AI system
- System
-
NCAR wind power forecasting system with the DICast adaptive forecast blending engine
- Model
-
Not assessed
- Release
-
Not assessed
- Application developer
-
National Center for Atmospheric Research (Research Applications Laboratory)
- Role
-
Not assessed
- Configuration
-
Not assessed
- Inputs and outputs
-
Not assessed
- Autonomy
-
Not assessed
- Workflow
-
Multiple numerical weather models blended with an adaptive statistical layer trained on each wind farm's history. Forecast updates every 15 minutes feeding unit-commitment decisions. A utility with large wind capacity and dispatchable coal and gas units to decommit.
- Human oversight
-
Not assessed
- Relation to AI
-
The forecasting system's contribution to savings is Xcel's own estimate relayed by NCAR authors and reporters. The AI character of the system rests on UCAR's reference to artificial-intelligence research and DICast's adaptive bias correction, not on an explicit machine-learning description. So both the AI status and the causal share are unverified.
- AI contribution
-
Not assessed
Benefits
- Outcome identifiers
-
Not assessed
- Benefit
-
cut mean wind forecast error from about 18% to 10.8% by 2014; Xcel estimated US$49m saved
- Category
-
Improved Task Accuracy
- Status
-
Not assessed
- Beneficiary groups
-
Organizations. Society.
- Beneficiaries
-
Not assessed
- Reach
-
Not assessed
- Distribution and access
-
Not assessed
- Duration
-
Not assessed
- Expected?
-
Yes
- Costs and harms
-
Not assessed
Measurement
- Studies
-
Not assessed
- Study design
-
Not assessed
- Measurement
-
Not assessed
- Setting and sample
-
Not assessed
- Measures
-
cut mean wind forecast error from about 18% to 10.8% by 2014; Xcel estimated US$49m saved
- Comparison
-
Xcel's wind forecasts in 2008, before NCAR's system
- Results
-
Reference: about 18 percent for 18 to 42 hours ahead. Result: 10.8 percent.
- Additional value from AI
-
Not assessed
- Limitations
-
Not assessed
- Denominators
-
Not assessed
- Materials for replication
-
Not assessed
Money
- Kind of value
-
saved through forecasting improvements, cumulative to 2015
- Amount
-
Estimate: US $49 million
- Currency and prices
-
Not assessed
- Period and unit
-
Not assessed
- Realization
-
Not assessed
- Recipients and payers
-
Xcel Energy customers
- Gross or net
-
Not assessed
- Included costs
-
Not assessed
- Comparison
-
Not assessed
- Method
-
Not assessed
- Return and payback
-
Not assessed
People and places
- Participants
-
Developer: National Center for Atmospheric Research (Research Applications Laboratory). Operator:
Xcel Energy. - Jurisdictions
-
Benefit:
United States. Developer: United States. Operator: United States. - Ownership and control
-
Not assessed
- Dependencies
-
Not assessed
- Cross-border connections
-
Not assessed
- Geopolitical implications
-
Not assessed
Follow-up
- Conditions for replication
-
Continuous observed generation data for bias correction. Operators who act on forecast confidence bands. An independent accounting of avoided production costs, absent here.
- Replication attempts
-
Not assessed
- Later findings
-
4 December 2017. By 2017 Xcel was still relying on the NCAR forecasts, which Haupt said saved about $10 million a year in production costs, while it planned coal retirements and more wind.
- Availability
-
Not assessed
- Institutional decision
-
Not assessed
- Grounds
-
Not assessed
- Action and result
-
Not assessed
- Success criterion
-
Not assessed
Evidence
- Source numbers
-
AIBD-0168-source-0001–0004
- Citations
-
Numbered references below.
- Language
-
Not assessed
- Source relationships
-
Developer, Independent
- Original evidence
-
Not assessed
- Access and versions
-
Not assessed
- Supported findings
-
Not assessed
- Evidence links
-
Not assessed
- Evidence strength
-
Not assessed
- Missing evidence
-
Not assessed
Grades
- Assessed benefits
-
Organizations. Society.
- Depth
-
Not assessed
- Reasons
-
Organizations: Xcel's US$49 million estimate and the error reduction to 10.8% are reported by the makers and the press without independent measurement. Society: Ratepayer savings and coal-unit decommitment are reported as estimates. No audited figure is documented.
- Observed or intended
-
Not assessed
- Rubric
-
Not assessed
- Assessor and date
-
Not assessed
- Alternative assessments
-
Not assessed
- Fill
-
Independently reported
Entry details
- Kind
-
Not assessed
- Related entries
-
Not assessed
- Field
-
Energy. Utility wind generation forecasting.
- Purpose
-
The system was built to forecast wind generation for Xcel's unit-commitment and trading decisions. The cost savings are its intended use.
- Dates
-
Date basis unassessed: 2015 (precision unassessed). Updated: 21 October 2026.
- Stage
-
Not assessed
- Version
-
SHA-256: 92c4d2686043…
References
- Other identifiers
-
Not assessed
- Category mappings
-
Not assessed
- Brief
-
Not assessed
- Revisions
-
5 revisions. Latest: 21 October 2026.
Sources
-
News report Developer
Sue Ellen Haupt and William P. Mahoney, 'Taming Wind Power With Better Forecasts', IEEE Spectrum, 22 October 2015
Xcel estimates that improvements in forecasting have saved customers US $49 million so far.
it constantly checks its predictions against what actually happened and adjusts its bias corrections and weightings.
-
News report Independent author
John Fialka and Benjamin Storrow, 'Xcel learns to forecast wind, then shuts coal plants', E&E News, 4 December 2017
the forecasts, preparations and timing meant the company could save $10 million in power production costs each year.
-
Press release Developer
UCAR/NCAR AtmosNews, 'Wind power wisdom', 7 February 2012
This effort is building on longstanding NCAR research in areas ranging from statistics to artificial intelligence, turbulence theory, and model development.