Methods and systems for adaptive non-contact / contact boring
AssigneeArcByt, Inc.
InventorsShivani Torres, Thomas Egan, Barzin Moridian
Life of this patent
Estimating the term…
Abstract
The systems and techniques described herein may allow for optimized boring through a variety of geologies. A plurality of different boring techniques may be utilized for boring through a geological formation, in order to suit the characteristics of various portions of the geological formation. The systems and techniques described herein includes determining geological features and adjusting operation of boring based on the geological features. In certain such embodiments, boring systems may include a bore head that includes a plurality of boring elements. Such boring elements may be contact and/or non-contact boring elements.
Claims15
A system comprising
- a bore head comprising a non-contact boring mechanism;
- a first sensor, configured to measure a first parameter associated with operations of the non-contact boring mechanism;
- and
- a controller, communicatively coupled to the first sensor and configured to perform operations comprising:
- causing the non-contact boring mechanism to operate in a first manner;
- receiving first data from the first sensor;
- determining a first boring parameter from the first data;
- and
- causing, based on the determined first boring parameter, the non-contact boring mechanism to operate in a second manner.
A method comprising:
- preparing first multi-head boring training data, the first multi-head boring training data comprising a plurality of boring scenarios for boring with a bore head comprising a non-contact boring mechanism and a contact boring mechanism, each boring scenario comprising:
- first geological composition composition data for a plurality of bore sites;
- first non-contact boring data indicating first non-contact boring portions of the plurality of bore sites;
- and
- first contact boring data indicating first contact boring portions of the plurality of bore sites;
- and
- providing the first multi-head boring training data to a machine learning device to train the machine learning device.
Dependencies are inferred from the claim text.