Prediction of Subsurface Fluid Flow via Physics Informed Neural Networks
Geoenergies such as underground gas storage and energy storage, geothermal energy and geologic carbon storage are key technologies on the way to the foreseeable energy transition. The reservoir characterization in these projects remains challenging since predictive modeling approaches face limitations in identifying the spatial distribution of distinct lithologies and their hydro-mechanical properties from downwell testing procedures. Pumping tests are usually carried out to infer permeability, but offer only few observation points in space and require large extrapolation through inversion. The application of Physics Informed Neural Networks (PINN) offers a promising solution which can seamlessly incorporate field data, while enforcing the accordance with physical laws in the domain of study. This concept is implemented via two distinct loss terms for both the physical constraints and for the observational data in the loss function of an Artificial Neural Network (ANN). The physics-informed loss term contains a mass balance equation consisting of a storage and a diffusion component. The process is considered to be purely hydraulic and Darcy flow is assumed. The observational loss term compares the output of the ANN to a set of training data. This set consists of the system’s initial fluid pressure, as well as of a fluid pressure time series at the domain boundaries and at the borehole location. Preliminary results suggest that our PINN model is able to forecast the spatiotemporal fluid pressure distribution in a 2D domain for a variety of pumping test schemes. In this way, we give a first impression of the opportunities that PINN applications offer in the field of reservoir modeling.
| Main Authors: | , , |
|---|---|
| Format: | comunicación de congreso biblioteca |
| Language: | English |
| Published: |
2023-03
|
| Subjects: | Physics Informed Neural Networks, Fluid flow, |
| Online Access: | http://hdl.handle.net/10261/310932 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| id |
dig-idaea-es-10261-310932 |
|---|---|
| record_format |
koha |
| spelling |
dig-idaea-es-10261-3109322023-06-08T05:49:14Z Prediction of Subsurface Fluid Flow via Physics Informed Neural Networks Walter, Linus Parisio, Francesco Vilarrasa, Víctor Physics Informed Neural Networks Fluid flow Geoenergies such as underground gas storage and energy storage, geothermal energy and geologic carbon storage are key technologies on the way to the foreseeable energy transition. The reservoir characterization in these projects remains challenging since predictive modeling approaches face limitations in identifying the spatial distribution of distinct lithologies and their hydro-mechanical properties from downwell testing procedures. Pumping tests are usually carried out to infer permeability, but offer only few observation points in space and require large extrapolation through inversion. The application of Physics Informed Neural Networks (PINN) offers a promising solution which can seamlessly incorporate field data, while enforcing the accordance with physical laws in the domain of study. This concept is implemented via two distinct loss terms for both the physical constraints and for the observational data in the loss function of an Artificial Neural Network (ANN). The physics-informed loss term contains a mass balance equation consisting of a storage and a diffusion component. The process is considered to be purely hydraulic and Darcy flow is assumed. The observational loss term compares the output of the ANN to a set of training data. This set consists of the system’s initial fluid pressure, as well as of a fluid pressure time series at the domain boundaries and at the borehole location. Preliminary results suggest that our PINN model is able to forecast the spatiotemporal fluid pressure distribution in a 2D domain for a variety of pumping test schemes. In this way, we give a first impression of the opportunities that PINN applications offer in the field of reservoir modeling. Peer reviewed 2023-06-08T05:49:14Z 2023-06-08T05:49:14Z 2023-03 comunicación de congreso EGU General Assembly 2022, Geophysical Research Abstracts, Vol. 24, Vienna, Austria, 23-27 May 2022 http://hdl.handle.net/10261/310932 en Sí open |
| institution |
IDAEA ES |
| collection |
DSpace |
| country |
España |
| countrycode |
ES |
| component |
Bibliográfico |
| access |
En linea |
| databasecode |
dig-idaea-es |
| tag |
biblioteca |
| region |
Europa del Sur |
| libraryname |
Biblioteca del IDAEA España |
| language |
English |
| topic |
Physics Informed Neural Networks Fluid flow Physics Informed Neural Networks Fluid flow |
| spellingShingle |
Physics Informed Neural Networks Fluid flow Physics Informed Neural Networks Fluid flow Walter, Linus Parisio, Francesco Vilarrasa, Víctor Prediction of Subsurface Fluid Flow via Physics Informed Neural Networks |
| description |
Geoenergies such as underground gas storage and energy storage, geothermal energy and
geologic carbon storage are key technologies on the way to the foreseeable energy transition. The
reservoir characterization in these projects remains challenging since predictive modeling
approaches face limitations in identifying the spatial distribution of distinct lithologies and their
hydro-mechanical properties from downwell testing procedures. Pumping tests are usually carried
out to infer permeability, but offer only few observation points in space and require large
extrapolation through inversion. The application of Physics Informed Neural Networks (PINN)
offers a promising solution which can seamlessly incorporate field data, while enforcing the
accordance with physical laws in the domain of study. This concept is implemented via two distinct
loss terms for both the physical constraints and for the observational data in the loss function of
an Artificial Neural Network (ANN). The physics-informed loss term contains a mass balance
equation consisting of a storage and a diffusion component. The process is considered to be
purely hydraulic and Darcy flow is assumed. The observational loss term compares the output of
the ANN to a set of training data. This set consists of the system’s initial fluid pressure, as well as
of a fluid pressure time series at the domain boundaries and at the borehole location. Preliminary
results suggest that our PINN model is able to forecast the spatiotemporal fluid pressure
distribution in a 2D domain for a variety of pumping test schemes. In this way, we give a first
impression of the opportunities that PINN applications offer in the field of reservoir modeling. |
| format |
comunicación de congreso |
| topic_facet |
Physics Informed Neural Networks Fluid flow |
| author |
Walter, Linus Parisio, Francesco Vilarrasa, Víctor |
| author_facet |
Walter, Linus Parisio, Francesco Vilarrasa, Víctor |
| author_sort |
Walter, Linus |
| title |
Prediction of Subsurface Fluid Flow via Physics Informed Neural Networks |
| title_short |
Prediction of Subsurface Fluid Flow via Physics Informed Neural Networks |
| title_full |
Prediction of Subsurface Fluid Flow via Physics Informed Neural Networks |
| title_fullStr |
Prediction of Subsurface Fluid Flow via Physics Informed Neural Networks |
| title_full_unstemmed |
Prediction of Subsurface Fluid Flow via Physics Informed Neural Networks |
| title_sort |
prediction of subsurface fluid flow via physics informed neural networks |
| publishDate |
2023-03 |
| url |
http://hdl.handle.net/10261/310932 |
| work_keys_str_mv |
AT walterlinus predictionofsubsurfacefluidflowviaphysicsinformedneuralnetworks AT parisiofrancesco predictionofsubsurfacefluidflowviaphysicsinformedneuralnetworks AT vilarrasavictor predictionofsubsurfacefluidflowviaphysicsinformedneuralnetworks |
| _version_ |
1781881367381934080 |