The deployment of various renewable energy technologies has increased rapidly as their associated costs have decreased significantly. Renewable resources such as wind and solar generate highly intermittent energy, which can be problematic for current electrical grids to integrate with a large amount. Geothermal battery energy storage, i.e., utilizing geological formations to store energy in the form of thermally heated brine, has been proposed to compensate for the increasing and intermittent renewable energy generation and to help stabilize U.S. grids. The low-temperature geothermal battery concept has been successfully applied to heat buildings by storing excess energy during low-use periods and recovering it during peak energy demand periods. However, grid-scale high-temperature geothermal storage for electricity generation still faces several questions including: 

1) What are the optimal formation characteristics (e.g., permeability) and operation conditions (e.g., injection temperature) suitable for energy storage? 

2) How long can the RTES run without performance degradation considering the Geochemistry effect? 

We use stochastic simulation with machine learning algorithms and site-specific geochemistry experiments informed modeling to answer the above questions. 

Related  Publications:

W. Jin, R. Podgorney, T. McLing (2020). THM Coupled Numerical Analysis on the Geothermal Energy Storage & Extraction in Porous Fractured Reservoir. In 54th US Rock Mechanics/Geomechanics Symposium. American Rock Mechanics Association. Link

Jin, W., Podgorney, R., McLing, T., & Carlsen, R. W. (2021, June). Geothermal Battery Optimization Using Stochastic Hydro-Thermal Simulations and Machine Learning Algorithms. In 55th US Rock Mechanics/Geomechanics Symposium. OnePetro. Link

Jin, W., Atkinson, T., Doughty, C., Neupane, G., Spycher, N., McLing, T., ... & Podgorney, R. (2022). Machine-learning-assisted high-temperature reservoir thermal energy storage optimization. Renewable Energy, 197,  384-397. Link 

W. Jin et al. (2022, June) Influence of mechanical deformation and mineral dissolution/precipitation on reservoir thermal energy storage. In 56th US Rock Mechanics/Geomechanics Symposium. American Rock Mechanics Association. Link

Many geomaterials exhibit strong orientation dependent mechanical behavior (anisotropy) due to bedding, layering or crack patterns. The goal of this research are

1. Propose a constitutive law  to capture the direction depended mechanical behavior in tension and in compression;

2. Implement this mechanical model into Finite Element code, coupling with fluid flow models, to realize fluid driving fracture propagation.

We have proposed and implemented

•  a nonlocal anisotropic damage model for transversely isotropic material,

• multiscale fracture propagation by transition from nonlocal damage to PPR coheisve fracture,

• Darcy's fluid flow in matrix and Cubic fluid flow along the fracture,

into a matlab toolbox.

Related  Publications:

W. Jin, C. Arson, (2018).  Anisotropic nonlocal damage model for materials with intrinsic transverse isotropy. International Journal of Solids and Structures, 139, 29-42.  Link

W. Jin, C. Arson, (2019). Fluid-driven transition from damage to fracture in anisotropic porous media: a multi-scale XFEM approach.  Acta Geotechnica, 15(1), 113-144. Link

Failure of quasi-brittle materials starts with diffuse micro-crack inception and growth, and evolves into localized macro fracture formation due to micro-crack coalescence. Accurately modelling this whole process has been a challenge for decades. We propose to couple a micromechanics based nonlocal anisotropic damage model with a cohesive zone model to capture the multiscale process. 

An anisotropic damage model is used to explicitly calculate the initiation and growth of micro-cracks. The free energy expression at the scale of the Representative Elementary Volume is calculated by homogenizing the deformation energy generated by displacement jumps at micro-crack faces. The evolution law of the micro-crack density is expressed in terms of nonlocal equivalent strains. 

We implement the model into Finite Element code. When damage in a FE exceeds a threshold value, the element is split and, using XFEM, a cohesive zone segment is inserted to represent the formation of a macro fracture. The critical crack density, above which micro-cracks interact, is determined by comparing the damaged elastic moduli predicted by the proposed nonlocal damage model (no crack interaction) with those predicted with Kachanov’s micromechanical model (cracks interact). The PPR traction separation law is used to model the evolution of the macro cohesive fracture. The cohesive strength is assigned as the stress at which micro-crack density reaches the critical value. The cohesive energy release rate is adjusted dynamically such that the total dissipated energy, which includes the energy dissipated by micro-crack propagation and by macro-fracture growth, is a constant value per unit area of fracture. 

Related  Publications:

W. Jin, C. Arson, (2018).  Nonlocal enrichment of a micromechanical damage model with tensile softening: advantages and limitations. Computers and Geotechnics, 94: 196-206.  Link

W. Jin, H. Xu, C. Arson, S. Busetti, (2017).  Computational model coupling mode II discrete fracture propagation with continuum damage zone evolution. International Journal for Numerical and Analytical Methods in Geomechanics, 41(2):223-250. Link

 W. Jin, C. Arson, (2018). From nonlocal enhanced microcrack damage to macro cohesive fracture - coupled computational tool using XFEM. Computer Methods in Applied Mechanics and Engineering, 357, 112617. Link

 Brittle materials such as concrete, rock, and ceramic composites, exhibit a complex mechanical behavior at the meso-scale, such as:

•  Stress-induced damage and stiffness anisotropy,

• Nonlinear stress-strain relationships,

• Volumetric dilation due to irreversible strain,

• Unilateral effects due to crack closure,

• Brittle-ductile transition with increasing confining stress,

• Different mechanical responses in tension and compression.

Physically,  all above effects can be explained by the nucleation and propagation of micro-cracks at the grain boundaries and/or from pore spaces. The goal of this project is to

1. formulate a constitutive model that captures the complex behavior with reasonable and physically meaningful material parameters;

2. Implement this model into Finite Element code without computational intensive resolution algorithms.

We first assume that the orientation of micro cracks can be discretely represented by 42 microplane directions, and they do not interact prior to the peak stress (Maximum stress reached during stress history). 

For stress controlled test, we adopt the theory of wing crack to capture the mixed mode micro crack propagation at micro-scale. Accompanied with proper projection of crack densities,  this novel idea captures increasing yield stress when confining stress increases for triaxial compression test without additional material parameter.  

For strain controlled test, we assume all crack propagate within the plane of crack surface.  Closed cracks propagate in pure mode II, whereas open cracks propagate in mixed mode (I/II). For each crack set, its crack density is controlled by independent evolution law. Thus, the elastic domain is at the intersection of the yield surfaces of the activated crack families, and it describes a non-smooth surface. I implemented the model into Abaqus User MATerial (UMAT) subroutine utilizing return mapping algorithm (closest point projection) at the Gauss point. 

Related  Publications:

W. Jin, C. Arson, (2017).  Discrete equivalent wing crack based damage model for brittle solids. International Journal of Solids and Structures, 110: 279-293.  Link

W. Jin, C. Arson, (2017).  Micromechanics based discrete damage model with multiple non-smooth yield surfaces: theoretical formulation, numerical implementation and engineering applications. International Journal of Damage Mechanics, 27(5), 611-639.  Link

 Fractures play a vital role in determining fluid transportation and mechanical response of rock masses, thus, it is crucial to properly use the limited in-situ data (e.g. fracture trace length on 2D exposure surface, fracture intersection with borehole in 1D) to construct 3D statistically equivalent Discrete Fracture Network (DFN). 

Based on the assumption that all fractures have circular shape, the mathematical relation between in-situ data (1D or 2D) and the statistical size distribution of fractures (3D) were obtained in the literature. However, the simple disk shape cannot accurately represent the geometry of complex fractures within rock masses. Assume fracture is elliptical, I derived the analytical relation to connect the 3D fracture size distribution to the orientation and trace length distribution on 2D exposed surfaces. Further, I proposed a method to estimate the size distribution of elliptical fractures from multiple parallel borehole loggings. Utilizing the method, I successfully inversed the DFN ahead of the mining face in Datong, China by collaborating research team in China University of Mining & Technology, Beijing. The constructed DFN was used to analyze coal methane flow and to design orientation and location of boreholes for hydraulic fracturing.

Related  Publications:

M. Gao,  W. Jin, R. Zhang, J. Xie, B. Yu, H. Duan, (2016).  Fracture size estimation using data from multiple boreholes. International Journal of Rock Mechanics and Mining Sciences, 86:29-41. Link

W. Jin, M. Gao, B. Yu, R. Zhang, J. Xie, Z. Qiu, (2015).  Elliptical fracture network modeling with validation in Datong Mine, China. Environmental Earth Sciences, 73(11):7089-7101. Link

W. Jin, M. Gao, R. Zhang, G. Zhang, (2014).  Analytical expressions for the size distribution function of elliptical joints. International Journal of Rock Mechanics and Mining Sciences, 70:201-211. Link