EnergyCAP is powerful energy efficiency software published by Good Steward Software. EnergyCAP helps you get a handle on your energy information and save money and energy ten ways. EnergyCAP is used by cities, counties, universities and colleges, Federal and state agencies, retail chains, property managers, manufacturers and businesses to: track and audit utility bills; identify billing, metering and building controls problems; assess the weather's impact on energy usage and cost; compare today's bills to a baseline year; perform rate/tariff analysis; process bills and campus chargebacks for payment; and produce hundreds of energy management reports, charts and benchmarks. See more at www.EnergyCAP.com.

Average daily temperature is the 'middle value' between the daily high and the daily low.

A degree day is a measure of relative heating and cooling energy required by buildings. It's calculated as the difference between the average daily temperature and the balance point temperature (60 degrees). When the average daily temperature is above the balance point, the result is cooling degree days; when below, the result is heating degree days.
You may ask, "To keep it simple, why not use average temperature instead of degree days?" The problem with average temperature is that highs and lows cancel each other out. A warm day (80 average temp) combined with a cold day (40 average temp) average 60. So do two mild days of 59 and 61. But in first case there are 20 CDD and 20 HDD while in the second there are 1 CDD and 1 HDD. Using degree days, you can see that the relative amount of energy required for the first set of days is much greater than for the second set of days. But if all you looked at was the average temperature, you would conclude that both sets of days were about the same.

The balance point temperature is the average daily outside temperature at which a building maintains a comfortable indoor temperature without heating or cooling. At this outside temperature, the indoor heat gains (due to people, lighting, equipment, etc) "balance" with heat loss through windows, walls, roof and ventilation.

Since the average daily outside temperature normally occurs at about 11am, here’s the question: On a typical day, if the outside temperature at 11am is 60, is the building being heated, cooled or neither? If heated, then the balance point should be set HIGHER than 60; if cooled, then the balance point should be set LOWER than 60; if neither, then a balance point setting of 60 is appropriate. See the next FAQ below for further discussion of balance point.

The 65-degree balance point standard was developed 75 years ago to help the gas industry predict heating loads in residences. Studies back then showed that when the average daily temperature fell below 65, residences began turning on the heat. To this day, many sources still track degree days using this standard, including the National Oceanic and Atmospheric Administration (NOAA).

Today's residences and commercial/institutional buildings are very different. Not only are walls, roofs and windows insulated much better, but also there are many more sources of internal heat gains (lights and equipment). Extensive use of degree day correlations by thousands of EnergyCAP users since 1982 has shown that a 60-degree balance point for modern buildings is almost universally more appropriate than 65.

Since residences have fewer sources of internal heat gains per square foot (occupants, lighting and equipment), you might find that 65 or even higher is a better balance point estimate, particularly in older residences that lack tight windows and high levels of insulation. In non-residential buildings, use a higher balance point (56-60+) for buildings that have low internal heat gains, high ventilation rates and poor insulation.

EnergyCAP uses a linear regression technique to determine (1) if a building's energy usage correlates with the weather in summer, winter, neither or both, (2) about how much of the energy usage is due to weather and how much is "base load" that is not weather sensitive (lighting, cooking, equipment, etc), and (3) how weather sensitive the building is (in usage per degree day). The regression is usage (dependent variable) vs. degree days (independent variable) and yields a 'predictor equation' in the form y=mx+b (predicted usage=slope x degree days + base load).

The cumulative degree day charts are designed to allow you to quickly compare one year's weather to another as it relates to building energy usage. By comparing the cumulative degree days as of the end of one year with another year, you can quickly see which year was more severe (more degree days) as it relates to heating and cooling needs. For example, if 2005 had 4,000 cumulative HDD and 2004 had 3,000 cumulative HDD, you can conclude that the 2005 weather was 33% colder and would have required about 1/3rd more heating energy. (EnergyCAP can help you calculate how much of a building's energy usage is consumed for heating/cooling vs. other non-weather uses.)

The two-year comparison chart allows you to compare one month between two years. The graph shows the comparison year (the later year) compared with a base year (the earlier year). The comparison percentage is in the red (above the line) when the comparison year's weather was more severe than the base (it was warmer in a summer month or colder in a winter month). The comparison percentage is in the green (below the line) when the comparison year's weather was milder than the base (it was cooler in a summer month or warmer in a winter month). This can help you to understand why a building's heating/cooling bills were more or less.

Note: When a month had less than 30 degree days, we set the percentage to zero. This is to prevent relatively insignificant weather variations from appearing to be large. For example, one year in Chicago in June had 2 HDD and another year had 6 HDD. Although the second year had a 200% increase in heating degree days, it's misleading to draw any conclusions from this low number of HDD because heating systems were probably turned off.

Weather data recording stations sometimes experience failures. When we obtain the weather data, we look for missing data. If a single day is missing, we set the average daily temperature equal to the prior day. This approach is used for up to seven days. If more than seven days in a row are missing (about 25% of a month), we discard the entire year for that station as being unreliable.

The best source of daily historical weather data is: www.AccuWeather.com

1. From the AccuWeather site, enter your zip code then click Go.

2. At upper left, click on Forecast, then click on Historical Weather.

3. Select the desired month. The degree days (65-degree base) are shown in the right columns.

A degree day forecast is a 14-day forecast of daily heating and cooling degree days. The forecast is created and continually updated by AccuWeather.com. Each morning at 6 am ET AccuWeather provides us with a forecast of mean daily temperature for hundreds of weather stations. We then convert the mean daily temperature to degree days using a balance point temperature of your choosing. (Default value is 60. See FAQ on Balance Point Temperature.) We display the forecasted heating and cooling degree days, the actual degree days for the same date range last year and the average degree days for the same date range in the last three years. The date range does not include today – it starts tomorrow and is for 14 days, inclusive of start and end dates. When the heating or cooling degree day total is zero (for example, the heating degree day forecast in late July) the charts will show zero days.

Energy usage within a home or non-residential building can be separated into two components: the energy usage that is sensitive to the weather and the energy usage that is not sensitive to the weather. In many cases, there will be no weather-sensitive component during the summer or winter (for example, a gas meter in the summer or an electric meter in the winter when the building has no electric heat).

If we know the non-weather component (expressed in usage per day) we can calculate the future non-weather usage. Example: my building uses 200 KWH of electricity per day for non-weather loads such as lighting and equipment. In the next 14 days my building will use about 2,800 KWH for non-weather loads.

If we know the weather component (expressed in usage per degree day) we can calculate the future weather usage. Example: my building uses 10 KWH of electricity per cooling degree day for A/C. In the next 14 days the forecast is for 200 cooling degree days, so my building will use about 2,000 KWH for A/C.

Adding the two together, the energy usage prediction is 4,800 KWH in the next 14 days.

Now comes the tricky part – how were the non-weather base load of 200 KWH/day and the weather factor of 10 KWH/degree day developed?

Using our EnergyCAP and FASER software:

We use single linear regression to statistically determine the non-weather base load and the weather factor. We first look at the start/end dates of each bill and add up the total heating and cooling degree days in the date range. We then reduce the degree days and usage to a per day basis and run a standard regression. This creates a regression line which is essentially a mathematical model of the energy usage of the meter in the form: y=mx+b where b is the y-intercept (the daily non-weather base load), m is the slope of the line (the weather factor in use/degree day), x is the independent variable (degree days) and y is the dependent variable (energy usage per day).

Using Excel:

You can use Excel’s built-in single linear regression function to correlate the dependent variable (energy usage) to the independent variable (degree days). Your results will be exactly the same as EnergyCAP. It is beyond our scope to describe the process in detail.

Using a simple estimation technique:
1. Gather your last 12 energy bills (on a meter-by-meter basis).
2. Look at the usage each month. Which best describes the trend:
a. High in summer, low in other months. (weather sensitive in summer only)
b. High in winter, low in other months. (weather sensitive in winter only)
c. High in winter and summer, low in spring/fall. (weather sensitive in both seasons)
d. About the same in all months, or wide variations between months. (not weather sensitive)

3. If a or b: Calculate the non-weather base load as the daily usage in the low season. Calculate the weather factor: total the usage in the high months, subtract the non-weather base load, then divide what’s left (the total weather usage) by the total of the degree days for those months. The result is the estimated usage per degree day.

If c: Use an approach similar to the above, only use the spring/fall bills to estimate the non-weather base load.

Using this website (future): We will soon be adding a free web-based calculation tool. You’ll be able to enter 12 months of energy bills and we’ll calculate and display the usage vs. weather correlation results.

The weather factor is an approximation of the energy used for heating and cooling on a per degree day basis. It is expressed in average usage per degree day (KWH per degree day of electricity, THERM or CCF per degree day of gas, etc.). Your summer and winter weather factors will usually be very different. Often, you will have a weather factor in one season but not in the other. You may have a high gas weather factor in the winter because you have gas heat but no gas weather factor in the summer; likewise you may have a summer electricity weather factor (due to A/C) but none in the weather (no electric heat).

The weather factor is an approximation. It is a “straight-line” estimate, meaning that it assumes that weather-sensitive energy usage is linear – the efficiency is the same whether the outside temperature is 50 or 20. In reality, heating and cooling systems are typically not linear. Efficiency at full-load conditions is often better than efficiency at part-load conditions due to the inherent losses of on/off cycling. Our degree day technique is a simplification that yields acceptable results in most circumstances and only requires monthly utility bills and daily mean temperature data. Any more rigorous and accurate calculation technique is far more expensive and complex.

EnergyCAP users: The weather factor for each meter is shown on report CAP05 (see Summer Use/Degree and Winter Use/Degree). It is also shown on the Cost Avoidance - Use vs. Weather chart. It is the slope of the green regression line. In the results table, it is labeled “Slope.” This is only statistically valid when the results table is green. If red, there is no valid usage vs. weather correlation and you should not attempt to use the results to predict future energy usage.

See the FAQ on How is Future Energy Predicted for an explanation of the calculation of the weather factor.

The non-weather base load is the component of energy usage that is not sensitive to the weather. Electricity used for lighting, hot water, cooking, equipment and “plug loads” (things that are plugged in) is not weather sensitive. Gas used for hot water, cooking, clothes drying and industrial processes is not weather sensitive. Each energy commodity within your building has a unique non-weather base load. The base load changes through the seasons – you probably use more electricity for lighting during the winter months than the summer months.

We express the non-weather base load in average usage per day (KWH per day of electricity, THERM or CCF per day of gas, etc.). Keep in mind that these values are approximations!

EnergyCAP users: The daily average non-weather base load is shown on the Cost Avoidance - Use vs. Weather chart. It is the y-intercept (the point at which the green regression line intersects the left vertical axis). In the results table, it is labeled “Inter.” This is only statistically valid when the results table is green. If red, there is no valid usage vs. weather correlation and you should not attempt to use the results to predict future energy usage.

See the FAQ on How is Future Energy Predicted for an explanation of the calculation of the non-weather base load.