Geothermal drilling industry faces various challenges such as poor overall drilling confidence and performance, lack of bottom hole awareness, lengthy tripping times, etc resulting in significant Non-productive Time (NPT) and unpredicted additional costs which makes the drilling process quite uncertain and expensive. Furthermore, there is a lack of digitisation and automation in the geothermal drilling industry and currently drillers mainly rely on personal skills and previous experiences.
The OptiDrill project aims to develop an innovative drilling advisory system utilising novel sensor systems and AI-based methods to predict and optimize the rate of penetration (ROP), drilled lithology, drilling problems, well completion, and enhancement and finally to unite those methods under one system to enable drilling process optimisation and intelligent decision making. Through the use of new technologies, the number of days taken to drill and complete wells will be greatly reduced, which in turn lowers costs and reduces risk and uncertainties. OPTIDRILL´s advisory system is based on a combination of enhanced monitoring systems, and multiple data-driven AI modules, each being responsible for either analysis, prediction, or optimisation of one aspect of the drilling, well completion, or well enhancement process.
The overall objectives of the project are
a) Digitalise the manual drilling data and text-based reports into one unified database
b) Instrumentalise the drilling process through the Implementation of the drill rig and BHA-compatible novel sensor strings
c) Implement novel system identification methods in the sensor string and monitoring systems
d) Employ the combination of machine learning and novel deep learning methods in drilling, well completion and enhancement modelling, performance prediction and optimisation
f) Predict and trigger detection of drilling problems through data-driven statistical and machine learning methods
g) Understand and predict the real-time lithology of the formation
e) Use Federated machine learning scheme in combination with self-learning machine learning algorithms to make the unified OptiDrill system
