Abstract
The paper presents the application of the contour integral method to solving mixed problems for a system of differential equations of heat and mass transfer under molecular and molar transport of energy and matter in an arbitrary three-dimensional domain. Furthermore, the contour integral method is used to construct the fundamental solution matrix of the aforementioned system in closed form, a compact formula representing the solution to the Cauchy problem is provided, and a scheme for applying the thermal potential theory method to solving a mixed problem that does not contain a time derivative in the boundary condition is indicated. Bibliography: 2.
Full Text
Preamble
This work, published in 1967, addresses boundary value problems for a specific class of partial differential equations. We consider the vector $x=(x_1, x_2, x_3)$ and an operator $A$ defined by coefficients $a_{ij}$ ($i, j = 1, 2, 3$). Following the methodology established in \cite{1}, we investigate the properties of the solution $v(x, t)$ within a domain $D$.
§ 1. Problem Statement and Fundamental Solutions
Consider the equation:
$$\begin{aligned} \Delta v - \lambda^2 v = \Phi(x) \end{aligned}$$
subject to the boundary conditions on $\Gamma$:
$$\begin{aligned} B(z, \lambda) v = \psi(z), \quad z \in \Gamma \end{aligned}$$
and the initial condition:
$$\begin{aligned} v(x, 0) = \Phi(x) \end{aligned}$$
We assume the following conditions hold:
1. The operator (1) satisfies the ellipticity conditions defined by I. G. Petrovsky.
2. The coefficients $a_k(z)$ and $P_k(z)$ ($k=0, 1$) are continuous on $\Gamma$, and the determinant $\det(a_0(z) + \lambda a_1(z)) \neq 0$.
3. The function $\Phi(x)$ is defined and sufficiently smooth in the domain $D$.
The characteristic equation associated with (1) is given by:
$$\delta(\mu) = \det(I \mu^2 + \sum a_i \mu + b) = 0$$
where $b_{ks}$ are coefficients related to the operator. Let $v_1, v_2, v_3$ be the roots of the characteristic polynomial:
$$v^3 + (a_{11} + a_{22} + a_{33})v^2 + (b_{11} + b_{22} + b_{33})v + \delta = 0$$
Under the assumption that the roots $v_k$ are distinct, the fundamental solution $P(x, \lambda)$ can be expressed via the roots of the characteristic equation. Specifically, we define the components $P_{mn}(x, \lambda)$ as:
$$P_{mn}(x, \lambda) = \frac{1}{8\pi} \sum_{k=1}^3 \frac{q_{mn}(v_k)}{\prod_{j \neq k} (v_k - v_j)} \frac{e^{-v_k |x|}}{|x|}$$
where $q_{mn}(v)$ are polynomials in $v$ determined by the coefficients of the original system.
§ 2. Construction of the Green's Function
To solve the boundary value problem, we introduce the potential $u_1(x, \lambda)$ and the integral representation:
$$u_1(z, \lambda) + \int_\Gamma K(z, y, \lambda) \mu(y, \lambda) d\Gamma_y = \psi(z, \lambda)$$
where the kernel $K(z, y, \lambda)$ is defined using the fundamental solution $P(z-y, \lambda)$ and the boundary operator coefficients. For sufficiently large $|\lambda| > R$, the kernel satisfies the estimate:
$$|K(z, y, \lambda)| \leq \frac{C}{|\lambda|} \exp(-\epsilon |\lambda| |z-y|)$$
This ensures the existence of a unique solution for the density $\mu(y, \lambda)$ via a Neumann series or Fredholm theory, depending on the domain geometry.
The Green's function $G(x, \xi, \lambda)$ for the problem is then constructed as:
$$G(x, \xi, \lambda) = P(x-\xi, \lambda) - Q(x, \xi, \lambda)$$
where $Q(x, \xi, \lambda)$ is the compensating term that ensures the boundary conditions are satisfied. We prove that for $x, \xi \in D$, the function $Q$ and its derivatives satisfy exponential decay estimates proportional to $\exp(-\epsilon |\lambda| |x-\xi|)$.
§ 3. Non-Stationary Solutions
The solution to the time-dependent problem (1)–(5) can be obtained using the inverse Laplace transform:
$$v(x, t) = \frac{1}{2\pi i} \int_{\gamma - i\infty}^{\gamma + i\infty} e^{\lambda t} \left[ \int_D G(x, \xi, \lambda) \Phi(\xi) dD_\xi \right] d\lambda$$
By analyzing the asymptotic behavior of $G(x, \xi, \lambda)$ in the complex $\lambda$-plane, we establish the existence and uniqueness of the solution $v(x, t)$.
For the case where $\Phi(x) = 0$ and the boundary conditions are non-homogeneous, the solution is represented as:
$$v_2(x, t) = \int_0 t d\tau \int_\Gamma Q(x-y, t-\tau) \mu(y, \tau) d\Gamma_y$$
where the density $\mu(z, t)$ satisfies a Volterra integral equation of the second kind:
$$\mu(z, t) = 2a^{-1}(z, t) \psi(z, t) + \int_0^t d\tau \int_\Gamma R(z, y, t-\tau) \mu(y, \tau) d\Gamma_y$$
The kernel $R$ is derived from the fundamental solution of the parabolic or hyperbolic system, ensuring the continuity of the solution up to the boundary.
References
- Ladyzhenskaya, O. A., Solonnikov, V. A., Uraltseva, N. N. Linear and Quasi-linear Equations of Parabolic Type. Nauka, 1967.
- Rikhl, M. L. Mathematical Methods in Physics. Nauka, 1964.