Joseph Peter Morris
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Joseph Peter Morris has a total of 215 co-authors. Here are the top co-authors by number of shared publications:
- Pengcheng Fu
Atmospheric, Earth, and Energy Division Lawrence Livermore National Laboratory Livermore CA USA
8 shared publications Randolph R. Settgast
Lawrence Livermore National Laboratory
7 shared publicationsFrederick J. Ryerson
Lawrence Livermore National Laboratory
6 shared publicationsJixiang Huang
Lawrence Livermore National Laboratory
6 shared publications- Christopher S. Sherman
Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)
5 shared publications Pengcheng Fu
5 shared publications
Latest Publications
- Open Access
- 23 November 2024
- Published by Elsevier in Geothermics
- Vol. 126
The EGS Collab project: Outcomes and lessons learned from hydraulic fracture stimulations in crystalline rock at 1.25 and 1.5 km depth
- 23 November 2024
- Published by Elsevier in Geothermics
- Vol. 126
Open AccessThe EGS Collab project: Outcomes and lessons learned from hydraulic fracture stimulations in crystalline rock at 1.25 and 1.5 km depth
- 23 August 2024
- Published by Elsevier in Engineering Fracture Mechanics
- Vol. 310
The damage Mechanics challenge Results: Participant predictions compared with experiment
- 23 August 2024
- Published by Elsevier in Engineering Fracture Mechanics
- Vol. 310
The damage Mechanics challenge Results: Participant predictions compared with experiment
- Open Access
- 1 January 2024
- Published by Elsevier in SSRN Electronic Journal
SMART – a Comprehensive Research and Development Program to Demonstrate Application of Machine Learning for Supporting CCS Deployment
The objective of the US Department of Energy’s SMART Initiative, i.e., Science-informed Machine Learning (ML) for Accelerating Real-Time Decisions in Subsurface Applications, is to showcase how the utilization of ML can significantly improve efficiency and effectiveness of field-scale commercial carbon storage operations. This paper will present the results from the current phase of SMART (field deployment) for demonstrating the applicability of ML-based tools and workflows for: (a) virtual learning during the pre-injection permitting phase, (b) advanced storage reservoir imaging to better characterize fractures and faults, and (c) dynamic storage reservoir modelling and optimization to inform operational decision making and visualization of system evolution. (a) Virtual Learning: The current state of the art in evaluating forecasting models submitted as part of the EPA Class VI permit applications is based on outputs from time consuming and detailed full physics simulations. In the SMART Initiative, we are developing an alternative “virtual learning” strategy that utilizes ML-based surrogate models for fast prediction. These ML-based models can provide a rapid assessment of the full suite of pre-existing model results using the SMART Model Explorer interface, and also evaluate the impact of various what-if scenarios on the Area of Review (i.e., outline of critical pressure contours and CO2 plume boundary) in seconds. We demonstrate performance of these tools using synthetic data from typical CCS projects. (b) Storage Reservoir Imaging: In the SMART Initiative, we are using Machine Learning in conjunction with passive seismic measurements during injection to correlate the location of fractures and faults, as well as estimate the orientation of movement along these features. This can help supplement the traditional interpretation that is fed into the static earth model to obtain a more detailed representation of these fracture and fault locations. In particular, the identification of the faults and fractures in basement rocks is important input to geomechanics models for predicting the potential for induced seismicity following injection. We demonstrate this using geophysical data from the Illinois Basin Decatur (IBDP) Project. (c) Dynamic Storage Reservoir Modeling and Optimization: Full-physics reservoir models serve a very important role in CO2 storage reservoir management by providing detailed predictions of pressure and saturation distributions in the project area as a function of time. Typically, these simulations are carried out using multi-million cell geologic framework models in a compositional fluid modelling framework, which may take hours for a single simulation. In the SMART Initiative, our approach significantly alleviates the computational challenges by using ML-based surrogate forward models for history matching and optimization that speeds up both steps considerably and allows for inputs to (near) real-time decision making. We demonstrate this using operational pressure and temperature data from the IBDP project.
- 1 January 2024
- Published by Elsevier in SSRN Electronic Journal
Open AccessSMART – a Comprehensive Research and Development Program to Demonstrate Application of Machine Learning for Supporting CCS Deployment
The objective of the US Department of Energy’s SMART Initiative, i.e., Science-informed Machine Learning (ML) for Accelerating Real-Time Decisions in Subsurface Applications, is to showcase how the utilization of ML can significantly improve efficiency and effectiveness of field-scale commercial carbon storage operations. This paper will present the results from the current phase of SMART (field deployment) for demonstrating the applicability of ML-based tools and workflows for: (a) virtual learning during the pre-injection permitting phase, (b) advanced storage reservoir imaging to better characterize fractures and faults, and (c) dynamic storage reservoir modelling and optimization to inform operational decision making and visualization of system evolution. (a) Virtual Learning: The current state of the art in evaluating forecasting models submitted as part of the EPA Class VI permit applications is based on outputs from time consuming and detailed full physics simulations. In the SMART Initiative, we are developing an alternative “virtual learning” strategy that utilizes ML-based surrogate models for fast prediction. These ML-based models can provide a rapid assessment of the full suite of pre-existing model results using the SMART Model Explorer interface, and also evaluate the impact of various what-if scenarios on the Area of Review (i.e., outline of critical pressure contours and CO2 plume boundary) in seconds. We demonstrate performance of these tools using synthetic data from typical CCS projects. (b) Storage Reservoir Imaging: In the SMART Initiative, we are using Machine Learning in conjunction with passive seismic measurements during injection to correlate the location of fractures and faults, as well as estimate the orientation of movement along these features. This can help supplement the traditional interpretation that is fed into the static earth model to obtain a more detailed representation of these fracture and fault locations. In particular, the identification of the faults and fractures in basement rocks is important input to geomechanics models for predicting the potential for induced seismicity following injection. We demonstrate this using geophysical data from the Illinois Basin Decatur (IBDP) Project. (c) Dynamic Storage Reservoir Modeling and Optimization: Full-physics reservoir models serve a very important role in CO2 storage reservoir management by providing detailed predictions of pressure and saturation distributions in the project area as a function of time. Typically, these simulations are carried out using multi-million cell geologic framework models in a compositional fluid modelling framework, which may take hours for a single simulation. In the SMART Initiative, our approach significantly alleviates the computational challenges by using ML-based surrogate forward models for history matching and optimization that speeds up both steps considerably and allows for inputs to (near) real-time decision making. We demonstrate this using operational pressure and temperature data from the IBDP project.
- Open Access
- 23 September 2022
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Atmospheres
- Vol. 127 (18)
Super‐Droplet Method to Simulate Lagrangian Microphysics of Nuclear Fallout in a Homogeneous Cloud
Nuclear detonations produce hazardous local and global particles or fallout. Predicting fallout size, chemical components, and location is necessary to inform officials and determine immediate guidance for the public. However, existing nuclear detonation fallout models prescribe the particle size distributions based on limited observations. In this work, we apply the super‐droplet method, which is a numerical modeling technique developed for cloud microphysics, to simulate size distributions of particles in a mushroom cloud formed post‐detonation of a nuclear device. We model fallout formation and evolution with homogeneous nucleation and condensation of a single species and a Monte Carlo coagulation algorithm. We verify the numerical methods representing coagulation and condensation processes against analytical test problems. Additionally, we explore several scenarios for the integral system mass and yield in equivalent kilotons (kt) of TNT (trinitrotoluene). The fallout size distribution median diameter dpg follows a scaling law based on the integral system mass mv0 kg and yield Y kt: nm. We test the effect of cloud turbulence, enhanced nucleation and growth, and vapor volatility with a sensitivity study. The range in median diameter predictions for simulations of historical tests performed over the Pacific encompass the measurements of particles sampled from the cloud caps. Predicted median particle size ranges up to 217, 123, 86, and 35 nm for historical tests with yields of 0.2, 0.7, 2, and 10 Mt, respectively. This work can be expanded in many different directions to build a more predictive model for fallout formation.
- 23 September 2022
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Atmospheres
- Vol. 127 (18)
Open AccessSuper‐Droplet Method to Simulate Lagrangian Microphysics of Nuclear Fallout in a Homogeneous Cloud
Nuclear detonations produce hazardous local and global particles or fallout. Predicting fallout size, chemical components, and location is necessary to inform officials and determine immediate guidance for the public. However, existing nuclear detonation fallout models prescribe the particle size distributions based on limited observations. In this work, we apply the super‐droplet method, which is a numerical modeling technique developed for cloud microphysics, to simulate size distributions of particles in a mushroom cloud formed post‐detonation of a nuclear device. We model fallout formation and evolution with homogeneous nucleation and condensation of a single species and a Monte Carlo coagulation algorithm. We verify the numerical methods representing coagulation and condensation processes against analytical test problems. Additionally, we explore several scenarios for the integral system mass and yield in equivalent kilotons (kt) of TNT (trinitrotoluene). The fallout size distribution median diameter dpg follows a scaling law based on the integral system mass mv0 kg and yield Y kt: nm. We test the effect of cloud turbulence, enhanced nucleation and growth, and vapor volatility with a sensitivity study. The range in median diameter predictions for simulations of historical tests performed over the Pacific encompass the measurements of particles sampled from the cloud caps. Predicted median particle size ranges up to 217, 123, 86, and 35 nm for historical tests with yields of 0.2, 0.7, 2, and 10 Mt, respectively. This work can be expanded in many different directions to build a more predictive model for fallout formation.
- Open Access
- 2 September 2022
- Published by Elsevier in International Journal of Greenhouse Gas Control
- Vol. 120
Deep learning-accelerated 3D carbon storage reservoir pressure forecasting based on data assimilation using surface displacement from InSAR
- 2 September 2022
- Published by Elsevier in International Journal of Greenhouse Gas Control
- Vol. 120
Open AccessDeep learning-accelerated 3D carbon storage reservoir pressure forecasting based on data assimilation using surface displacement from InSAR
- 7 August 2022
- Published by Springer Nature in Transport in Porous Media
- Vol. 144 (3) , 797-823
Effects of Heterogeneous Fracture Aperture on Multiphase Production from Shale Reservoirs
- 7 August 2022
- Published by Springer Nature in Transport in Porous Media
- Vol. 144 (3) , 797-823
Effects of Heterogeneous Fracture Aperture on Multiphase Production from Shale Reservoirs
- 22 June 2022
- Published by Springer Nature in Rock Mechanics and Rock Engineering
- Vol. 55 (9) , 5453-5466
Controlling Hydraulic Fracture Growth Through Precise Vertical Placement of Lateral Wells: Insights from HFTS Experiment and Numerical Validation
- 22 June 2022
- Published by Springer Nature in Rock Mechanics and Rock Engineering
- Vol. 55 (9) , 5453-5466
Controlling Hydraulic Fracture Growth Through Precise Vertical Placement of Lateral Wells: Insights from HFTS Experiment and Numerical Validation
- Open Access
- 1 January 2022
- Published by Elsevier in SSRN Electronic Journal
Deep Learning-Accelerated 3d Carbon Storage Reservoir Pressure Forecasting Based on Data Assimilation Using Surface Displacement from Insar
Fast forecasting of the reservoir pressure distribution during geologic carbon storage (GCS) by assimilating monitoring data is a challenging problem. Due to high drilling cost, GCS projects usually have spatially sparse measurements from few wells, leading to high uncertainties in reservoir pressure prediction. To address this challenge, we use low-cost Interferometric Synthetic-Aperture Radar (InSAR) data as monitoring data to infer reservoir pressure build up. We develop a deep learning-accelerated workflow to assimilate surface displacement maps interpreted from InSAR and to forecast dynamic reservoir pressure. Employing an Ensemble Smoother Multiple Data Assimilation (ES-MDA) framework, the workflow updates three-dimensional (3D) geologic properties and predicts reservoir pressure with quantified uncertainties. We use a synthetic commercial-scale GCS model with bimodally distributed permeability and porosity to demonstrate the efficacy of the workflow. A two-step CNN-PCA approach is employed to parameterize the bimodal fields. The computational efficiency of the workflow is boosted by two residual U-Net based surrogate models for surface displacement and reservoir pressure predictions, respectively. The workflow can complete data assimilation and reservoir pressure forecasting in half an hour on a personal computer.
- 1 January 2022
- Published by Elsevier in SSRN Electronic Journal
Open AccessDeep Learning-Accelerated 3d Carbon Storage Reservoir Pressure Forecasting Based on Data Assimilation Using Surface Displacement from Insar
Fast forecasting of the reservoir pressure distribution during geologic carbon storage (GCS) by assimilating monitoring data is a challenging problem. Due to high drilling cost, GCS projects usually have spatially sparse measurements from few wells, leading to high uncertainties in reservoir pressure prediction. To address this challenge, we use low-cost Interferometric Synthetic-Aperture Radar (InSAR) data as monitoring data to infer reservoir pressure build up. We develop a deep learning-accelerated workflow to assimilate surface displacement maps interpreted from InSAR and to forecast dynamic reservoir pressure. Employing an Ensemble Smoother Multiple Data Assimilation (ES-MDA) framework, the workflow updates three-dimensional (3D) geologic properties and predicts reservoir pressure with quantified uncertainties. We use a synthetic commercial-scale GCS model with bimodally distributed permeability and porosity to demonstrate the efficacy of the workflow. A two-step CNN-PCA approach is employed to parameterize the bimodal fields. The computational efficiency of the workflow is boosted by two residual U-Net based surrogate models for surface displacement and reservoir pressure predictions, respectively. The workflow can complete data assimilation and reservoir pressure forecasting in half an hour on a personal computer.