An unconditionally stable finite difference formula for a linear second order one space dimensional hyperbolic equation with variable coefficients
Introduction
We consider the one space dimensional linear second order hyperbolic equation with variable coefficients of the formdefined in the region Ω × [0 < t < T], where Ω = {x∣0 < x < 1} and T is a positive integer.
The initial conditions consist ofand boundary conditions consist ofwhere u = u(x, t). We assume that A(x, t) > 0 in the region Ω × [0 < t < T]. For α, β, A are constants and α > β > 0, A = 1, the equation above represents a telegraph equation. We shall assume that the initial and boundary conditions are given with sufficient smoothness to maintain the order of accuracy of the difference scheme under consideration.
A rectilinear grid with sides parallel to coordinate axes is superimposed on Ω × [0 < t < T] with spacing h > 0 and k > 0 in x- and t-directions, respectively. λ = (k/h) > 0 is the mesh ratio parameter. The internal grid points (x, t) are given by x = xl = lh, 0 ⩽ l ⩽ N + 1 with (N + 1)h = 1 and t = tj = jk, 0 ⩽ j ⩽ M + 1 with (M + 1)k = T, where N and M are positive integers. Again let and be the exact and approximate values of u at the grid point (l, j) ≡ (xl, tj), respectively.
Difference schemes of O(k2 + h2) and O(k2 + h4) for the one-space dimensional linear hyperbolic equations with constant and variable coefficients were discussed in [1], [2], [3]. In most of the cases, these schemes are stable for 0 < λ < 1. Recently, Mohanty et al. [4], [5] have proposed three level implicit unconditionally stable difference schemes of O(k2 + h2) for the linear hyperbolic equations in two and three space dimensions, respectively. To the author’s knowledge no unconditionally stable difference scheme for the linear hyperbolic equation (1) is known in the literature so far. In this article, we describe a new three level implicit difference scheme (see Fig. 1) of O(k2 + h2) for the linear hyperbolic differential equation (1), which is stable for 0 < λ < ∞. The most important feature of our scheme is that we solve only tri-diagonal equations and that fictitious points are not needed at each time step along the boundary.
Section snippets
Mathematical details of stability analysis
Now, we discuss the finite difference method for the hyperbolic equation (1) and its stability analysis.
We need the following approximations for the partial derivatives of u. Let
At the grid point (l, j), the proposed differential equation (1) may be discretized bywhere , , , , etc.
Computational result
In order to test the efficiency and viability of the chosen method, we have solved the proposed differential equation (1) for different values of α(x, t) > 0, β(x, t) > 0 and A(x, t) > 0. The exact solution values for all cases are given by u = e−2t sinh x as the test procedure. The initial and boundary conditions and right hand side function f(x, t) can be obtained using the exact solution. The proposed scheme (11) is an implicit three level scheme. To start any computation, it is necessary to know the
Concluding remarks
In this article, we have discussed a new three level implicit difference scheme (11) of O(k2 + h2) for the difference solution of the linear hyperbolic equation (1). The scheme is stable for 0 < λ < ∞ and applicable to singular equations. We have also discussed an explicit scheme of O(k2) for the difference solution of u at t = k, which is also applicable to singular equations. The usefulness of the proposed method is exhibited from the numerical results. In all cases, we have chosen α(x, t) > β(x, t) > 0.
References (7)
- et al.
Fourth order approximations at first time level, linear stability analysis and the numerical solution of multi-dimensional second order nonlinear hyperbolic equations in polar coordinates
J. Comput. Appl. Math.
(1998) An operator splitting method for an unconditionally stable difference scheme for a linear hyperbolic equation with variable coefficients in two space dimensions
Appl. Math. Comput.
(2004)- et al.
A note on the operator compact implicit method for the wave equation
Math. Comput.
(1978)