Gproms量纲分析和推理

In the thesis we describe a dimensional analysis and inference system for a strongly typed language for simulation of physical systems, gPROMS. We show how gPROMS can benefit from such a system and we believe that it will increase the physical correctness of the simulation models.
Our proposed dimensional inference system is based on generating dimensional constraint equations from the model equations using formal inference rules. The resulting constraints constitute a linear system of equations which is solves by means of linear algebra. If the system of equations is solvable the simulation model is proven to be dimensionally consistent.
We have implemented our dimensional inference algorithms in Java in the form of a stand-alone tool. The tool will be used by the industry to quickly locate physical errors in simulation models resulting from dimensionally inconsistent equations.