Tesis doctoral de Roberto Villafáfila Robles
Power systems are changing due to economic. Social , political and environmental lacts. On the other hand. Technological improvements have permitted he deployment 01 new type 01 generators within power systems. Aboye all at distribution networks. Such new generators connected al distribution networks are know as emph{distributed generation (dg)) . This means a shift irom traditional vertical to horizontal power systems. Moreover, mosl 01 his new generation is based on renewable sources and then, non-completely dispatchable due to their output powers vary stochastically according to heir prime mover furthermore. As distribution networks are spread out over limi1ed area. 1he dependence be1ween the such generators may have a significative impact. la large-scale integration 01 dg units in distribution systems might affect distribution network operators on their daily routine. Apart irom non ~ispatchable dg and their dependencies. Loads have a stochastic behavior, too. This situation needs a new approach in modeling in order to take into ilccount the inherent uncertainty 01 such elements on power system operation and planning lor assessing a proper power quali1y in steady-state pperation traditional deterministic analysis is limited when there is uncertainty beca use it only considers snapshots 01 the systems. On the other hand. probabilistic analysis provide a better understanding 01 uncertainty and its risk because variables are represented by probability lunctions. irhe present dissertation analyzes the probabilistic modeling 01 stochastic generators and their impact on distribution power systems, and has three main contributions. Firstly, an improving deterministic approach lor analyzing distribution systems with renewable source based dg units, namely wind power. Secondly, the modeling 01 stochastic generation and loads, and their correlation is discussed and performed. Finally. A monte cario simulation is performed to assess the impact on voltage profile in a real distribution system with dg units. And a control 01 voltage with the collaboration of dg units is proposed. irhe first contribution describes a new algorithm that improves deterministic load fiow lor distribution systems where several types of wind turbines are onnected. The effects on voltage profile and power generation 01 wind turbines are analyzed through simulations in a test network. The results show hat reactive power 01 fixed-speed wind turbines is sensitive to voltage variations at the point 01 connection. On the other hand, it is demonstrated that dg units can reduce power losses. irhe second contribution discuses the probabilistic modeling 01 generation and load in normal operation. Linearization. Norma lit y and independence have been usual assumptions lor probabilistic anal ysis. This might lead to an underestimation 01 the impacts. On one hand. Monte cario method avoid linearization 01 load flow equations. On the other hand. Methods as joint normal translorm and extreme stochastic bounds permit to estimate ~ependence between random variables. Even though when the correlations are not linear, as it is the case 01 wind power. Former lacts have been ested and used lor the probabilistic modeling 01 a real distribution system with dg units. irhe final contribution deals with the control 01 voltage in distribution networks in normal operation. Traditionally. Distribution network operators control ftoltage in normal by tap-changers installed at translormers considering voltage drop effect. However, dg units changes voltage profile and tap hangers may be affected. In order to avoid possible interferences wi th on-ioad tap-changer operation. A droop control strategy 01 reactive power 01 dg units in coordination with the tap-changer operation is proposed and simulated. It results in a flat voltage profile. Later. It is proposed and tested that dg units provide voltage support lor limiting tap-changer operations.
Datos académicos de la tesis doctoral «Probabilistic modelling in normal operation and the control of distribution systems with renewable source based dg units«
- Título de la tesis: Probabilistic modelling in normal operation and the control of distribution systems with renewable source based dg units
- Autor: Roberto Villafáfila Robles
- Universidad: Politécnica de catalunya
- Fecha de lectura de la tesis: 18/09/2009
Dirección y tribunal
- Director de la tesis
- Andreas Sumper
- Tribunal
- Presidente del tribunal: mircea Chindris
- alfred Rufer (vocal)
- sergio Herraiz jaramillo (vocal)
- antoni Sudria andreu (vocal)