Emulating the Global Change Analysis Model with Deep Learning
Authors:
Andrew Holmes,
Matt Jensen,
Sarah Coffland,
Hidemi Mitani Shen,
Logan Sizemore,
Seth Bassetti,
Brenna Nieva,
Claudia Tebaldi,
Abigail Snyder,
Brian Hutchinson
Abstract:
The Global Change Analysis Model (GCAM) simulates complex interactions between the coupled Earth and human systems, providing valuable insights into the co-evolution of land, water, and energy sectors under different future scenarios. Understanding the sensitivities and drivers of this multisectoral system can lead to more robust understanding of the different pathways to particular outcomes. The…
▽ More
The Global Change Analysis Model (GCAM) simulates complex interactions between the coupled Earth and human systems, providing valuable insights into the co-evolution of land, water, and energy sectors under different future scenarios. Understanding the sensitivities and drivers of this multisectoral system can lead to more robust understanding of the different pathways to particular outcomes. The interactions and complexity of the coupled human-Earth systems make GCAM simulations costly to run at scale - a requirement for large ensemble experiments which explore uncertainty in model parameters and outputs. A differentiable emulator with similar predictive power, but greater efficiency, could provide novel scenario discovery and analysis of GCAM and its outputs, requiring fewer runs of GCAM. As a first use case, we train a neural network on an existing large ensemble that explores a range of GCAM inputs related to different relative contributions of energy production sources, with a focus on wind and solar. We complement this existing ensemble with interpolated input values and a wider selection of outputs, predicting 22,528 GCAM outputs across time, sectors, and regions. We report a median $R^2$ score of 0.998 for the emulator's predictions and an $R^2$ score of 0.812 for its input-output sensitivity.
△ Less
Submitted 11 December, 2024;
originally announced December 2024.
Topology Optimization through Differentiable Finite Element Solver
Authors:
Liang Chen,
Herman M. H. Shen
Abstract:
In this paper, a topology optimization framework utilizing automatic differentiation is presented as an efficient way for solving 2D density-based topology optimization problem by calculating gradients through the fully differentiable finite element solver. The optimization framework with the differentiable physics solver is proposed and tested on several classical topology optimization examples.…
▽ More
In this paper, a topology optimization framework utilizing automatic differentiation is presented as an efficient way for solving 2D density-based topology optimization problem by calculating gradients through the fully differentiable finite element solver. The optimization framework with the differentiable physics solver is proposed and tested on several classical topology optimization examples. The differentiable solver is implemented in Julia programming language and can be automatically differentiated in reverse mode to provide the pullback functions of every single operation. The entire end-to-end gradient information can be then backed up by utilizing chain rule. This framework incorporates a generator built from convolutional layers with a set of learnable parameters to propose new designs for every iteration. Since the whole process is differentiable, the parameters of the generator can be updated using any optimization algorithm given the gradient information from automatic differentiation. The proposed optimization framework is demonstrated on designing a half MBB beam and compared to the results with the ones from the efficient 88-line code. By only changing the objective function and the boundary conditions, it can run an optimization for designing a compliant mechanism, e.g. a force inverter where the output displacement is in the opposite direction of the input.
△ Less
Submitted 20 September, 2020;
originally announced September 2020.
