Drilling radar attitude angle prediction method based on BP neural network

The borehole radar attitude angle is directly predicted through the BP neural network, which solves the problem of sensor installation error that is difficult to correct in the existing technology and realizes high-precision attitude angle measurement.

CN120671758APending Publication Date: 2025-09-19SHANXI YANGMEISI JIAZHUANG COAL IND CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202510732107.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-06-03
Publication Date
2025-09-19

AI Technical Summary

Technical Problem

In the attitude angle measurement of borehole radar, due to the installation error and alignment error of the three-axis magnetic sensor and acceleration sensor, the three-dimensional spatial error is difficult to correct, affecting the measurement accuracy.

Method used

A BP neural network-based method is used to ignore the sensor installation error. The attitude angle of the borehole radar is directly predicted through the measurement data after ellipsoid fitting. The error back propagation algorithm is used to optimize the network performance and obtain the optimal attitude angle prediction network.

Benefits of technology

The influence of installation errors was effectively calibrated, and the accuracy of borehole radar attitude angle measurement was improved. The absolute error of inclination angle was less than 0.1°, and the absolute error of azimuth angle was less than 1°.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120671758A_ABST
    Figure CN120671758A_ABST
Patent Text Reader

Abstract

The invention relates to a drilling radar attitude angle prediction method based on a BP neural network, and belongs to the technical field of drilling radar attitude angle prediction. The method comprises the following steps: acquiring an attitude angle prediction training sample; acquiring a hidden layer node number; constructing a borehole radar attitude angle prediction network according to the obtained hidden layer node number; training a borehole radar attitude angle prediction network by using the training sample; the method comprises the following steps of: processing acquired training data layer by layer through an input layer and a hidden layer, comparing an error between an actual output value and an expected output value in a propagation process, and adjusting a connection weight and a threshold value between each layer and each node by utilizing an error back propagation algorithm, so as to continuously optimize the network performance; an optimal drilling radar attitude angle prediction network is obtained; wherein the attitude angle comprises an inclination angle, a rotation angle and an azimuth angle; and performing drilling radar attitude angle prediction by using the trained drilling radar attitude angle prediction network. According to the invention, the influence of installation errors can be effectively calibrated, and the measurement precision is improved.
Need to check novelty before this filing date? Find Prior Art