Downhole drilling represents one of the most challenging operational scenarios for condition monitoring due to the limited access to the downhole information as a result of harsh environmental conditions. Recent progress in downhole sensing technologies enables high frequency dynamic measurements for drilling tool monitoring. Downhole vibration data offers indispensable insight into drilling performance evaluation and downhole tool diagnostics, particularly for rotary percussion drilling tools for ROP improvement. Additionally, integrating downhole vibration data analysis and surface drilling data builds a foundation for advanced digital drilling diagnostics. This work details a novel, efficient, and open-source workflow to process and analyze a high-frequency large-dataset of downhole vibration data to investigate the features of downhole vibration when drilling with a hammer tool.
The unique dataset utilized in this study consists of three-axis vibration data sampled at 1,500 Hz. A novel workflow was developed to handle the large dataset efficiently and consists of 1) data conversion from Big-endian format to decimal, 2) data optimization and slicing into 1-hour segments, 3) data analysis in the time domain, 4) data filtering, and 5) frequency domain analysis. The efficient open-source workflow does not require a high-performance workstation and can be deployed in the field. The analysis results are integrated with surface drilling parameters to describe the hammer's performance comprehensively.
The analysis results include the maximum vibration amplitude, vibration modes, main vibration frequencies, and vibration energy content. The workflow is first applied to a vibration dataset from a yard test to establish the performance benchmark for the hydraulic hammer. The workflow then analyses downhole vibration data and combines surface drilling data. The algorithm can analyze a 1-hour dataset (>500 MB) in less than a minute. The results highlight that the hydraulic hammer operates in multiple modes depending on the mud flow rate, WOB, and formation strength. The results offer precious unprecedented insight into the performance modes of the hydraulic hammer. The time and frequency domain analyses generate essential parameters to support the tool development and performance investigation. Future use-cases include in-situ downhole tool evaluation and digital drilling evaluation using AI. The main challenges faced during the study are relatively limited sampling frequency, drilling surface data quality, and data reliability.
To our best knowledge, this is the first systematic study on high frequency vibration monitoring of a rotary percussion drilling tool in a filed application. The paper introduced a novel analysis method applied to a large, unique downhole vibration dataset (>50GB/100 hours) acquired during drilling using a hydraulic hammer. The developed workflow is efficient and open-source and allows fast, in-depth analysis of high-resolution vibration data. Insights gathered allow advanced tool design, downhole problem solving and drilling performance evaluation.