Methods and systems for predicting formation thermal properties

A method for characterizing thermal properties of a subsurface formation, comprising:

• i) obtaining a plurality of formation rock samples;
• ii) performing laboratory measurements on the plurality of formation rock samples, wherein the laboratory measurements measure a predefined set of petrophysical properties of each formation rock sample in the plurality of formation rock samples;
• iii) performing additional laboratory measurements on the plurality of formation rock samples, wherein the additional laboratory measurements measure a predefined set of thermal properties of each formation rock sample in the plurality of formation rock samples;
• iv) generating and storing a training dataset derived from the petrophysical properties of the plurality of formation rock samples measured in ii) and the thermal properties of the plurality of formation rock samples measured in iii);
• v) using the training dataset of iv) to train a ML model that predicts label data representing the predefined set of thermal properties given input data representing the predefined set of petrophysical properties of an arbitrary formation rock sample;
• vi) validating the ML model trained in v);
• and
• vii) in the event of successful validation of the ML model in vi), using the ML model to predict label data representing the predefined set of thermal properties for at least one new formation rock sample given input data representing the predefined set of petrophysical properties of the least one new formation rock sample.