Abstract
In this invention, we introduce a novel physics guided ML bit wear model by coupling the physics and data driven wear model. In the model, the large amount of unknown wear states between the sharp and very end dull state are estimated using a physics-based wear model with empirical rock and drilling knowledge. With the physics guidance, the number of labeled data is significantly increased, and the AI wear model accuracy is improved. To account for the accuracy of the labels from physics-model with empirical knowledge, a loss weight factor is used to adjust the training process.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Baker Hughes Company, "Physics Guided AI ML Model for Bit Wear Prediction", Technical Disclosure Commons, (May 17, 2024)
https://www.tdcommons.org/dpubs_series/7023