Summary of Multi-step Embed to Control: a Novel Deep Learning-based Approach For Surrogate Modelling in Reservoir Simulation, by Jungang Chen et al.
Multi-Step Embed to Control: A Novel Deep Learning-based Approach for Surrogate Modelling in Reservoir Simulation
by Jungang Chen, Eduardo Gildin, John Killough
First submitted to arxiv on: 16 Sep 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper introduces a novel deep learning-based surrogate model, called multi-step embed-to-control (E2C) model, for constructing proxy models with improved long-term prediction performance. The proposed approach builds upon existing E2C and E2CO methods by incorporating multiple forward transitions in the latent space using Koopman operator, allowing the model to consider a sequence of state snapshots during training phrases. This novel method also redesigns the loss function to accommodate these multiple transitions and respect underlying physical principles. The paper demonstrates the efficacy of the proposed framework through comparative analysis with conventional E2C models on a two-phase (oil and water) reservoir model under a waterflooding scheme, showing significant improvements in long-term simulation scenarios. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper creates a new way to make simpler models that are faster and cheaper to use. These “proxy” models can be used to predict what will happen over time, like how much oil or water will be left in a reservoir. The old ways of making these proxy models didn’t work well for long-term predictions because they only looked at one step at a time. This new method looks ahead and considers many steps at once, which makes it more accurate. The researchers tested their new method on a special kind of problem called a “two-phase” reservoir model and found that it worked much better than the old way. |
Keywords
» Artificial intelligence » Deep learning » Latent space » Loss function
