EVALUATION OF WIND ENERGY SOURCES INFLUENCE ON COMPOSITE GENERATION AND TRANSMISSION SYSTEMS RELIABILITY Carmen Lucia Tancredo Borges carmen@dee.ufrj.br João Paulo Galvão joaopaulo@mercados.com.br Federal University of Rio de Janeiro COPPE – EE PO Box 68504, 21941 – 972 Rio de Janeiro – Brazil Abstract – This paper presents some results of studies performed to evaluate the impact of using wind energy based generation on power systems composite reliability evaluation. The reliability evaluation is based on non- sequential Monte Carlo simulation for consideration of both generation and transmission systems failure events. The stochastic models for the conventional generation units and the transmission elements are the two state or multiple states Markov models, usually adopted in reliabil- ity studies. For the wind energy based generation units, a special model is developed based on a combination of the wind velocity characteristic and the multiple power output levels associated with the velocity of the wind turbine. Three different impact studies is performed: wind genera- tion penetration degree, wind generation farm location and a comparison with other energy sources. The test system used is the IEEE RTS-79 modified to include the representation of the wind plants. The results obtained allow for the quantitative and qualitative analysis of wind generation influence on composite system reliability for several generation level and connection point scenarios. Keywords: Composite Reliability Evaluation, Wind Energy Systems, Statistical Clustering Technique, Monte Carlo Simulation. 1 INTRODUCTION A significant increase in the use of renewable energy sources for power system generation expansion is being observed all over the world in the last years. A number of strategic and commercial issues combined are re- sponsible for this increasingly interest, mainly [1]: - Pollutant gazes emission reduction; - Energy sources matrix diversification; - Electric sector deregulation and competition pol- icy; - Reduced construction time and capital cost of smaller plants; - Increase in energetic efficiency and rational use of energy; - Need for expansion of the generation system; - Availability of modular generation plants ; - Transmission costs reduction by allocation of gen- eration closer to the load. In Brazil, particularly, government programs stimu- late the investment in generation plants based on alter- native energy, such as biomass, wind energy, photo- voltaic energy, etc. Renewable sources, however, tend to have smaller energy density when compared to fossil fuels and, for that reason, the plants are smaller and geographically more distributed. The wind farms, in addition, need to be located in places with a favorable wind pattern and the smaller ones are, in general, con- nected to distribution and sub-transmission systems. They are not centrally dispatched by the Independent System Operator but, instead, by the Independent Pro- ducers according to the availability and the velocity of the wind. The large uncertainty present in wind farms generation influences strongly on power systems reli- ability. This paper presents some results of studies per- formed to evaluate the impact of using wind energy based generation on power systems composite reliabil- ity evaluation. The reliability evaluation is based on non-sequential Monte Carlo simulation for considera- tion of both generation and transmission systems failure events. The stochastic models for the conventional generation units and the transmission elements are the two state or multiple states Markov models, usually adopted in reliability studies. For the wind energy based generation units, however, a special model was devel- oped based on a combination of the wind velocity char- acteristic and the multiple power output levels associ- ated with the velocity of the wind turbine. The hourly based wind velocity curve is aggregated into a number of discrete wind velocity levels by the application of statistical clustering techniques [2]. The technique applied identifies the more significant veloc- ity classes in the curve and represent the entire curve in terms of these hierarchically clustered wind levels and their transitions to other levels. This is done in order to translate all possible wind velocity values into a statisti- cally representative and feasible number of levels for analysis. The clustered wind characteristic will provide different power generation levels according to the wind - power generation characteristic of the wind turbine. Therefore, the wind energy based generation unit is represented by a multiple states Markov model with transition rates and limiting state probabilities derived from the wind characteristic for the farm location re- gion. Three different studies were performed in order to evaluate the impact on composite reliability of the fol- 15th PSCC, Liege, 22-26 August 2005 Session 29, Paper 2, Page 1