A fuzzy inference procedure has been implemented as a MatLab program and tested for well log data classification using data from a public dataset coming from the KTB project (Germany). Rock property value distributions have been precisely emulated by fitting the envolvent of its transformed histograms with unnormalized double gaussian functions. A new final lithology confidence operator has been introduced and applied to adress the combination of individual confidences from different properties. Besides, a doubting class has been proposed for the final decision stage, as we think that it is more valuable an “in doubt” output that a wrong classification. The decision criteria is based on a relative confidence threshold, which value is entered by the end user. Results show that this fuzzy logic based method is suited to rapidly and reasonably suggest a lithology column from well log data, provided that there is enough a priori information and that the hypothesis of representativity holds between lithologies from the train and test datasets. This methodology will be a very useful tool to aid in the interpretation of well log data from future CO2 storage pilot sites.


Article metrics loading...

Loading full text...

Full text loading...

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error