Research And Application On General Dynamic Process Simulation System And Its Core Technologies

Dynamic technology is developed on the basis of steady-state simulation technology, and thus it is more complex and needs higher technical requirements. In present, it is necessary to develop dynamic process simulation for chemical and refinery plants by using the large complex system due to their higher complexity and requirement of automation level. Dynamic process simulation system can be used to train operators, shorten the time of start-up, save costs of material, help chemical engineers and designers to achieve process optimization, control optimization and extend safe operation time. Thus it is of great significance to develop a chemical dynamic process simulation system with independent intellectual property rights.The major research contents and results of this thesis are as follows:(1) In the software development of dynamic process simulation system, using object and runtime environment of software are determined. It was proposed that the core codes were written using C/C++, while the communication system was written by Java since Java is better than other languages in terms of its portability and memory management.(2) A chemical dynamic process simulation system was developed by system intergration. Process model subsystems had been built including node model, heat exchanger model, flash model, distillation model, reactor model, etc.. DCS control system had been built, including proportional control model, integral control model, differential control model and interlock control model. In interlock logic, only basic logic models (AND, OR, NOT, etc.) were built and programmed, and all complex logics were implemented by simple combination of basic logics. Four types of data transmission, i.e. long connection, short connection, synchronous connection, asynchronous connection, were analyzed and compared. The asynchronous long connection based on TCP/IP protocol was determined as communication mode Because of the huge data in dynamic simulation system and the requirement of frequent data transfer.(3) After analyzing two common methods for solving equilibrium, a new method, homotopy continuation, was proposed to solve ill-posed problems in equilibrium model. By adding adaptive parameters, homotopy equation was established. By adaptive parameters correction, Numerical continuation was achieved from homotopy equations to equilibrium equations. Using this new method, normal boiling points of pure components were calculated with relative error smaller than 1%. For the ternary system, the maximum relative error of bubble point temperature and composition are 1.4%, and 4.8%, respectively. The excellent results justify the applicability of this new method for solving equilibrium. Bubble point temperature for binary system was calculated by new method and compared with common method. By keeping results with the same number of significant digits, the computing speed of the new method is much faster than the common method. Also the influence of initial composition and initial temperature on equilibrium calculation was analyzed. The initial composition does not affect equilibrium calculation for the two methods. However, the initial temperature shows different influence, namely, the common method is sensitive to initial temperature and divergence may occur with inappropriate initial temperature. In contrast, the new method does not require an initial value, and the iterative calculation can be converged in a wide range of domains.(4) A new algorithm for flash model was built based on homotopy continuation method. The average relative error is about 3% for the calculated compositions of vapor and liquid phases by using this method. It proves the correctness of this method. The results by this method show that the effects meet or exceed the commercial software.(5) According to interactive feature in dynamic process simulation system, a mathematic model of pipeline network was established. Automatic identification and modeling of pipeline network were achieved by adjacency matrix. In order to solve ill-posed problem of model solution, a new algorithm for solving pipeline network model was proposed. The amount of calculation is effectively reduced and the computational efficiency is improved.(6) The dynamic sequential modular approach was adopted as the whole process solving strategy. The traditional sequential modular approach was improved for solving multi-scale and time lag units in dynamic process simulation. This method improves computational efficiency and numerical stability.(7) Based on the characteristics of mechanistic model, a mathematical model of real component method was established for dynamic simulation of the refining process. Characterization of petroleum was achieved. The component fraction equation was obtained, and it was solved by Newton iterative method. The modeling and solution for dynamic simulation of the refining process were determined.(8) Using the above model, a dynamic simulation system was established by chemical dynamic process simulation system and related algorithms for the diesel hydrotreating process. Compared with actual values, the maximum and minimun relative error of the simulation results are 6% and 1%, respectively, for each equipment, which meet the practical needs of engineering. The true boiling point curve calculated for refining diesel is consistent with actual data. The dynamic response trends of PID controller model and interlock model are consistent with the actual ones. All of the above results justified the applicability of the mathematical model of real component method and the validity of chemical dynamic process simulation system developed in the present work. And it has good results in computing speed, convergence and numerical stability.