The Design Of Logging While Drilling Tool Monitoring System Based On Power Supply EMI Noise Analysis

During the operation of a logging-while-drilling tool,it is necessary to monitor and control the working state of the electronic circuit of the logging tool,to ensure that all circuits of LWD are in a normal and stable working mode.In this paper,an online monitoring method of circuit power working states of LWD tool based on EMI signal characteristic analysis is designed and implemented,which is used to solve the problem of circuit power monitoring of LWD tool working in high temperature and vibration environment.An EMI signal acquisition and data processing circuit based on STM32F746 ZG and MAX10M50 SC is designed to collect and process the EMI signal of the LWD tool.The signal acquisition circuit is composed of signal conditioning,analog-to-digital conversion,first in first out memory consisting of MAX10M50 SC,and time sequence control module.The EMI signal of the power supply is collected continuously at a sampling rate of 25 MHz.The data processing circuit is composed of STM32F746 ZG.Real-time signal processing methods such as Kalman filter algorithm and fast Fourier transform are used to carry out digital denoising and feature extraction of EMI signals,and the feature values are converted into dimensionless expressions by DC removal and normalization processing.The characteristic parameters represent different power working states,and their time stability provides a stable basis for the input of the classifier.The inner space of the drill collar of the LWD tool is narrow and difficult to maintain.Thus,a non-invasive single-point sensing method is introduced,which only installs a single sensor on the power supply port.A lightweight convolutional neural network classification algorithm is deployed on the embedded platform to identify the power EMI signal and execute abnormal operation judgment.In this paper,the designed circuit monitoring method has complete data acquisition,feature extraction,and the power supply EMI feature recognition functions.Using learning algorithm to implement identification and positioning on circuit boards of LWD tool,which can effectively meet the requirements of circuit board working state detection in the process of downhole operation.This method will reduce the economic consequences of circuit board faults.Also,it provides a reference for field experiments in future work.