Hmm - I'm usually quite good at data representation - but I'm struggling to see the point here - it seems that one non zero intercept has been exchange for another. Has anyone else got a handle on what is trying to be said here?

Posted By: goodevansHmm - I'm usually quite good at data representation - but I'm struggling to see the point here - it seems that one non zero intercept has been exchange for another. Has anyone else got a handle on what is trying to be said here?

It seemed fairly straightforward to me. There is exactly one temperature that is correct for use as the base temperature for each building, unique to that building. The base temperature is said to be the lowest at which the building 'balances' with no need for heating. The base temperature can be determined by plotting actual heating consumption against that predicted using degree days, and adjusting the base temperature used to calculate the degree days until the graph is a straight line. Any other base temperature results in a graph with a curved section near the origin.

Posted By: gyrogearDoes this mean that the average outside temp in France is 2.5°C higher than in UK ?

Given that France is entirely to the south of the UK, I'd be surprised if it wasn't warmer in France.

Without going into too much detail, and based on what we found out on here a few years back, that report is just showing the weakness of HDD.
I did a chart somewhere showing that, at my location, windspeed had a greater affect on energy usage.

After a lot of digging about, I found that the 'average' temperature was calculated on the median, not the mean. This is for historic reasons when only 2 temperature readings were taken every day, but can cause some skewing of the data at the extremes i.e. at the lower readings (not bothered by the higher ones as no heating needed).

HDDs are only a rough calculations and cannot be applied to all and every building, no matter how much you tinker with the base temperature.

Here in cold (and, in the summer, hot) Montreal, 18C is used as the basis for heating and cooling degree-days. If the temperature is below 18C, you need some heating to keep your internal temperature at 21C and, if it averages 18C or more, you will need some cooling to maintain your internal temperature at 21C (21C is the temperature used by the energy audit programs here and is used by such tools as hot2000 to figure out the heating and cooling load - of course, the source of that heat can (and often is) purely internal gains from appliances, lighting etc. - but these are all factored in anyway). 15.5C is used in the UK because people seem to like cold houses!!

Paul in Montreal.

The base temperature is irrespective of what your interior temperature is - it's just an apples-to-apples comparison of different locations. All houses have a negative energy requirement when the exterior temperature is at the HDD base and the interior is at the external temperature. Even if your internal temperature is higher than the HDD base, there still could be a negative energy requirement (due to solar and other incidental gains). However, you can plot some graphs of the HDD number versus the supplied heating energy to find out how efficient your building is (assuming you can hold everything else constant). One challenge is that you might have two days with identical HDD numbers, but one of them is sunny and the other is dull - so the amount of heating energy you have to supply on those days to maintain your constant interior temperature (and we use 21C here in Canada) could be quite different. HDD numbers are more away to compare climates at different locations. FWIW Montreal has about 10x more HDDs per year than CDDs.

Paul in Montreal.

Hello! I've done a lot of work for the Degree Days.net website so I have a fair amount of experience with degree days. I came across this thread and thought I might be able to help clarify a few things for anyone finding it in future.

When using degree days to compare the weather at different locations (e.g. between different cities), it makes sense to choose a sensible base temperature and stick to it for your comparisons. But, when using degree days to analyze energy data from a particular building, it generally makes sense to choose the most appropriate base temperature for that particular building. Buildings vary a lot, and consequently so do their base temperatures. I wrote an article recently that explains the main factors that determine the base temperature of a building:

https://www.degreedays.net/base-temperature

I see the traditional base-temperature defaults such as 15.5 C for the UK and 65 F for the US as a hangover from the days when degree days were disseminated in print. Building base temperatures would have varied back then too, but it just wasn't practical to publish degree days in scores of different base temperatures for scores of different locations to suit everyone optimally. One-size-fits-all base temperatures were the best practical option available at the time. They still linger on today, but times are different now, it's now much easier for people to find degree days in whatever base temperatures they want.

I agree with the process described in the enmanreg.org post that djh started this thread with. It's essentially a simplified version of the process used by the regression tool on the Degree Days.net website (which tests regressions of energy data against HDD and/or CDD in different base temperatures to find the ones with the best statistical fit).

A positive intercept is pretty common as many buildings have non-weather-dependent energy consumption that is metered together with heating/cooling energy consumption. A negative intercept is generally a bad sign, however, indicating that either the base temperature is wrong, or that there are some anomalies in the energy data, or that something odd was happening in the building.

On SteamyTea's comments about how degree days are calculated: actually it depends on the source you get the data from. The original and correct way to calculate degree days is to use the full detail of the temperature data available (which, for most good weather stations, is typically at least hourly). This creates degree days that capture all the known temperature variations within each day. It is, however, a fairly intensive calculation, so many sources use a simpler approximation method instead. Approximating degree days using daily max/min temperatures is common and the accuracy of the figures suffers, particularly for periods over which the temperature was near to the base temperature. Naturally I would recommend using accurately-calculated degree days over approximated degree days whenever possible.

Posted By: DamonHD(I'm a big user and fan of DegreeDays.net, please be kind to bromley.)
Just as a note, I treat my home's heating base temperature as 12C as that has a good fit to observed behaviour.

Snap.
I came to that conclusion following analysis of several years of my own data (i.e. heating energy used [easy as we are all electric] and temperatures from my own weather station).

The colder the temp the higher the correlation between heating demand & HDD12. As we get to the change over and the daily average hovers around that figure, things get more complicated (a bit of sunshine can make a lot of difference). In practice we tend to find that we can turn off the heating when the average daily temp is reliably 10C or greater.