Anyone who has built a house knows how challenging it can be. Countless decisions have to be made before you even start: What do I want the house to look like? What materials should I make it out of? What is the best type and quality of insulation? Which windows should I choose? What kind of heating system will keep it warm in the colder months? Do I need a ventilation system? And what are the best systems to install for the various building services? It really is one question after another – and many of them come back to the issue of energy. People’s awareness of how much energy we consume is continuously evolving, and this is reflected in political efforts to move towards a greener, more sustainable energy economy. These developments have a clear impact on the construction of new buildings. As the complexity of the building process increases, so too does the popularity of simulation software, especially software programs that specialize in thermal and energetic building simulations. "Professional planners use these programs to perform the key calculations required for building energy assessments before construction even starts," says Matthias Kersken, a researcher at the Fraunhofer Institute of Building Physics IBP in the field of energy efficiency. Modern buildings make use of an increasingly wide array of supply and storage technologies, and Kersken notes that this is creating buildings that behave more and more dynamically, increasing the complexity of the interactions between the various thermal transfer and storage processes. This can potentially lead to heat regulation problems and overheating in the summer. "That’s why it’s so important for the construction industry to have building simulation programs that are able to supply reliable predictions." But how do you know if the simulation results are an accurate reflection of how the finished building will behave in terms of energy performance?
"Right now, building simulation programs are validated by running mutual comparisons of their calculation results," says the Fraunhofer IBP expert. This method of comparing results from different programs is widely accepted and very successful, especially when it comes to revealing programming errors. "But there is legitimate criticism that these kinds of tests are not based on real world measurements. That makes it impossible to run any kind of systematic check as to whether the absolute level of the calculation results corresponds to reality." Kersken and his colleagues set out to change all that by signing up for the research program
IEA ECB ANNEX 58 "Reliable building energy performance characterization based on full scale dynamic measurement"
organized by the
International Energy Agency (IEA)
. Germany’s contribution to this initiative is funded by the
Federal Ministry for Economic Affairs and Energy
. The task of the Fraunhofer IBP researchers participating in this project was to generate real world data sets. "That sounds simpler than it actually is", says Kersken. "As well as obtaining high-quality sets of measurements, you also have to accurately record and document all the other conditions involved, for example the weather, the way in which the building was constructed, and any changes that occur, such as the position of the blinds." Fortunately Fraunhofer IBP has its outdoor test site near Munich where the so called
are located. The two identical houses, which are designed in the style of a typical German single-family home, are equipped with extensive systems for taking measurements and monitoring and controlling the building systems. As well as making it easy to collect all the required data, the twin house set-up also made it possible to run different scenarios in the two identical houses to compare outcomes. Both of these points were a key prerequisite for this aspect of ANNEX 58.
But before the team could start gathering the measurements required to validate the 11 participating simulation programs, Kersken and his colleagues had to get everything ready. That meant painstakingly documenting a long list of specifications for both twin houses, including their location, building geometry, and data on the optical and energy characteristics of the glazing and frames as well as information on structural components, blind characteristics, ventilation and heating systems, airtightness measurements, ground reflectance, and interior and exterior photos of the buildings. "The documentation we prepared even included aspects such as accurate calculations of each individual
," Kersken explains. "And on top of that we obviously had to carefully check and calibrate each and every measurement sensor to ensure that the data we collected wasn’t corrupted by faulty measurements." All in all the preparation phase took around two months before the team was able to start the actual measuring phase in the summer of 2013. The Fraunhofer IBP researchers conducted a total of four measuring phases preceded by an initialization phase. In each phase, the team ran through different scenarios in the twin houses: "The submodels of the building simulation programs involved in the validation process essentially simulate thermal storage, thermal conduction by components and materials, interaction between individual rooms, ventilation and infiltration, solar heat gain, shading, heating and the calculation of climatic conditions, as well as the distribution of solar radiation on inclined surfaces and the long-wave background temperature of the sky," says Kersken. "So we took measurements of all those things." To achieve that, the sensors collected data every second of the day and night which was then stored in the
database specially developed by Fraunhofer IBP. At the same time, Fraunhofer IBP’s very own
supplied regular measurements that were also incorporated in the IMEDAS™ database.
Even after they had completed all the measurements, it still wasn’t time for the Fraunhofer IBP scientists to make the data available to the 21 participating validation teams from 13 different organizations. "First we carried out what we call a blind validation in which the teams only receive part of the information," Kersken explains. Based on the calculated parameters of the twin houses, the weather and climate measurements, and the experiment specifications, the validation teams were asked to use their simulation programs to calculate the predicted figures for heating demand and indoor air temperatures on an hour by hour basis. Following this initial stage, the validation teams received the full sets of measurement data and the results of the blind validation comparisons. "That allowed us to detect deviations stemming from user and input errors, program-based model deviations, and measurement uncertainties," says Kersken.
Once all the user errors had been eliminated, the teams were able to assign any remaining deviations in the heating energy and temperature figures over time to the algorithms and model assumptions used in their simulation programs. "Having completed the comparison phase, that finally enabled us to rate the reliability of the dynamic thermal/energy building simulation programs involved in the project," says Kersken in summary. Detailed documentation on the study and the ANNEX reports will be ready by early 2016 and will be made freely available
on the Internet
"The most sensible step to take next would be to generate these same kinds of data sets in a study with real building users," says Matthias Kersken. "That’s obviously even more challenging, but it would make the simulation programs even more accurate – and that would ultimately make the energy saving predictions even more reliable."