Weather Forecasting for Load Forecasts - PowerPoint PPT Presentation

Loading...

PPT – Weather Forecasting for Load Forecasts PowerPoint presentation | free to download - id: 412a55-MmNiY



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Weather Forecasting for Load Forecasts

Description:

Weather Forecasting for Load Forecasts DTN/Meteorlogix Outline Corporate Overview Forecast Improvement Goals The Meteorlogix Forecast System Statistical Weather ... – PowerPoint PPT presentation

Number of Views:37
Avg rating:3.0/5.0
Slides: 23
Provided by: JamesB166
Learn more at: http://www.ercot.com
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Weather Forecasting for Load Forecasts


1
Weather Forecasting for Load Forecasts
  • DTN/Meteorlogix

2
Outline
  • Corporate Overview
  • Forecast Improvement Goals
  • The Meteorlogix Forecast System
  • Statistical Weather Models
  • Manual Input from Meteorologists
  • Multiple data output
  • Benefits
  • Forecast Summary
  • Forecast Preview for the 6-10 Day Period

3
DTN/Meteorlogix Corporate OverviewCommercial
weather leader
  • DTN is a leading provider of proprietary
    business-to-business real-time information
    services enabling its customers to make Smarter
    Decisions
  • Offices in Omaha, Minneapolis and Boston
  • 110,000 subscribers across Agricultural, Energy
    and Weather Markets
  • Provide weather information and services to 700
    utilities
  • Stable organization with over 675 employees with
    strong technology skills and vertical market
    expertise.
  • 95 Degreed Meteorologists
  • Worlds largest commercial weather service
    provider, delivering weather information to the
    Energy industry since 1946
  • Industry leading technology
  • State of the art weather forecasting system
  • Proprietary load forecast information (Effective
    Degree Days)
  • Comprehensive GIS weather support

4
Forecast Improvement GoalsReduced Mean Absolute
Errors
5
Making Better Forecasts
  • Leverage all available resources
  • Improvements in numerical weather models
  • New technology
  • Experienced forecast staff
  • Combine these into a Forecast System
  • Measure and quantify the results

6
Making Better ForecastsNumerical Weather
Prediction
  • There are many models
  • No one model is right all of the time
  • Model skill scores continue to improve
  • Forecast system needs to take advantage of these
    improvements

7
Making Better ForecastsThe DTN/Meteorlogix
Forecast System
  • Automatically create a very good first guess
    forecast
  • DICast statistical forecasts
  • Incorporate an ensemble of many weather models,
    and use statistical methods to optimize the
    forecast
  • Integrating manual input and experience into the
    forecast
  • Graphical Forecast Editing
  • Focus only where manual input adds value
  • Tools to monitor and manage forecasts
  • Real-time verification and feedback
  • Trend and bias analysis

8
Statistical Weather Forecasts
  • An ensemble of high-resolution models, MOS, and
    Dynamic MOS
  • Updated hourly, using current observations
  • Self-learning, error-correction
  • Proven to have lower errors than any individual
    forecast component

9
Statistical ForecastsBetter than any single
model
10
Manual Input
  • Graphical editing allows forecasters to see the
    way they think - spatially
  • Focus on limited areas, and specific times
  • Incorporates terrain and local effects
  • Insures consistency

11
DTN/Meteorlogix Forecast System Output
  • Forecast values are available anywhere in the
    CONUS
  • Points
  • Areas
  • Multiple parameters
  • Temperature, dew point, wind, clouds, etc.
  • There are always 15 days of forecast data
  • Hourly
  • Daily

12
DTN/Meteorlogix Forecast System Benefits
  • Timeliness
  • Forecasts are updated every hour
  • Quality
  • All available forecast data is used in each
    forecast
  • Makes for the best, most consistent day-in,
    day-out, forecast
  • Meteorological experience is focused on adding
    value
  • Extreme events
  • Precipitation and its effects
  • Reliability
  • Forecast products are always current and up to
    date
  • Scalability
  • Can provide high quality forecasts even without
    observations

13
Forecast Summary
  • Weather forecasts are improving
  • The way in which weather forecasts are created is
    rapidly evolving
  • New forecast processes are making more and
    different kinds of data available
  • Are there ways to take advantage of new weather
    data in load forecasts?

14
High Impact WeatherApril 17, 2006
All ERCOT Locations All ERCOT Locations All ERCOT Locations All ERCOT Locations
  AvgHrlyErr MaxErr MinErr
Day 1 Average 2.1 1.7 2.8
Day 2 Average 2.3 1.7 2.7
Day 3 Average 3.4 2.8 3.7
Day 4 Average 3.7 4.2 3.7
Day 5 Average 3.7 4.2 3.8
Day 6 Average 4.7 5.5 4.2
Day 7 Average 4.3 5.4 4.1

15
High Impact WeatherApril 17, 2006
RptDate RptDay Station AvgHrlyErr ActMax FcstMax MaxErr ActMin FcstMin MinErr
2006-04-17 1 KABI 2 98 98 0 60 63 3
2006-04-17 1 KACT 2 97 96 -1 72 67 -5
2006-04-17 1 KAUS 3 99 96 -3 73 65 -8
2006-04-17 1 KBRO 1.4 91 93 2 73 73 0
2006-04-17 1 KCRP 1.3 90 92 2 71 71 0
2006-04-17 1 KDFW 3.1 101 96 -5 69 67 -2
2006-04-17 1 KGLS 1.7 80 83 3 73 73 0
2006-04-17 1 KIAH 0.8 92 92 0 72 70 -2
2006-04-17 1 KINK 3.1 99 98 -1 55 53 -2
2006-04-17 1 KJCT 2.4 101 100 -1 56 58 2
2006-04-17 1 KLFK 1.6 91 92 1 70 67 -3
2006-04-17 1 KLRD 1.5 107 106 -1 73 75 2
2006-04-17 1 KMAF 2.2 93 94 1 58 56 -2
2006-04-17 1 KMWL 3.5 101 97 -4 54 62 8
2006-04-17 1 KSAT 1.3 99 98 -1 72 70 -2
2006-04-17 1 KSJT 3.3 99 100 1 55 61 6
2006-04-17 1 KSPS 2 101 99 -2 56 59 3
2006-04-17 1 KTYR 1.5 92 93 1 70 68 -2
2006-04-17 1 KVCT 2.4 90 92 2 71 70 -1
Day 1 Average   2.1     1.7     2.8
16
High Impact WeatherApril 17, 2006
RptDate RptDay Station AvgHrlyErr ActMax FcstMax MaxErr ActMin FcstMin MinErr
2006-04-17 2 KABI 2.5 98 98 0 60 62 2
2006-04-17 2 KACT 2.6 97 96 -1 72 67 -5
2006-04-17 2 KAUS 2.4 99 97 -2 73 67 -6
2006-04-17 2 KBRO 2.5 91 94 3 73 74 1
2006-04-17 2 KCRP 1.1 90 91 1 71 71 0
2006-04-17 2 KDFW 3 101 97 -4 69 67 -2
2006-04-17 2 KGLS 2 80 83 3 73 73 0
2006-04-17 2 KIAH 0.7 92 92 0 72 71 -1
2006-04-17 2 KINK 2 99 100 1 55 57 2
2006-04-17 2 KJCT 2.8 101 99 -2 56 58 2
2006-04-17 2 KLFK 1.8 91 93 2 70 67 -3
2006-04-17 2 KLRD 1.4 107 107 0 73 75 2
2006-04-17 2 KMAF 3.1 93 96 3 58 59 1
2006-04-17 2 KMWL 2 101 99 -2 54 61 7
2006-04-17 2 KSAT 2 99 99 0 72 69 -3
2006-04-17 2 KSJT 4.8 99 98 -1 55 61 6
2006-04-17 2 KSPS 3.3 101 96 -5 56 61 5
2006-04-17 2 KTYR 1.5 92 93 1 70 68 -2
2006-04-17 2 KVCT 3 90 92 2 71 70 -1
Day 2 Average   2.3     1.7     2.7
17
High Impact WeatherJanuary 12th 13th, 2007
18
High Impact Weather Model Data Jan 13th, 2007
  • Model disagreement Jan 12th Jan 13th, 2007
  • KTYR Tyler, TX
  • ETA MOS GUIDANCE    1/13/2007  1200 UTC 
  • DT /JAN  13/JAN  14                /JAN 
    15                /JAN  16 
  • HR   18 21 00 03 06 09 12 15 18 21 00 03 06 09 12
    15 18 21 00 06 12 
  • N/X                    36          36         
    28          34    25 
  • TMP  42 41 39 39 41 41 40 36 33 34 34 35 34 34 33
    28 30 33 31 29 27 
  • DPT  42 41 38 35 38 39 37 35 33 34 34 30 26 25 25
    22 21 21 21 20 20 
  • CLD  OV OV OV OV OV OV OV OV OV OV OV OV OV OV OV
    OV OV OV OV OV OV 
  • WDR  33 34 35 01 35 02 36 35 36 34 32 34 33 35 35
    34 35 34 34 34 35 
  • WSP  06 09 08 12 09 05 05 05 04 07 11 12 15 12 15
    14 16 16 12 11 09 
  • GFS MOS GUIDANCE    1/13/2007  1200 UTC 
  • DT /JAN  13/JAN  14                /JAN 
    15                /JAN  16 
  • HR   18 21 00 03 06 09 12 15 18 21 00 03 06 09 12
    15 18 21 00 06 12 
  • N/X                    49          53         
    36          38    29 

19
High Impact WeatherObserved Temperatures
20
High Impact WeatherResults and Conclusions
  • Forecasts did not handle the rapid drop in
    temperatures well
  • Average hourly error of 5.0 for ERCOT locations
  • Enhanced weather editing mode installed soon
  • Easier editing
  • Quicker results in products
  • Better response to rapid temperature change

21
High Impact WeatherImprovement
22
DTN/Meteorlogix Contact Information
  • Richard Wilson
  • Director of Energy Services
  • DTN/Meteorlogix
  • Phone (781) 932-3539
  • Email richard.wilson_at_dtn.com
  • Jeremy Duensing
  • Quality Assurance Manager
  • DTN/Meteorlogix
  • Phone (952) 882-4554
  • Email jeremy.duensing_at_dtn.com
About PowerShow.com