This work proposes a nonlinear finite element method whose nodal values preserve bounds known for the exact solution. The discrete problem involves a nonlinear projection operator mapping arbitrary nodal values into bound-preserving ones and seeks the numerical solution in the range of this projection. As the projection is not injective, a stabilisation based upon the complementary projection is added in order to restore well-posedness. Within the framework of elliptic problems, the discrete problem may be viewed as a reformulation of a discrete obstacle problem, incorporating the inequality constraints through Lipschitz projections. The derivation of the proposed method is exemplified for linear and nonlinear reaction-diffusion problems. Near-best approximation results in suitable norms are established. In particular, we prove that, in the linear case, the numerical solution is the best approximation in the energy norm among all nodally bound-preserving finite element functions. A series of numerical experiments for such problems showcase the good behaviour of the proposed bound-preserving finite element method.
翻译:一种保持节点界限的有限元法
翻译后的摘要:
本文提出了一种非线性有限元法,其节点值保持精确解已知的界限。离散问题涉及一个非线性投影算子,可以将任意节点值映射到保持界限的值,并以该投影的范围寻求数值解。由于投影不是单射的,因此添加了基于互补投影的稳定化,以恢复良好的解析性。在椭圆问题的框架下,离散问题可以被视为离散障碍问题的改进,通过Lipschitz投影将不等式约束并入其中。给出了线性和非线性反应扩散问题的示范,并建立了在适当的范数下的近最佳逼近结果。特别地,在线性情况下,数值解是所有保持节点界限的有限元函数中在能量范数下最佳逼近。一系列这类问题的数值实验展示了所提出的保持节点界限的有限元法的良好行为。