当前位置: X-MOL 学术J. Process Control › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Application of Koopman operator for model-based control of fracture propagation and proppant transport in hydraulic fracturing operation
Journal of Process Control ( IF 4.2 ) Pub Date : 2020-07-01 , DOI: 10.1016/j.jprocont.2020.05.003
Abhinav Narasingam , Joseph Sang-Il Kwon

Abstract This work explores the application of the recently developed Koopman operator approach for model identification and feedback control of a hydraulic fracturing process. Controlling fracture propagation and proppant transport with precision is a challenge due in large part to the difficulty of constructing approximate models that accurately capture the characteristic moving boundary and highly-coupled dynamics exhibited by the process. Koopman operator theory is particularly attractive here as it offers a way to explicitly construct linear representations for even highly nonlinear dynamics. The method is data-driven and relies on lifting the states to an infinite-dimensional space of functions called observables where the dynamics are governed by a linear Koopman operator. This work considers two problems: (a) fracture geometry control, and (b) proppant concentration control. In both cases, an approximate linear model of the corresponding dynamics is constructed and used to design a model predictive controller (MPC). The manuscript shows that in the case of highly nonlinear dynamics, as observed in the proppant concentration, use of canonical functions in the observable basis fails. In such cases, a priori system knowledge can be leveraged to choose the required basis. The numerical experiments demonstrate that the Koopman linear model shows excellent agreement with the real system and successfully achieves the desired target values maximizing the oil and gas productivity. Additionally, due to its linear structure, the Koopman models allow convex MPC formulations that avoid any issues associated with nonlinear optimization.

中文翻译:

Koopman算子在水力压裂作业中基于模型的裂缝扩展和支撑剂输送控制中的应用

摘要 这项工作探讨了最近开发的 Koopman 算子方法在水力压裂过程的模型识别和反馈控制中的应用。精确控制裂缝扩展和支撑剂输送是一项挑战,这在很大程度上是因为难以构建近似模型来准确捕捉该过程所表现出的特征移动边界和高度耦合的动力学。Koopman 算子理论在这里特别有吸引力,因为它提供了一种为甚至高度非线性动力学显式构建线性表示的方法。该方法是数据驱动的,并依赖于将状态提升到称为可观察量的无限维函数空间,其中动态由线性 Koopman 算子控制。这项工作考虑了两个问题:(a)裂缝几何控制,(b) 支撑剂浓度控制。在这两种情况下,都会构建相应动力学的近似线性模型并用于设计模型预测控制器 (MPC)。手稿表明,在高度非线性动力学的情况下,如在支撑剂浓度中观察到的那样,在可观察基础中使用规范函数失败。在这种情况下,可以利用先验系统知识来选择所需的基础。数值实验表明,Koopman 线性模型与实际系统显示出极好的一致性,并成功实现了所需的目标值,从而最大限度地提高了油气产能。此外,由于其线性结构,Koopman 模型允许凸 MPC 公式,避免与非线性优化相关的任何问题。构建相应动力学的近似线性模型并用于设计模型预测控制器 (MPC)。手稿表明,在高度非线性动力学的情况下,如在支撑剂浓度中观察到的那样,在可观察基础中使用规范函数失败。在这种情况下,可以利用先验系统知识来选择所需的基础。数值实验表明,Koopman 线性模型与实际系统显示出极好的一致性,并成功实现了所需的目标值,从而最大限度地提高了油气产能。此外,由于其线性结构,Koopman 模型允许凸 MPC 公式,避免与非线性优化相关的任何问题。构建相应动力学的近似线性模型并用于设计模型预测控制器 (MPC)。手稿表明,在高度非线性动力学的情况下,如在支撑剂浓度中观察到的那样,在可观察基础中使用规范函数失败。在这种情况下,可以利用先验系统知识来选择所需的基础。数值实验表明,Koopman 线性模型与实际系统显示出极好的一致性,并成功实现了所需的目标值,从而最大限度地提高了油气产能。此外,由于其线性结构,Koopman 模型允许凸 MPC 公式,避免与非线性优化相关的任何问题。
更新日期:2020-07-01
down
wechat
bug