Can dynamic thermal models reliably predict home energy demands; empirical validation and uncertainty analysis?
Ozlem Duran Loughborough University
The purpose of the project is to search the accuracy of the dynamic simulation programs, such as EnergyPlus, by using the empirical validation and uncertainty analyses.
The aim of the project is to understand the absolute accuracy of the Energy Plus and the BREDEM model by comparing their predictions with the measured energy demands and internal temperatures of chosen houses.
The building energy models are the virtual models, which are developed to calculate and evaluate the energy consumption of the buildings. Earlier days, these models are done by basic mathematical calculations. By the time, regarding computational development, energy simulation programs are created, which can visualise and calculate the energy performance of a building under realistic conditions. To date, energy simulation programs can model a building as a three dimensional exact copy. After modelling phase, the inputs are implemented to this model to make it closer to the reality. These inputs are the properties of the buildings, such as; materials, orientation, the gaps and the cracks of the envelope. Or the usage information, like; how many people live in, what kind of appliances are used, how many hours the building is heated, how efficient the heating system is. The simulation phase follows the modelling phase, by which various results can be found out according to the requirements of the work, such as; total energy consumption of the building, air temperature, heating system demand, comfort level of the occupancies, performance of the building envelope.
Although the capabilities of the simulation programs progress rapidly, there are still problems that are needed to be searched. In this project one of these problems is argued. For the case that a stock of buildings has to be modelled and the input data is limited, can a simulation program such as EnergyPlus find out reliable results? This question has to be answered because the answer of the question will affect the approach that will be taken for the SECURE Project. In SECURE Project, the building stock of the UK is aimed to be modelled. For the purpose of answering this question, a specific building will be chosen, modelled and simulations will be run. In order to search the accuracy of the program, the simulation results will be compared with the real consumption values. This process is called empirical validation. The more the simulation results are close to the real consumption, the more the program will be accepted as validated. But as mentioned, the input data is limited and for some inputs, the exact values are not measured. Thus, some assumptions have to be made. But anything that is assumed rather than measured would change the predicted results. Therefore, to understand the effect of assumptions on the results, an analysis method, named uncertainty analysis, will be applied to the results. In each simulation, an input will be increased in a small margin, such as 1% and the change percentage of the result will be checked. This correlation will give the information about the sensitivity of the input, in other words; the more an input is sensible, the more the result will be affected. When the most sensible inputs and their sensibility levels are known, models can be calibrated. The simulation results of calibrated models, which are in an accurate range, can be accepted as validated.
The aim of the project is to understand the absolute accuracy of the Energy Plus and the BREDEM model by comparing their predictions with the measured energy demands and internal temperatures of chosen houses.
This aim will be achieved by pursuing the following objectives;
– Understanding the theory of Energy Plus and principles behind empirical validation
-The evaluation of BREDEM models and measurements
-Modelling the chosen case studies
-The uncertainty analyses