Graden avlegges ved Universitetet i Sørøst-Norge (USN), Fakultet for teknologi, naturvitenskap og maritime fag.
Electric generators in hydropower systems are designed to handle variations in power usage and to withstand basic faults in the electric grid. Moreover, there are strict restrictions on how hydrogenerators can be operated to avoid overheating with resulting short lifetime. A hydrogenerator can always provide much more short-term than it's capable of providing in long-term. It is precisely in the short period of time that the power system has the biggest challenges.
In this work, the possibility to break the strict operational restrictions from authorities has been considered. This gives more freedom to handle variations in power consumption and grid faults without tripping the hydrogenerator. The advantage is improved possibilities to keep the grid and the hydrogenerator running through short term problems. However, this can not be done in such a way that the hydrogenerator is destroyed. To avoid damaging the generator, a model of temperature evolution of the generator has been developed and validated. This model gives a quantitative description of how generator operation and grid conditions influence the temperature.
Next, a new control strategy has been developed which attempts to keep the hydrogenerator/grid working in an optimal way during power variations/grid faults, while simultaneously making sure that the temperature of the hydrogenerator is kept under control, to minimize thermal reduction of generator lifetime.
More specifically, a single hydrogenerator has been considered, connected to an infinite bus. A low order thermal model is used, and system states are estimated from measurements using an Unscented Kalman Filter (UKF). The state estimates from the UKF has been used in a nonlinear model predictive controller (NMPC); the NMPC has been integrated with the standard generator controller.
Since the hydropower generator's golden age in the 20th century, few innovative steps have been provided to increase the utilization of the hydrogenerator. However, in this work, a new thermal prediction algorithm for the flexible control of an air-cooled hydrogenerator is proposed.
Two popular science articles from this work were recently published:
Teknisk Ukeblad (in Norwegian):