CN-121976788-A - Drilling process state monitoring method based on interval information granulation

CN121976788ACN 121976788 ACN121976788 ACN 121976788ACN-121976788-A

Abstract

The invention discloses a drilling process state monitoring method based on interval information granulation, which relates to the technical field of geological drilling engineering and mainly comprises the following steps of preprocessing original drilling process data to obtain a drilling data time sequence; the method comprises the steps of carrying out multi-region information granulation on a drilling data time sequence to obtain a low-dimensional time sequence, obtaining a drilling operation state monitoring model according to the low-dimensional time sequence by using a principal component analysis method with a kernel function, and carrying out drilling operation state monitoring by using the drilling operation state monitoring model. By implementing the drilling process state monitoring method based on interval information granulation, the running state of a drilling system can be monitored in a refined mode and sensitively perceived.

Inventors

FAN HAIPENG
HUANG CHENG
WU MIN
CAO WEIHUA
Lu chengda
DU SHENG

Assignees

中国地质大学（武汉）

Dates

Publication Date: 20260505
Application Date: 20251224

Claims (10)

1. The drilling process state monitoring method based on interval information granulation is characterized by comprising the following steps of: s1, preprocessing original drilling process data to obtain a drilling data time sequence; s2, performing multi-region information granulation on the drilling data time sequence to obtain a low-dimensional time sequence; S3, obtaining a drilling operation state monitoring model by using a principal component analysis method with a kernel function according to the low-dimensional time sequence; and S4, monitoring the drilling running state by using the drilling running state monitoring model.
2. The interval information granulation-based drilling process state monitoring method according to claim 1, wherein the preprocessing includes abnormal data detection and missing value completion.
3. The method for monitoring the state of a drilling process based on interval information granulation according to claim 1, wherein step S2 specifically comprises: s21, carrying out sliding window segmentation on the drilling data time sequence to obtain a time sequence segment; S22, granulating interval information of the time sequence segments to obtain information particles; And S23, obtaining optimal multi-granularity interval information particles by using a PSO algorithm according to the information particles, and obtaining a drilling data time sequence according to the obtained optimal multi-granularity interval information particles.
4. The interval information granulation-based drilling process state monitoring method according to claim 1, wherein the interval information granulation formula is: , Wherein, the Representing information particles; As the lower limit of the information grain, Is the upper limit of the information grain; Representing the number of non-overlapping time series segments; Representing the number of data points contained in the information grain; representing the time series length; is a multiplicative function; And Respectively representing the rationality and the specificity of the granularity structure; is in the numerical form of information particles; And Respectively represent the optimal upper limit And an optimal lower limit Is a target function value; And Respectively representing objective functions selected from the optimal upper limit and the optimal lower limit of the granularity interval; and (3) reasonable interval information particles representing the drilling data construction time sequence.
5. The method for monitoring the state of a drilling process based on interval information granulation according to claim 1, wherein step S3 specifically comprises: s31, mapping the low-dimensional time sequence to a feature space; s32, carrying out feature decomposition by utilizing a Gaussian kernel function according to the feature space to obtain a main component matrix; s33, obtaining a main component contribution degree according to the main component matrix, and reserving main components with the main component contribution degree larger than a preset contribution degree; and S34, extracting a principal component vector of the drilling data after interval granulation according to principal components with the principal component contribution degree larger than the preset contribution degree, so as to obtain a drilling operation state monitoring model.
6. The interval information granulation-based drilling process status monitoring method according to claim 1, wherein the drilling operation status monitoring model is as follows: Wherein, the Represent the first A plurality of operating state principal component vectors; the number of the main components is reserved; is a combination coefficient.
7. The method for monitoring the state of a drilling process based on interval information granulation according to claim 1, wherein step S4 specifically comprises: S41, calculating Huo Linte statistic and error square statistic by using the operation state principal component vector selected from the drilling operation state monitoring model; s42, obtaining a main component space control limit and a residual space control limit according to the historical operation data in the normal operation state and the preset confidence coefficient of the kernel density estimation algorithm; S43, judging whether the drilling process is normally operated according to Huo Linte statistic and error square statistic of the real-time operation state data, the main component space control limit and the residual space control limit.
8. The interval information granulation-based drilling process status monitoring method according to claim 7, wherein the Huo Linte statistic and the square error statistic are calculated by the following formula: In the formula, Huo Linte statistics; Is the square statistic of error; to extract the principal component vector of the on-line data according to the drilling operational state monitoring model, Is the first A principal component vector; Selecting a diagonal matrix formed by the eigenvalues according to the sorting; reconstructing a residual error matrix for a feature vector of the online data; to monitor the number of variables.
9. The interval information granulation-based drilling process state monitoring method according to claim 7, wherein the calculation formulas of the principal component space control limit and the residual space control limit are: Wherein, the P is the number of main components; inputting sample data for a kernel density estimation algorithm; Is the corresponding confidence Lower degree of freedom is F distribution values of (2); is the control limit of the residual space; Controlling limits for residual space Scale parameters of (a); Is of degree of freedom of Confidence is The chi-square distribution quantiles; Is a degree of freedom; to estimate the mean value of the error square statistic of the sample, To estimate the mean variance of the error square statistic for the samples.
10. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the drilling process status monitoring method based on interval information granulation as claimed in any one of claims 1-9.

Description

Drilling process state monitoring method based on interval information granulation Technical Field The invention relates to the technical field of geological drilling engineering, in particular to a drilling process state monitoring method based on interval information granulation. Background The human demand for main resource energy and mineral products is kept high, and ensuring the safety of the resource energy is one of the keys of economic sustainable development, and deep geological drilling is an essential and indispensable method in resource exploration. Through the drilling activities deep in the stratum, scientists and geologists can accurately know key information such as the properties of rock, the distribution of mineral deposits, the flow of underground water and the like, and the precious data has an irreplaceable effect on making a scientific and reasonable mineral development plan. Deep geological resources and unconventional resources in China are rich in energy source and have great development potential. However, as the depth of the borehole increases, the wellbore temperature is very likely to break through 150 ℃, especially 6000 to 800 m, and is very likely to break through 200 ℃. The difficulty of fault diagnosis and early warning is further increased by dynamic change of drilling multiple working conditions, concealment of faults in holes and the like under complex geological environment, so that the safety risk of the drilling process is high. According to statistics, the abnormal operation time caused by drilling faults accounts for 15% -20% of the total drilling time. The accidents such as oil leakage and the like caused by the overlarge pressure in the drill rod and the failure of the blowout prevention valve cause serious economic loss. Therefore, the development of the fault diagnosis and safety early warning method in the drilling process has a key effect on the aspects of reducing the accident rate and improving the drilling efficiency. With the continuous advancement of industrial informatization and intelligence, many studies on data-driven industrial process monitoring and process modeling are gradually rising. Numerous scientific researchers and drilling enterprises at home and abroad are adopting various technical means to monitor and analyze the drilling process and analyze the working performance of the drilling process, thereby providing basis for safe production, cost reduction and efficiency enhancement. However, the prior art lacks high-quality data processing and filtering algorithms and multivariate statistical analysis methods, cannot extract important features (distance, similarity, correlation and the like) contained in operation data, cannot establish a working state monitoring model to judge the current working state and mine working performance improvement potential, has low efficiency and lacks safety lines in drilling operation level, and is a technical problem to be solved urgently, how to extract the important features contained in the operation data and establish the working state monitoring model to judge the current working state, mine the working performance improvement potential and maintain the high-efficiency and safety drilling operation level. Disclosure of Invention The invention aims to provide a drilling process state monitoring method based on interval information granulation, which can carry out fine monitoring and sensitive perception on the running state of a drilling system. The invention provides a drilling process state monitoring method based on interval information granulation, which comprises the following steps: s1, preprocessing original drilling process data to obtain a drilling data time sequence; s2, performing multi-region information granulation on the drilling data time sequence to obtain a low-dimensional time sequence; S3, obtaining a drilling operation state monitoring model by using a principal component analysis method with a kernel function according to the low-dimensional time sequence; and S4, monitoring the drilling running state by using the drilling running state monitoring model. The invention also provides a computer program product comprising a computer program which, when executed by a processor, implements the steps of the above-described interval information granulation based drilling process state monitoring method. The drilling process state monitoring method based on interval information granulation has the following beneficial effects: The invention models the drilling operation state based on interval information granulation in the drilling process state monitoring process, specifically, adopts K Nearest Neighbor (KNN) and a moving average filtering algorithm to detect abnormal values of original drilling process monitoring data, carries out filtering smoothing and normalization treatment on corrected drilling data, overcomes adverse effects of the abnormal data on the operation state modeling, compens