Using machine learning to predict aquifers from groundwater chemistry data

A non-transitory, computer-readable medium storing one or more instructions executable by a computer system to perform operations comprising:

• replacing, in groundwater chemistry data from aquifers, missing measurements with calculated equivalents;
• splitting the groundwater chemistry data into a model building/optimizing dataset and a model validation dataset;
• normalizing the groundwater chemistry data;
• performing a dimensionality reduction on the groundwater chemistry data, comprising identifying principal components, wherein identifying principal components comprises applying Principal Component Analysis (PCA) on the groundwater chemistry data to reduce the dimensionality of the dataset by reducing components to three principal components having highest influences;
• generating and training, using the model building/optimizing dataset and machine learning, an aquifer prediction model;
• validating, using the model validation dataset, performance of the aquifer prediction model;
• predicting, during production of a new well and using the aquifer prediction model and groundwater chemistry data for the new well, locations of aquifers;
• and
• changing, based on the locations and the groundwater chemistry of the aquifers, settings, targets, and parameters of equipment used for drilling the aquifers.

A computer-implemented system, comprising:

• one or more processors;
• and
• a non-transitory computer-readable storage medium coupled to the one or more processors and storing programming instructions for execution by the one or more processors, the programming instructions instructing the one or more processors to perform operations comprising:
• replacing, in groundwater chemistry data from aquifers, missing measurements with calculated equivalents;
• splitting the groundwater chemistry data into a model building/optimizing dataset and a model validation dataset;
• normalizing the groundwater chemistry data;
• performing a dimensionality reduction on the groundwater chemistry data, comprising identifying principal components, wherein identifying principal components comprises applying Principal Component Analysis (PCA) on the groundwater chemistry data to reduce the dimensionality of the dataset by reducing components to three principal components having highest influences;
• generating and training, using the model building/optimizing dataset and machine learning, an aquifer prediction model;
• validating, using the model validation dataset, performance of the aquifer prediction model;
• predicting, during production of a new well and using the aquifer prediction model and groundwater chemistry data for the new well, locations of aquifers;
• and
• changing, based on the locations and the groundwater chemistry of the aquifers, settings, targets, and parameters of equipment used for drilling the aquifers.