Sensitivity Analysis In Building Environment Based On Existing Data And Energy Models

With the development of economy, the proportion of building energy consumption is increasing. In the energy saving analysis of new building and existing building, sensitivity analysis can determine the main factors affecting energy consumption or carbon emission in order to evaluate the reliability of the different energy saving measures. The method of sensitivity analysis in theory provides analysis basis for the uncertainty and optimization of building energy consumption. In practice, sensitivity analysis can guide project of energy conservation and emission reduction. Hence, the sensitivity analysis methods are widely used in the region of building energy assessment.Sensitivity analysis method combines the energy modeling software with mathematical statistics software effectively, which not only improves the calculation speed,but also enhances the precision of energy computation. In the analysis of building energy consumption, there are a large number of variables and the interactions between variables are particularly complex. The global sensitivity analysis method can overcome the shortcomings in order to explore the relationship between input variables and energy consumption. At present, global sensitivity analysis methods are used rarely in the region of domestic building energy consumption, but mostly used in the operation capability of weapon equipment and the application of tsunami wave. In the region of international building energy consumption, there is lack of comparison of different sensitivity analysis methods affecting building energy consumption. This paper can be divided into two parts:sensitivity analysis based on existing energy use data and based on building energy models.The first part is based on the existing data of building energy use. This study chooses the energy consumption data of London as the research object in order to compare the characteristic of several sensitivity analysis methods, including Genizi,CAR(correlation-adjusted marginal correlation), PCC(partial correlation coefficient), and SRC(standardized regression coefficient).The results indicate that the land area for domestic buildings is the only dominant variable influencing gas use. The higher the council tax band is, the much more influencing the gas and electricity is. SRC is not suitable for the correlated factors in urban energy assessment. The results of other three methods are similar.The second part chooses an office building as object in order to compare the characteristics of four global sensitivity analysis: SRC(standardized regression coefficient),Morris design, extended FAST(Fourier Amplitude Sensitivity Test), and TGP(treed Gausian process) method. The results indicate that the TGP method is the best choice in terms of both accuracy and computational cost. Note that the TGP method needs more time to calculate the sensitivity index although it requires only moderate time for running building energy models. If choosing the same sampling method, the time of running building energy models would be the same. The more the number of building energy models is, the more time for running building energy models. It is also found that the time for running building energy models is longer than that of calculating sensitivity index. In building energy analysis, at least two different methods for sensitivity analysis are proposed to be used to provide more robust results.