Open Equation Modeling For Process Systems Based On Object-Oriented Method

Computer integrated automation system plays an important role in process industries, in which process modeling and optimization are core technique for improving operation of process systems and achieving true potential of equipment. Nowadays, with numerical theories and computer technology progressing rapidly, open equation technique based on rigid mechanism is well recognized as state of arts in process system engineering. However, traditional methods of open equation modeling have great limitations in the maintenance and reuse of models, especially for large-scale modeling problems. The building and maintaining of process models with thousands and tens of thousands variables and equations are nightmare for chemical engineers. Legacy models could not be inherited and reused easily. Some new and efficient way to make models easily constructed and well organized is demanded heavily. In this thesis, open equation technique, general algebraic modeling languages and object-oriented method are deeply investigated and integrated. A new modeling approach is presented based on this. The main work could be stated as follows: 1. The histories of research work and industrial implementation of process simulation are systematically reviewed. The basic idea of open equation technique is given. It is pointed out that flexibility, extensibility and model-reuse is vital for developing and maintaining models easily. 2. Algebraic modeling language and mathematical programming software are discussed respectively. Studies on six kinds of optimization software show that GAMS is the best of all for formulation and solving of nonlinear simulation and optimization problems. The architecture of GAMS is investigated and its syntax is illustrated in an example. The advantage and disadvantage of using GAMS for open equation modeling is indicated. 3. The idea of object-orientation is expatiated and introduced for open equation modeling, in which 00 basics such as capsulation, inheritance and polymorphism are captured and adopted. The framework structure of the new modeling environment is proposed and designed. Some details such as communication and conjunction among classes are discussed. 4~ The new modeling approach is used for formulating simulation and optimization problems of chemical processes. The taxonomy, definition, and organization of chemical engineering class library are proposed and implemented. The validity, usability and efficiency are demonstrated in modeling of an adiabatic flash process.