High-Frequency Drilling Data Analysis to Characterize Water-Ice on the Moon

Abstract

This paper discusses the evolution of a pattern recognition algorithm that utilizes the high-frequency drilling data to characterize the water-ice on the Moon. The algorithm developed here can estimate the moisture content of a grout sample by analyzing the trend of drilling data. Such an algorithm can be used by NASA and private organizations in near-future to identify and produce water from Lunar Poles. An auger based rotary drilling rig with a drilling data acquisition system was designed and fabricated. The data acquisition system records drilling parameters like RPM, drilling depth, torque, and weight on bit at 1000Hz. The data is then filtered to remove electromagnetic interference. The drilling tests were conducted on meticulously designed grout block samples which contained the lunar soil simulant with specific geotechnical properties, replicating lunar subsurface at a specific pressure, temperature condition, and moisture contents. Simultaneously, the UCS of the grout block was tested in a lab to correlate the drilling data to specific grout strength. The drilling data recorded in the homogenous samples and the data collected during UCS tests were used to develop and validate a pattern recognition algorithm. The algorithm was tested on layered grout samples containing layers with varying strength. The UCS estimated by this algorithm was then correlated to the moisture content using the lab observations and published literature. The algorithm could distinctly detect the boundaries between different layers, estimate the UCS, and moisture content in real-time. The high-frequency drilling data was also used to identify different dysfunctions and optimize drilling operations. Trends of MSE, RPM, and torque were analyzed to detect auger choking, inefficient cuttings transport, and drilling vibrations. On the Moon, this algorithm can be invaluable in optimizing the drilling operations, estimating the spatial distribution of various subsurface layers improving our understanding of the subsurface lunar stratigraphy. It can also be used to estimate the quantity of water available on the lunar poles, which will be essential in planning the manned and robotic lunar missions in the coming years.

A Mission Simulating the Search for Life on Mars with Automated Drilling, Sample Handling, and Life Detection Instruments Performed in the Hyperarid Core of the Atacama Desert, Chile

“…9 was based on patterns recognized from drilling many sites in Atacama with the ARADS drill. In a study focused on developing methods for interpreting drill data to identify icy deposits on the moon, a suite of cement and icy samples with different compressive strengths were drilled and the data used to train a machine learning algorithm to recognize the materials drilled (Joshi et al ., 2020 ; Joshi, 2021 ). Machine learning automates pattern recognition but is dependent on sufficient training data to achieve accurate results.…”