Hybrid physics-AI method for accurate and fast mud-weight window calculations
A system and method for determining a mud-weight window is disclosed. The method includes determining, for each of a plurality of borehole-formation training models, a fracture mud weight using a first physics-based method and a collapse mud weight using a second physics-based method, and training, using the plurality of borehole-formation training models, an artificial intelligence (AI) network to predict the fracture and collapse mud weights for a borehole formation model. The method further includes determining, using the trained AI network, an AI fracture and an AI collapse mud weight from an observed borehole-formation model, and a search window surrounding each AI mud weight. The method further includes predicting a final fracture mud weight within the fracture search window using the first physics-based method and a final collapse mud weight within the collapse search window using the second physics-based method, and determining the mud-weight window bounded by the two final mud weights.
1 . A method for determining a mud-weight window, comprising:
determining, using a computer processor, for each of a plurality of borehole-formation training models, a fracture mud weight using a first physics-based method and a collapse mud weight using a second physics-based method;
training, using the computer processor and the plurality of borehole-formation training models, an artificial intelligence (AI) network to predict the fracture mud weight and the collapse mud weight for a borehole-formation model;
determining, using the computer processor and the trained AI network, an AI fracture mud weight and an AI collapse mud weight from an observed borehole-formation model;
determining, using the computer processor, a fracture search window surrounding the AI fracture mud weight and a collapse search window surrounding the AI collapse mud weight;
predicting, using the computer processor, a final fracture mud weight within the fracture search window using the first physics-based method and a final collapse mud weight within the collapse search window using the second physics-based method; and
determining the mud-weight window bounded by the final fracture mud weight and the final collapse mud weight,
wherein the determining the fracture search window comprises:
determining a first fracture potential for the AI fracture mud weight using a physics-based method,
determining a fracture search window extent based on a first predetermined percentage of the AI fracture mud weight,
setting a first bound of the fracture search window equal to the AI fracture mud weight, and
setting a second bound of the fracture search window equal to the AI fracture mud weight minus a product of a sign of the first fracture potential and the fracture search window extent.
2 . The method of claim 1 , further comprising:
mixing a mud with a mud-weight within the mud-weight window by adding an amount of solid material with an amount of fluid; and
drilling a borehole while pumping the mud through an interior channel of a drillstring and through at least one nozzle of a drill bit.
3 . The method of claim 1 , wherein each of the plurality of borehole-formation training models comprises at least one of a geometry, a stress and a mechanical strength parameter, a thermal parameter, and a chemoelectrical parameter.
4 . The method of claim 1 , wherein each physics-based method comprises an iterative physics-based simulation and a root-finding process.
5 . The method of claim 1 , wherein the training of the AI network comprises tuning at least one parameter of the AI network to minimize an error metric between an AI network predicted mud weight and a physics-based predicted mud weight for each of the plurality of borehole-formation training models.
6 . The method of claim 1 where the first predetermined percentage is 10%.
7 . The method of claim 1 , wherein determining the collapse search window comprises:
determining a first collapse potential for the AI collapse mud weight using a physics-based method;
determining a collapse search window extent based on a second predetermined percentage of the AI collapse mud weight;
setting a first bound of the collapse search window equal to the AI collapse mud weight; and
setting a second bound of the collapse search window equal to the AI collapse mud weight plus a product of a sign of the first collapse potential and the collapse search window extent.
8 . The method of claim 7 where the second predetermined percentage is 10%.
9 . A non-transitory computer readable medium storing instructions executable by a computer processor, the instructions comprising functionality for:
determining for each of a plurality of borehole-formation training models, a fracture mud weight using a first physics-based method and a collapse mud weight using a second physics-based method;
training, using the plurality of borehole-formation training models, an artificial intelligence (AI) network to predict the fracture mud weight and the collapse mud weight for a borehole-formation model;
determining, using the trained AI network, an AI fracture mud weight and an AI collapse mud weight from an observed borehole-formation model;
determining a fracture search window surrounding the AI fracture mud weight and a collapse search window surrounding the AI collapse mud weight;
predicting a final fracture mud weight within the fracture search window using the first physics-based method and a final collapse mud weight within the collapse search window using the second physics-based method; and
determining the mud-weight window bounded by the final fracture mud weight and the final collapse mud weight,
wherein the determining the fracture search window comprises:
determining a first fracture potential for the AI fracture mud weight using a physics-based method;
determining a fracture search window extent based on a first predetermined percentage of the AI fracture mud weight;
setting a first bound of the fracture search window equal to the AI fracture mud weight; and
setting a second bound of the fracture search window equal to the AI fracture mud weight minus a product of a sign of the first fracture potential and the fracture search window extent.
10 . The non-transitory computer readable medium of claim 9 , wherein each of the plurality of borehole-formation training models comprise at least one of a geometry, a stress and a mechanical strength parameter, a thermal parameter, and a chemoelectrical parameter.
11 . The non-transitory computer readable medium of claim 9 , wherein each physics-based method comprises an iterative physics-based simulation and a root-finding process.
12 . The non-transitory computer readable medium of claim 9 , wherein the training of the AI network comprises tuning at least one parameter of the AI network to minimize an error metric between an AI network predicted mud weight and a physics-based predicted mud weight for each of the plurality of borehole-formation training models.
13 . The non-transitory computer readable medium of claim 10 where the first predetermined percentage is 10%.
14 . The non-transitory computer readable medium of claim 9 , wherein determining the collapse search window comprises:
determining a first collapse potential for the AI collapse mud weight using a physics-based method;
determining a collapse search window extent based on a second predetermined percentage of the AI collapse mud weight;
setting a first bound of the collapse search window equal to the AI collapse mud weight; and
setting a second bound of the collapse search window equal to the AI collapse mud weight plus a product of a sign of the first collapse potential and the collapse search window extent.
15 . The non-transitory computer readable medium of claim 14 where the second predetermined percentage is 10%.
16 . A system, comprising:
a computer processor configured to:
determine, for each of a plurality of borehole-formation training models, a fracture mud weight using a first physics-based method and a collapse mud weight using a second physics-based method,
train, using the plurality of borehole-formation training models, an artificial intelligence (AI) network to predict the fracture mud weight and the collapse mud weight for a borehole-formation model,
determine, using the trained AI network, an AI fracture mud weight and an AI collapse mud weight from an observed borehole-formation model,
determine a fracture search window surrounding the AI fracture mud weight and a collapse search window surrounding the AI collapse mud weight,
predict a final fracture mud weight within the fracture search window using the first physics-based method and a final collapse mud weight within the collapse search window using the second physics-based method, and
determine the mud-weight window bounded by the final fracture mud weight and the final collapse mud weight; and
a drilling system configured to:
mix a mud with a mud weight within the mud-weight window by adding an amount of solid material with an amount of fluid, and
drill a borehole while pumping the mud through an interior channel of a drillstring and through at least one nozzle of a drill bit,
wherein the determining the fracture search window comprises:
determining a first fracture potential for the AI fracture mud weight using a physics-based method;
determining a fracture search window extent based on a first predetermined percentage of the AI fracture mud weight;
setting a first bound of the fracture search window equal to the AI fracture mud weight; and
setting a second bound of the fracture search window equal to the AI fracture mud weight minus a product of a sign of the first fracture potential and the fracture search window extent.
17 . The system of claim 16 , wherein determining the collapse search window comprises:
determining a first collapse potential for the AI collapse mud weight using a physics-based method;
determining a collapse search window extent based on a second predetermined percentage of the AI collapse mud weight;
setting a first bound of the collapse search window equal to the AI collapse mud weight; and
setting a second bound of the collapse search window equal to the AI collapse mud weight plus a product of a sign of the first collapse potential and the collapse search window extent.