**3. Stochastic Simulations**

After the reference building models were validated, stochastic simulations were conducted to consider di fferent building design variations for the urban building energy consumption database development. Eight di fferent design parameters, such as the building envelope's thermal properties, infiltration, heating and cooling schedules, lighting power density, and equipment power density, were considered to cover di fferent constructions, building design scenarios and the occupant's energy usage patterns.

To di fferentiate building vintage, three levels (high, medium, and low) of building envelopes were studied by grouping U factors of di fferent parts (external wall, slab, roof, and glass). The building geometry was kept constant to represent the most common configuration in Wuhan.

To reflect the actual energy saving behavior of the occupants and better capture di fferent energy usage patterns, thirteen heating and cooling schedules were proposed. The schedule information was derived based on statistical analysis of the actual building energy data from CRECS. The data from the CRECS energy consumption survey was ranked from low to high. Level of 5% means the top 5% from the ranking. It represents the most e fficient energy usage, in terms of heating and cooling hours per day and days per year. Level of 95% represents the least e fficient energy usage (bottom 5% from the ranking). It is assumed that the lighting and plug/equipment loads are coupled with heating/cooling schedules, since people's energy saving behavior is consistent. Other energy usage profiles can be interpolated.

Figure 8 shows the stochastic cases of the residential building. In total, 117 design scenarios were considered. The combination of residential building baseline model is highlighted in yellow. In a similar way, 117 small o ffice and 117 large o ffice EnergyPlus models were generated to cover di fferent energy use intensity scenarios for o ffice buildings.

**Figure 8.** 117 stochastic scenarios of the residential building model.
