arXiv subjects

Zhangxin Chen

Publications and source records attributed to Zhangxin Chen.

At least 19 recordsLinked to original sources

arXiv2024

Analysis of a FEM-MCM Discretization for the 2D/3D stochastic closed-loop geothermal system

This paper develops a new 2D/3D stochastic closed-loop geothermal system with a random hydraulic conductivity tensor. We use the finite element method (FEM) and the Monte Carlo method (MCM) to discrete physical and probability spaces, respectively. This FEM-MCM method is effective. The stability for velocity and temperature is rigorously proved. Compared with the deterministic closed-loop geothermal system, a same optimal error estimate for approximate velocity and temperature is obtained. Furthermore, a series of numerical experiments were carried out to show this method has better stability and accuracy results.

math.NA↗

arXiv2024

A high-order, high-efficiency adaptive time filter algorithm for shale reservoir model based on coupled fluid flow with porous media flow

In this paper, a third-order time adaptive algorithm with less computation, low complexity is provided for shale reservoir model based on coupled fluid flow with porous media flow. The algorithm combines the three-step linear time filters method for simple post-processing and the second-order backward differential formula (BDF2), is third-order accurate and provides, at no extra computational complexity. At the same time, the time filter method can also be used to damp non-physical oscillations inherent in the BDF2 method, ensuring stability. We proves the variable time stepsize second-order backward differential formula plus time filter (BDF2-TF) algorithm's stability and the convergence properties of the fluid velocity u and hydraulic head $ϕ$ in the $L^2$ norm with an order of $O(k_{n+1}^3 + h^3)$. In the experiments, the adaptive algorithm automatically adjusts the time step in response to the varying characteristics of different models, ensuring that errors are maintained within acceptable limits. This algorithm addresses the issue that high-order algorithms may select inappropriate time steps, resulting in instability or reduced precision of the numerical solution, thereby enhancing calculation accuracy and efficiency. We perform three-dimensional numerical experiments to verify the BDF2-TF algorithm's effectiveness, stability, and third-order convergence. Simultaneously, a simplified model is employed to simulate the process of shale oil extraction from reservoirs, further demonstrating the algorithm's practical applicability.

math.NA↗

arXiv2023

A Local Parallel Finite Element Method for Super-Hydrophobic Proppants in a Hydraulic Fracturing System Based on a 2D/3D Transient Triple-Porosity Navier-Stokes Model

A hydraulic fracturing system with super-hydrophobic proppants is characterized by a transient triple-porosity Navier-Stokes model. For this complex multiphysics system, particularly in the context of three-dimensional space, a local parallel and non-iterative finite element method based on two-grid discretizations is proposed. The underlying idea behind utilizing the local parallel approach is to combine a decoupled method, a two-grid method and a domain decomposition method. The strategy allows us to initially capture low-frequency data across the decoupled domain using a coarse grid. Then it tackles high-frequency components by solving residual equations within overlapping subdomains by employing finer grids and local parallel procedures at each time step. By utilizing this approach, a significant improvement in computational efficiency can be achieved. Furthermore, the convergence results of the approximate solutions from the algorithm are obtained. Finally, we perform 2D/3D numerical experiments to demonstrate the effectiveness and efficiency of the algorithm as well as to illustrate its advantages in application.

math.NA↗

arXiv2023

Low Rank Optimization for Efficient Deep Learning: Making A Balance between Compact Architecture and Fast Training

Deep neural networks have achieved great success in many data processing applications. However, the high computational complexity and storage cost makes deep learning hard to be used on resource-constrained devices, and it is not environmental-friendly with much power cost. In this paper, we focus on low-rank optimization for efficient deep learning techniques. In the space domain, deep neural networks are compressed by low rank approximation of the network parameters, which directly reduces the storage requirement with a smaller number of network parameters. In the time domain, the network parameters can be trained in a few subspaces, which enables efficient training for fast convergence. The model compression in the spatial domain is summarized into three categories as pre-train, pre-set, and compression-aware methods, respectively. With a series of integrable techniques discussed, such as sparse pruning, quantization, and entropy coding, we can ensemble them in an integration framework with lower computational complexity and storage. Besides of summary of recent technical advances, we have two findings for motivating future works: one is that the effective rank outperforms other sparse measures for network compression. The other is a spatial and temporal balance for tensorized neural networks.

cs.LG↗

arXiv2022

Tucker-O-Minus Decomposition for Multi-view Tensor Subspace Clustering

With powerful ability to exploit latent structure of self-representation information, different tensor decompositions have been employed into low rank multi-view clustering (LRMVC) models for achieving significant performance. However, current approaches suffer from a series of problems related to those tensor decomposition, such as the unbalanced matricization scheme, rotation sensitivity, deficient correlations capture and so forth. All these will lead to LRMVC having insufficient access to global information, which is contrary to the target of multi-view clustering. To alleviate these problems, we propose a new tensor decomposition called Tucker-O-Minus Decomposition (TOMD) for multi-view clustering. Specifically, based on the Tucker format, we additionally employ the O-minus structure, which consists of a circle with an efficient bridge linking two weekly correlated factors. In this way, the core tensor in Tucker format is replaced by the O-minus architecture with a more balanced structure, and the enhanced capacity of capturing the global low rank information will be achieved. The proposed TOMD also provides more compact and powerful representation abilities for the self-representation tensor, simultaneously. The alternating direction method of multipliers is used to solve the proposed model TOMD-MVC. Numerical experiments on six benchmark data sets demonstrate the superiority of our proposed method in terms of F-score, precision, recall, normalized mutual information, adjusted rand index, and accuracy.

cs.LG↗

arXiv2021

Electrical Conductivity of Nanoscale Hydrophobic Porous Media: Surface Charge Density and Heterogeneous Pore Structure

Electrical conductivity is an inherent property of a hydrophobic porous media (HPM) and has critical applications. This research aims to provide a solution for predicting the electrical conductivity of nanoscale HPM with heterogeneous pore structure. Molecular dynamics (MD) simulations are compared with the modified Poisson-Boltzmann (MPB) model for understanding ionic charge density distributions in nanopores. The effective medium approximation (EMA) participates in calculating the effective conductance and conductivity of the nanoscale HPM. The results show that the surface charge density affects the ionic density profiles in the hydrophobic nanopores. As the pore size increases, the conductance increases. As the molarity of the aqueous electrolyte solution (AES) decreases, the conductance decreases. A phenomenon related to the conductance saturation occurred when the molarity of AES is very low. The effective conductance of an HPM increase as the coordination number increases. Finally, based on the calculated effective conductance and the heterogeneous pore structure parameters, the electrical conductivity of a nanoscale HPM is calculated.

cond-mat.soft↗

arXiv2020

Adaptive generalized multiscale approximation of a mixed finite element method with velocity elimination

In this paper, we propose offline and online adaptive enrichment algorithms for the generalized multiscale approximation of a mixed finite element method with velocity elimination to solve the subsurface flow problem in high-contrast and heterogeneous porous media. We give the theoretical analysis for the convergence of these two adaptive methods, which shows that sufficient initial basis functions (belong to the offline space) leads to a faster convergence rate. A series of numerical examples are provided to highlight the performance of both these two adaptive methods and also validate the theoretical analysis. Both offline and online adaptive methods are effective that can reduce the relative error substantially. In addition, the online adaptive method generally performs better than the offline adaptive method as online basis functions contain important global information such as distant effects that cannot be captured by offline basis functions. The numerical results also show that with a suitable initial multiscale space that includes all offline basis functions corresponding to relative smaller eigenvalues of local spectral decompositions in the offline stage, the convergence rate of the online enrichment is independent of the permeability contrast.

math.NA↗

arXiv2020

Generalized multiscale approximation of a multipoint flux mixed finite element method for Darcy-Forchheimer model

In this paper, we propose a multiscale method for the Darcy-Forchheimer model in highly heterogeneous porous media. The problem is solved in the framework of generalized multiscale finite element methods (GMsFEM) combined with a multipoint flux mixed finite element (MFMFE) method. %In the MFMFE methods, appropriate mixed finite element spaces and suitable quadrature rules are employed, which allow for local velocity elimination and lead to a cell-centered system for the pressure. We consider the MFMFE method that utilizes the lowest order Brezzi-Douglas-Marini ($\textrm{BDM}_1$) mixed finite element spaces for the velocity and pressure approximation. The symmetric trapezoidal quadrature rule is employed for the integration of bilinear forms relating to the velocity variables so that the local velocity elimination is allowed and leads to a cell-centered system for the pressure. %on meshes composed of simplices and $h^2$-perturbed parallelograms. We construct multiscale space for the pressure and solve the problem on the coarse grid following the GMsFEM framework. In the offline stage, we construct local snapshot spaces and perform spectral decompositions to get the offline space with a smaller dimension. In the online stage, we use the Newton iterative algorithm to solve the nonlinear problem and obtain the offline solution, which reduces the iteration times greatly comparing to the standard Picard iteration. Based on the offline space and offline solution, we calculate online basis functions which contain important global information to enrich the multiscale space iteratively. The online basis functions are efficient and accurate to reduce relative errors substantially. Numerical examples are provided to highlight the performance of the proposed multiscale method.

math.NA↗

arXiv2019

A Scalable Thermal Reservoir Simulator for Giant Models on Parallel Computers

This paper introduces the model, numerical methods, algorithms and parallel implementation of a thermal reservoir simulator that designed for numerical simulations of thermal reservoir with multiple components in three dimensional domain using distributed-memory parallel computers. Its full mathematical model is introduced with correlations for important properties and well modeling. Various well constraints, such as fixed bottom hole pressure, fixed oil, water, gas and liquid rates, constant heat transfer model, convective heat transfer model, heater model (temperature control, rate control, dual rate/temperature control), and subcool (steam trap), are introduced in details, including their mathematical models and methods. Efficient numerical methods and parallel computing technologies are presented. The simulator is designed for giant models with billions or even trillions of grid blocks using hundreds of thousands of CPUs. Numerical experiments show that our results match commercial simulators, which confirms the correctness of our methods and implementations. SAGD simulation with 15106 well pairs is also presented to study the effectiveness of our numerical methods. Scalability testings demonstrate that our simulator can handle giant models with over 200 billion grid blocks using 100,800 CPU cores and the simulator has good scalability.

cs.CE↗

arXiv2018

Numerical Simulations of Polymer Flooding Process in Porous Media on Distributed-memory Parallel Computers

Polymer flooding is a mature enhanced oil recovery technique that has been successfully applied in many field projects. By injecting polymer into a reservoir, the viscosity of water is increased, and the efficiency of water flooding is improved. As a result, more oil can be recovered. This paper presents numerical simulations of a polymer flooding process using parallel computers, where the numerical modeling of polymer retention, inaccessible pore volumes, a permeability reduction and polymer absorption are considered. Darcy's law is employed to model the behavoir of a fluid in porous media, and the upstream finite difference (volume) method is applied to discretize the mass conservation equations. Numerical methods, including discretization schemes, linear solver methods, nonlinearization methods and parallel techniques are introduced. Numerical experiments show that, on one hand, computed results match those from the commercial simulator, Schlumberger-Eclipse, which is widely applied by the petroleum industry, and, on the other hand, our simulator has excellent scalability, which is demonstrated by field applications with up to 27 million grid blocks using up to 2048 CPU cores.

cs.CE↗

arXiv2018

A New In-Situ Combustion Simulator for Parallel Computers

As a competitive recovery method for heavy oil, In-Situ Combustion (ISC) shows its great potential accompanied by technological advances in recent years. Reservoir simulation will play an indispensable role in the prediction of the implementation of ISC projects. With the computational complexity, it is imperative to develop an effective and robust parallel in-situ combustion simulator. In this paper, a mathematical model for In Situ Combustion is proposed, which takes full consideration for related physical phenomena, including multi-dimensional multi-component three-phase flow, heat convection and conduction, chemical reactions, and mass transfer between phases. In the mathematical model, different governing equations and constraints are involved, forming a complicated PDE (partial differential equation) system. For physical and chemical behaviors, some special treatments for the ISC simulator are discussed and applied. Also, a modified PER (Pseudo-Equilibrium Ratio) method is proposed in the thesis. A fully implicit scheme is applied, and discretization is implemented with the FDM (Finite Difference Method). In solving nonlinear systems, the Newton Method is introduced, and both numerical and analytical Jacobian matrices are applied. Due to the complexity of an ISC problem, an appropriate decoupling method must be considered. Thus the Gauss-Jordan transformation is raised. Then, with certain preconditioners and iterative solvers, a numerical solution can be obtained. The results of different models are given, which are validated with the results from CMG STARS. Also, the scalability of parallelization is proved, indicating the excellent performance of parallel computing. This accurate, efficient, parallel ISC simulator applies to complex reservoir models.

cs.CE↗

arXiv2018

Numerical Simulation of Two-phase Flow in Natural Fractured Reservoirs Using Dual Porosity Method on Parallel Computers

The two-phase oil-water flow in natural fractured reservoirs and its numerical methods are introduced in this paper, where the fracture is modeled by the dual porosity method. Efficient numerical method, including the finite difference method, CPR-FPF preconditioners for linear systems and effective decoupling method, are presented. Parallel computing techniques employed in simulation are also presented. Using those numerical methods and parallel techniques, a parallel reservoir simulator is developed, which is capable of simulating large-scale reservoir models. The numerical results show that our simulator is accurate and scalable compared to the commercial software, and the numerical methods are also effective.

cs.CE↗

arXiv2018

Development of A Scalable Platform for Large-scale Reservoir Simulations on Parallel computers

This paper presents our work on designing a parallel platform for large-scale reservoir simulations. Detailed components, such as grid and linear solver, and data structures are introduced, which can serve as a guide to parallel reservoir simulations and other parallel applications. The main objective of platform is to support implementation of various parallel reservoir simulators on distributed-memory parallel systems, where MPI (Message Passing Interface) is employed for communications among computation nodes. It provides structured grid due to its simplicity and cell-centered data is applied for each cell. The platform has a distributed matrix and vector module and a map module. The matrix and vector module is the base of our parallel linear systems. The map connects grid and linear system modules, which defines various mappings between grid and linear systems. Commonly-used Krylov subspace linear solvers are implemented, including the restarted GMRES method and the BiCGSTAB method. It also has an interface to a parallel algebraic multigrid solver, BoomerAMG from HYPRE. Parallel general-purpose preconditioners and special preconditioners for reservoir simulations are also developed. Various data structures are designed, such as grid, cell, data, linear solver and preconditioner, and some key default parameters are presented in this paper. The numerical experiments show that our platform has excellent scalability and it can simulate giant reservoir models with hundreds of millions of grid cells using thousands of CPU cores.

cs.CE↗

arXiv2017

ProLanGO: Protein Function Prediction Using Neural~Machine Translation Based on a Recurrent Neural Network

With the development of next generation sequencing techniques, it is fast and cheap to determine protein sequences but relatively slow and expensive to extract useful information from protein sequences because of limitations of traditional biological experimental techniques. Protein function prediction has been a long standing challenge to fill the gap between the huge amount of protein sequences and the known function. In this paper, we propose a novel method to convert the protein function problem into a language translation problem by the new proposed protein sequence language "ProLan" to the protein function language "GOLan", and build a neural machine translation model based on recurrent neural networks to translate "ProLan" language to "GOLan" language. We blindly tested our method by attending the latest third Critical Assessment of Function Annotation (CAFA 3) in 2016, and also evaluate the performance of our methods on selected proteins whose function was released after CAFA competition. The good performance on the training and testing datasets demonstrates that our new proposed method is a promising direction for protein function prediction. In summary, we first time propose a method which converts the protein function prediction problem to a language translation problem and applies a neural machine translation model for protein function prediction.

q-bio.QM↗

arXiv2017

Load Balancing using Hilbert Space-filling Curves for Parallel Reservoir Simulations

The goal of load balancing (grid partitioning) is to minimize overall computations and communications, and to make sure that all processors have a similar workload. Geometric methods divide a grid by using a location of a cell while topological methods work with connectivity of cells, which is generally described as a graph. This paper introduces a Hilbert space-filling curve method. A space-filling curve is a continuous curve and defines a map between a one-dimensional space and a multi-dimensional space. A Hilbert space-filling curve is one special space-filling curve discovered by Hilbert and has many useful characteristics, such as good locality, which means that two objects that are close to each other in a multi-dimensional space are also close to each other in a one dimensional space. This property can model communications in grid-based parallel applications. The idea of the Hilbert space-filling curve method is to map a computational domain into a one-dimensional space, partition the one-dimensional space to certain intervals, and assign all cells in a same interval to a MPI. To implement a load balancing method, a mapping kernel is required to convert high-dimensional coordinates to a scalar value and an efficient one-dimensional partitioning module that divides a one-dimensional space and makes sure that all intervals have a similar workload. The Hilbert space-filling curve method is compared with ParMETIS, a famous graph partitioning package. The results show that our Hilbert space-filling curve method has good partition quality. It has been applied to grids with billions of cells, and linear scalability has been obtained on IBM Blue Gene/Q.

cs.DC↗

arXiv2017

Decoupled Block-Wise ILU(k) Preconditioner on GPU

This research investigates the implementation mechanism of block-wise ILU(k) preconditioner on GPU. The block-wise ILU(k) algorithm requires both the level k and the block size to be designed as variables. A decoupled ILU(k) algorithm consists of a symbolic phase and a factorization phase. In the symbolic phase, a ILU(k) nonzero pattern is established from the point-wise structure extracted from a block-wise matrix. In the factorization phase, the block-wise matrix with a variable block size is factorized into a block lower triangular matrix and a block upper triangular matrix. And a further diagonal factorization is required to perform on the block upper triangular matrix for adapting a parallel triangular solver on GPU.We also present the numerical experiments to study the preconditioner actions on different k levels and block sizes.

math.NA↗

arXiv2017

A Parallel Simulator for Massive Reservoir Models Utilizing Distributed-Memory Parallel Systems

This paper presents our work on developing parallel computational methods for two-phase flow on modern parallel computers, where techniques for linear solvers and nonlinear methods are studied and the standard and inexact Newton methods are investigated. A multi-stage preconditioner for two-phase flow is applied and advanced matrix processing strategies are studied. A local reordering method is developed to speed the solution of linear systems. Numerical experiments show that these computational methods are effective and scalable, and are capable of computing large-scale reservoir simulation problems using thousands of CPU cores on parallel computers. The nonlinear techniques, preconditioner and matrix processing strategies can also be applied to three-phase black oil, compositional and thermal models.

cs.CE↗

arXiv2017

Scalable linear solvers for sparse linear systems from large-scale numerical simulations

This paper presents our work on designing scalable linear solvers for large-scale reservoir simulations. The main objective is to support implementation of parallel reservoir simulators on distributed-memory parallel systems, where MPI (Message Passing Interface) is employed for communications among computation nodes. Distributed matrix and vector modules are designed, which are the base of our parallel linear systems. Commonly-used Krylov subspace linear solvers are implemented, including the restarted GMRES method, the LGMRES method, and the BiCGSTAB method. It also has an interface to a parallel algebraic multigrid solver, BoomerAMG from HYPRE. Parallel general-purpose preconditioners and special preconditioners for reservoir simulations are also developed. The numerical experiments show that our linear solvers have excellent scalability using thousands of CPU cores.

cs.MS↗