Microbiome based systems, apparatus and methods for the exploration and production of hydrocarbons
There are provided methods, systems and processes for the utilization of microbial and related genetic information for use in the exploration, determination, production and recovery of natural resources, including energy sources, and the monitoring, control and analysis of processes and activities.
1. A system comprising:
a database storing extraction data corresponding to one or more sample materials obtained at a resource production field, the extraction data including microbiome data corresponding to the one or more sample materials; and
a computer system including one or more memories and one or more processors, one or more computer instructions incorporated into the one or more memories to configure the one or more processors to perform operations for improving a predictive accuracy of predictive microbiome data with respect to a target environmental parameter at the resource production field, the operations comprising:
accessing the database to obtain the microbiome data;
selecting a predictive model from a plurality of predictive models based on an expected prediction error of the predictive model being lower than an expected prediction error of other models of the plurality of predictive models, the expected prediction error of the predictive model being based on repeated fittings of the predictive model to different subsets of the microbiome data or other related microbiome data;
generating the predictive microbiome data from the first subset of the microbiome data, the generating of the predictive microbiome data including applying the predictive model using a machine-learned combination of feature inputs and model parameters; and
causing an interactive display of a visualization of the predictive microbiome data in a graphical user interface to assist an operator in creating a plan for directing an activity at the resource production field.
2. The system of claim 1 , further comprising selecting an abundance of genetic material identified in the microbiome data as one of the feature inputs based on the machine-learned combination indicating that the abundance discriminates between classifications of data points within the microbiome data, the classifications relating to the target environmental parameter.
3. The system of claim 1 , wherein the plurality of models is selected from a class of models based on structures inherent in the first subset of the microbiome data.
4. The system of claim 1 , wherein a constraint inherent in the microbiome data includes a diversity of data and the class of models includes generative models appropriate for the diversity.
5. The system of claim 1 , wherein the target environmental parameter relates to hydrocarbon exploration or production at the resource production field.
6. The system of claim 1 , wherein the target environmental parameter relates to at least one of subsurface flow communication or reservoir connectivity.
7. The system of claim 1 , wherein the target environmental parameter relates to at least one of oil saturation or permeability of a well zone.
8. The system of claim 1 , wherein the target environmental parameter relates to wettability of a reservoir.
9. The system of claim 1 , wherein the target environmental parameter relates to at least one of viscosity, temperature, pressure, porosity, or compressibility of oil or water in a reservoir.
10. A method comprising:
performing, using one or more processors, operations for improving a predictive accuracy of predictive microbiome data with respect to a target environmental parameter at a resource production field, the operations comprising:
accessing a database to obtain the microbiome data, the database storing extraction data corresponding to one or more sample materials obtained at the resource production field, the extraction data including the microbiome data;
selecting a predictive model from a plurality of predictive models based on an expected prediction error of the predictive model being lower than an expected prediction error of other models of the plurality of predictive models, the expected prediction error of the predictive model being based on repeated fittings of the predictive model to different subsets of the microbiome data or other related microbiome data;
generating the predictive microbiome data from the microbiome data, the generating of the predictive microbiome data including applying the predictive model using a machine-learned combination of feature inputs and model parameters; and
causing an interactive display of a visualization of the predictive microbiome data in a graphical user interface to assist an operator in creating a plan for directing an activity at the resource production field.
11. The method of claim 10 , further comprising selecting an abundance of genetic material identified in the microbiome data as one of the feature inputs based on the machine-learned combination indicating that the abundance discriminates between classifications of data points within the microbiome data, the classifications relating to the target environmental parameter.
12. The system of claim 10 , wherein the plurality of models is selected from a class of models based on structures inherent in the first subset of the microbiome data.
13. The system of claim 10 , wherein a constraint inherent in the microbiome data includes a diversity of data and the class of models includes generative models appropriate for the diversity.
14. A non-transitory machine-readable storage medium embodying instructions that, when executed by one or more processors, cause the one or more processors to perform operations for improving a predictive accuracy of predictive microbiome data with respect to a target environmental parameter at a resource production field, the operations comprising:
accessing a database to obtain the microbiome data, the database storing extraction data corresponding to one or more sample materials obtained at the resource production field, the extraction data including the microbiome data;
selecting a predictive model from a plurality of predictive models based on an expected prediction error of the predictive model being lower than an expected prediction error of other models of the plurality of predictive models, the expected prediction error of the predictive model being based on repeated fittings of the predictive model to different subsets of the microbiome data or other related microbiome data;
generating the predictive microbiome data from the microbiome data, the generating of the predictive microbiome data including applying the predictive model using a machine-learned combination of feature inputs and model parameters; and
causing an interactive display of a visualization of the predictive microbiome data in a graphical user interface to assist an operator in creating a plan for directing an activity at the resource production field.
15. The non-transitory machine-readable storage medium of claim 14 , further comprising selecting an abundance of genetic material identified in the microbiome data as one of the feature inputs based on the machine-learned combination indicating that the abundance discriminates between classifications of data points within the microbiome data, the classifications relating to the target environmental parameter.
16. The non-transitory machine-readable storage medium of claim 15 , wherein the plurality of models is selected from a class of models based on structures inherent in the first subset of the microbiome data.
17. The non-transitory machine-readable storage medium of claim 14 , wherein a constraint inherent in the microbiome data includes a diversity of data and the class of models includes generative models appropriate for the diversity.