In situ machining of large workpieces is made possible by moving mobile machines to workpiece locations. To guide and enhance on site machining performance, dynamics of the assembled system must be predicted beforehand. Beforehand prediction of assembled response requires dynamics of the mobile machine in its free-free configuration be coupled with response of base/workpiece measured at location. Since obtaining the free-free response of the machine is non-trivial, this paper presents a substructure decoupling scheme that instead extracts the machine's response from known dynamics of the machine mounted on a base, and from information of a substructural base system. Simulation driven investigations show the decoupling scheme to be robust even under the vagaries of measurement noise and changes in the residual substructural base system. Extracted dynamics of the machine can be subsequently synthesized with known dynamics of other bases/workpieces for a priori prediction of assembled dynamics to guide first-time-right on site machining.