Additive manufacturing (AM) of large-scale polymer and composite parts using robotic arms integrated with extruders has received significant attention in recent years. Despite the contributions of great technical progress and material development towards optimizing this manufacturing method, different failure modes observed in the final printed products have hindered its application in producing large engineering structures used in aerospace and automotive industries. We report failure modes in a variety of printed polymer and composite parts, including fuel tanks and car bumpers. Delamination and warpage observed in these parts originate mostly from thermal gradients and residual stresses accumulated during material deposition and cooling. Because printing large structures requires expensive resources, process simulation to recognize the possible failure modes can significantly lower the manufacturing cost. In this regard, accurate prediction of temperature distribution using thermal simulations is the first step. Finite element analysis (FEA) was used for process simulation of large-scale robotic AM. The important steps of the simulation are presented, and the challenges related to the modeling are recognized and discussed in detail. The numerical results showed reasonable agreement with the temperature data measured by an infrared camera. While in small-scale extrusion AM, the cooling time to the glassy state is less than 1 s, in large-scale AM, the cooling time is around two orders of magnitudes longer. © 2022 by the authors