考虑热储参数不确定性影响的干热岩热开采能力随机评价方法研究

The evaluation of heat extraction capacity of hot dry rock (HDR) is one key problem in the exploration of hot dry rock geothermal energy. Most of the popular evaluation methods are based on deterministic analysis, in which the uncertainty of reservoir parameters is neglected, leading to the actual operation and overall performance of the HDR system deviate from the original design substantially. To address with this problem, the applicant proposes to carry out the following research tasks: (1) Using the random field theory, we develop an accurate and efficient random filed generation method for the representative reservoir parameters with uncertainty based on the Karhunen–Loève (K-L) expansion considering the optimal truncation of eigenvalue and eigenfunction. (2) Inspired by the “multilevel grid”, we propose a novel stochastic method to evaluate the heat extraction capacity of HDR based on a new full multigrid multilevel quasi-Monte Carlo (FMG-MLQMC) model, in which the sampling of thermal-hydraulic-mechanics-chemical (THMC) coupling field can be implemented efficiently with sample recycling technique；(3) Based on above research, the reservoir temperature decline, production temperature and total heat extraction are investigated in detail with and without consideration of uncertainty of reservoir parameters during the whole lifespan of HDR under given completion and operation parameters. Then we carefully reveal the influence of reservoir parameters uncertainty on the evaluation of heat extraction capacity of HDR system in a prescribed confidence interval. By combining these three aspects of research, we expect to provide a theoretical foundation for accurate and scientific evaluation of the heat extraction capacity of HDR system.

干热岩热开采能力评价是干热岩型地热能开发的关键环节。现有评价方法多为确定性分析方法，未考虑热储地质参数不确定性的影响，导致干热岩系统前期设计与实际运行偏差较大，无法达到预期效果。本项目针对这一科学问题，开展如下研究：(1)采用随机场理论，发展基于最优截断Karhunen–Loève展开的干热岩热储典型地质参数三维随机场准确高效生成方法； (2)基于“多层次网格”思想，构建全多重网格—多层次拟蒙特卡罗干热岩热开采能力随机评价模型，实现对热储热-流-固-化多场的高效抽样和样本回收利用；(3)根据所发展方法和模型，在固定的开发方案下，研究分析考虑和不考虑热储地质参数不确定性影响时干热岩全生命周期热储温降、采出温度、生产井采热量等评价指标的差异，在一定置信区间（可靠性）内定量揭示热储地质参数不确定性对干热岩系统热开采能力评价的影响。本项目研究可为干热岩热开采能力的科学评价提供理论支撑。

干热岩热开采能力评价是干热岩型地热能开发的关键环节。现有评价方法多为确定性分析方法，未考虑热储地质参数不确定性的影响，导致干热岩系统前期设计与实际运行偏差较大。本项目针对这一科学问题，研究了干热岩热储多孔介质渗流参数随机场生成的Karhunen–Loève(K-L)展开法，发展了考虑多孔介质参数不确定性影响的全多重网格-多层次蒙特卡罗干热岩渗流高效随机评价方法，建立了基于POD-DEIM和Multigrid-DEIM低阶模型的多孔介质可压缩渗流高效计算方法，主要结论有:(1)K-L展开法生成渗流参数随机场时截断项数与特征值贡献率、协方差函数参数、网格疏密等因素密切相关，特征值贡献率越大、网格划分越密、随机场刻画越细致，截断项数越大。(2)基于全多重网格-多层次蒙特卡罗的干热岩渗流高效随机评价方法采用全多重网格方法求解干热岩热储多孔介质渗流方程，采用多层次蒙特卡罗方法在不同网格层次上进行抽样和样本回收利用，渗流速度的均值、方差的抽样计算效率较传统多层次蒙特卡罗方法提高了约10~40倍。(3)基于POD-DEIM和Multigrid-DEIM低阶模型的多孔介质可压缩渗流高效计算方法，与传统基于有限差分的半隐顺序求解方法求解多孔介质渗流压力、达西速度和压缩因子的计算精度基本相当，但计算效率比传统方法提高约2个数量级；POD基函数和离散经验插值点数目的选取对计算加速效果具有重要影响，但POD基函数数目的影响更大，需综合考虑计算精度和计算量进行确定。本项目研究可为干热岩热开采能力的科学评价提供理论支撑。