Reanalysis for wind and solar electricity simulations: challenges and lessons learned in the Renewables.ninja project - 2018-07-17.key
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Reanalysis for wind and solar electricity simulations: challenges and lessons learned in the Renewables.ninja project Stefan Pfenninger 1st International Symposium on Regional Reanalysis Dept of Environmental Systems Science 17.7.2018 Bonn, Germany With Iain Staffell, Imperial College London
100 GW Why variability is problematic GW 8080 GW Electricity Power generation and consumption 60 60 GW Demand Conventional (coal, gas) 40 40 GW Germany, May 2014 PV 20 20 GW Wind 0 10. May 0 GW 04:00 08:00 12:00 16:00 20:00 11. May 04:00 08:00 12:00 16:00 20:00 12. May 04:00 08:00 12:00 16:00 20:00 10. May 11. May 12. May Conventional power plants Solar Wind Wind Offshore Water Biomass Electricity Consumption Hard coal Lignite Nuclear Pumped hydro Natural gas Other • Demand and generation must match second by second Agora Energiewende; Current to: 16.07.2016, 19:10 • Electricity is difficult to store • Demand is not very controllable • Power stations are not like light bulbs • Everything must be carefully coordinated and scheduled https://www.agora-energiewende.de/en/topics/-agothem-/Produkt/produkt/76/Agorameter/ 5
Power System Localised Long-term Dispatch Wind and Solar Modelling Resource Profiles Long-term Multi-scenario European Power System Model x5 70 Five-fold increase in variability Total Generation Costs 65 Calm/Dull of CO2 emissions (€bn per year) 60 and total generation 55 Windy/Sunny costs by 2030 50 Calm/Dull Power system 45 Windy/Sunny functions well with 40 high penetrations 2015 2030 of renewables Collins et al. (2018). Joule [press embargo until 11:00am EST 26 July 2018] 6
Variable weather = variable curtailment Curtailment in ENTSO-E vision 3 (~35% of EU electricity from solar and wind) Collins et al. (2018). Joule [press embargo until 11:00am EST 26 July 2018] 7
Variability of CO2 emissions will rise 0% Emissions intensity variability in 2030 under ENTSO-E vision 3 (30% VRE) 15% Collins et al. (2018). Joule [press embargo until 11:00am EST 26 July 2018] 8
Virtual Wind Farm Model 25 m/s 20 Take hourly Interpolate wind speeds 15 from grid from a points to site 10 reanalysis of actual wind (originally 5 farms MERRA) 0 100 100% Height above ground (m) 80 Extrapolate Convert NASA data 80% wind speeds from wind Load factor 60 Log wind profile 60% to hub height speed to Extrapolated speed 40 at hub height with place- 40% power output and time- 20% Farm using whole- 20 Turbine specific farm power 0% 0 parameters 0 10 20 30 curve 0 2 4 6 8 10 12 Wind speed (m/s) Wind speed (m/s) Staffell and Pfenninger (2016) 10
Bias correction: not optional Average wind capacity factors in Europe Original model Actual data (uncorrected MERRA-1) Bias correction Staffell and Pfenninger (2016) 11
Bias correction using electric generation data Hourly weather conditions Bias-correct (wind speeds, sunlight, …) Solar and wind electricity Measured electricity simulation models generation data Hourly estimates of solar/wind power generation Validate Pfenninger and Staffell (2016), Staffell and Pfenninger (2016) 12
Bias-corrected wind simulations Simulating the UK wind fleet with the MERRA reanalysis: R² = 0.95 6 British Fleet Output (GW) 5 Actual Simulated 4 3 2 1 0 Jul-13 Aug-13 Sep-13 Oct-13 Nov-13 Dec-13 Jan-14 Staffell and Pfenninger (2016); actual data from Elexon & National Grid 13
25 years of daily variability Simulated UK wind and PV fleets, 1991-2015 14
3. Comparing reanalyses (solar PV only, for now) Image source: http://xarray.pydata.org/ 15
lower resolution Reanalysis and satellite datasets more coverage higher resolution less coverage Urraca et al., 2018: “We conclude that ERA5 and COSMO-REA6 have reduced the gap between reanalysis and satellite-based data” (for solar irradiance data) Urraca et al. (2018). Solar Energy Pfenninger and Staffell. Work in progress. 16
Germany: PV capacity factors (2012) 7-day mean capacity factor Pfenninger and Staffell. Work in progress. 17
Single PV system: SARAH best, COSMO-REA6 not bad Pfenninger and Staffell. Work in progress. Post-processed COSMO-REA6 data from Frank et al. (2018). Solar Energy 18
4. The ninja project 19
www.renewables.ninja Goal: provide easy access to validated wind and PV simulations worldwide. >1500 users from >250 institutions in 65 countries. Pfenninger and Staffell (2016). Energy; Staffell and Pfenninger (2016). Energy 20
renewables.ninja 21
renewables.ninja 22
External validation — National level PV Wind Moraes et al. (2018). Applied Energy 23
External validation — NUTS-2 level Moraes et al. (2018). Applied Energy 24
External validation “The lack of a trust-worthy source of data for comparison makes the evaluation of data quality challenging in many countries.” “[…] the chain of methods used to convert wind speeds and solar radiation into power outputs are decisive in this process, and the use of reanalysis data is promising” Moraes et al. (2018). Applied Energy 25
Renewables.ninja: Upcoming improvements • NUTS-2 level data for Europe should be online by end of this week • Global country and sub-country level data (e.g. U.S. states, Chinese provinces) later this year • New generation of reanalyses and better bias correction 26
1. 2. 3. 4. Why weather How we use Comparing The ninja matters reanalysis reanalyses project Discussion • What is ground truth? Reference energy datasets → better reanalysis validation and bias correction. • Which reanalyses have which strengths? Or, how to choose the right reanalysis for a particular task? • What improvements for energy applications are easily possible within existing / next-gen reanalyses? stefan.pfenninger@usys.ethz.ch | www.renewables.ninja | www.callio.pe Own publications available on www.pfenninger.org/publications 27
You can also read