A Wind Power Forecasting System to Optimize Power Integration

Abstract

The National Center for Atmospheric Research (NCAR) has developed a wind prediction system for Xcel Energy, the power company with the largest wind capacity in the United States. The wind power forecasting system includes advanced modeling capabilities, data assimilation, nowcasting, and statistical post-processing technologies. The system ingests both external model data and observations. NCAR produces a deterministic mesoscale wind forecast of hub height winds on a very fine resolution grid using the Weather Research and Forecasting (WRF) model, run using the Real Time Four Dimensional Data Assimilation (RTFDDA) system. In addition, a 30 member ensemble system is run to both improve forecast accuracy and provide an indication of forecast uncertainty. The deterministic and ensemble model output plus data from various global and regional models are ingested by NCAR’s Dynamic, Integrated, Forecast System (DICast®), a statistical learning algorithm. DICast® produces forecasts of wind speed for each wind turbine. These wind forecasts are then fed into a power conversion algorithm that has been empirically derived for each Xcel power connection node. In addition, a ramp forecasting technology fine-tunes the capability to accurately predict the time, magnitude, and duration of a ramping event. This basic system has consistently improved Xcel’s ability to optimize the economics of incorporating wind energy into their power system.

Wind Power Ramp Events Prediction with Hybrid Machine Learning Regression Techniques and Reanalysis Data

“…Additionally, and because of their non-linearity, a feasible way to solve this problem consists of using numerical models and methods, leading to the so-called Numerical Weather Prediction (NWP) models [48][49][50][51]. To put it simply, in an NWP model, the prediction of future states of the atmosphere is made by numerically modeling the dynamics and other physical processes of the atmosphere.…”

“…To put it simply, in an NWP model, the prediction of future states of the atmosphere is made by numerically modeling the dynamics and other physical processes of the atmosphere. The NWP model predictions are initialized from analyses, which represent the observed state of the atmosphere, on a three-dimensional grid, by combining observational data with an earlier prediction [49][50][51]. Enhancements of NWPs can be done by including some physical aspects of the terrain such as the roughness, orography and obstacles [52].…”

“…Within the field of application of wind energy, NWP models are able to provide, relatively quickly, wind data with high spatial and temporal resolutions. They are able to provide forecasts ranging from the next few hours up to several days ahead [49][50][51].…”

The importance of forecasting regional wind power ramping: A case study for the UK

“…At relatively short lead times (minutes to hours), forecasts can be made using simple statistical methods such as ARMA (auto-regressive moving average) [9] or more complicated data-driven methods such as artificial neural networks (ANN) [10,11]. Forecasts for the next few hours up to several days ahead rely on numerical weather prediction (NWP) models [12,13,14]. NWP model forecasts are initialised from analyses, which represent the observed state of the atmosphere on a three-dimensional grid by blending observational data with an earlier forecast.…”

Short‐term forecasting of categorical changes in wind power with Markov chain models

“…9 In addition, NWP models produce deterministic forecasts, which require several hours to run. Improvements in the uncertainty information contained in NWP models by running ensembles in which initial conditions are perturbed are a popular current area of research, [10][11][12][13] but this increases the length of time required to obtain results. On the other hand, many sophisticated statistical models for short-term forecasting of wind are now available.…”