Forecasting tools: Improving the science behind enhanced model predictions
New forecasting tools aim to have a positive impact by improving the short-term predictability of wind energy and in the process improve maintenance and operations.
By John R. Johnson
There is no question that enhanced weather data and forecasting tools greatly improve the predictability of wind production, especially in flat areas such as the Texas Great Plains in the southern U.S., where wind turbines are commonplace.
Forecasting becomes especially difficult, however, in complex landscapes and mountainous terrain where wind energy development has some of its highest potential. Atmospheric challenges encountered in complex geographies worldwide often have a negative impact on wind variability, the reliability of short-term forecasting, and the overall performance of wind energy projects.
The U.S. Department of Energy is hoping to solve these issues by funding a $2.5m wind study that will use advanced meteorological equipment to analyze specific environmental characteristics that affect wind flow patterns in the Columbia River Gorge region of Washington and Oregon in the northwestern U.S. This area was chosen because it experiences nearly all of the identified atmospheric phenomena known to impact wind energy.
Improving weather models
The goal of the 39-month project is to improve the wind industry's weather models for short-term wind forecasts, especially for those issued less than 15 hours in advance.
“Complex terrain creates substantial forecast challenges for wind plants in most regions,” says Jack Peterson, manager of energy operations support at Southern California Edison, which produced 7.5 billion kilowatt-hours of renewable energy from wind sources in 2013.
“We have seen many situations where the forecasts are dramatically different at neighboring wind farms with only slight elevation changes. Improving the science behind forecasts is an important step and will greatly benefit the industry by removing some of the challenges we face,” he says.
The DOE’s Wind Forecasting Improvement Project 2 (WFIP2) is targeted at enhancing the reliability of wind forecasting around the world. The end result could help to significantly reduce the cost of grid integration while helping wind farm operators optimize performance through more effective short-term modeling of wind variability.
Atmospheric phenomena in complex terrain
The DOE has tasked Colorado-based environmental and industrial measurement firm Vaisala with conducting the comprehensive three-phase study of atmospheric phenomena in complex terrain. The end goal includes enhancing the widely used Weather Research and Forecasting (WRF) model and the National Oceanic and Atmospheric Administration's (NOAA) Rapid Refresh (RAP) and High Resolution Rapid Refresh (HRRR) models.
“It’s rare to have such a commitment from the DOE and NOAA,” says Dr. James McCaa, manager of advanced applications at Vaisala and principal investigator for WFIP2. “This is an exciting study and represents a sustained commitment from DOE to get to the bottom of some of the challenges in forecasting.
“One of the cool things about this study is that it has several gigawatts of installed wind energy capacity and lots of complicated terrain, so we can directly see the benefits of what we are doing and measure the improvements.”
Following a 12-month design and planning phase, Vaisala and the WFIP2 team -- which includes major utilities like Southern California Edison, Cowlitz County Public Utility District, Bonneville Power Administration and Portland General Electric -- will deploy extensive measurement equipment for 18 months. Beginning this summer, the project will analyse the specific environmental characteristics affecting wind flow patterns, ranging from soil moisture and surface temperatures to the unique topographical features of mountain-valley regions.
Enhanced model predictions
These observations will then be used to update and improve the computational and atmospheric physics that support current forecasting models. Enhanced model predictions produced during the third phase of the project will then be compared with baseline forecasts produced by existing models to evaluate the success of the initiative.
“Wind energy is an inherently variable resource,” says McCaa. “However, as modeling and forecasting techniques improve, we are increasingly able to calculate and predict that variability, a key factor for developers and operators worldwide as they make crucial long-term investment decisions. In mountainous areas, where atmospheric phenomena and unique topography expose some weaknesses in current models, there is still work to be done to enhance the reliability of forecasting. The Columbia River Gorge provides an ideal study site.”
Wind power changes rapidly, and in many cases the wind can stop blowing instantly, creating a huge void in the required electricity that must be ramped up in some other manner. Knowing that a change in wind output is coming – even the probability of it in a certain time range – will allow wind operators to prepare and respond in an effective way. Another benefit is the positive impact that improving the short-term predictability of wind energy can have on maintenance and operations.
“This study can certainly have an impact on O&M procedures,” says McCaa. “O&M is an area where the wind industry is really maturing, with our ability to manage the operation of the equipment in ways that prolong its life span. This project can help in that regard.”
Aside from the utility partners, others involved in the research project include the University of Colorado Boulder, Sharply Focused, Lockheed Martin, Texas Tech University, the University of Notre Dame and Iberdrola Renewables.
According to CU-Boulder project leader Julie Lundquist, an assistant professor in the Department of Atmospheric and Oceanic Sciences (ATOC), better forecasting will help integrate renewably generated electricity into the power grid, leading to lower energy costs for consumers.
“We will be making an unprecedented number of measurements in and around wind farms specifically to understand the meteorology in complex terrain as a way to improve the reliability of forecasts,” said Lundquist. “We also will be measuring how large wind farms impact the weather in their local environments.”