What do the technical terms mean?

Statistics in Heating / Air-conditioning / Data-gathering

  1. Statistics and Use
  2. Ideas about measuring temperatures
  3. What is a kilowatt hour?


Tip: To find your term quickly use the windows search function by pressing 'CTRL+F' and continue searching with 'F3'.

1. Statistics and Use

Annual heating work
Unit: kWh [work (kWh) = power (kW) * time (h)]
When do you actually need a given output at a given time from your heating system?
The German DIN standard demands that the size of your boiler should be according to the lowest mean daily outside temperature occurring in your region over the last 20 years. This is so that the boiler can cope at these temperatures. Since these low temperatures go on for only a few days, in the course of the year your boiler will run as follows:
Annual heating work at ext. temp. ºC Comment
7% +15...+10 +15 heating threshold ==> full shut-down
20% +10...+5  
36% +5...0 main running time
22% 0...-5  
9% -5...-10  
6% -10...-15 -10...-16 max. output limit
[Source: Volkhard Neuhoff, Giersch GmbH]
What does this tell us?
  • Nearly 2/3 of the annual heating work (63%) is done at outside temperatures in the range of +15ºC to 0ºC.
  • Maximum output is only required 6% of the running time. The system is always designed on the basis of that maximum output - or a bit more: the 'safety' margin!
  • From experience, we can say that small systems are twice as big as necessary. (Fatal overkill :-)
  • We are heating suplementary with 2/3 of our annual electric current up to four times more expensive than gas or oil. This 'loss' of 25...30 kWh/(m²*a) will always be mishold.
You can save energy when adapting heating systems merely by choosing the right size of boiler. It is possible to save more by using a modulating or at least a two-stage burner (oil), so that the small stage can deliver the 36% of the annual heating work with fewer switch-ons and thus not so many standby losses.
So our radical strategy is: cut the DIN size in half for residential homes!

[ contents ][ continue ]
Annual Utilisation Ratio AUR
Unit: none
Ratio of heat used (boiler output) to heat delivered (burner input) over a heating year. This includes as variable, time and the statistics of use. In systems which include heat quantity counters it is also possible to measure the AUR. The usual value is in the range 50...70%. Yous pay 100%, you get only 50...70%!
Sceptics can look at Environmental madness in the office.

[ contents ][ continue ]
Boiler efficiency
Unit: none
is the ratio of actual output at nominal boiler load to burner output i.e.:
  eta b = 100 - off-gas loss - radiation loss (in percent)
According to EN 303/DIN 4702 the degree of boiler efficiency is measured under permanent full load To=80C, Ti=60C.
Monentary value (no connection to Annual Utilisation Ratio AUR).

[ contents ][ continue ]
Burning efficiency
Unit: none
Burning efficiency determines off-gas losses and is subject to legal regulation in Germany. In terms of numbers, off-gas losses are the greatest losses you suffer in a boiler: 5...11%, with an air/off-gas chimney 1...4%). However, this loss only occurs during the approx. 1500 hours in which the burner is annually running, i.e. approx. 17% of the year, whereas radiation/standby losses add up to 8760 hours (occuring due to heat loss of the boiler to the surroundings). The efficiency of the boiler will be reduced by one percentage point for every 15K extra off-gas temperature.
      Burning efficiency impacts AUR (Annual Utilisation Ratio) only slightly: A 1% change in the quantity of CO2 will change the off-gas loss by about 6%, yet the change in boiler efficiency will be less than 1%. The AUR figures are only relevant in Germany. In the course of a year, radiation loss impacts boiler efficiency 6 times as much as off-gas losses.
Over the year, enhanced boiler insulation and an air/off-gas system produce greater benefit than the last AUR percentage point.

[ contents ][ continue ]
Unit: ºK*d
To obtain a unit for heat consumption in the heating period, the term 'degree-days' [DD] was introduced. According to the German standard VDI 2067, degree-day is the the product of the number of heating days and the difference between mean room temperature and mean outside temperature. The smallest measurement interval is the mean value for the day.
The number of degree-days is a statistical quantity, the shortest measurement interval is arrived at from daily mean values. Appropriate with monthly reports.
Example: In Aachen, the 25-year degree-day mean from 1973 to 1998 is 3498, while the 10-year degreeday mean from 1999 to 2000 is 3366.
Gt = total (ti-tam)

Gt   No. of degree-days
z    No. of heating days in a heating period
ti   mean room temperature (20ºC)
tam  mean outside temperature of a heating day
This is how simple degreeday numbers are!
It may seem terribly complicated, but once you get the hang of it, it's easy:
Have a look at page one of our annual report
  1. Look at column 'GradTagsZ - GT 15/20' (number of degree-days).
  2. You will see under January the value 484.1: the number of degree-days for January 2002.
  3. These 484.1ºdays are divided by the number of heating days in that month: 31
  4. The result is 15.6K, i.e. the difference in inside and outside temperature in January.
  5. The proof is on the left: Ta=4.4K. (Result of 20º-15.6º)
    In the row of 'Januar' is shown [Ti-Ta]=16.35, because the value is calculated every second and not via statistical values! A clear proof that it is always necessary to run through the same databases.

Have a look at: the town of Zurich is leading in the field to publish the number of degree-days. In Germany we are still in the Stone Age. You might like to pester the German Transport Minister if you'd like this to be changed!

[ contents ][ continue ]
Unit: none
... is the momentary ratio between use and input ('eta' for short, value inferior to 1).
For a business man: profit against cost.
Important is the period of time over which the measurement is taken:

[ contents ][ continue ]
Energy consumption characteristic value
Unit: kWh/m²*a
Primary energy consumed in one year related to a fixed value of the building (usually the heated or air-conditioned useful area). This is defined in VDI 3807. The smallest measurement interval is an annual mean value.
Type of building Year Without hot water: kWh/(m²a)
Air-conditioned office area 1970 800
2000 200-300
Multi-occupied residential building 1975 200-350
1992 70-100
Residential house 1982 203-226 (mean)
1992 60-80 (requirement)
Low-energy house 2000 <50-35 (requirement)
Zero-energy house 2002 <15 (requirement)

The shortest measurement interval is a year, result is a year's mean. If you feel like looking at the numbers, look here.

[ contents ][ continue ]
Energy certificate, Primary energy certificate
Unit: kWh/(m²*a)
In the year 2006, EU guidelines will make energy certificates mandatory. The energy certificate will therefore be an aid in countering the greenhouse effect, unless oil consumption is cut by oil price increases before the certificates arrive :-))
dena-sticker 2003
Classes of the dena label:
The characteristic values refect primary energy consumption.
The energy certificate will put an end to all sorts of heating, energy, and building certificates. The final version is currently being field-tested in 20 different versions. The version in Saxony already looks promising.
      On the left you can see the classifications of the future certificate. The dena figures are about 10...20% higher than the previous characteristic values for heating energy consumption and include insulation and heating mode.

[ contents ][ continue ]
Heating day
Unit: d
Heating days (according to definition) are days on which the mean outside temperature is below 15ºC. In Europeen latitudes, the daily fluctuation rate is approx. 10ºC. Therefore, over a given 24-hour period, daytime temperature may be 20ºC, and the night-time temperature 10ºC.
The shortest measuring period is 24 hours, which means that ananlysis is only meaningful for periods of at least one week.

[ contents ][ continue ]
Heating degree-days
see Degree-day

[ contents ][ continue ]
Mean value calculation
Each value captured by a sensor is a momentary value. What the result represents is a value which depends on the sampling frequency of the device: if the device only captures a single value within the interval, we know very little about events during the entire interval. One value per second achieves a more accurate result, when the interval was selected as 15 minutes: the mean is more accurate.
      Example of data gathering: Temperature data are normally saved every 5-10 minutes. We save the data every second to obtain an accurate 15-minute mean from 900 momentary values, or a 60-minute mean from 3600 momentary values. To achieve accurate process monitoring we form real 2-minute mean values from the per second data.

[ contents ][ continue ]
Measurement interval, period
[Lat.: intervallum] This is the period or value range over which the measurements applied in the calculation of mean and other statistical characteristic values are taken. In principle, the intervals can be freely selected, but a number of particular intervals have proved to be practical. 2 minutes is a good timestamp for slow processes. A 15 minutes-mean is applied by the energy suppliers. 1 hour, 1 day, 1 week, etc... are known time blocks according to which we divide up our time.
      Mean values only become available when the recording of data over a given period in order to obtain these mean values has been completed. Thus, the mean value from one day only becomes available on the following day, although the time stamp conventionally shows the middle of the measuring period, in this case, day. A day mean value is the value of the measurements taken over 24 hours, not the mean value during daylight!
      Example: Gathering data of the outside temperature for 1.1.1999 is completed in the first second of the new day, but the mean of 3.5ºC is assigned to 1.1.1999 at 12:00.

[ contents ][ continue ]
Mean value calculation does not provide a stisfactory rerference point. To obtain more meaningful results what we need is a value which is more reliable over a longer period than any single value currently being measured. For this reason we use a long-term mean value (e.g. 1 day, 1 month or 1 year).
  • Burner running time
    The most important parameter which tells us how well a heating system is adapted to its object is the mean duration of burning per switch-on.
                                   Burner running time/year
    running time/switch-on [min] = ------------------------ * 60
                                   no. of burner switch-ons
    This gives you the most important statistical quantity (in minutes).
    Which values are good and which values are bad?
    1500 hours of burner running time per year result in
    •  2000 starts lasting 45 mins (normally unachievable wit conventional controlls)
    • 10000 starts  lasting 9 mins
    • 20000 starts lasting 4.5 mins (national German average)
    • 40000 starts lasting 2.25 mins (in such a system your Annual Utilisation Ratio will only reach 30 to 40%! More information under burner switch-ons)
    • even 100000 starts is something technicians have experienced...

  • Heating work
    Savings in heating energy are only comparable after first eliminating the fluctuations by normalizing the number of degree-days. The annual comparisons should be normalized via a long-term floating mean (e.g. 10 years):
    Bnorm = Byear * ------
    Gt   No. of degree-days
    B    fuel consumption
  • Examples: Optimisation in our demonstration system Obermühle

[ contents ][ continue ]
Normalized Utilisation Ratio NUR
Unit: %
Load steps acording to German DIN 4702-T8
[Source: Viessmann]
Utilisation Ratio is the ratio between heating output and firing input over a heating year. This includes as variables, time and statistical quantities of utilisation (hence the name).
(Efficiency is the current ratio between use and [energy] input).
The Normalized Utilisation Ratio (graph shown left) is defined in the German Standard DIN 4702-T8 as a boiler comparison procedure. To standardize the measurement of boilers, the total annual heating work of the boiler (work = output * time) is broken down into five partial output blocks of equal area depending on outside temperature. In conventionally controlled systems this measured NUR is never attained.

Table to determination of the five partial utilisation ratios according to German Standard DIN 4702-T8

Relative boiler output in % Temperature of heating medium
System temperature 75/60 System temperature 40/30
dT [K] Tout [C] Tin [C] dT [K] Tout [C] Tin [C]
13 2 27 25 2 23 21
30 5 37 32 3 26 23
39 6 42 36 4 28 24
48 7 46 39 5 30 25
63 10 55 45 7 33 26

Although the DIN specifies the above table, it does not specify in detail the test procedure and this represents a basic lacuna in NUR test bench procedures. This enables each manufacturer to decide when the values are reached, which is exactly what they do. Our own inquries have shown that not one of the three leading German manufacturers have binding guidelines for standardisation of these values, which are so important commercially. All published values are thus either 'manufactured' or random. It might be a good idea for the boiler manufacturers to get inspiration from their colleagues in the chemical industry, where quality assurance is firmly based on Good Manufacturing Practice (GMP).
      Did you ever run a boiler at these temperatures? The level of condensation would lead to water running out at the bottom... Have a look at the low dT values!
      The German Normalized Utilisation Ratio is due to be replaced: in future, only boiler efficiency at 30% nominal output should be of relevance. Let's hope that all manufacturers will have to measure in the same way. Of course, this won't mean that the same values will be achievable when the boilers are actually running, which begs the question as to whether the NUR is not a long-term flop, or a way of leading consumers down the garden path.

Outer temperature
Unit: ºC
For meteorologists, the air temperature measured approx. 2 meters above the ground using a temperature-sensitive resistance (usually a PT 100 or PT 1000).
      (In many places, the outer temperatures at meteorological stations used to be measured three times a day and the evening value was included two times in the mean value. The difference to the actual mean was said to be relatively small :-)
      The heating control device should measure the wall temperature, not the air temperature.
More about measurement accuracy.
[ contents ][ continue ]

2. Ideas about measuring temperatures

No need to get confused about current temperature values!

When setting boilers and controllers, you always have to measure and track temperatures: Learn to read your data correctly!

Outside temperature
Perhaps you have already looked at our temperature extremes and you have seen that settings and planning are never based on the infrequently occuring extremes. The extremes form the day mean values only when it is as cold as it was on 1.1.1997 for days on end.
So remember, when setting a controller, to set it for day mean value. At any given time, the day mean value can be higher or lower than the current temperature value.
Boiler temperature
When the burner is running, the boiler temperature indicator only tells half the story: the sensor ist installed at the outlet, i.e. the hottest zone, of the boiler. The boiler back flow may be stone cold. This means that if there is no boiler circulation pump, the mean value will be generally lower than the indicator shows!

How to correctly measure technically produced temperatures.

The ongoing calculation of mean values is not always appropriate, as the example of the off-gas temperature shows.
      When you are measuring continously, the mean off-gas temperature is also being formed at those times when the burner is off. In this case, the day's mean off-gas temperature can be described as 'low with a number of peaks over the running time'. The 200C peaks are averaged away and 'lost' in the mean boiler temperature level. But our aim is to measure the mean off-gas temperature only when the burner is on. We may also be interested in the cool-down phase up to the point when the boiler temperature is reached.
      We decided to calculate in the operating log (in a second step) only the losses. We did this by having a varying excluding condition: the off-gas temperature is only included in the mean calculation only when it is above the boiler water temperature, which is when the boiler is heating the chimney.
Example: (in a new window) Daily heating report

[ contents ]

3. What is a kilowatt hour?

A unit which quantifies work. Work costs money. Work = effort * time.

A state-run telephone company is normally more expensive than private providers. Do you know that in 2005 you were paying up to 20 cents per kilowatt hour of electricity, bur only about 5 cents for a kilowatt hour produced by gas or oil?
Poor efficiency of only 50% in heating systems with standard controllers will double your cost to 10 cents/kWh. Not convinced?
How is a kilowatt hour produced?

  • from half a shovel of coal
  • from 100cm heating oil or 1/10m gas
  • from 45 minutes of solar radiation during the middle of the summer of 1m of earth.
  • from the muscle power generated by working out on a home trainer for about 10 hours
What do you get from a kilowatt hour?
  • a load of washing at 60C using an economic washing machine
  • run a 300-litre refrigerator for 2 days
  • run a 10-watt energy-saving lamp for 100 hours
  • a standard 100-watt bulb only gives 10 hours of running time.
  • you can work for 5 hours on your computer with no artificial light.
  • hot plates will waste this energy in 30 minutes,
  • electrical instantaneous water heater only 3 minutes!
...in terms of hot water
Your 10-minute morning shower will cost you about 120 litres of water, which adds up to 64.5 cts if heated electrically or 62.8cts if heated with natural gas or oil. A continuous-flow heater is 99% efficient, the heating and boiler system only ~50%.
...in terms of heating
A one-family dwelling rated at 100kWh/(m*a) with a living area of 200m which is actually quite economic uses 2000 litres of oil annually. These are burned in approx. 1500 hours, i.e. 1 litre of oil will last 45 minutes. However, most one-family dwellings consume 4000 litres of oil, which is 1 litre every 22 minutes.
Normal annual utilisation ratios are 50% to 60%, with the best systems currently achieving an Ho-related figure of 92%
  • 1 kWh in a low-energy house - 50kWh/(m*a): 9 minutes heated
  • 1 kWh in an economic house - 100kWh/(m*a): 4.5 minutes heated
  • 1 kWh in an average one-family dwelling in 2005 - 200kWh/(m*a): 2.2 minutes heated

| Homepage | · | Expertise | Save | Electrical... | Heating... | Stastics - Top |

eMail © System Integration Beitzke
page written 03.1.1999, last change 13:23 29.6.2014, Sonntag