**3. Methodology**

This work presents a bottom-up approach for modeling and forecasting end-use energy consumption and demand in Kuwait's residential buildings up until 2040. The methodology relies on information pertaining to the energy consumption of specific household equipment and appliances, where factors such as quantity, operating hours, and power requirements are accumulated and extrapolated to a national scale to ultimately estimate the usage patterns in Kuwait. Therefore, energy consumption and demand are calculated at the individual level and aggregated to estimate the national consumption and demand. In this model, end-uses were broken into air conditioning, lighting, appliances, and space heating and water heating, and further sub-categorized by different technologies. Moreover, each end-use category was further broken down by different equipment and appliances with corresponding data on diffusion rates and energy efficiency ratings. The rate of diffusion was based on data obtained from surveys and the available literature [23,24]. The driver variables of this model were based on macroeconomic variables such as population, household size and income and engineering variables like unit energy consumption, and efficiency ratings. Figures 4 and 5 illustrate the modeling structure.

The initial step is to model the quantity of equipment owned and the present initial stock. The sales and stock turnover are then derived from first purchases and replacements. The first purchases are driven by a growth in population and increase in ownership, while replacements are calculated based on the age of equipment and a retirement function. Next, the average unit energy consumption (*UEC*) and unit power demand (*UPD*) per equipment are derived and the total energy consumption and peak demand are modeled using the following general equations:

$$\text{Total Energy Consumption} (y) = \sum\_{i=1}^{L} \text{Stock}(y, i) \times \text{LEC}(y - i),\tag{1}$$

$$\text{Peak Load Demand}(y) = \sum\_{i=1}^{L} \text{Stock}(y, i) \times \text{LPD}(y-i),\tag{2}$$

where *Stock* (*y*, *i*) represents the quantity of equipment of vintage (*i*) remaining annually in year (*y*). The variable *UEC* (*y*, *i*) on the other hand, denotes the unit energy consumption at the corresponding year of purchase (*y* − *i*) and the *UPD* (*y*, *i*) is the unit demand power during the peak time. Finally, the overall useful life of the equipment is represented by *L*. Due to the lack of published information, acquiring data on the sales volumes of equipment, efficiency ratings, ownership details, and daily consumption patterns is not at all feasible for the state of Kuwait. This analysis therefore utilized an array of surveys that included national statistics and numerous reports published by the government [2,22,23].

**Figure 4.** End-use energy consumption model structure.

**Figure 5.** Energy demand model structure.
