Introduction to Finite Difference Methods （介绍了有限差分方法）.pdfVIP

Introduction to Finite Difference Methods Since most physical systems are described by one or more differential equations, the solution of differential equations is an integral part of many engineering design studies. Your introductory course in Differential Equations has shown you how to solve many simple ODEs, and we have already reviewed some of the analytical methods studied in Differential Equations as part of this course -- and we will do several more examples before the end of the semester. In addition, Lesson 11 will focus on two specific numerical techniques for solving ODEs within Matlab. In particular, we will use Matlab’s ode45 routine to directly solve initial value problems (IVPs), and we will use this same ODE solver within an iterative method, called the Shooting Method, to solve a wide range of boundary value problems (BVPs). However, these methods are only useful for differential equations with a single independent variable -- that is, they can solve ordinary differential equations (ODEs). For problems involving more than one independent variable -- partial differential equations (PDEs) -- the most common solution method involves discretizing the original differential equation (DE) and converting it into a series of recursive and/or algebraic equations. Basically, this requires that one discretize each of the independent variables, essentially breaking the problem into a number of finite volumes or elements. Then, the continuous DE is evaluated at each discrete point or node in the system. Finally, upon approximating the discrete derivatives in the latter equations with finite difference approximations (as developed in the Lesson 4 Lecture Notes), one can completely convert the original continuous differential equations into a series of difference equations -- and it is the discrete difference equations that are evaluated on the computer. The so-called Finite Di