Loading presentation...

Present Remotely

Send the link below via email or IM


Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.



MSCS 6090 Seminar, 3/8/2010

George Corliss

on 19 April 2017

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of GasDay

We revel in the abnormal
Forecasting Customer Demand for Natural Gas
George Corliss, George.Corliss@Marquette.edu
Presented: UW Whitewater Physics, 10/19/2012
Marquette University GasDay
Ron Brown, PhD, Director
Dr. Vitullo’s charge:
1) Basic problems and questions being addressed  
2) Detailed discussion of one problem or aspect
3) Role of collaborators and clients
4) Ethical concerns and responsibilities
5) Style and division of work (who does what;
are there regular meetings, for instance)   
6) How are results of the research
communicated to clients or published
Demand for Natural Gas
Collaborators: Ron Brown, Tom Quinn,
Steve Vitullo, and about 200 other students
What do we do?
License software to utilities
Forecast natural gas demand
Research studies
How do we do it in a university?
Business vs. mission?
With students
Dec., 2006
Wisconsin Ave., Dec., 2006
What's the point?
We know winter
Gas keeps us warm
How much do we need?
Natural Gas Transmission System
Home Computer of 2004 - Popular Mechanics 1954?
Scientists from the RAND Corp. have created this model to illustrate how a "home computer" could look in the year 2004. However, the needed technology will not be economically feasible for the average home. Also, the scientists readily admit that the computer will require not yet invented technology to actually work, but 50 years from now, scientific progress is expected to solve these problems. With teletype interface and the Fortran language, the computer will be easy to use. -- Hoax: http://www.snopes.com/inboxer/hoaxes/computer.asp#photo
What is GasDay?
Rolling 8-day natural gas load forecasting service
that is licensed to natural gas utilities
New program features and model improvements
are released each year

Students develop the model improvements
Students develop the software
Students work with customers to deliver
and support GasDay
Where is GasDay used?
How does GasDay work?
8-day temperature and wind forecast
Temperature and wind for the previous day
(typically based on 21 hrs actual, 3 hrs forecast)
Actual temperature and wind for 2 days ago
Flow from 2 days ago
Forecast flow 8 days
Other time series analysis
Gas Hour, Month, Year
GasDay USA
GasDay w Econometrics
Peak day studies
Grow the data
Surrogate data
Measurement Scene Investigator
Heating oil forecaster
Custom studies
Surrogate Data
Data from one area (donor) is transformed
to act like data for another (recipient)
Unusual days from donor areas are chosen
as surrogate data
Use surrogate data plus native data to
make forecast models
Measurement Scene Investigator
Given historical meter readings
Detect outliers
Exception report:
How do we do it in a university?
Business vs. mission?
Value >> price
Customers, not us
With students
Entrepreneurship with Math?
$500+K / yr licensing revenues
Negotiate contracts
Deliver software product
Customer support
Sales & marketing
Intellectual property
Staff turn-over
Lab space
$500+ K / yr external funds
Supports grad & undergrad students
Involves lots more
~2 masters students / yr
PhD / ~3 yr
Conference presentations
Research in Math?
One customer’s experience:
Run-outs last heating season down from
several hundred down to 20
Delivered more oil than the previous year
with 20% fewer trucks and men
So confident in the service that they’re placing
a huge bet on it. If you run out, they’ll give
you a free tank of oil (over $1,000!)
Heating Oil Forecaster
Fit teaching and research mission of university
Knowledge of start-up business environment
Concept of customer/supplier relationships
100% risk-free deal
Reaction time incompatible with business/
industry/investor environment
Institutional incentives for individuals to
commercialize technology
Pro-actively facilitate deals
vs. operate strictly in a reactive mode
External linkages to investment community
Internal commercialization linkages
Oversight of technology transfer
Objective, experienced, industry-based board of advisors
We revel in the abnormal
GasDay lab - Olin 534/540
George Corliss
Ron Brown, Director; Tom Quinn
Rohan Kennedy, Detecting Outliers and Meter Anomalies in Natural Gas
Customer Flow Data, MS, 2006
Brian Marx, Forecasting Daily Processing Tomato Harvest Tonnage in
California, MS, 2004
Susanto Halim, Selection of Inputs for Generating Combinatorial Daily
Natural Gas Demand Forecasts, MS, 2004
David Diggs, Multiple-step Financial Time Series Prediction with Portfolio
Optimization, MS, 2004
Everton Walters, A New Approach to System Identification and the
Characterization of Natural Gas Pipelines Based on Genetic Algorithms,
MS, 2002
Hui Li Esther Lim, Computational Intelligence Models for Short Term
Natural Gas Demand Forecasting, MS, 2002
Richard Lukas, Development of Adaptive Online Fuzzy Arbitrator for
Forecasting Short-Term Natural Gas Usage, PhD, 2001
Lance Hilbelink, A Tale of Ten Cities: Consolidating Weather Information
in Gas Load Forecasting, MS, 1999
Avinash Taware, Forecasting and Identification Methods Applied to Gas
Load Estimation Problems, MS, 1998
YOU?, topic To Be Determined
Come and see
Tian Gao, Wenyan Min, Hermine Akouemo Kengmo Kenfack, topics to be determined
Anisha D'Silva, Estimating the Extreme Low Temperature Event Using Nonparametric
Methods, MS, 2012
Bo Peng, The Impact of Additional Weather Inputs on Gas Load Forecasting,
MS, 2012
Steve Vitullo, Disaggregating Time Series Data for Energy Consumption by
Aggregate and Individual Customer, PhD, 2011
Tsuginosuke Sakauchi, Applying Bayesian Forecasting to Predict New Customers'
Heating Oil Demand, MS, 2011
Samson Kiware, Detection of Outliers in Time Series Data, MS, 2010
Sidhartha Tenneti, Identification of Non-Temperature-Sensitive Natural Gas
Customers and Forecasting their Demand, MS, 2009
Brian Marx, Fitting a Continuous Profile to Hourly Natural Gas Flow Data, PhD, 2007
Meng He, Annual Transformation Algorithm for Hourly Data in Natural Gas
Consumption, MS, 2007
Steven Vitullo, Disaggregating Interval Time-Series Data Applied to Natural
Gas Flow Estimation, MS, 2007
WISN Chanel 12, Nov. 19, 2007
Students save money for We Energies’ customers
Myra Sanchick, Fox 6 News, Sep. 29, 2009
We Energies rate increase hearing
Journal/Sentinel, Sunday, Jan. 10, 2010
Lab learns to predict gas demand
Jason Fischer of WisconsinEye's
Business Profile, Feb. 17, 2010
Copyright 2012 George Corliss, Marquette Univ.
Copyright 2012 George Corliss, Marquette Univ.,
Copyright 2012 George Corliss, Marquette Univ., Milwaukee WI
Full transcript