nearly exact Boundary Element Method
...a novel approach to solve Poisson's equation numerically
neBEM
Home Application Solver Publication Gallery AboutUs Contact
neBEM The importance of inverse laws has been acknowledged in various branches of science and technology for a long time. Whenever, a particular physical phenomenon is modeled using sources or sinks, the inverse laws come into play. These laws are found to be crucially important in gravitation, electromagnetics, ideal fluid dynamics, Stoke's flow, acoustics, optics, thermodynamics and many other fields. In fact, a large part of the classical physics, when assumed non-dissipative, can be described by some form of the inverse laws such as the Laplace's and Poisson's equations. These two linear second order partial differential equations have been considered to be among the most important differential equations in the whole of classical physics. While the effect of point sources and sinks can be easily computed, it has not been possible to obtain closed form expressions for computing the effects of distributed sources, except for very simple cases. But, since in many of the real-life problems the singularities are found to be distributed on surfaces of various shapes and sizes, it has been customary to represent them using the simple shapes for which closed form expressions are known, or simply by assuming the surface to be composed of a large number of point sources. These approximations, besides being computationally rather expensive, turn out to be significantly restricted and inaccurate.

The ISLES (inverse square law exact solutions) library allows us to compute the influence of distributed singularities on rectangular and triangular flat elements. As a result, the neBEM (nearly exact Boundary Element Method) solver based on this library, can solve the Poisson's equation in 3D efficienty to a very high degree of accuracy. Due to the availablity of both rectangular and triangular elements, the solver is flexible enough to solve problems involving complicated geometries.