Posters

Abstract

Since the European energy system faces an unprecedented crisis, the long-term financial viability of any renewable energy project is one of the most important but also difficult factor to estimate. For that purpose, developers often turn to Power Purchase Agreements (PPAs) in order to warrantee a stable energy price in time, which ensures a reliable estimation of the financial profitability of the project. With that in mind, energy production hourly series for long periods of time become of great interest, especially for different exceedance levels. Financial analysis of a project is typically done from a long-term hourly or monthly energy production estimate, preferably for P90. However, this estimate will not be the P90 energy production value for each separate hour, but rather a distribution of the energy production of a long-term representative year at P90. Nonetheless, the form in which the production is distributed is of great importance in order to determine the profit yielded by the project according to market prices. Generally, the P90 hourly distribution is done based on the behavior of the P50 hourly production series. The hourly series for P90 is then constructed by adjusting the value for each hour proportionally, so that the addition of all hours in a year is equal to the 1-year P90 energy production value. Even if this process is done thoroughly, the resulting series will never accurately describe the behavior of the hourly series for P90, as it applies the global variability value for each hour and month, dismissing the variability differences among seasons, months or hours. By contrast, if the P90 energy production value was calculated for each hour separately, the resulting hourly production series would be noticeably lower, and its addition would not equal the 1-year energy production value for P90. That is because the probability of P90 hours happening consecutively during a year is lower than 10%, and hence, would not match a 1-year P90 production, but rather the value for a higher exceedance level. In spite of that, the hourly production series would be a precise distribution of the 1-year production value at that exceedance level. Following that reasoning, a procedure could be applied to find an exceedance level that, when applied to the hourly energy production, results in an hourly series that could be added up to a P90 long-term 1-year energy production. With this method, the resulting hourly production values in the series would be equiprobable, and describe rigorously the expected project performance for reduced periods of time. In conclusion, we intend not only to bring into light the risks of prorating the P50 behavioral pattern in order to calculate hourly production at greater exceedance levels, but also to propose a solution that will produce accurate long-term hourly production series for P90. The resulting hourly production series could then be crossed with the energy market price and deliver a trustworthy financial estimation, that can aid the adjustment of PPAs, and avoid inaccurate hourly and monthly production values that would yield incorrect results.