A Comparison of Triangular Mass-lumped Finite Elements for 2D Wave Propagation
W. A. Mulder
Abstract
W. A. Mulder
Abstract
Mass-lumped continuous finite elements provide accurate solutions of the second-order acoustic or elastic wave equation in complex geological settings, in particular in the presence of topography and large impedance contrasts. Elements of higher polynomial degree have better accuracy than those of lower degree but are also more costly. Here, the performance of elements of degree 1 to 6 is compared for a simple acoustic test problem. The element of degree 6 is new. The numerical test confirm the increase of accuracy with the polynomial degree of the basis functions. In terms of computational cost, the element of degree 4 performs best in the specific example. For higher degrees, the cost of having extra nodes and a more restrictive stability constraint on the time step can no longer be compensated by having a smaller number of larger triangular elements. Only at very high accuracy, the new element of degree 6 wins.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Mass-lumped continuous finite elements provide accurate solutions of the second-order acoustic or elastic wave equation in complex geological settings, in particular in the presence of topography and large impedance contrasts. Elements of higher polynomial degree have better accuracy than those of lower degree but are also more costly. Here, the performance of elements of degree 1 to 6 is compared for a simple acoustic test problem. The element of degree 6 is new. The numerical test confirm the increase of accuracy with the polynomial degree of the basis functions. In terms of computational cost, the element of degree 4 performs best in the specific example. For higher degrees, the cost of having extra nodes and a more restrictive stability constraint on the time step can no longer be compensated by having a smaller number of larger triangular elements. Only at very high accuracy, the new element of degree 6 wins.
Key concepts: Degree (music), Degree of a polynomial, Finite element method, Polynomial, Constraint (computer-aided design), Stability (learning theory), Simple (philosophy), Mathematical analysis