Excitation system – simulation, tuning and evaluation

Abstract: Swedish Electrical and Power Control AB (Svea Power) offers excitation systems for electrical production plants. An excitation system controls the electrical performance of the generator’s three phase outputs and must meet requirements set out in the business agency “Svenska Kraftnäts (SvK:s) föreskrifter, SvKFS 2005:2”. This thesis is based on three of the requirements in SvKFS2005:2. First, there is a requirement on the automatic voltage regulator (AVR) response time, when performing a step in the AVR reference signal. Another requirement states that the AVR must be equipped with an attenuation function, which attenuates angular oscillations occurred in the rotor when the power load is changing. The third requirement states that a production plant should manage a voltage variation (following a short circuit profile) with the network connection retained, at the generator voltage terminals. These three requirements are important when designing an excitation system and it would be preferable to do simulations of an excitation system with respect to these three requirements. Therefore, the aim of this thesis is to create a simulation program to simulate and justify that the requirements are fulfilled when installing a new excitation system on a certain generator.    A graphical simulation tool with excitation control, capable of electrical simulations on a Single Machine- Infinite Bus (SMIB) model was defined in a preparatory study.  As a result of the preparatory study, a program called ATPDraw was selected to be the simulation tool.   A simulation program in ATPDraw, containing a Single Machine- Infinite Bus system equipped with an automatic voltage regulator (AVR) and a power system stabilizer (PSS) was created. Inductances and other fundamental parameters for the SMIB model can be filled in to match different kinds of generators and grid characteristics. The AVR and PSS in the simulation program were created according to Svea power standards.    The simulation program was validated by inserting parameters from real plants into the program and comparing the simulated curves with curves obtained from tests performed on real generators. i.e. the simulation program is a copy of the real plant and it is exposed to the same tests. Three different types of real production units were simulated to tune and verify the simulation program.    The results show that simulated values for the AVR step response correspond very well to the real measurements, after adding a factor that compensates for errors occurring from cold field windings in the real units. The rotor oscillation simulations correspond very well to two units. The third unit is located among other generators and is probably therefore affected by those, because the error is eliminated when the generator moment of inertia is heavily increased. The SMIB model is hence too simple in such a case.    Fault-ride-through tests are not performed by Svea Power and thus there are no real results available to compare with. Therefore, the fault-ride-through simulations are verified against another report where short circuits are simulated. The system setup in the compared report is a bit more advanced than the SMIB model and the results deviates slightly from each other.    The conclusion is that the model is reliable and accurate for AVR step tests, if one compensates for cold field windings. The rotor oscillation simulations are accurate when there is only one unit that is operative at the production plant. The fault-ride-through test is reliable if the simulated generator synchronism to the grid is retained. Because of uncertainty and lack of verifications, it is considered as not reliable if it is not retained.