BPS and IDA ICE Encyclopedia
Previous (What makes good BPS?) | Start | Next (The challenge of complex and multifaceted problems)
What makes good BPS?
Requirements on good BPS
Use the best validated physical models
BPS models are generally based on fundamental laws of physics. The most important thing is first and foremost that these fundamental laws (such as the conservation of energy or mass) are fulfilled. But every pyhsical model can't consider every detail, but rather contains approximations. To ensure that these approximations do not lead to significant errors in the results, the models used should always be carefully validated before they are built into a BPS software.
There are three types of validation of simulation models. The first is the comparison of the results with those of equivalent simulators. This requires a very precise description of all possible input values for test examples that are as relevant as possible. The most important example of such peer-to-peer validation is the American standard ASHRAE 140, which has now also been adopted in the European standards EN 52016 and 52017. These standards replace the old European validation standards EN 13791, 15255 and 15265, in which the simulation results had to be compared with those of a reference model defined in the standard.
Peer-to-peer validation
The second type of validation is validation against a reference model. In many countries, the majority of building simulation programs are validated in accordance with European standards and will therefore soon be switching to the peer-to-peer validation of ASHRAE 140, whereas for example in Germany validation still has to be carried out in accordance with VDI 6007, which requires the same results as with the n-k node model specified in the standard. Validation against such often very unphysical models can be almost infinitely time-consuming because a fundamentally different model has to be adapted to the reference model.
Validation against a reference model
The most useful, but also the most difficult type of validation is the comparison against measurements. This requires not only a careful measurement of the results to be compared, but also all the input values required for the calculation, such as details of the building envelope and measurement of the weather data. Measurement data that is both accurate and complete enough to be used for meaningful validation is very rare. Measurements in well-monitored laboratory test cells are easier than in real buildings exposed to uncontrollable conditions.
Validation against measurements on a laboratory test cell
The most interesting thing for the user is probably the comparison of his simulation results with measured values on the real building. In the example of the following figure, the purchased district heating for an office building is compared. It shows that a few key parameters are sufficient to calibrate the model correctly.
Vlaidation against measurements on a real building
Be efficient and flexible
The requirements of planning practice are always controversial. The user wants the software to be handeled in a simple way, but he still wants detailed models. The simulation must run quickly, but deliver accurate results. And the software has to protect him from errors, but provide great flexibility.
simple vs. detailed
quick vs. accurate
error-free vs. flexible
If one concentrates only on the requirements on the left (simple, quick and error-free), the result is an imprecise procedure that provides hardly any information about the real system. However, if you concentrate only on the requirements on the right (detailed, accurate and flexible), the result is far too complicated for the user. Simulation software must therefore be as open and flexible as possible for the parameterization of the model and at the same time offer the user efficient input aids that guarantee correct parameterization in standard situations.