Research on the Imaging and Accuracy Analysis for Fish Head Detection by Using Directional Borehole Radar

Abstract

Oilfield casing is crucial for effective oil extraction and storage. Casing failures often occur at various depths during drilling, significantly affecting production. Conventional geophysical methods are inadequate in drilling environments, posing challenges for accurate casing failure detection. This article presents a directional detection technique for “fish head” identification using borehole radar, alongside imaging and accuracy analysis. The working principle of the fish head radar detection system is discussed, including necessary hardware and software components. A pulsewidth modulation (PWM) control algorithm enables omnidirectional data acquisition in “blind hole” conditions. The study investigates the impact of radar rotation speed and detection movement speed on results. Utilizing a gray-level co-occurrence matrix, the analysis focuses on features such as energy, contrast, homogeneity, and correlation to quantitatively assess the pipeline response area in radar images, identifying optimal rotation and movement speeds of 0.12 m/s and a PWM duty cycle of 50%–70%. Field experiments for fish head detection were conducted in the Daqing oilfield with two radar antennas of different frequencies. Results show a deviation between the set and actual antenna angles within 1°, achieving an accuracy of 95.5%. Casing breaks were detected at 0° and 235° in the simulated well, with maximum detection depths of 6 m at 500 MHz and 25 m at 100 MHz. These findings validate the capability of borehole radar omnidirectional scanning for precise anomaly detection around oil wells, providing technical support for identifying casing break orientations.