As a virtual digital model that can reflect physical entities or systems, digital twins are revolutionizing industry. The first prerequisite for the construction of digital twins is the establishment of high-precision and complex entities or system models. A 47-components numerical system is established for the core engine test rig main test system by using the finite volume modularization modeling method. A comprehensive solution to the system-level valve-spool/orifice throttling modeling, the key issue of the fluid pipeline system modeling, is presented, and the algorithms of throttling and mixing are deepened and expanded. The full-process simulation study on two tests of normal-temperature 1400 s and low-temperature 1240 s shows that the combined regulation of five regulator valves and the change of cold source directly decide dynamic change of the system in each stage; the simulation reveals the phenomena such as the gas cylinder cooling with deflation, the air cooling when expanding from main pipeline to two branch pipelines, shunting flow by branch pipeline, and the cold and hot gases mixing; the overall variation trends of the simulation curves are consistent with those of all the experimental curves of the test rig normal-temperature/low-temperature air supply lines, exhaust bypass, and engine main line in two operating conditions, and the maximum error between simulation curves and test curves of pressure, total pressure, and total temperature is less than 12%. The numerical system can be used for the construction of virtual models of digital twins, and the modeling method provides a feasible solution to the key technology of digital twins.