Xu Liu

Dr. Xu Liu

orcid

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
26 June 2025 in Sustainability
Verified
Journal Article
Open access
PDF file
XQXiaojun Qian
Xiaojun Qian
SYShaohua You
Shaohua You
RWRuizhe Wang
Ruizhe Wang
YYYunzhi Yue
Yunzhi Yue
Qinzhuo LiaoQinzhuo Liao
Qinzhuo Liao
XLXu Liu
Xu Liu
The transition to a sustainable energy future hinges on the development of reliable large-scale hydrogen storage solutions to balance the intermittency of renewable energy and decarbonize hard-to-abate industries. Underground hydrogen storage (UHS) in salt caverns emerged as a technically and economically viable strategy, leveraging the unique geomechanical properties of salt formations—including low permeability, self-healing capabilities, and chemical inertness—to ensure safe and high-purity hydrogen storage under cyclic loading conditions. This review provides a comprehensive analysis of the advantages of salt cavern hydrogen storage, such as rapid injection and extraction capabilities, cost-effectiveness compared to other storage methods (e.g., hydrogen storage in depleted oil and gas reservoirs, aquifers, and aboveground tanks), and minimal environmental impact. It also addresses critical challenges, including hydrogen embrittlement, microbial activity, and regulatory fragmentation. Through global case studies, best operational practices for risk mitigation in real-world applications are highlighted, such as adaptive solution mining techniques and microbial monitoring. Focusing on China’s regional potential, this study evaluates the hydrogen storage feasibility of stratified salt areas such as Jiangsu Jintan, Hubei Yunying, and Henan Pingdingshan. By integrating technological innovation, policy coordination, and cross-sector collaboration, salt cavern hydrogen storage is poised to play a pivotal role in realizing a resilient hydrogen economy, bridging the gap between renewable energy production and industrial decarbonization.

24 October 2017 in Geophysical Journal International
Verified
Journal Article
Open access
PDF file
XLXu Liu
Xu Liu
SGStewart Greenhalgh
Stewart Greenhalgh
MWMoosoo Won
Moosoo Won
MGMark Greenhalgh
Mark Greenhalgh
A method is suggested to express the effective bulk modulus of the solid frame of a poroelastic material as a function of the saturated bulk modulus. This method enables effective Biot theory to be described through the use of seismic dispersion measurements or other models developed for the effective saturated bulk modulus. The effective Biot theory is generalized to a poroviscoelastic model of which the moduli are represented by the relaxation functions of the generalized fractional Zener model. The latter covers the general Zener and the Cole–Cole models as special cases. A global search method is described to determine the parameters of the relaxation functions, and a simple deterministic method is also developed to find the defining parameters of the single Cole–Cole model. These methods enable poroviscoelastic models to be constructed, which are based on measured seismic attenuation functions, and ensure that the model dispersion characteristics match the observations.

21 March 2023 in International Journal of Numerical Methods for Heat & Fluid Flow
Verified
Journal Article
Closed access
MKM.R. Anil Kumar
M.R. Anil Kumar
PKPradeep Kaswan
Pradeep Kaswan
NSNantu Sarkar
Nantu Sarkar
XLXu Liu
Xu Liu
MKManjeet Kumari
Manjeet Kumari
Purpose: The purpose of this article is to investigate the propagation characteristics (such as particle motion, attenuation and phase velocity) of a Rayleigh wave in a nonlocal generalized thermoelastic media. Design/methodology/approach: The bulk waves are represented with Helmholtz potentials. The stress-free insulated and isothermal plane surfaces are taken into account. Rayleigh wave dispersion relation has been established and is found to be complex. Due to the presence of radicals, the dispersion equation is continuously computed as a complicated irrational expression. The dispersion equation is then converted into a polynomial equation that can be solved numerically for precise complex roots. The extra zeros in this polynomial equation are eliminated to yield the dispersion equation’s roots. These routes are then filtered for inhomogeneous wave propagation that decays with depth. To perform numerical computations, MATLAB software is used. Findings: In this medium, only one mode of Rayleigh wave exists at both isothermal and insulated boundaries. The thermal factors of nonlocal generalized thermoelastic materials significantly influence the particle motion, attenuation and phase velocity of the Rayleigh wave. Originality/value: Numerical examples are taken to examine how the thermal characteristics of materials affect the existing Rayleigh wave’s propagation characteristics. Graphical analysis is used to evaluate the behavior of particle motion (such as elliptical) both inside and at the isothermal (or insulated) flat surface of the medium under consideration.

30 August 2019 in Geophysical Journal International
Verified
Journal Article
Closed access
XLXu Liu
Xu Liu
SGStewart Greenhalgh
Stewart Greenhalgh
SUMMARY: The Cole–Cole mechanical model can match a simple seismic attenuation curve over a broad frequency range, but it is not a suitable model for replicating complicated seismic attenuation dispersion curves which exhibit multiple peaks or display pronounced asymmetry. In this case, we use the General Fractional Zener (GFZ) Model, which comprises multiple Cole–Cole elements, to approximate the attenuation observations. The observations here represent the arbitrary (frequency-dependent) dispersion behaviour from actual measurements (phase velocities and/or dissipation factors) or from some physical dissipation mechanism(s) such as local induced fluid flow in effective Biot theory. The key parameters of these viscoelastic models, which include the stress and strain relaxation times and the fractional derivative orders, are determined with a simulated annealing method. Instead of searching for the relaxation times directly, we search for the Zener peak attenuation points and corresponding frequencies, each of which corresponds to a pair of relaxation times. We show that just two fractional Zener elements can sometimes provide a satisfactory approximation to the observations over the entire frequency range. A simple deterministic method is developed to extract the parameters of the single element Zener model using phase velocity observations. As a special case of the GFZ model, we found and proved the constancy of the width of the attenuation curves at the half maximum amplitude point (FWHM) for all Zener models which is critical to the design of reasonable observation frequencies. We stress and demonstrate that ignoring the frequency-dependence of Q may result in significant discrepancies of calculated waveforms with observed or predicted values.

29 March 2020 in Geophysics
Verified
Journal Article
Closed access
HLHuijian Li
Huijian Li
SGStewart Greenhalgh
Stewart Greenhalgh
SCShijun Chen
Shijun Chen
XLXu Liu
Xu Liu
BLBo Liu
Bo Liu
The central frequency shift technique for estimating wave attenuation in seismic exploration assumes a quasi-symmetric amplitude spectrum and has its limitations in low quality factor (Q) regions. The asymmetry of the wavelet amplitude spectrum becomes more pronounced during wave propagation, so using a constant parameter n (the asymmetry index estimated) in the modified frequency-weighted exponential formula method to estimate Q still leads to errors and can degrade the results of inverse Q-filtering. We have derived a new Q-estimation scheme that does not require constant parameter fitting that also works for strongly asymmetric receiver spectra. It is based on forming a synthetic wavelet as the geometric mean of the source spectrum (approximated by the near-source receiver spectrum) and the subsequent attenuated (receiver) wavelet spectrum. The changing centroid frequency and variance of this new wavelet as a function of traveltime can automatically adapt to a changing asymmetric spectrum caused by attenuation. It yields more stable and accurate results, especially in low-Q regions. This adopted approach is successfully applied to synthetic data and vertical seismic profile field survey data and proves to be superior to previous frequency-shift methods.

16 March 2020 in Geophysics
Verified
Journal Article
Closed access
XLXu Liu
Xu Liu
SGStewart Greenhalgh
Stewart Greenhalgh
José M. CarcioneJosé M. Carcione
José M. Carcione
Seismic waves in an attenuative porous medium are generally inhomogeneous waves, which have different directions of propagation and attenuation. The dissipation factors (1/Q) of inhomogeneous waves are strongly dependent on the degree of wave inhomogeneity and cannot be expressed correctly with the usual 1/Q expressions valid only for homogeneous waves. We have used the differing definitions of 1/Q for inhomogeneous waves (i.e., the ratio of the time-averaged dissipated energy density to the time-averaged strain energy density or time-averaged total energy density) and the complex form of the energy balance equations of poroviscoelastic media to derive concise and explicit expressions for the dissipation factors. They are given as simple functions of the material parameters and the wave inhomogeneity parameter for inhomogeneous SV-waves and fast and slow P-waves. The isotropic, poroviscoelastic medium under consideration is upscaled from effective Biot theory for a double-porosity solid, which is the most general theory to describe wave propagation in a reservoir. We find that, if the inhomogeneity parameter is infinite (i.e., the inhomogeneity angle is 90°) for all three Biot waves, then the dissipation factors only depend on the ratio of the imaginary to the real part of the complex shear modulus. Our explicit expressions for the dissipation factors of poroviscoelastic materials also are reduced to obtain their counterparts for viscoelastic media as a special case. The inhomogeneous waves in an example poroviscoelastic material are used to demonstrate that the 1/Q values of the three Biot waves strongly depend on the inhomogeneity parameters and furthermore the different definitions may cause significant differences of 1/Q values. We find that the dissipation factor of fast P-waves may decrease with the increasing degree of inhomogeneity, which contradicts previously published results.