T marked by manpower Human mark The primary smoothing data P1 is divided into an A1 part and an A2 part, wherein the A1 part represents the primary smoothing data P1 of the accident-related parameter value P before the accident occurs, and the A2 part represents the primary smoothing data P1 of the accident-related parameter value P after the accident occurs. Then, difference items DI of adjacent two primary smoothed data P1 are added in the A1 section and the A2 section, respectively, and DI i ＝P1 i -P1 i-1 (i.gtoreq.2), wherein,
DI i representing the added i-th difference term,
P1 i representing the value of the i-th primary smoothed data,
P1 i-1 representing the value of the i-1 st primary smoothed data.
And combining the processing result of the primary smoothing processing, extracting starting point data SD, end point data CD and accident data AD of accident occurrence according to the change trend of primary smoothing data P1 of the accident related parameter value P, wherein the accident data AD is data positioned between the starting point data SD and the end point data CD. When an accident occurs in petroleum drilling, the trend of the change in the value of the relevant parameter for determining the drilling accident tends to decrease or increase depending on the type of the accident. For example, when the drilling accident is lost circulation, the change trend of the total pool volume, the riser pressure and the outlet/inlet flow of the relevant parameters for the drilling accident is judged to be a descending trend, and the story is that the relevant parameter value P directly takes the positive value of the corresponding total pool volume, the riser pressure and the outlet/inlet flow of the relevant parameters; when the drilling accident is a kick, determining that the change trend of the total pool volume, the riser pressure and the outlet/inlet flow of the related parameters for the accident is an ascending trend, and taking the negative value of the corresponding total pool volume, the riser pressure and the outlet/inlet flow of the related parameters as the story value P; when the drilling accident is a drilling tool piercing accident, the change trend of the pressure of the vertical pipe is a descending trend, and the related parameter value P of the story takes the positive value of the pressure of the vertical pipe. Thus, when an accident occurs in petroleum drilling, the accident-related parameter value P is in a descending trend, so that the accident-related parameter value P corresponding to the period of time when the accident actually occurs is continuously descending, and the last primary smooth data corresponding to the difference item DI less than or equal to 0 in the part A1 is taken as the starting point data SD, and the first primary smooth data corresponding to the difference item DI less than or equal to 0 in the part A2 is taken as the end point data CD. Of course, during specific operation, the accident-related parameter value P may also be in an ascending trend, and at this time, the accident-related parameter value P corresponding to the time period of the actual accident is continuously ascending, so the primary smooth data corresponding to the last unified difference item di≡0 in the A1 part is the start data SD, and the primary smooth data corresponding to the difference item di≡0 in the first one in the A2 part is the end data CD.
Judging the data length of the accident data AD, and carrying out the following processing according to the data length of the accident data AD:
in the primary smoothing process, in order to avoid that the smoothing degree is lower due to the fact that the local polynomial regression fitting operation includes an overlong time period, a fitting curve obtained after smoothing is rough and is not suitable for accurately estimating the drilling accident time, so that when the data length of the accident data AD is greater than or equal to 10, accident related parameter values P ' corresponding to the accident data AD are intercepted from the accident related parameter values P, secondary smoothing is conducted on the accident related parameter values P ', namely, the local polynomial regression fitting operation is conducted on the accident related parameter values P ' corresponding to the accident data AD so as to update the starting point data SD, the end point data CD and the accident data AD, and the updated result is saved and the next step is conducted. Further, in order to make the history data more concentrated and to make the start point data SD and the end point data CD as accurate as possible, the value of τ may be increased, for example, τ=0.2. Preferably, after performing the second-level smoothing processing on the accident-related parameter value P 'corresponding to the accident data AD, the second-level smoothing data P1 of the accident-related parameter value P' is obtained new Further, a difference term DI_new of the two adjacent two-level smooth data P1_new is obtained, and DI_new=P1_new (i+1) -P1_new i Wherein, the method comprises the steps of, wherein,
P1_new i a value representing the i-th secondary smoothed data,
P1_new (i+1) representing the value of the i+1st secondary smoothed data.
In order to obtain more accurate starting time and ending time of drilling accidents, the starting point data SD and the end point data CD are redetermined according to the processing result of the secondary smoothing processing, wherein the starting point data SD is the secondary smoothing data corresponding to the value range of the last difference item DI in the A1 part is [ -0.04, -0.03], and the end point data CD is the secondary smoothing data corresponding to the difference item DI more than or equal to 0 in the first part in the A2 part.
In the primary smoothing process, in order to avoid overfitting caused by overlarge fitting parameter theta values adopted in the local polynomial regression fitting operation, that is, to avoid the situation that the fitting is not only free of noise, but also fluctuation of the noise is fitted to cause very short accident time, so that the data length of accident data AD between starting point data SD and end point data CD of accident occurrence extracted after fitting is very short, and even less than 10. In addition, in general, the time interval between two adjacent pieces of data in the same set of historical data is typically 5s to 10s, and the time period for a drilling accident to occur is typically several minutes to several hours; only in a very small number of cases will the length of time that a drilling accident occurs, i.e. the data length of the accident data, be within 10. Therefore, when the data length of the accident data AD is smaller than 10, that is, the number of data pieces included in the accident data AD is smaller than 10, the concentration degree parameter τ of the history data is adjusted, the primary smoothing is performed again on the accident-related parameter value P, so as to change the data length of the accident data AD obtained by the primary smoothing, when the data length of the accident data AD obtained by the processing is greater than or equal to 10, the accident-related parameter value P ' corresponding to the accident data AD is intercepted from the accident-related parameter value P, and the secondary smoothing is performed on the accident-related parameter value P ', that is, the local polynomial regression fitting operation is performed on the accident-related parameter value P ' corresponding to the accident data AD to update the start point data SD, the end point data CD and the accident data AD, and the updated result is saved and the next step is performed; when the data length of the processed accident data AD is still smaller than 10, the occurrence time of the drilling accident is considered to be extremely short, so that the processing result is saved and the next step is performed. Since the data length of the accident data AD obtained by the primary smoothing is too small, the concentration of the history data, i.e., the value of τ, can be reduced when the primary smoothing is performed again on the accident-related parameter value P, for example, τ=0.02, so that the accident data AD obtained by the primary smoothing performed again covers more data, i.e., the data length of the accident data AD is increased.
Determining final accident data AD based on the updated result after the secondary smoothing or the processing result of the primary smoothing again f Actual time T of accident Implementation And the actual end time T of the accident Real terminal 。
In addition, the invention also provides a petroleum drilling accident time marking system, which marks the petroleum drilling accident time by adopting any petroleum drilling accident time marking method. Preferably, the petroleum drilling accident time marking system marks the petroleum drilling accident time and then obtains the final accident data AD f Summarizing and outputting, and finally obtaining accident data AD f Including the accident occurrence time T of well drilling well number and manual marking Human mark Time T of actual occurrence of accident Implementation And the actual end time T of the accident Real terminal 。
Because the change of the total well volume can be used for judging various well drilling accidents such as lost circulation, well kick, oil invasion, gas invasion and the like, the time of the well drilling accidents such as lost circulation, well kick, oil invasion, gas invasion and the like can be effectively detected by calculating the change of the total well volume. Meanwhile, in order to verify the marking precision of the petroleum drilling accident time marking method, taking the total pool volume of drilling as an accident related parameter value P as an example, the actual accident occurrence time T of the drilling accident is marked by adopting the petroleum drilling accident time marking method Implementation And the actual time of reception T of the accident Real terminal 。
Verification example 1
A lost circulation accident occurs on a certain drilling well (the drilling well number is JH 01) on a certain day, and the accident occurrence time T is marked manually Human mark 15:20:00 pm. Historical data of the day 13:00 to 18:00 of drilling is extracted, and plotted according to total pool volume and corresponding timeThe resulting total cell volume change curve l is shown in FIG. 2. The actual accident occurrence time T of the raise-log accident obtained by adopting the petroleum drilling accident time marking method Implementation The actual accident receiving time T is 14:55:10 Real terminal 16:23:42. As can be seen in connection with fig. 2, the total well volume of the well begins to drop in the event of a lost circulation event; when the lost circulation accident is over, the total pool volume of the well stops descending; and the accident occurrence time T is marked manually Human mark Significantly later than the time at which the total pool volume begins to drop. It follows that compared with the manually noted accident occurrence time T Human mark The actual accident occurrence time T is marked by the method for marking the oil drilling accident time Implementation And the actual end time T of the accident Real terminal The method is more in line with the starting time point of the drilling accident, and simultaneously gives the ending time point of the drilling accident.
Verification example 2
A lost circulation accident occurs on a certain drilling well (the drilling well number is JH 02) on a certain day, and the accident occurrence time T is marked manually Human mark Is 12:32:00 pm. Historical data of 11:00 to 13:00 of the drilling day are extracted, and the total pool volume change trend is obtained according to the total pool volume and the corresponding time in a drawing mode, wherein the total pool volume change trend is shown in figure 3. The actual accident occurrence time T of the raise-log accident obtained by adopting the petroleum drilling accident time marking method Implementation 11:25:55, actual accident receiving time T Real terminal 12:33:35. As can be seen in connection with fig. 3, the total well volume of the well begins to drop in the event of a lost circulation event; when the lost circulation accident is over, the total pool volume of the well stops descending; and the accident occurrence time T is marked manually Human mark Far behind the time at which the total cell volume begins to drop and near the time at which the total cell volume stops dropping. It follows that compared with the manually noted accident occurrence time T Human mark The actual accident occurrence time T is marked by the method for marking the oil drilling accident time Implementation And the actual end time T of the accident Real terminal The method is more in line with the starting time point of the drilling accident, and simultaneously gives the ending time point of the drilling accident.
In summary, the method for marking the oil drilling accident time automatically marks the occurrence time and the ending time of the drilling accident, can effectively correct the condition of inaccurate manual marking, and can give the actual occurrence time and the actual ending time of the accident accurate to seconds. Therefore, after the occurrence time and the end time of the drilling accident are marked by using the marking method of the petroleum drilling accident time, the essential cause of the drilling accident can be accurately obtained in the later analysis, so that the petroleum drilling accident early warning system can accurately predict the occurrence of the petroleum drilling accident in the subsequent drilling construction process, and the prediction efficiency and the prediction accuracy of the petroleum drilling accident are effectively improved.

Claims (8)

1. The method for marking the petroleum drilling accident time is characterized by comprising the following steps of:
step 1, collecting historical data in the petroleum drilling construction process, wherein the historical data are grouped according to corresponding drilling well numbers JH, accident-related parameter values P in each group of historical data are sequenced according to corresponding collection time t, the data sequence numbers are recorded as i, i=1, 2 and 3 … …, the accident-related parameter values P refer to values for judging related parameters of drilling accidents, and when the change trend of the related parameter values for judging the drilling accidents is a descending trend, the accident-related parameter values P are positive values of the related parameters for judging the drilling accidents; when the change trend of the related parameter value for judging the drilling accident is an ascending trend when the accident happens, the accident related parameter value P is a negative value of the related parameter for judging the drilling accident;
step S2, performing primary smoothing processing on the accident-related parameter value P in each set of history data acquired in the step S1 to obtain primary smoothing data P1 of the accident-related parameter value P,
according to the accident occurrence time T marked by manpower Human mark Dividing the primary smoothing data P1 into an A1 part and an A2 part, wherein the A1 part represents the primary smoothing data P1 of the accident-related parameter value P before the accident occurs, and the A2 part represents the accidentPrimary smoothed data P1 of the post-generation accident-related parameter value P, and a difference term DI of two adjacent primary smoothed data P1 is added in the A1 part and the A2 part, and DI i-1 ＝P1 i -P1 i-1 (i.gtoreq.2), wherein,
DI i-1 represents the added i-1 st difference term,
P1 i representing the value of the i-th primary smoothed data,
P1 i-1 a value representing the i-1 st primary smoothed data;
step S3, extracting starting point data SD, end point data CD and accident data AD of accident occurrence according to the change trend of primary smooth data P1 of the accident related parameter value P, wherein the starting point data SD is the primary smooth data of the last part of the A1, which corresponds to the difference item DI is less than or equal to 0, and the end point data CD is the primary smooth data of the first part of the A2, which corresponds to the difference item DI is more than or equal to 0; the accident data AD is data located between the start point data SD and the end point data CD;
step S4, judging the data length of the accident data AD, when the data length of the accident data AD is greater than or equal to 10, intercepting an accident related parameter value P 'corresponding to the accident data AD from the accident related parameter value P, performing secondary smoothing on the accident related parameter value P' so as to update the starting point data SD, the end point data CD and the accident data AD, and storing an update result and performing the next step S5;
when the data length of the accident data AD is smaller than 10, performing primary smoothing processing on the accident-related parameter value P again, and when the data length of the processed accident data AD is greater than or equal to 10, intercepting the accident-related parameter value P 'corresponding to the accident data AD from the accident-related parameter value P, performing secondary smoothing processing on the accident-related parameter value P', so as to update the starting point data SD, the end point data CD and the accident data AD, and storing the updated result and performing the next step S5; when the data length of the processed accident data AD is still smaller than 10, storing the processing result and performing the next step S5;
step S5, determining final accident data AD according to the updated or processed result in step S4 f Actual time T of accident Implementation And the actual end time T of the accident Real terminal 。
2. The method according to claim 1, wherein in the step S2, smoothing is performed by using a local polynomial regression fitting method.
3. The method for marking time of petroleum drilling accident according to claim 2, wherein the fitting is performed by using a weighted least square method pair.
4. The method for marking time of an oil well drilling event according to claim 3, wherein when smoothing the event related parameter value P by using a local polynomial regression fitting method, the sigma is used i w i (P i -θi-b) 2 The value of (1) is the smallest, and primary smooth data P1 is obtained i =θi+b, wherein,
w i an accident-related parameter value P representing the ith data in the history data i Weights at fitting time, and w i ＝exp(i-i c 2 /2τ 2 )，i c Data sequence number of data representing the central position, τ represents the concentration degree of the history data, and τ=0.08 when the accident-related parameter value P is subjected to primary smoothing; τ=0.02 when the primary smoothing process is performed again on the accident-related parameter value P; when performing secondary smoothing processing on the accident-related parameter value P' corresponding to the accident data AD, τ=0.2;
θ and b represent fitting parameters.
5. The method according to claim 4, wherein in the step S4, when the data length of the accident data AD is greater than or equal to 10, the start point data SD and the end point data CD are redetermined according to the processing result of the secondary smoothing processing, and the start point data SD is the last secondary smoothing data corresponding to the difference term DI in the A1 part when the value range is [ -0.04, -0.03], and the end point data CD is the first secondary smoothing data corresponding to the difference term DI not less than 0 in the A2 part; and when the data length of the accident data AD is smaller than 10, adjusting the concentration degree parameter tau of the historical data, and carrying out primary smoothing on the accident-related parameter value P again.
6. The method according to claim 5, wherein the accident related parameter value P corresponds to a related parameter for determining an accident, which is a total pool volume, an outlet/inlet flow, or a riser pressure of the well.
7. A petroleum drilling accident time marking system, which is characterized in that the petroleum drilling accident time marking system marks the petroleum drilling accident time by adopting the petroleum drilling accident time marking method according to any one of claims 1-6.
8. The system for marking time of oil drilling event as claimed in claim 7, wherein the system for marking time of oil drilling event is adapted to the final event data AD after marking time of oil drilling event f Summarizing and outputting the final accident data AD f Including the accident occurrence time T of well drilling well number and manual marking Human mark Time T of actual occurrence of accident Implementation And the actual end time T of the accident Real terminal 。
CN201910263364.1A 2019-04-02 2019-04-02 Petroleum drilling accident time marking method Active CN110009151B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910263364.1A CN110009151B (en) 2019-04-02 2019-04-02 Petroleum drilling accident time marking method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910263364.1A CN110009151B (en) 2019-04-02 2019-04-02 Petroleum drilling accident time marking method

Publications (2)

Publication Number Publication Date
CN110009151A CN110009151A (en) 2019-07-12
CN110009151B true CN110009151B (en) 2023-06-16

Family

ID=67169796

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910263364.1A Active CN110009151B (en) 2019-04-02 2019-04-02 Petroleum drilling accident time marking method

Country Status (1)

Country Link
CN (1) CN110009151B (en)

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6820702B2 (en) * 2002-08-27 2004-11-23 Noble Drilling Services Inc. Automated method and system for recognizing well control events
US9022140B2 (en) * 2012-10-31 2015-05-05 Resource Energy Solutions Inc. Methods and systems for improved drilling operations using real-time and historical drilling data
JP2017084268A (en) * 2015-10-30 2017-05-18 株式会社東芝 Accident occurrence forecast system and accident occurrence forecast method
CN107894969B (en) * 2017-09-13 2020-11-24 山东大学 Transformer latent fault early warning method based on trend analysis
CN109508827B (en) * 2018-11-14 2022-04-29 西南石油大学 Drilling accident early warning method based on time recursion neural network

Also Published As

Publication number Publication date
CN110009151A (en) 2019-07-12

Similar Documents

Publication Publication Date Title
CN110867196B (en) Machine equipment state monitoring system based on deep learning and voice recognition
CN110018670B (en) Industrial process abnormal working condition prediction method based on dynamic association rule mining
CN110008301B (en) Regional geological disaster susceptibility prediction method and device based on machine learning
CN115829422B (en) Industrial equipment operation abnormal state identification method based on big data
CN109960232B (en) Method for selecting leading auxiliary parameter and method for equipment maintenance pre-diagnosis
CN111080982A (en) Dam safety intelligent monitoring and early warning system and method based on multiple sensors
JP2015011027A (en) Method for detecting anomalies in time series data
US11544554B2 (en) Additional learning method for deterioration diagnosis system
CN106796157A (en) Abnormal method in detection distribution network, particularly drinking water distribution network
CN117371337B (en) Water conservancy model construction method and system based on digital twin
CN116708038B (en) Industrial Internet enterprise network security threat identification method based on asset mapping
CN108376293B (en) ZJ17 cigarette equipment maintenance intelligent prediction method based on fuzzy mathematics improved analytic hierarchy process
CN112149750A (en) Water supply network pipe burst identification data driving method
CN110398775B (en) Tunnel water burst disaster micro-seismic event signal fluctuation first arrival pickup method and system
CN114021620B (en) BP neural network feature extraction-based electric submersible pump fault diagnosis method
CN110009151B (en) Petroleum drilling accident time marking method
CN107844679B (en) Sectional type blast furnace molten iron silicon content prediction method and device
CN113128797A (en) Method and device for monitoring abnormal business indexes
CN110956330A (en) Method and system for predicting line loss of power transmission line based on multidimensional influence quantity
CN115856092A (en) Method for determining rock crack initiation stress based on acoustic emission data and stress data
CN116255896A (en) Landslide early warning method and device based on satellite and electronic equipment
CN114862612A (en) Mine data preprocessing method based on Bagging improvement
CN111178576B (en) Operation optimization method based on refining device operation data
CN110376436B (en) Multi-scale noise power spectral line spectrum detection method
CN114155914A (en) Detection and correction system based on metagenome splicing error

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant