Title: Introducing Climate Variability in Energy Systems Modelling
A new peer-reviewed article by Clim2power project partners Filipa Amorim, Sofia G. Simoes, Gildas Siggini and Edi Assoumou is now available on the journal Energy. This paper, which is an output of the ongoing research within Clim2Power’s Portuguese case study, provides insights on the relevance of using a highly detailed spatial and temporal modelling tool in order to adequately capture climate variability to support strategic decision-making. The authors assess the relevance of using a highly detailed spatial and temporal modelling tool for studying the future evolution of the power system, considering the needs for integrating large shares of variable Renewable Energy Systems.
Abstract: This paper presents the ongoing research within Clim2Power project Portuguese case study. Its main goal (as a first step) is to show the relevance of using a highly detailed spatial and temporal modeling tool of the Portuguese electricity system in order to be able to adequately capture climate variability in the planning of the system up to 2050. To do so, we consider seasonal and intraday hydro, wind and solar resources variability in a large TIMES energy system model, in the eTIMES_PT model. Existing hydro, wind and thermal powerplants are modelled individually, whereas new plants are modelled at municipality level. The importance of introducing climate variability is assessed by modeling six scenarios: a reference case and both “humid” and “dry” hydropower scenarios. Each of these is also modelled with CO2 emissions cap by 2050. Results show that hydropower electricity generation variations are within range of those referred in literature by other authors. However, in this work, we are able to capture higher variations within seasons and time of day. Also, the analysis enables to account for the combined variability of hydro, PV and wind resources. This variability will subsequently consider data from seasonal forecasts and climate projections.
The full article is available through this link: https://doi.org/10.1016/j.energy.2020.118089
Title: Less Information, Similar Performance: Comparing Machine Learning-Based Time Series of Wind Power Generation to Renewables.ninja
Members of the Clim2power consortium Johann Baumgartner, Katharina Gruber, Sofia Simoes, Yves-Marie Saint-Drenan and Johannes Schmidt published an peer-reviewed article on the journal Energies as part of the special issue Modelling of Variable Renewable Generation: Wind and Solar Photovoltaic Power Plant. The paper assessed how time series generated by machine learning models (MLMs) compare to Renewables.ninja models (RN) in terms of their ability to replicate the characteristics of observed nationally aggregated wind power generation for Germany. The authors demonstrated that MLM models show a similar performance to RN, even when some information is unavailable.
Abstract: Driven by climatic processes, wind power generation is inherently variable. Long-term simulated wind power time series are therefore an essential component for understanding the temporal availability of wind power and its integration into future renewable energy systems. In the recent past, mainly power curve-based models such as Renewables.ninja (RN) have been used for deriving synthetic time series for wind power generation, despite their need for accurate location information and bias correction, as well as their insufficient replication of extreme events and short-term power ramps. In this paper, we assessed how time series generated by machine learning models (MLMs) compare to RN in terms of their ability to replicate the characteristics of observed nationally aggregated wind power generation for Germany. Hence, we applied neural networks to one wind speed input dataset derived from MERRA2 reanalysis with no location information and two with additional location information. The resulting time series and RN time series were compared with actual generation. All MLM time series feature an equal or even better time series quality than RN, depending on the characteristics considered. We conclude that MLM models show a similar performance to RN, even when information on turbine locations and turbine types is unavailable
The full article is available through this link: https://doi.org/10.3390/en13092277
Image caption: Scatterplot of modelled generation time series compared with observations (Renewables.ninja time series abbreviated as RN and machine learning model time series abbreviated as MLM1 and MLM2 versus observations on the X-axis)