Power system optimization in the presence of large amount of renewables

AN Island is formed when a subtransmission system is disconnected from the main grid and remains operational as an island entity. Islanding is either due to intentional events, e.g. maintenance or due to unintentional events e.g. faults and their effects on the switching actions. In both intentional and unintentional cases, some issues can be considered for islanding such as: loss of control over voltage and frequency, safety issues, excessive transient stresses and reactive power problems. In this thesis, a procedure to guaranty a stable island area considering as much as possible connected load in the subtransmission system is proposed. Wind Farms will be among the most important contribution to the achievement of EU goals, as it is estimated that their power production will keep growing and will overtake by up to 30% of the production generated. In this context, a test system contains a large number of wind turbine connected to subtransmission system that could be tested in island conditions applying different scenarios on it. Wind turbines that are modeled in this project, do not participate in the network voltage control and they work at unitary power factor. Such plants cannot be simply modeled by a conventional synchronous or asynchronous machine for two reasons: first, the mechanical and electrical parts and the controls of the wind generators are different from the conventional ones; and second, wind farms usually contain a large number of generators (synchronous, induction) with various characteristics. The lack of voltage regulation ability makes most wind turbine generators need to absorb reactive power from grid with active power injected into grid; this can result in voltage deviation specifically after islanding in subtransmission grids. With wind power installation increasing, the voltage problem will become more serious. In this thesis, a new constraint is designed to assure that after islanding there is enough reactive reserve to balance the reactive power of the connected loads and to maintain the voltages in a desirable range. The main idea of the research project is to define the possibility, for the subtransmission network or a part of it, to survive from a disconnection from the bulk power system. Also, the goal of the project is to define: which part of the subtransmission system is likely to survive after a disconnection from the bulk power system; which emergency control actions would be necessary to guaranty its stable islanding operation. In this way an innovative method is proposed to determine in advance the possible control actions needed in case of a total or partial disconnection of the subtransmission grid to guarantee a stable islanding operation thus avoiding its blackout. Therefore, the control actions are determined and the results are distributed among the Substation Automation Systems (SAS) to be carried out whenever necessary. As the control actions should be applied immediately after the time of the islanding, they need to be computed off-line, e.g. every 10 minutes, and transmitted to the local SAS and applied in case of islanding, depending on the topology of the new formed island. To define the emergency control actions, the SAS collects the information about the network state and determines the possible emergency control actions and then sends the emergency control actions to relevant breakers of loads and generators. The island feasibility function is designed to maximize the loads that can be kept in service after any islanding, considering both any available controllable resources and the significant renewable energy generation. The procedure of islanding mathematically can be expressed in two optimization problems; one maximizes loads and the other one maximizes the sum of the loads and generation in the subtransmission systems. The objective functions subject to constraints that are related to both, real and reactive power of the network. The constraints are relevant to: balance the long term load fluctuation; the mitigation of the initial imbalance between the total load and total generation, when islanding occurs; balance reactive power in islanded area. The optimization model is a mixed integer non linear programming model and has been solved using both GAMS and the COINBONMIN solver and also, Genetic Algorithm (GA) in MATLAB. Then, the output of the optimization procedure has been verified by means of dynamic simulations within Digsilent. The procedure has been tested on two 150 kV subtransmission networks. Combination of parameters has been studied and eventually the results from the mathematical model above described have been validated by the dynamic simulations. Moreover, it is concluded from the results that the dynamic simulation results are very sensitive to such parameters, thus making it necessary to adopt as far as possible, the real and actual data of the system in order to obtain accurate and really applicable results. The main features have been tested and the conclusions are: in almost all cases, the feasibility procedure finds a solution, and allows to continue to supply significant amount of load. This conclusion is dependent on the dynamic response of the system that is likely to be much different from case to case. numerical results are evaluated by dynamic simulations that are successfully carried out and frequency and voltage behavior of island area are desirable.a constraint related to reactive power in subtransmission grid is applied locally to the island area that is divided in small areas due to the local behavior of reactive power; the results show the reactive power constraint finds good solutions for islanding conditions.

Nel corso degli ultimi anni, la penetrazione di generazione rinnovabile (RES) non programmabile è andata sempre aumentando. Certamente, l'integrazione di una grande quantità di RES in un sistema elettrico è una sfida, ma la tecnologia fornisce molti strumenti per utilizzare e gestire queste fonti con criteri e risultati che si possono in qualche modo assimilare a quanto ottenibile dalle fonti tradizionali. Un esempio riguarda l’impiego di tale generazione per consentire, se opportunamente coordinata, il funzionamento in isola di una porzione di rete di subtrasmissione. Nel presente lavoro, si presenta una procedura di valutazione (off-line) che, da una parte, indaga l'attuabilità del funzionamento in isola e, dall'altra, ne pianifica le azioni di controllo in emergenza. Questa procedura definisce, in caso di funzionamento in isola, quali generatori e carichi potrebbero rimanere connessi, in maniera tale da garantire il bilancio di potenza attiva e reattiva della rete. Per ogni possibile insieme di eventi che porterebbero la rete a funzionare in isola, la procedura determina e comunica al sistema di automazione locale l'insieme delle azioni di controllo discrete (apertura degli interruttori) da attuare. I risultati, che includono anche la risposta dinamica del sistema, sono presentati con riferimento sistemi di subtrasmissione reali.