Q. How often are your forecast products updated?
A. Our 7-day weather and electricity demand forecasts are updated every morning. The date and time of update are indicated near the top of the main page.
Q. How did you develop your power demand forecast models?
A. We collected 10 years of weekly electric output data over each of the 9 geographical regions of the contiguous United States. We also collected daily high, low, and average temperature data over the same regions and processed population-weighted weekly temperature data. After adjusting the output data to account for the year-to-year demand growth due to non-weather factors such as population growth and economy expansion, correlation analysis were performed on the electric output and the temperature data and non-linear relationships between output and temperature were obtained for the 9 regions and for 4 seasons. The relationships give power demand forecast when temperature is given as input.
Q. On a day-to-day basis, is temperature the dominant factor that determines power demand?
A. Yes, on a day-to-day basis, temperature is the dominant factor that causes fluctuation in power demand. Especially during a winter cold front or a summer heat wave, daily power demand can change 30% from one day to another, and the exact magnitude of the change depends on the size of the region.
Q. On the electricity demand anomaly map sometimes I see an anomaly value of up to 50%, but on the state or regional forecast chart the magnitude is not as large. Why?
A. The anomaly maps show the anomaly magnitude at every point, while the state and regional charts show the AVERAGE magnitude of all the points in that area. The averaging process cause the state and regional value to be smaller than the maximum value found among all the points within that area.
Q. How should I interpret the regional forecast charts?
A. The 9 regional forecast charts show power demand forecasts for the next 7 days. The unit of the power demand is Billion Kilowatt-hour per week (GWh/week), which means that if the same temperature condition of that day lasts one week, the weekly power demand for that region is the value given by the vertical axis. To convert the value to daily demand, you may simple divide the weekly number by 7. However, there is a caveat here. By dividing by 7, you are assuming that each day of the week has the same power demand if temperature is the same regardless whether that day is a weekday or a weekend or a holiday. In fact, under the same weather conditions, daily power demand during weekend and holidays is lower than during a workday due to closing of business. So you have to keep this in mind when you convert weekly values into daily values.
Q. How should I interpret the state and local forecast charts?
A. The state and local forecasts are made relative to the so-called normal demand. They do not use the unit of GWh because our historical electric output data are on the regional level only. The normal demand is defined as the power demand for an area on a specific day of the year under the expected weather. In other words, when we talk about normal demand, we have to specify which area (e.g., which state, or which power company) and what date. The normal demand can be computed using the power demand model for the region in which the area is located based on the historical average temperature for that area on that date. The actual demand can be computed using the forecast temperature. Notice that by using the regional model both the actual demand and the normal demand have the unit of GWh. Dividing the actual demand by the normal demand cancels out the unit and gives us the relative demand forecast.
Q. Does your forecast for a particular company on a particular day depends on how that company operates its plants?
A. No, our demand forecast depends only on the weather. Suppose that company closes down all its power plants, we still predict the same power demand. Demand represents how much the customers need, regardless of whether the supply can meet that need.
Q. How do you make forecasts for individual companies? Do you have their data? Are they involved in the forecasting process.
A. We do not have the output data of the individual companies, nor do we have any data on the engineering factors of their power plants. And these companies are not involved in any way in the process of making the forecasts. In making forecasts for a company, we assume that the company has the same customer mix as that of the region in which it is located, so that the regional model can be used. But we have to use the forecast and historical temperatures for the service area of that company so that the demand forecast is relevant.
more Q & A to come ...
Any redistribution of information found at
this site strictly prohibited.
Copyright © 2011 Climaton Research Co. All rights reserved.