The rotor of a steam turbine is a main element for converting thermal energy into mechanical energy, and its lifetime affects the safety of the whole unit. The units in power plants face several constraints during operation. For instance, the start-up time of a steam turbine is affected by thermal stress. Especially, due to the heavy variation of operation environments, some features of a rotor such as thermal stress, apparent expansion and thermal strain get worse. If they exceed the limitation, the corresponding elements are damaged or even destroyed, causing fatal accidents.
For rapid start-up of a unit, it is necessary to find out its appropriate start-up curve, and it should be operated accordingly to improve the safety and economic efficiency of the plant.
Optimization of turbine start-up in power plants has been studied in many aspects so far. Most of the models were built for real-time simulations. They reflected the features of stage groups of turbine using real-time operation database and various equations for quickness, and in most cases, the efficiency of stage groups were considered to be invariable. In order to determine the thermal stress exerted on the rotor during the start-up and shut-down of a turbine, it is necessary to accurately evaluate the inlet and outlet parameters of the nozzle and blades at each stage, not stage groups. Therefore, detailed calculations for each stage should be conducted to determine the pressure and enthalpy at the nozzle and blades, and their relative internal efficiency.
Rim Ju Yong, a researcher at the Faculty of Thermal Engineering, built a model of a 210MW subcritical condensing steam turbine for a power plant to determine the parameters mentioned above, and composed a special program with VC. Then, according to the operation regulation of turbine, he framed various cold start-up schedules. After that, he calculated the steam parameters on the parts of all stages by using the special program, and studied the temperature field and stress field by the finite element method (FEM).
Taking the calculation results as sample data, he established a regression model to determine the stress at the risk point of rotor. Based on the regression model, he determined the optimal cold start-up schedule using genetic algorithm (GA).
The results show that the fatigue lifetime increases by 20% with the decrease in the maximum von Mises stress by 1.4%, and the start-up period of the optimized schedule is reduced by 124min compared to the original one.