Methods and systems for logging while drilling and optimized telemetry using artifical intelligence
A method includes obtaining, with an edge computing device disposed in a bottom hole assembly of a drilling system, logging while drilling data. The method further includes determining, with the edge computing device, at least one interpreted quantity of a subsurface using one or more artificial intelligence (AI) models processing the logging while drilling (LWD) data. The method further includes determining, with the edge computing device, an optimal set of data transfer parameters for a telemetry system governed by a set of data transfer parameters, to telemeter the at least one interpreted quantity, where the set of data transfer parameters are continually updated to the optimal set of data transfer parameters while the logging while drilling data is being acquired. The method further includes transmitting, with the edge computing device, one or more guidance control parameters to a geosteering system based on the at least one interpreted quantity.
1 . A method, comprising:
obtaining, with an edge computing device disposed in a bottom hole assembly of a drilling system, logging while drilling (LWD) data acquired while drilling a well through a subsurface with the drilling system;
obtaining a tool uncertainty for each of one or more LWD logs comprised by the LWD data;
determining, with the edge computing device, at least one interpreted quantity of the subsurface using one or more artificial intelligence (AI) models processing the LWD data, wherein the tool uncertainty for each LWD log is propagated through the one or more AI models to obtain an uncertainty estimate for the at least one interpreted quantity;
determining, with the edge computing device, an optimal set of data transfer parameters for a telemetry system of the drilling system configured with a set of data transfer parameters, to telemeter the at least one interpreted quantity,
wherein the set of data transfer parameters are continually updated to the optimal set of data transfer parameters while the logging while drilling data is being acquired; and
transmitting, with the edge computing device, one or more guidance control parameters to a geosteering system of the drilling system based on the at least one interpreted quantity and the uncertainty estimate.
2 . The method of claim 1 , wherein the one or more guidance control parameters define a wellbore path for a wellbore drilled by the drilling system.
3 . The method of claim 1 , wherein the at least one interpreted quantity is a lithology of the subsurface as a function of depth of the well.
4 . The method of claim 1 , wherein the one or more AI models comprises a neural network.
5 . The method of claim 1 , wherein determining the optimal set of data transfer parameters comprises adjusting the set of data transfer parameters with a reinforcement learner while monitoring a bandwidth and a latency of telemetered data using the telemetry system configured with the set of data transfer parameters.
6 . The method of claim 1 , further comprising:
obtaining, with the edge computing device, LWD metadata corresponding to the LWD data,
wherein the at least one interpreted quantity comprises a first subsurface property and the one or more AI models comprises a first AI model,
wherein the first subsurface property is determined by processing at least one LWD log comprised by the LWD data with the first AI model,
wherein the first AI model is selected to process the at least one LWD log based on the LWD metadata.
7 . The method of claim 5 , wherein the reinforcement learner is a Q-learner.
8 . A non-transitory computer-readable memory comprising computer-executable instructions stored thereon that, when executed on a processor, cause the processor to perform steps comprising:
obtaining logging while drilling (LWD) data acquired while drilling a well through a subsurface with a drilling system;
obtaining a tool uncertainty for each of one or more LWD logs comprised by the LWD data;
determining at least one interpreted quantity of the subsurface using one or more artificial intelligence (AI) models processing the LWD data, wherein the tool uncertainty for each LWD log is propagated through the one or more AI models to obtain an uncertainty estimate for the at least one interpreted quantity;
determining an optimal set of data transfer parameters for a telemetry system of the drilling system configured with a set of data transfer parameters, to telemeter the at least one interpreted quantity,
wherein the set of data transfer parameters are continually updated to the optimal set of data transfer parameters while the logging while drilling data is being acquired; and
transmitting one or more guidance control parameters to a geosteering system of the drilling system based on the at least one interpreted quantity and the uncertainty estimate.
9 . The non-transitory computer-readable memory of claim 8 , wherein the one or more guidance control parameters define a wellbore path for a wellbore drilled by the drilling system.
10 . The non-transitory computer-readable memory of claim 8 , wherein the at least one interpreted quantity is a lithology of the subsurface as a function of depth of the well.
11 . The non-transitory computer-readable memory of claim 8 , wherein the one or more AI models comprises a neural network.
12 . The non-transitory computer-readable memory of claim 8 , wherein determining the optimal set of data transfer parameters comprises adjusting the set of data transfer parameters with a reinforcement learner while monitoring a bandwidth and a latency of telemetered data using the telemetry system configured with the set of data transfer parameters.
13 . The non-transitory computer-readable memory of claim 8 , wherein the computer-executable instructions, when executed on the processor, further cause the processor to perform steps:
obtaining LWD metadata corresponding to the LWD data,
wherein the at least one interpreted quantity comprises a first subsurface property and the one or more AI models comprises a first AI model,
wherein the first subsurface property is determined by processing at least one LWD log comprised by the LWD data with the first AI model,
wherein the first AI model is selected to process the at least one LWD log based on the LWD metadata.
14 . A system, comprising:
a drilling system comprising a bottom hole assembly comprising:
a drill bit to drill a well in a subsurface,
a suite of logging while drilling (LWD) tools comprising at least one LWD tool proximate the drill bit,
an edge computing device configured to obtain LWD data from the suite of LWD tools while drilling the well through the subsurface with the drilling system and to obtain a tool uncertainty for each of the at least one LWD tool, and
a telemetry system configured with a set of data transfer parameters and configured to telemeter data from the bottom hole assembly to a surface;
wherein the edge computing device is further configured to:
determine at least one interpreted quantity of the subsurface using one or more artificial intelligence (AI) models processing the LWD data, wherein the tool uncertainty for each LWD tool is propagated through the one or more AI models to obtain an uncertainty estimate for the at least one interpreted quantity,
determine an optimal set of data transfer parameters for the telemetry system to telemeter the at least one interpreted quantity, and
transmit, one or more guidance control parameters to a geosteering system of the drilling system based on the at least one interpreted quantity and the uncertainty estimate.
15 . The system of claim 14 , wherein the one or more guidance control parameters define a wellbore path for the well drilled by the drilling system.
16 . The system of claim 14 , wherein the at least one interpreted quantity is a lithology of the subsurface as a function of depth of the well.
17 . The system of claim 14 , wherein determining the optimal set of data transfer parameters comprises adjusting the set of data transfer parameters with a reinforcement learner while monitoring a bandwidth and a latency of telemetered data using the telemetry system configured with the set of data transfer parameters.