Jia Meng, Robert Sorensen Troy, Martina Hammerling Dorit
Department of Applied Mathematics and Statistics, Colorado School of Mines, Golden, Colorado 80401, United States.
Energy Emissions Modeling and Data Lab, The University of Texas at Austin, Austin, Texas 78712, United States.
ACS Sustain Resour Manag. 2024 Dec 18;2(1):72-81. doi: 10.1021/acssusresmgt.4c00333. eCollection 2025 Jan 23.
We propose a generic, modular framework to optimize the placement of point-in-space continuous monitoring system sensors on oil and gas sites aiming to maximize the methane emission detection efficiency. Our proposed framework substantially expands the problem scale compared to previous related studies and can be adapted for different objectives in sensor placement. This optimization framework is comprised of five steps: (1) simulate emission scenarios using site-specific wind and emission information; (2) set possible sensor locations under consideration of the site layout and any site-specific constraints; (3) simulate methane concentrations for each pair of emission scenario and possible sensor location; (4) determine emissions detection based on the site-specific simulated concentrations; and (5) select the best subset of sensor locations, under a given number of sensors to place, using genetic algorithms combined with Pareto optimization. We demonstrate the practicality and effectiveness of our framework through its application to an oil and gas emission testing facility with a large search space of possible sensor locations; a setting which is computationally infeasible to solve with commonly used mixed-integer linear programming. Additionally, a case study illustrates the successful application of our algorithm to an operating oil and gas site, showcasing its real-world applicability and effectiveness.
我们提出了一个通用的模块化框架，用于优化油气田点空间连续监测系统传感器的布局，旨在最大限度地提高甲烷排放检测效率。与之前的相关研究相比，我们提出的框架显著扩大了问题规模，并且可以适用于传感器布局的不同目标。这个优化框架由五个步骤组成：（1）使用特定场地的风和排放信息模拟排放情景；（2）在考虑场地布局和任何特定场地限制的情况下设置可能的传感器位置；（3）针对每对排放情景和可能的传感器位置模拟甲烷浓度；（4）根据特定场地模拟浓度确定排放检测；（5）在给定要放置的传感器数量的情况下，使用遗传算法结合帕累托优化选择最佳的传感器位置子集。我们通过将其应用于一个具有大量可能传感器位置搜索空间的油气排放测试设施，证明了我们框架的实用性和有效性；这种设置使用常用的混合整数线性规划来求解在计算上是不可行的。此外，一个案例研究说明了我们的算法在一个运营中的油气田的成功应用，展示了其在现实世界中的适用性和有效性。
