CO2 capture and storage is a viable solution in the effort to mitigate global climate change. Deep saline aquifers, in particular, have emerged as promising storage options, owing to their vast capacity and widespread distribution. However, the task of proficiently monitoring and simulating CO2 behavior within these formations poses significant challenges. To address this, we introduce the physics-constraint neural network for CO2 storage (CO2PCNet), a model specifically designed for simulating and monitoring CO2 storage in deep saline aquifers during injection and post-injection periods. Recognizing the significant challenges in accurately modeling the distribution and movement of CO2 under varying permeability conditions, the CO2PCNet integrates the principles of physics with the robustness of deep learning, serving as a powerful surrogate model. The architecture of CO2PCNet starts with an encoder that adeptly processes spatial features from overall mole fraction (zCO2) and pressure fields (Pl), capturing the complex dynamics of a CO2 trajectory. By incorporating permeability information through a conditioning step, the network ensures a faithful representation of the influences on CO2 behavior in subsurface conditions. A ConvLSTM module subsequently discerns temporal evolutions, reflecting the real-world progression of CO2 plumes within the reservoir. Lastly, the decoder precisely reconstructs the predictive spatial profile of CO2 distribution. CO2PCNet, with its integration of convolutional layers, recurrent mechanisms, and physics-informed constraints, offers a refined approach to CO2 storage simulation. This model offers the potential of utilizing advanced computational methods in advancing CCS practices.

中文翻译：

---

用于注入期间和注入后深层咸水层中 CO2 储存的物理约束神经网络


CO2 捕获和封存是缓解全球气候变化的可行解决方案。尤其是深咸含水层，由于其巨大的容量和广泛的分布，已成为有前途的储存选择。然而，熟练监测和模拟这些地层中的 CO2 行为的任务带来了重大挑战。为了解决这个问题，我们引入了用于 CO2 封存的物理约束神经网络 （CO2PCNet），该模型专门用于模拟和监测注入期间和注入后深层咸水层中的 CO2 储存。CO2PCNet 认识到在不同渗透率条件下准确模拟 CO2 分布和移动的重大挑战，将物理原理与深度学习的稳健性相结合，作为一个强大的替代模型。CO2PCNet 的架构从一个编码器开始，该编码器熟练地处理来自总摩尔分数 （zCO2） 和压力场 （Pl） 的空间特征，从而捕获 CO2 轨迹的复杂动力学。通过调节步骤整合渗透率信息，该网络可确保忠实地表示地下条件下对 CO2 行为的影响。随后，ConvLSTM 模块识别时间演变，反映储层内 CO2 羽流的真实进展。最后，解码器精确地重建了 CO2 分布的预测空间剖面。CO2PCNet 集成了卷积层、递归机制和物理学约束，为 CO2 封存仿真提供了一种精致的方法。该模型提供了利用高级计算方法来推进 CCS 实践的潜力。