CN-121995532-A - Method for inhibiting stratum thickness interpolation oscillation based on improved KNN

CN121995532ACN 121995532 ACN121995532 ACN 121995532ACN-121995532-A

Abstract

The invention discloses a method for inhibiting stratum thickness interpolation oscillation based on improved KNN, and belongs to the technical field of geological modeling. Aiming at the problem that a conventional interpolation method is easy to generate a pseudo stratum without geological basis near a stratum pinch-out line, the method comprises the steps of firstly extracting a drilling position, stratum existence labels and thickness, then constructing a weighted distance formula based on an improved KNN algorithm through calculating spatial anisotropy weights, intelligently partitioning a region to be predicted into an interpolation effective region and an ineffective region, then performing spatial interpolation on stratum thickness only in the effective region, forcibly setting the thickness to zero in the ineffective region, and finally fusing the two results to generate a final thickness model. The invention fundamentally inhibits interpolation oscillation, ensures the clear stratum boundary, has more reasonable model result and obviously improves the reliability of resource evaluation.

Inventors

TIAN FENGLIANG
ZENG XIANGYU
Xin Fengyang
WANG ZHONGXIN
ZHAO MING
ZHOU HAO
DONG ZHIQIANG
WANG HAIYANG

Assignees

中煤科工集团沈阳设计研究院有限公司

Dates

Publication Date: 20260508
Application Date: 20251225

Claims (5)

1. A method for suppressing formation thickness interpolation oscillations based on improved KNN, comprising the steps of: s1, data preparation, namely acquiring drilling data, extracting position information of each drilling hole, a target stratum existence tag and thickness to form a sample feature set for spatial analysis and interpolation; S2, intelligently judging areas, namely dividing a research area into an interpolation effective area and an interpolation ineffective area by utilizing position information and a presence label in the sample feature set based on an improved KNN algorithm; S3, assigning a partition thickness, namely performing spatial interpolation on the target stratum thickness in the interpolation effective area based on the position information and the thickness of the sample point with the presence label of 1 in the sample feature set; And S4, generating a model, namely fusing an interpolation result of the interpolation effective area and a zero thickness result of the interpolation ineffective area to generate a final thickness model.
2. The method for suppressing formation thickness interpolation oscillation based on improved KNN according to claim 1, wherein the step S1 of obtaining borehole data, extracting positional information of each borehole, a target formation presence tag, and a thickness specifically includes: for each drilling hole, acquiring an opening coordinate of the drilling hole as the position information, wherein the opening coordinate at least comprises an east coordinate and a north coordinate; Traversing lithology layering records of the target stratum, judging whether the target stratum is contained, if the target stratum exists, calculating the thickness of the target stratum, marking the existence label as 1, and if the target stratum does not exist, setting the thickness as zero, marking the existence label as 0.
3. The method for suppressing formation thickness interpolation oscillation based on improved KNN according to claim 1, wherein the dividing the investigation region into an interpolation effective region and an interpolation ineffective region in step S2 specifically comprises: S21, calculating pearson correlation coefficients of the east and north coordinate directions and the presence labels based on the position information and the presence labels of all samples in the sample feature set, and calculating normalized weights of the east and north directions according to the pearson correlation coefficients; s22, constructing an anisotropic distance formula, and calculating a weighted Euclidean distance between the point to be predicted and each sample point by using the normalized weight; S23, for each to-be-predicted point in a research area, searching K nearest neighbor sample points by using the weighted Euclidean distance, and calculating the probability of stratum existence at the point based on the existence labels of the K nearest neighbor samples; s24, according to a preset probability threshold, marking points with the stratum existence probability larger than or equal to the threshold into an interpolation effective area, and marking points with the stratum existence probability smaller than the threshold into an interpolation ineffective area.
4. The method for suppressing formation thickness interpolation oscillation based on improved KNN according to claim 1, wherein the performing spatial interpolation in the interpolation effective area in step S3 specifically comprises: s31, screening sample points with the existence label of 1 from the sample feature set to form an effective sample subset for thickness interpolation; S32, based on the position information and the thickness value of each sample point in the effective sample subset, establishing a spatial interpolation model by adopting a Kriging interpolation method; s33, calculating an initial thickness estimated value of each point to be predicted, which is marked into the interpolation effective area, by using the spatial interpolation model.
5. The method for suppressing formation thickness interpolation oscillations based on improved KNN according to claim 1, characterized in that said forcing the target formation thickness to zero in said interpolation null zone in step S3 comprises, for each point to be predicted divided into said interpolation null zone, assigning its target formation thickness value directly to zero, whatever its estimated value is calculated based on any interpolation method.

Description

Method for inhibiting stratum thickness interpolation oscillation based on improved KNN Technical Field The invention belongs to the technical field of geological engineering and computer-aided modeling, and particularly relates to a method for inhibiting stratum thickness interpolation oscillation based on improved KNN. Background In the field of geologic modeling, a modeling method based on stratum thickness interpolation has obvious advantages compared with the traditional modeling according to altitude. The absolute space coordinates of the stratum interface are directly interpolated according to the altitude modeling, and the method mixes the post-construction deformation signals with the original deposition characteristics. In complex formation areas, strong formation inclinations can lead to severe distortion of the apparent thickness of the formation. And the original deposition thickness of the stratum is restored firstly based on true thickness modeling, so that the structure continuity is better. However, deterministic interpolation methods, including kriging, have the mathematical nature of constructing a statistically best-fit surface of spatial data that has continuous and smooth characteristics. However, such mathematical smoothness requirements tend to induce numerical oscillations in the data boundaries or sparse areas of control points, resulting in severe distortion of the geologic characterization. When such surfaces are applied to geological properties such as coal seam thickness, which have strict physical boundaries, two types of problems may arise: (1) Negative thickness problem. At the coal seam pinch or erode boundary, the true thickness should be zero. However, due to the smoothness requirement of the interpolation function, the interpolation function can generate gradual transition when crossing the boundary, so that a negative thickness value is calculated in a coal-free area, and the basic physical law of non-negative stratum thickness is violated. (2) No pseudo stratum problem of geological basis. The algorithm may generate coal seams with significant thickness by mathematically extrapolating the evidence in ancient elevations or eroded areas not covered by the deposition system. Such pseudo-formations lack of geological basis can seriously affect accurate assessment of the paleogeography of depositions and the potential of coal resources. Aiming at the problem, the conventional method comprises the steps of firstly forcing a region with a negative thickness to be 0 thickness, but the method cannot cope with the problem (2), secondly triangulating a sample point and constraining the thickness of a coal seam in the region according to the thickness of the coal seam of the sample point, wherein the method cannot extrapolate, and interpolation oscillation outside a triangular net cannot be effectively constrained. Disclosure of Invention In order to overcome the defects in the prior art, the invention aims to provide a method for restraining formation thickness interpolation oscillation based on improved KNN. The technical scheme adopted by the invention is that the method for inhibiting the stratum thickness interpolation oscillation based on the improved KNN is characterized by comprising the following steps: s1, data preparation, namely acquiring drilling data, extracting position information of each drilling hole, a target stratum existence tag and thickness to form a sample feature set for spatial analysis and interpolation; S2, intelligently judging areas, namely dividing a research area into an interpolation effective area and an interpolation ineffective area by utilizing position information and a presence label in the sample feature set based on an improved KNN algorithm; S3, assigning a partition thickness, namely performing spatial interpolation on the target stratum thickness in the interpolation effective area based on the position information and the thickness of the sample point with the presence label of 1 in the sample feature set; And S4, generating a model, namely fusing an interpolation result of the interpolation effective area and a zero thickness result of the interpolation ineffective area to generate a final thickness model. In the above scheme, the step S1 of obtaining the borehole data, extracting the position information of each borehole, the target formation existence tag, and the thickness specifically includes: for each drilling hole, acquiring an opening coordinate of the drilling hole as the position information, wherein the opening coordinate at least comprises an east coordinate and a north coordinate; Traversing lithology layering records of the target stratum, judging whether the target stratum is contained, if the target stratum exists, calculating the thickness of the target stratum, marking the existence label as 1, and if the target stratum does not exist, setting the thickness as zero, marking the existence label as 0. In the above schem