A Machine Learning-Based Geothermal Drilling Optimization System Using EM Short-Hop Bit Dynamics Measurements
- E-Spectrum Technologies, Inc., San Antonio, TX (United States)
- Texas A & M Univ., College Station, TX (United States)
This report describes the work conducted on an SBIR Phase 1 project to develop a machine learning based geothermal drilling optimization system, referred to as the Geothermal Drilling Optimization System (GDOS), to help drillers make more effective decisions and optimize the Rate of Penetration (ROP) achieved during geothermal drilling operations. The concept behind the system is to utilize a high-temperature, high-data-rate Short-hop Data Analytic (SDA) downhole telemetry system, coupled with a high-temperature, directional-drilling EM telemetry tool to telemeter at-bit, dynamic data to the surface for near-real-time analysis by a machine learning based data analytics software application. This surface software application, referred to as the Dynamic Advisory System (DAS), is a real-time data analytics engine that compares pattern identifier metric bit-data, and other downhole data received from the SDA, with surface-measured data obtained through the rig’s Well Information Telemetry System (WITS) infrastructure. The DAS information is presented to the driller as a graphical display of a recommended optimal WOB/rotary speed combination that produces the highest penetration rate with minimal bit vibration, wear, and damage. The SDA component of the system consists of an at-bit, high-speed, EM short-hop link, which telemeters data around the mud motor, to an adjacent, collar mounted commercial EM MWD telemetry system. Work accomplished on the SDA during Phase 1 included the design of a high-temperature compatible embedded controller prototype of the short-hop telemetry link electronic hardware, including the construction of a scale model TX/RX coil pair and scaled drill collar mandrel. This scale model coil pair was tested in-situ, utilizing a salt-water test tank to simulate variable earth-load impedances, to show that the empirical test data matched a theoretical earth/drill-string model developed in Matlab during the Phase 1 effort. A System Design Document (SDD) was developed to describe the architecture and operational design of the Dynamic Advisory System software application. The SDD included a detailed library of bit-dysfunctions that the DAS will detect and analyze in real-time. Feasibility of the DAS engine was demonstrated by selecting one bit-dysfunction, Stick/Slip, and modeling it to determine if the DAS can detect it based on bit-data received from a downhole near-bit tool. Modeling results showed that Stick/Slip dysfunctions could be detected with an accuracy of up to 89% while drilling.
- Research Organization:
- E-Spectrum Technologies, Inc., San Antonio, TX (United States)
- Sponsoring Organization:
- USDOE Office of Science (SC), Engineering & Technology. Office of Small Business Innovation Research (SBIR) and Small Business Technology Transfer (STTR) Programs
- DOE Contract Number:
- SC0019866
- OSTI ID:
- 1842454
- Type / Phase:
- SBIR (Phase I)
- Report Number(s):
- DOE-ESTI-19866
- Country of Publication:
- United States
- Language:
- English
Similar Records
Raytheon Downhole Information System: Electromagnetic lithospheric information system for geothermal drilling. Interim report
An MWD downhole assistant driller