Random selection of observation cells for proxy modeling of reactive transport modeling

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

• randomly selecting observation cells from a model of process-based reactive transport modeling (RTM);
• incorporating the observation cells into a neural network for proxy modeling;
• receiving, through a user interface, selections of training parameters for training a set of parameter-specific proxy models;
• training, using the selections of training parameters, a set of parameter-specific proxy models represented by a neural network, wherein each parameter-specific proxy model corresponds to a specific RTM parameter from a set of RTM parameters;
• performing tests using the set of parameter-specific proxy models, wherein each test tests a specific one of the parameter-specific proxy models;
• generating predictions using the set of parameter-specific proxy models, wherein generating the predictions includes using nested prediction strategies including:
• predicting, using the specific RTM parameter, a critical observation variable;
• and
• predicting, using the critical observation variable, another observation variable;
• performing 3-dimensional interpolation of the observation cells;
• and
• generating, using results of multiple execution iterations of the parameter-specific proxy model, a progress sub-panel in a user interface, wherein the progress sub-panel identifies at least a number of completed iterations and a performance value corresponding to a mean squared error achieved by each iteration.

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:
• randomly selecting observation cells from a model of process-based reactive transport modeling (RTM);
• incorporating the observation cells into a neural network for proxy modeling;
• receiving, through a user interface, selections of training parameters for training a set of parameter-specific proxy models;
• training, using the selections of training parameters, a set of parameter-specific proxy models represented by a neural network, wherein each parameter-specific proxy model corresponds to a specific RTM parameter from a set of RTM parameters;
• performing tests using the set of parameter-specific proxy models, wherein each test tests a specific one of the parameter-specific proxy models;
• generating predictions using the set of parameter-specific proxy models, wherein generating the predictions includes using nested prediction strategies including:
• predicting, using the specific RTM parameter, a critical observation variable;
• and
• predicting, using the critical observation variable, another observation variable;
• performing 3-dimensional interpolation of the observation cells;
• and
• generating, using results of multiple execution iterations of the parameter-specific proxy model, a progress sub-panel in a user interface, wherein the progress sub-panel identifies at least a number of completed iterations and a performance value corresponding to a mean squared error achieved by each iteration.