**3. Case Study and Results**

## *3.1. Case Description*

To compare the environmental cost of residential buildings during the operational stage, two sites were selected as case study buildings. One is in Beijing, and the other is in Xiamen, which represent two different climate zones in China. According to the construction organization flow chart, the construction period is 2 years. The major construction materials include concrete, rebar, steel tube, cement mortar, wood, aluminum, glass and alkyd paint. The specific information and corresponding direct costs of the two case buildings are shown in Table 8.

**Table 8.** Comparison of the case study buildings in Beijing and Xiamen.


Data and specifications required for this study are obtained from the structural drawings of the buildings, acceptable LCI database and other archived literature. The materials used for construction are specified in the bill of quantities and can be obtained from the contractors.

#### *3.2. Results Analysis*

Based on the calculation method mentioned above, per capita energy consumption level of residents in Xiamen and Beijing were calculated based on China's Yearbook. The environmental costs of the two case studies are shown in Figures 2–5.

Most of the input data used in this case study comes from actual utility bills. However, due to the inevitable limitations of the input data, corresponding assumptions were made during the analysis. Considering the variability of critical input variables, sensitivity analysis of key parameters was conducted. Sensitivity analysis is the measurement of changes in one or more uncertainties to determine the extent to which changes in each factor affect the expected objective [34]. In this paper, the single-factor sensitivity analysis method is used to quantitatively describe the importance degree of input variables when only one parameter changes by 1%. The calculation formula is as shown in Equation (9).

$$E\_i = \Delta \mathbb{C}\_{\text{er}} / \Delta F\_i \tag{9}$$

where *Ei* is the sensitivity parameter of the variable *Fi*; Δ*Cei* is the corresponding rate of change in environmental costs (%); Δ*Fi* is the rate of change of the variable *Fi*, taken as 1%.

To find the critical input variables, the sensitivity analysis results are shown in Figure 2. For both of the case study buildings, the sensitivity coefficient of the unit virtual abatement cost of CO2 is the largest, equaling 0.67, which means that CO2 has the largest influence on the final environmental cost. Additionally, the unit virtual abatement cost of N2O, CH4 and NOx are also key parameters that may lead to significant changes in the outcome, with values of 0.12, 0.45 and 0.32 respectively. The environmental cost results are not sensitive to the unit virtual abatement cost of CO, COD, dust, NH4 <sup>+</sup>, SO2, solid waste and VOC.

**Figure 2.** Sensitivity coefficients of the unit virtual abatement costs.

Additionally, the data quality indicators (DQI) method [35] (see Table 9) and Monte Carlo simulation were used in this case study to analyze the LCA data quality and uncertainty of the results. The engineering quantity data are all from the engineering quantity list, and the emission factor data are from a database. According to the standard deviation provided in Eco-invent [36], the distribution type of the LCI data is selected as lognormal distribution, and the uncertainty is shown in Table 10. Using the Monte Carlo simulation, the variability of environmental scores associated with the ratio of green construction measures cost to each subengineering fees, transportation distance and the average fuel consumption for each vehicle can be estimated. The selected variables are assumed to be uniform distribution or lognormal distribution (see Table 10), and 10,000 iterations were carried out based on previous studies [37].


**Table 9.** Data quality indicators (DQI) and uncertainty.

**Table 10.** Data uncertainty of each parameter in the calculation.


Figures 3 and 4 show that the added variability did not significantly change the average values nor did it change the ranking of the four stages in terms of *Cva*, *Ced* and *Cgc*. The minimum, average and maximum total environmental costs are 412, 616 and 827 CNY/m2, respectively, in the Xiamen case study building, while they are 489, 673 and 899 CNY/m2, respectively, in the Beijing case study building. The coefficient of variation of the material production stage is the largest, followed by the operation and maintenance stage, while that of the demolition stage is the smallest.

The average value of the case study building in Xiamen is shown in Figure 3, where the total environmental cost is 616.29 CNY/m2, of which the biggest contributor to environmental cost is material production stage reaching 330.96 CNY/m2, followed by operation stage, 199.40 CNY/m2. The environmental cost of demolition stage is negative, which indicates that the recycled material can bring positive environmental benefit. For the case study building in Beijing (shown in Figure 4), the total environmental cost is 672.80 CNY/m2. The environmental cost of material production stage is 307.42 CNY/m2, followed by operation stage, 247.07 CNY/m2, which is slightly higher than that of the Xiamen case building's operation stage. This is possibly because energy consumption of heating is excluded for the case study building in Xiamen, which is located in a hot-summer and warm-winter zone where heating in the winter is not necessary. For the both case study buildings, construction

stage is the third largest contributor to the environmental cost of the life cycle. During this stage, the green construction cost accounts for the largest percentage of the total environmental cost, about 65%. Demolition stage has the minimum environmental cost. For the both case study buildings, *Ced* accounts for about 69% of the total environmental cost during material production stage, operation stage and demolition stage.

(**c**) *Cgc* (**d**) Total environmental cost

**Figure 4.** Environmental cost of case study building in Beijing (unit: CNY/m2).

According to the results of both case studies (see Figure 5), a default discount rate was selected as 7%; the environmental cost in life cycle achieves an indispensable 14% share of the direct cost. However, the existing direct cost of the life cycle often neglects environmental cost, resulting in a great warp between calculation results and actual results. In some research, environmental costs are roughly assumed as 10% of direct costs. Since the percentage adopted is less than the result of this case study, this would lead to an error. In order to consider the variability of discount rate, therefore, this study

assumes discount rates of 7%, 12% and 17%. The uncertainty analysis of the two cases shows that the ratio of environmental cost to direct life cycle cost decreases as the discount rate increases, but the change does not exceed 5%.

**Figure 5.** Ratio of environmental cost to direct cost considering different discount rate.
