Stability of power systems with large amounts of distributed generation

University dissertation from Stockholm : Elektrotekniska system

Abstract: This four-part dissertation is essentially concerned with some theoretical aspects of the stability studies of power systems with large penetration levels of distributed generation. In particular, in Parts I and II the main emphasis is placed upon the transient rotor angle and voltage stability. The remaining two parts are devoted to some system-theoretic and practical aspects of identification and modeling of aggregate power system loads, design of auxiliary robust control, and a general qualitative discussion on the impact that distributed generation has on the power systems.One of the central themes of this dissertation is the development of analytical tools for studying the dynamic properties of power systems with asynchronous generators. It appears that the use of traditional tools for nonlinear system analysis is problematic, which diverted the focus of this thesis to new analytical tools such as, for example, the Extended Invariance Principle. In the framework of the Extended Invariance Principle, new extended Lyapunov functions are developed for the investigation of transient stability of power systems with both synchronous and asynchronous generators. In most voltage stability studies, one of the most common hypotheses is the deterministic nature of the power systems, which might be inadequate in power systems with large fractions of intrinsically intermittent generation, such as, for instance, wind farms. To explicitly account for the presence of intermittent (uncertain) generation and/or stochastic consumption, this thesis presents a new method for voltage stability analysis which makes an extensive use of interval arithmetics.It is a commonly recognized fact that power system load modeling has a major impact on the dynamic behavior of the power system. To properly represent the loads in system analysis and simulations, adequate load models are needed. In many cases, one of the most reliable ways to obtain such models is to apply a system identification method. This dissertation presents new load identification methodologies which are based on the minimization of a certain prediction error.In some cases, DG can provide ancillary services by operating in a load following mode. In such a case, it is important to ensure that the distributed generator is able to accurately follow the load variations in the presence of disturbances. To enhance the load following capabilities of a solid oxide fuel plant, this thesis suggests the use of robust control. This dissertation is concluded by a general discussion on the possible impacts that large amounts of DG might have on the operation, control, and stability of electric power systems.