Increasing complexity of mathematical models demands techniques of model order
reduction (MOR) that enable an efficient numerical simulation. For example, network approaches yield systems of differential algebraic equations (DAEs) in electric circuit simulation. Thereby, miniaturization and increasing packing densities result in extremely large DAE systems. MOR methods are well developed for linear systems of ordinary differential equations (ODEs), whereas the nonlinear case represents still an open field of research. We consider MOR for semi-explicit systems of DAEs. Techniques for ODEs can be generalized to semi-explicit DAEs by a direct or an indirect approach. In the direct approach, we introduce artificial parameters in the system. Accordingly, MOR methods for ODEs can be applied to the regularized system. Numerical simulations using the constructed MOR strategies are presented.