|
| |
Economic and Demographic Variables
System scale changes over time based on growth in service are population,
changes in economic conditions, and changes in technology. Examples of economic
indicators are income measures, unemployment measures, and price measures.
Demographic variables include the total population and its composition by
cohort.
Regional economic indicators are measured on a quarterly or annual basis, and
are only available some time after the fact. Often, economic data will not be
measured on a geographic basis completely consistent with a service area of
interest. While these variables do not tell us a lot about how load varies hour
to hour, they are important indicators of the overall size of the system. Also,
forecasts of these variables are normally available from government agencies and
econometric forecasting services.
Once the system is calibrated, it will perform well for several months or even
longer, based on its scale self-recalibration facilities. However, to project
forward for periods of six months or more, it is useful to have a relationship
between the system size and the major economic and demographic variables, to
facilitate planning scenarios like, "what happens if it is a harsh winter but
the region is prosperous and growing".
A technical problem is that economic measures are denominated by quarter, and
the typical TESLA model is specified on an hourly basis. To solve this problem,
we use latent variable methods to combine the hourly and daily information with
longer-term measures of economic conditions.
|
