A load rejection disconnects the generator from the electrical grid. The resulting power excess accelerates the turbo set. Reacting to the load rejection, the turbine governor rapidly closes the steam admission valves. The remaining entrapped steam expands, thereby continuing to power the turbine. Thus the turbine speed rises till a dynamic equilibrium of accelerating and braking forces is reached. Thereafter the turbine speed decreases. If the maximally attained turbine speed remains below the trip threshold, immediate re-synchronization to the electrical grid is possible. Consequently, a forced outage of the steam turbine can be avoided and operational reliability is increased. Furthermore, functional safety requirements demand that the maximum turbine speed remains below test speed under all failure conditions. Accordingly, steam turbine design has to account for the impact of overspeed for a reliable and safe operation of the turbo set.
In order to manage load rejection requirements for steam turbine operation, the design engineer applies standard rules and overspeed calculation methods. These rules limit standardized overspeed estimation by defining maximum steam volumes, valve closing times, and I&C reaction times, as well as type and number of non-return valves.
A more thorough turbine overspeed investigation is necessary for several reasons, such as to evaluate this behavior under undesired failure conditions e.g. failure of non-return valves or blocking of control valves. A second justification for this investigation would be to predict changes resulting from turbine modifications — e.g. turbine upgrade or change at I&C systems.
In this paper, basic and advanced overspeed calculation tools are illustrated and compared, with respect to required effort as well as accuracy of prediction. It is shown how system parameters which are most sensitive with respect to overspeed can be identified and their influence assessed. Thus, firstly it is already possible to identify and improve critical overspeed behavior during design. Secondly, the impact of particular failures can be accurately predicted, thus allowing for due implementation of appropriate counter measures.
The methods, presented in this paper, were developed by the authors and their predecessors at SIEMENS AG for large steam turbo sets with a power range between 100 MW and 1500 MW.