1: \begin{abstract}

2: We present a method to extend the finite element library FEniCS to solve problems with domains in \eirik{dimensions above three by constructing} tensor product finite elements. This methodology only requires that the high dimensional domain is structured as a Cartesian product of two lower dimensional subdomains. In this study we consider linear partial differential equations, though the methodology can be extended to non-linear problems. The utilization of tensor product finite elements allows us to construct a global system of \eirik{linear algebraic} equations that only \eirik{relies on} the finite element infrastructure of the lower dimensional subdomains contained in FEniCS. 

3: \eirik{We demonstrate the effectiveness of our methodology} in three \eirik{distinctive} test cases.  The first test case is a Poisson equation  \eirik{posed in} a four dimensional domain which is a Cartesian product of two unit squares \eirik{solved using the classical Galerkin finite element method}. The second test case is the wave equation in space-time, where the \eirik{computational} domain is a Cartesian product of a two dimensional space grid and a one dimensional time interval. \eirik{In this second case we also employ the Galerkin method. } Finally, the third test case is an advection dominated advection-diffusion equation where the global domain is a Cartesian product of two one dimensional intervals. The streamline upwind Petrov-Galerkin method \eirik{is applied to ensure discrete stability}. In all three cases, optimal convergence rates are achieved with respect to $h$ refinement. 

4: \end{abstract}