One-day ahead Power Forecasting is more and more required on the energy markets, and its accuracy is more and more crucial since it affects the net income of operators. 1. Weather Numerical Prediction, including a meso scale downscaling, provides a global prediction. A RANS CFD-tools is used for the micro-scale downscaling, providing a precise wind forecast at each wing generator hub. 2. To improve the reliability of this forecast, especially in the short term range, the use of "fresh" SCADA data is performed. Attention is focused on the Active Power, but other signals such as temperature and local wind characteristics can be taken into account. 3. In order to erase systematic errors and bias from the downscaled NWP based forecast (1.), as well as to mix it with the persistent model (2.), an Artificial Neural Network is trained using long term history. This paper explains first the method used and the choices made, especially concerning the Machine Learning parameters. A second part presents some results on some real cases, with different time horizons.
Automating Google Workspace (GWS) & more with Apps Script
Short term power forecasting Awea 2014
1. 0:00 0:30 1:00 1:30 2:00 2:30 3:00 3:30 4:00
0
5 %
10%
15%
20%
25%
Scada Horizon [H]
RMSEnormalizedbyParkNominalPower
Full approach
Raw persistence
Persistence + ANN
Raw NWP/CFD
NWP/CFD + ANN
Short term Power Forecasting: Mix of CFD Down scaling,
SCADA Data and Machine Learning
Carsten Albrecht , Julien Berthaut-Gerentes
Introduction: position of the problem Data sets
Deterministic Forecast: physical approach
Benchmark
Stochastic Approach: Machine Learning
References Presenter
Our deterministic approach starts from a Numerical Weather Prediction, which
predicts the “meteorological” wind near the wind farm area. Then, a Computational
Fluid Dynamics tool makes a micro down scaling to provide the future wind
characteristics at the exact location of each wind turbine. Then, thanks to the power
curves of the machines, the air density provided by the NWP, and a planning
maintenance given by the user, the final Power Output is calculated. This step
includes the interaction between the wind turbines, through out the Jensen wake
model.
Usually speaking, Forecast systems are classified along several axes:
• Intraday/Extraday, or Short Term/ Very Short Term. This has to do with the look
ahead time (horizon): from 0 to several hours for Very short term (Intraday) ,
from some hours to some days for Short term (Extraday).
• Deterministic/Stochastic. These two opposite concepts use different scientific
tools: physics and mechanics for deterministic approach, statistical tools and
machine learning for stochastic approach.
• Numerical Weather Prediction / Online Data: there are basically two sources of
data in order to perform a power forecasting. The first one is meteorological
predictions, based on global measurements and heavy numerical calculations,
the second is online measurement (for instance from SCADA system).
For each axis, one concept generally excludes the other. Intraday (Very Short term) is
commonly Stochastic with online measurements while Extraday (Short term) is
usually Deterministic based on NWP data.
This work aims to breakdown these classifications, proposing a unique tool based on
the unification of all these techniques.
Our Machine Learning procedure is based on an Artificial Neural Network. The
architecture is a supervised feed-forward fully connected network. It is trained using a
backpropagation learning process, and selected thanks to a genetic algorithm.
The optimization of input variables brought the following final set:
• Raw Power Forecast: the power provided by the deterministic approach
• Scada Production: the instant output power measured in the real wind farm
• Scada Horizon: the time lag between the last measurement and the
forecasted time
• Hour: the forecast time within the day (in order to take into account the
night/day effects, maybe omitted in the deterministic approach)
A usual question raised in Machine Learning concerns the historical data necessary
to train the network. In this case, the historical NWP meteorological data has to be
collected, as well as long term measurements on site (Observations). An automatic
batch for re-running the deterministic approach provides the Raw Power Forecast
historical data.
The NWP runs are provided twice a day and they cover
several days. This provides several forecasts for each
time step, with distinct NWP-horizons. In order to match
with observations (measured data), a two dimensional
vision of time has to be adopted: Observation time vs
NWP-horizon.
With the addition of online measurements, with
their own horizons, this two dimensional vision
of time has to be changed in a 3D time vision:
Observation time vs NWP-horizon vs Scada-
horizon. A single observation can be predicted
with different combinations of NWP and Scada
horizons.
This process replicates several times the same data (a single measurement is
duplicated along NWP and Scada horizon axes), which makes a strong correlation
between different data points. Thus, constructing the learning/test/validation by pure
randomization results in data snooping (data within different sets are strongly
correlated). A way to avoid this failure is to regroup the data by day before splitting
them into different sets randomly.
We compare the results of the whole process, with those form simpler approaches:
• pure persistence (forecasted power = measured power)
• persistence + ANN Learning,
• pure NWP/CFD (with no calibration at all)
• NWP/CFD + ANN.
This approach is tested on a large wind farm
(99 MW, 66 WTG) on a complex terrain. We
have about 1 year historical data. There are 4
NWP runs a day, with a 15 min time step,
leading to 2,5 millions of data points (250k=37
days for validation, 750k=108 days in test set,
1M5= 216 days in the learning set). The
results are based on the Normalized Root
Mean Square Error, binned by Scada
horizons.
Observation date
NWPHorizon
Observation date
NWPHorizon
Observation date
NWPHorizon
Observation date
NWPHorizon
Observation date
NWPHorizon
Observation date
NWPHorizon
Observation date
NWPHorizon
Observation date
NWPHorizon
Observation date
NWPHorizon
- ANN Power Forecast
Raw Power Forecast –
Scada Production –
Scada Horizon –
Hour (cyclic) –
2 hidden Layers
For horizon less than 30min, Raw persistence is still preferred. For all longer horizon,
the full model has a smaller error than other solutions. Numerical solutions
(NWP/CFD) beats online measurements after 3~4 hours.
GFS model Wind generator
characteristics
Global
wind
Local
winds
Raw Power
Forecast
1. Optimal combination of CFD modelling and statistical learning fr short-term wind power forecasting, Jérémie JUBAN,
AWEA 2013, Chicago
2. NWP Research and Its Applications for Power Systems in China, Dr Rong Zhu, ICEM 2013, Toulouse
3. Implementation of a Fast Artificial Neural Network Library (FANN), S. Nissen, University of Copenhagen (DIKU), 2003.
Dipl. Inf Carsten Albrecht
Managing director
albrecht@al-pro.de
Carsten Albrecht has been employed for four years by the company
ENERCON - the largest German manufacturer of wind energy converters. In
2001 he founded his own company AL-PRO. In July 2005 he became the
first chairman of the wind consult advisory board of the Federal Association
for Wind Energy in Germany (in short "BWE"). Since 2011, he distributes
Meteodyn softwares.