© Fluent Inc. 1/10/2018L1 Fluids Review TRN Solution Methods.

Slides:



Advertisements
Similar presentations
SolidWorks Flow Simulation
Advertisements

Fluent Overview Ahmadi/Nazridoust ME 437/537/637.
Chapter 8 Elliptic Equation.
Optimization of thermal processes
MTM 144 ASPEN FLUENT Timeschedule ProgramLecturesTeaching assistance Fluent INTRO4h2h GAMBIT4h8h FLUENT6h14h ASPEN6h4h
MULTISCALE COMPUTATIONAL METHODS Achi Brandt The Weizmann Institute of Science UCLA
Geometric (Classical) MultiGrid. Hierarchy of graphs Apply grids in all scales: 2x2, 4x4, …, n 1/2 xn 1/2 Coarsening Interpolate and relax Solve the large.
Lectures on CFD Fundamental Equations
Introduction to numerical simulation of fluid flows
Coupled Fluid-Structural Solver CFD incompressible flow solver has been coupled with a FEA code to analyze dynamic fluid-structure coupling phenomena CFD.
Numerical methods for PDEs PDEs are mathematical models for –Physical Phenomena Heat transfer Wave motion.
Lecture 5 - Solution Methods Applied Computational Fluid Dynamics
© 2011 Autodesk Freely licensed for use by educational institutions. Reuse and changes require a note indicating that content has been modified from the.
Instructor: André Bakker
FUNDAMENTAL EQUATIONS, CONCEPTS AND IMPLEMENTATION
Tutorial 5: Numerical methods - buildings Q1. Identify three principal differences between a response function method and a numerical method when both.
© Fluent Inc. 9/5/2015L1 Fluids Review TRN Solution Methods.
Lecture Objectives Review SIMPLE CFD Algorithm SIMPLE Semi-Implicit Method for Pressure-Linked Equations Define Residual and Relaxation.
ME451 Kinematics and Dynamics of Machine Systems Numerical Solution of DAE IVP Newmark Method November 1, 2013 Radu Serban University of Wisconsin-Madison.
1 Discretization of Fluid Models (Navier Stokes) Dr. Farzad Ismail School of Aerospace and Mechanical Engineering Universiti Sains Malaysia Nibong Tebal.
© Fluent Inc. 10/26/ Introductory FLUENT Notes FLUENT v6.0 Jan 2002 Fluent User Services Center Solver Settings.
Module 4 Multi-Dimensional Steady State Heat Conduction.
Lecture Objectives: Explicit vs. Implicit Residual, Stability, Relaxation Simple algorithm.
© Fluent Inc. 11/24/2015J1 Fluids Review TRN Overview of CFD Solution Methodologies.
Using the Segregated and Coupled Solvers
Lecture Objectives Review Define Residual and Relaxation SIMPLE CFD Algorithm SIMPLE Semi-Implicit Method for Pressure-Linked Equations.
Parallel Solution of the Poisson Problem Using MPI
MECH4450 Introduction to Finite Element Methods Chapter 9 Advanced Topics II - Nonlinear Problems Error and Convergence.
FALL 2015 Esra Sorgüven Öner
Discretization Methods Chapter 2. Training Manual May 15, 2001 Inventory # Discretization Methods Topics Equations and The Goal Brief overview.
Discretization for PDEs Chunfang Chen,Danny Thorne Adam Zornes, Deng Li CS 521 Feb., 9,2006.
ERT 216 HEAT & MASS TRANSFER Sem 2/ Dr Akmal Hadi Ma’ Radzi School of Bioprocess Engineering University Malaysia Perlis.
Lecture Objectives: - Numerics. Finite Volume Method - Conservation of  for the finite volume w e w e l h n s P E W xx xx xx - Finite volume.
Model Anything. Quantity Conserved c  advect  diffuse S ConservationConstitutiveGoverning Mass, M  q -- M Momentum fluid, Mv -- F Momentum fluid.
Computational Fluid Dynamics Lecture II Numerical Methods and Criteria for CFD Dr. Ugur GUVEN Professor of Aerospace Engineering.
Transfer Functions Convenient representation of a linear, dynamic model. A transfer function (TF) relates one input and one output: The following terminology.
Introduction to the Finite Element Method
Finite Difference Methods
Convection-Dominated Problems
Chapter 30.
Objective Numerical methods SIMPLE CFD Algorithm Define Relaxation
Transfer Functions Chapter 4
Chapter 9: Differential Analysis of Fluid Flow
J. Murthy Purdue University
Lecture Objectives: Review Explicit vs. Implicit
Nodal Methods for Core Neutron Diffusion Calculations
Implementing Finite Volume Methods
Linearized Block Implicit (LBI) Method applied to Quasi-1D Flow
Lecture 19 MA471 Fall 2003.
Fluent Overview Ahmadi/Nazridoust ME 437/537/637.
Convergence in Computational Science
Introduction to Multigrid Method
Objective Review Reynolds Navier Stokes Equations (RANS)
Lecture Objectives: Advance discretization methods
Objective Unsteady state Numerical methods Discretization
Finite Volume Method for Unsteady Flows
Topic 6 FluidFlowEquations_Introduction
Supported by the National Science Foundation.
Convergence in Numerical Science
Objective Numerical methods Finite volume.
Lecture Objectives Review for exam Discuss midterm project
Introduction to Fluid Dynamics & Applications
topic4: Implicit method, Stability, ADI method
topic16_cylinder_flow_relaxation
topic4: Implicit method, Stability, ADI method
Comparison of CFEM and DG methods
Topic 8 Pressure Correction
topic4: Implicit method, Stability, ADI method
Akram Bitar and Larry Manevitz Department of Computer Science
Presentation transcript:

© Fluent Inc. 1/10/2018L1 Fluids Review TRN Solution Methods

© Fluent Inc. 1/10/2018L2 Fluids Review TRN Overview u Properties of Numerical Solution Methods u FVM and FEM solution methods u Characteristics of solution algorithms u Equations solvers u Underrelaxation u Convergence

© Fluent Inc. 1/10/2018L3 Fluids Review TRN Numerical Solution Methods (1) u The important components of a numerical solution method are: 1.Mathematical model of flow n e.g. equations of motion- unsteady and steady, compressible and incompressible, 2D and 3D, turbulence, etc. 2. Discretization Method n Approximation of the differential equations by a system of algebraic equations s Finite Difference Method (FDM) s Finite Volume Method (FVM) s Finite Element Method (FEM) 3. Coordinate system n cartesian or cylindrical, curvilinear orthogonal and non-orthogonal coordinate systems

© Fluent Inc. 1/10/2018L4 Fluids Review TRN Numerical Solution Methods (2) 4. Numerical Grid n The solution domain is subdivided by the grid. The algebraic conservation equations for the variables are computed on a finite number of control volumes or elements in the domain. n Types of Grids s Structured grids s Multi-block-structured grids s Unstructured grids 5. Finite Approximations n Discretizing the solution domain gives rise to errors from the approximation of the continuous differential functions n FDM - approximate the derivatives through the Taylor series expansion n FVM - approximate the surface and volume integrals n FEM - choose weighting functions Unstructured surface grid for vehicle aerodynamic analysis.

© Fluent Inc. 1/10/2018L5 Fluids Review TRN Numerical Solution Methods (3) 6. Solution Criteria and Convergence Criteria n This is the topic of this lecture n Methods of solving the system of algebraic equations n The nonlinear nature of the governing equations requires an iterative solution method. Convergence criteria determine when to terminate the iterative process. Accuracy and efficiency are considered.

© Fluent Inc. 1/10/2018L6 Fluids Review TRN Properties of Solution Methods l Consistency l Stability l Convergence l Conservation l Boundedness l Realizability l Accuracy

© Fluent Inc. 1/10/2018L7 Fluids Review TRN FVM - Solution Algorithms u The discretized form of the governing conservation equations can be written as: n where nb denotes the cell neighbors of cell P u In a 2D structured grid, the face P has four neighbors (E,W,N,S). In a 3D grid, a cell has six neighbors. u In an unstructured grid, the number of neighbors depends on the cell shape and mesh topology. u The above algebraic equation is written for each transport variable, that is, velocity, temperature, species concentration and turbulence quantities. n P E W N S j i e s w

© Fluent Inc. 1/10/2018L8 Fluids Review TRN FVM - Solution Algorithms u The solution of the Navier-Stokes equations is complicated by the lack of an independent equation for pressure. Pressure is linked to all three momentum equations u The pressure-velocity coupling algorithm SIMPLE (Semi-Implicit Pressure Linked Equations), and it’s variants, are used. u Concept: l the momentum equations are used to compute velocity l a pressure equation is derived from the continuity equation l a discrete pressure correction equation is derived from the discrete forms of the pressure and momentum equations l the pressure correction equation is updated with pressure and a mass flux balance through a mass correction

© Fluent Inc. 1/10/2018L9 Fluids Review TRN Finite Volume Solution Methods u The Finite Volume Solution method can either use a “segregated” or a “coupled” solution procedure. u The solution procedure of each method is the same.

© Fluent Inc. 1/10/2018L10 Fluids Review TRN Segregated Solution Procedure Update properties. Solve momentum equations (u, v, w velocity). Solve pressure-correction (continuity) equation. Update pressure, face mass flow rate. Solve energy, species, turbulence, and other scalar equations. Converged? Stop No Yes

© Fluent Inc. 1/10/2018L11 Fluids Review TRN Coupled Solution Procedure Solve continuity, momentum, energy, and species equations simultaneously. Converged? Stop NoYes Solve turbulence and other scalar equations. Update properties.

© Fluent Inc. 1/10/2018L12 Fluids Review TRN Unsteady Solution Procedure u Same procedure for segregated and coupled solvers: Execute segregated or coupled procedure, iterating to convergence Take a time step Requested time steps completed? NoYes Stop Update solution values with converged values at current time

© Fluent Inc. 1/10/2018L13 Fluids Review TRN FVM - Linear Equation Solvers  Consider the system of algebraic equations for variable  l The above system of equations is arranged in a matrix and solved iteratively. u For a structured grid, the coefficient matrix is banded. Special line-by-line iterative techniques such as the Line Gauss-Seidel (LGS) method may be used. l LGS method involves solving the equations in a “line” simultaneously. l The equations are set-up in a tri-diagonal matrix solved via Gaussian elimination u For an unstructured grid, no line structure exists. Point-iterative methods are used, e.g., the Point Gauss-Seidel (PGS) technique. u LGS/PGS locally reduce errors but can miss long-wavelength errors. Multigrid acceleration will speed up the LGS/PGS convergence.

© Fluent Inc. 1/10/2018L14 Fluids Review TRN FVM - Line Gauss-Seidel (LGS) Method u The LGS method is used on structured grids and involves the following steps: l simultaneously solve the equations in the sweep direction l march to next row or column Line to be solved Values from previous sweep Values from previous iteration Flow Marching direction sweeping direction

© Fluent Inc. 1/10/2018L15 Fluids Review TRN FVM - The Multigrid Solver u The LGS and PGS solvers both transmit the influence of near-neighbors effectively and are less effective at transmitting the influence of far away grid points and boundaries, thereby, slowing convergence. u “Multigrid” solver accelerates convergence for: l Large number of cells l Large cell aspect ratios  x/  y > 20 l Large differences in thermal conductivity l Such as in conjugate heat transfer u General concept of multigrid is the same for structured and unstructured grids, although the implementation is different.

© Fluent Inc. 1/10/2018L16 Fluids Review TRN The Multigrid Concept (1) u Multigrid solver uses a sequence of grids going from fine to coarse. u Influence of boundaries and far-away points more easily transmitted to interior on coarse meshes than on fine meshes. l In coarse meshes, grid points are closer together in the computational space and have fewer computational cells between any two spatial locations. u Fine meshes give more accurate solutions.

© Fluent Inc. 1/10/2018L17 Fluids Review TRN The Multigrid Concept (2) u The solutions on the coarser meshes is used as a starting point for solutions on the finer meshes. l Coarse-mesh solution contains influence of boundaries and far neighbors. l These effects felt more easily on coarse mesh. l Accelerates convergence on fine mesh. u Final solution obtained for original (fine) mesh. l Coarse mesh calculations: l only accelerates convergence l do not change final answer fine mesh corrections summed equations (or volume-averaged solution) coarse mesh

© Fluent Inc. 1/10/2018L18 Fluids Review TRN  For stability the change in a variable  p value from iteration to iteration is reduced by an “under-relaxation” factor,  : For example, an under-relaxation of 0.2 restricts the change in  P to 20% of the computed change of  for one iteration. FVM - Under-relaxation u Equation set being solved is non-linear. u Equation for one variable may depend on other variables, e.g., l Temperature l Mass fraction

© Fluent Inc. 1/10/2018L19 Fluids Review TRN u Residual at point P is defined as: u An overall measure of the residual in the domain is: u Residuals can be scaled relative to the starting residual FVM - Residuals and Convergence u At convergence: l All discrete conservation equations (momentum, energy, etc.) are obeyed in all cells to a specified tolerance. l The solution no longer changes with additional iterations. l Mass, momentum, energy and scalar balances are obtained. u “Residuals” measure imbalance (or error) in conservation equations.

© Fluent Inc. 1/10/2018L20 Fluids Review TRN Finite Element Solution Methods u We seek a solution to the equation of the form: K(u) u = F u A solution method is made up of two parts l Algorithm: solution organization scheme l Equation solver: solves linear system of equations u We shall consider two algorithms and two equation solvers

© Fluent Inc. 1/10/2018L21 Fluids Review TRN FEM Algorithms and Equation Solvers u Algorithms: l fully-coupled l segregated u Equation solvers: l Gaussian elimination l Iterative methods: n non-symmetric equation systems n symmetric equation systems (pressure eqns.)

© Fluent Inc. 1/10/2018L22 Fluids Review TRN Fully-Coupled Algorithm (1) u The most common solution scheme is the so-called Newton-Raphson iteration, or Newton’s method for short u First, re-write the equation as: R(u) = K(u) u - F u Using a Taylor series expansion and some further manipulations, we arrive at:

© Fluent Inc. 1/10/2018L23 Fluids Review TRN Fully-Coupled Algorithm (2) u Advantages: l converges very rapidly u Disadvantages: l requires good initial guess l calculation of J -1 (u i ) is expensive u Alternatives: l Modified Newton-Raphson: evaluate J -1 (u i ) only once l Quasi-Newton: update J -1 (u i ) in a simple manner graphic representation of Newton’s method

© Fluent Inc. 1/10/2018L24 Fluids Review TRN Segregated Algorithm (1) u K(u) u = F is never formed u Rather, it is decomposed into a set of decoupled equations: l K u u - C x p = f u u momentum equation l K v v - C y p = f v v momentum equation l C x T u + C y T v = 0continuity equation l K T T = f T energy (scalar) equation u No explicit equation for pressure! l Replace continuity equation with Poisson-type pressure matrix equation (derived from manipulating discretized momentum and continuity eqn’s) u The pressure can be calculation in a number of ways

© Fluent Inc. 1/10/2018L25 Fluids Review TRN Segregated Algorithm (2) u Pressure projection method l given the current values of u, v and T, obtain an approximate pressure bo solving a discrete pressure equation l relax the pressure, i.e.: l using pnew, solve the momentum equations and energy equation using the newly computed velocities, solve for the pressure correction,  p adjust the velocity field (so that it obeys the incompressibility constraint) using  p u Advantage: less memory use u Disadvantage: more iterations Each equation set can be solved iteratively (inner iteration) or simulaneously (Gaussian elimination) uvpTuvpT outer iteration

© Fluent Inc. 1/10/2018L26 Fluids Review TRN Equation Solvers u Iterative l Non-symmetric equation systems: n Conjugate gradient squared n GMRES l Symmetric equation systems (pressure): n Conjugate gradient n Conjugate residual u Gaussian elimination

© Fluent Inc. 1/10/2018L27 Fluids Review TRN Underrelaxation u Two forms are used l explicit (similar to FVM approach) n carries some “history” forward n used with fully-coupled method n also used for pressure in segregated method l implicit n alters the weighting term for matrix diagonal n used for other equations (not pressure) with segregated method

© Fluent Inc. 1/10/2018L28 Fluids Review TRN Convergence u Various quantities can be used to judge convergence of an FEM solution u The more commonly used are: l Relative change in solution between iterations ||U i - U i-1 || / ||U i || < tolerance l Relative numerical accuracy (R is residual vector) ||R i || / ||R 0 || < tolerance