Load Composition

This document describes the load composition methodology for estimating the load on a feeder. This methodology is important to identifying the response of the feeder to voltage changes.

The methodology is implemented Load composition.xls spreadsheet locating in the Load Composition download. If you want to load different TMY data files, you will also need to extract the TMY files from the TMY folders using the Load TMY Data button on the Conditions worksheet.

= Aggregation method =

Load composition is the term used to describe the breakdown of end-use load according to the nature of the load. The current load composition breakdown is as follows.


 * 1) Electronic $$P_E$$ loads are those that use power electronics.  These loads typically have constant power requirements, but often have adverse harmonic characteristics that make them appear to have poor power factors.
 * 2) Motor A $$P_A$$  are three-phase induction motors that drive constant torque loads, such as industrial and commercial compressors and refrigerators.
 * 3) Motor B $$P_B$$ are three-phase induction motors that drive speed-squared loads with high inertia, such as fans.
 * 4) Motor C $$P_C$$ are three-phase induction motors that drive speed-squared loads with low inertia, such as pumps.
 * 5) Motor D $$P_D$$ are single-phase induction motors that drive constant torque load such as residential A/C compressors, refrigerators and heat-pumps.
 * 6) ZIP $$I_p$$ is the real part of constant current loads.
 * 7) ZIP $$I_q$$ is the reactive part of the constant current loads.
 * 8) ZIP $$P_p$$ is the real part of constant power loads (often denoted as P).
 * 9) ZIP $$P_q$$ is the reactive part of constant power loads (often denoted as Q).
 * 10) ZIP $$Z_p$$ is the resistance part of constant impedance loads (often denoted as G)
 * 11) ZIP $$Z_q$$ is the reactance part of constant impedance loads (often denoted as B)

Each load type has a contribution from residential, commercial, industrial and agricultural components. The residential includes single-family homes and multi-family buildings. Commercial loads include small and large offices, small and large retail, hotels, and motels. Industrial and agricultural loads are idiosyncratic and are not modeled in detail.

Taken together the combined power factor of the constant loads is

$$ PF = \begin{cases} P_q + I_q + Z_q = 0 & 1 \\ P_q + I_q + Z_q > 0 & \frac{P_p+I_p+Z_p}{\sqrt{(P_p+I_p+Z_p)^2+(P_q+I_q+Z_q)^2}} \\ P_q + I_q + Z_q < 0 & -\frac{P_p+I_p+Z_p}{\sqrt{(P_p+I_p+Z_p)^2+(P_q+I_q+Z_q)^2}} \end{cases} $$

The total power (magnitude) $$P_{total}$$ of the composite load is

$$ P_{total} \approx |P_E| + |P_A| + |P_B| + |P_C| + |P_D| + \sqrt{(P_p+I_p+Z_p)^2+(P_q+I_q+Z_q)^2} $$

= Climate =

TMY2 climate data is used to determine the weather conditions for any particular month, day of week, and hour of day. In the TMY2 file, the following data is used


 * 1) Dry Bulb Temperature [1/10 °C], which is converted to &deg;F;
 * 2) Wind speed [tenths m/s], which is converted to miles/hour;
 * 3) Relative humidity [%];
 * 4) Diffuse Horizontal Radiation [Wh/m^2], which is converted to Btu/sf.h; and
 * 5) Direct normal Radiation [Wh/m^2], which is converted to Btu/sf.h.

In addition, the following data is either extracted from the TMY2 data or looked up elsewhere


 * 1) Latitude, which is obtained using the city;
 * 2) Heating design temperature [F], which is the minimum TMY2 observation;
 * 3) Cooling design temperature [F], which is the maximum TMY2 observation; and
 * 4) Peak solar radiation [Btu/sf.h], which is the maximum direct normal observation.

= Residential loads =

Single and multi family dwellings are represented by typical loads, which are used to characterize a population of homes on a feeder. If multiple characteristics are needed, multiple models must be used and the results summed before computing the composite load. The load contribution to the feeder is always multiplied by the number of dwellings having those characteristics.

Single-family dwellings
The basic characteristics of single-family residences are shown in Table 1.


 * Notes :
 * 1) $$\dot V_{thermal} \approx 1.877 Floorarea \sqrt{Buildingeight|T_{in}-T_{out}|/T_{in}}/Airvolume$$ in puV/h
 * 2) $$\dot V_{wind} \approx Floorarea \times Windspeed \times 10^{-5}$$ in puV/h
 * 3) $$Q_{peaksolar} = Peaksolar \times Windowarea \times Shading \times Exposurefraction/8$$ in Btu/h with $$Exposurefraction = \begin{cases} Solarelevation > 0 & : \sqrt{2} \cos(Solarelevation) \sin(Solarelevation) \\ Solarelevation \le 0 & : 0 \end{cases}$$
 * 4) $$H_{vent} = 0.182 Ventilationrate \times Airvolume$$
 * 5) $$UA_{wall} = Wallarea(1-Windowwallratio)/Wallrvalue\,\!$$
 * 6) $$UA_{roof} = Flooarea/Roofrvalue\,\!$$
 * 7) $$UA_{window} = Wallarea \times Windowallratio \times Windowrvalue$$
 * 8) $$Solarexposure = Windowarea (1-Externalshading)Exposurefraction/8 \,\!$$

The end-use electricity is used to determine what fraction of the end-use load ends up as electric load, as shown in Table 2.

The system efficiency is used to determine that operating load of the end-uses, as shown in Table 3.

The installed capacity is used to determine the total end-use load capacity per unit floor area, as shown in Table 4.



Before computing the diversified load, the end-use load shapes for the 5 demand-based end-uses (cooking, hotwater, lighting, plugs and washing) are used to determine the fraction of the load operating at a given time. For end-use load shapes are used for winter/summer and weekend/weekday conditions, as shown in Figures 1-4.

The ELCAP end-use load shape are rescaled according to the daily energy use estimates, as shown in Table 5. The estimates used a approximately 50% of the original daily ELCAP consumption.

The non-demand end-uses (heating, cooling and refrigeration) are computed directly from the power density and the end-use duty-cycle (if any)

$$ Q_x = DC_x \times PD_x \times Floorarea/1000 $$

The final diversified end-use load is computed by looking up the rescaled ELCAP demand for the season (winter/summer) and day type (weekday/weekend), multiplying by the power density and the floor area. The final diversified load is weighted between the winter and summer values based on the day of year.

$$ Q_{winter} = Q_{ELCAP_{winter}}\frac{E_{daily_{winter}}}{E_{ELCAP_{winter}}} \times PD_x \times Floorarea/1000 $$

$$ Q_{summer} = Q_{ELCAP_{summer}}\frac{E_{daily_{summer}}}{E_{ELCAP_{summer}}} \times PD_x \times Floorarea/1000 $$

$$ \begin{align} Q_{diversified}& = Q_{winter}(1-|\sin(\frac{3.14}{12}(month-1.5))|) \\ & + Q_{summer}|\sin(\frac{3.14}{12}(month-1.5))| \end{align} $$

Finally, the end-use load composition is determined by multiplying by the end-use composition matrix for single-family residential dwellings:

Multi-family dwellings
Multi-family buildings are very similar to single family dwellings, except that some of the default parameters are different or calculated differently. Specifically
 * Floor area : The floor area per dwelling unit is used as the basic parameter. When combined with the floors per building and units per floor' this give a rough approximation of the total building floor area (excluding conditioned circulation space).
 * Floor to floor height : The floor height is used to compute the total building height and air volume.

The load composition matrix for multi-family buildings is shown in Table 6.

= Commercial loads =

All commercial load composition models are developed using the California End-Use Survey (CEUS) results. These results are largely valid for the WECC, but care should be take to account to differences in construction types and building codes in regions not covered by the survey.


 * Note : The CEUS data used is for the whole state of California. There is CEUS data available for specific utilities, but that was not deemed helpful for load compositions that would apply WECC-wide.

The CEUS data from the the Itron CEUS results website was used to obtain the commercial load tables. The exp16day data set is used because it is more compact than the 8760 data set, but provides end-use load shapes for each season, day type, and hour.

The load densities for the following end-uses are estimated for each building type.
 * Heating
 * Cooling
 * Ventilation
 * Water heating
 * Cooking
 * Refrigeration
 * Exterior lighting
 * Interior lighting
 * Office equipment
 * Miscellaneous
 * Process equipment
 * Motors
 * Air compression

The basic method for determining commercial load composition is to estimate the hourly load density (W/sf) using the hourly CEUS energy data. The load densities are then multiplied by the average building floor area to yield the load (MW).

The heating and cooling loads are interpolated based on the building's balance temperature. The heating load is multiplied by the heating duty cycle

$$ \rho_{heating} = \begin{cases} T_{out}< T_{balance}& : \frac{T_{balance}-T_{out}}{T_{balance}-T_{design_{heating}}} \\ T_{out}\ge T_{balance}& : 0 \end{cases} $$

and similarly the cooling load is multiplied by the cooling duty cycle

$$ \rho_{cooling} = \begin{cases} T_{out}> T_{balance}& : \frac{T_{out}-T_{balance}}{T_{design_{cooling}}-T_{balance}} \\ T_{out}\le T_{balance}& : 0 \end{cases} $$

Small office
Each small office load vector is multiplied by the small office end-use composition map to determine the end-use load composition for small offices, as shown in Table 7.

Large office
Each large office load vector is multiplied by the large office end-use composition map to determine the end-use load composition for large offices, as shown in Table 8.

Retail
Each retail load vector is multiplied by the retail end-use composition map to determine the end-use load composition for retail buildings, as shown in Table 9.

Lodging
Each lodging load vector is multiplied by the lodging end-use composition map to determine the end-use load composition for lodging buildings, as shown in Table 10.

Grocery
Each grocery load vector is multiplied by the grocery end-use composition map to determine the end-use load composition for grocery stores, as shown in Table 11.

Restaurant
Each restaurant load vector is multiplied by the restaurant end-use composition map to determine the end-use load composition for restaurants, as shown in Table 12.

School
Each school load vector is multiplied by the school end-use composition map to determine the end-use load composition for schools, as shown in Table 13.

Health
Each health load vector is multiplied by the health end-use composition map to determine the end-use load composition for health care facilities, as shown in Table 14.

= Industrial and agricultural loads =

Industrial and agricultural loads are considered idiosynchratic and must be entered directly as an end-use load composition on their respective worksheets.

= Analysis results =

A number of analysis results are provided with the worksheets. The Feeders analysis enumerates the load component compositions for residential, commercial, and mixed (50/50) feeders. The Loadshapes worksheets provides daily load component shapes for Portland OR. The sensitivity analysis provides the sensitivities of component loads to temperature.

Feeder compositions
The feeder component compositons were computed for winter peak, typical shoulder, and summer peak conditions at 6:00, 9:00, 15:00, and 18:00 hours for a 100% single-family residential feeder, 50% single-family residential/small-office commercial mixed feeder, and a 100% small-office commercial feeder in each of the cities for which climate data was available. The tables were generated using Version 1.6.3.

Load component shapes
Figure 1 - Daily clustomer type shape (Portland OR, weekday, summer peak)

Figure 2 - Daily customer type composition (Portland OR, weekday, summer peak)

Figure 3 - Daily load component shape (Portland OR, weekday, summer peak)

Figure 4 - Daily Customer component composition (Portland OR, weekday, summer peak)

Sensitivities analysis
The load composition sensitivity analysis computes the changes in important output values with respect to changes in certain input values. The output values considered load (MW) and composition (%) for each of the building types. The input values considered are the temperature (F) and the number of buildings.

In the spreadsheet, the sensitivity analysis is performed only when the Update sensitivities button on the Composition worksheet is pressed.

The following load sensitivities were calculated with Version 1.6.2 using a –1 &deg;F perturbation on peak cooling conditions.


 * Notation : The value 0 indicates that no change was detected. The value 0.000 indicates the change was less than 0.0005.  The value - indicates that the difference is between two very small or zero values.

= Other sources =


 * California Statewide Residential Appliance Saturation Study; Volume 2, Study Result; Final Report; June 2004