The Value of Real-Time Wind Power Forecasting
The rapid growth of wind power is offering new challenges for transmission grid operators throughout the world, especially in the U.S. As wind power levels comprise an ever larger portion of utility supply portfolios, rapid changes in the speed and direction of wind resources often cause wind farms to deviate significantly from scheduled deliveries. When this happens, grid operators must tap into volatile spot market to keep the transmission grid in balance.
The value of spot market purchases required to balance the grid can cost from -100/MegaWatt hour (MWh) (during times of oversupply) up to $6,000/MWh (during times of supply shortages), depending upon wholesale supply conditions. Near real-time forecasting data for wind could be worth millions of dollars – both earned and saved –- but so far, few of the handful of firms that provide wind resource assessment and forecasting have developed a coherent business strategy to tap this emerging market for their products and services.
When compared to simple reliance upon climatology, an advanced forecasting system will reduce short-term forecasting errors by 40 to 60%. The magnitude of potential revenue losses can reach almost $4/MWh.
The error rate in day-ahead wind forecasts for any single wind farm is between 12% and 20%. But if these forecasts are aggregated across an entire region, error rates drop to roughly 5%. But even that level of error carries large financial impacts.
In Europe, where wind penetration is greatest, the costs of the variability of wind has largely been socialized, and therefore grid operators have not had as strong of an incentive to keep track of near real-time wind data.
In the U.S., forecast markets vary considerably by region. The Midwest Independent System Operator (MISO) appears to be developing the most advanced market for wind forecasting services, as it changes its market rules to fully integrate wind into its scheduling markets. The new market redesign will include a “dispatchable intermittent” category where any wind farm would be treated virtually the same as traditional generation, with only one special provision. Instead of dispatching wind at the amount that generators offered into the market, MISO will dispatch these resources up to the maximum MW output figures provided by a real-time forecasting system.
Current best bets for near-term are “nodal markets” where spot prices are segmented according to different transmission line nodes. In these nodal markets, generation units are typically committed to provide power every 15 minutes; power is dispatched every five minutes, offering multiple opportunities to buy and sell. To date, the California Independent System Operator (CAISO), MISO, Pennsylvania, New Jersey and Maryland (PJM) ISO and now ERCOT are all in varying degrees of nodal market development. To give a sense of scale, the Texas transmission market will feature 4,000 different nodes.
Ironically, the main advantage of real-time wind data may not be when the wind dies down but rather when the wind speed increases to the point that the turbines have to shut-down. In this case, the entire wind farm would be producing at maximum capacity, and then rather suddenly come to a complete halt, potentially creating a nightmare for grid operators.
By 2015, the costs of real-time forecasting should come down to the point of allowing broad networks of sensors to increase accuracy at a competitive price. The going rate for wind forecasting services at present is $2,500 per month per wind farm. At that rate, the current wind forecasting market in the U.S. represents roughly a $20 million annual investment. Yet that figure should escalate significantly as wind capacity continues to grow and the value of real-time wind resource data also increases as MISO, ERCOT, NYISO and other control areas look to integrate wind more directly into scheduling protocols.
Globally, wind forecasting market may be as large as a $100 million annual enterprise, but given the diversity of market designs, there is considerable uncertainty. It is safe to say that this global market will grow in lockstep with future wind capacity additions, which grew at a 31% annual rate in 2009.