Modeling of renewable energy based embedded generation in long term planning

Abstract

The share of Non-Conventional Renewable Energy (NCRE) penetration in Sri Lanka has been increased to a significant level in the recent past. As at 3 1 s t of August 2010, around 215 MW of NCRE based power plants were connected to the Sri Lanka national grid. This accounts to 6% of total energy and 7.8% percent of total capacity of the system. According to the Energy Policy - 2006, the government's Endeavor is to reach 10% of total energy in the national grid from NCRE by 2015. As a result, modeling and studying of NCRE is vital especially in generation planning. Though the present practice of using lumped NCRE thermal equivalent model is simple, it can cause a lot of inaccuracies owing to the volatile nature of wind and hydro sectors.

The issue of non-availability of suitable NCRE models was addressed by this research and separate NCRE models were prepared for each of the three major types of NCRE technologies; mini hydro, wind and biomass power plants, in Sri Lanka. Each NCRE technology was separately treated and modeled considering the seasonal variation of energy availability. Models were tested in generation planning software, Wien Automatic System Planning (WASP) to ascertain the accurate functioning of models. The mini hydro model prepared has five stage probabilistic distributions on a monthly basis. Past generation data of about 74 existing plants were used to validate the model. Predictability and adaptability of the model with WASP were tested and accurate dispatch of mini hydro power plants by the model was confirmed. Due to unavailability of reliable and continuous long term historic generation data, the wind model was prepared on measured wind speed data. After refinement of data, two modeling approaches were initially considered namely, a Run-of-River (ROR) type hydro equivalent wind model and the Modified Load Duration Curve method. Based on the results of error vs. commitment analysis, a five-stage probabilistic ROR type hydro equivalent wind model was selected as the best fitted model for the study. Biomass power plants were modeled as a conventional thermal equivalent plant.

NCRE models prepared were then used for three scenario studies to examine their adaptability in both committed and least cost basis. Various methods to interpret results of scenarios were conducted to verify that there is no curtailment of results due to the new NCRE models. The break even cost and avoided cost of generation of each NCRE technology were also calculated.