I wrote earlier on the “saga” of electrifying a multi-unit apartment building in San Francisco. I described the 3-year process along with all the bumps, delays and mistakes that occurred along the way. Early in the electrification project, I did electrical load calculations using the procedures recommended by the National Electric Code (NEC). These calculations were used to justify upgrading the building-level electric service to 400 amperes. After the building was electrified, I installed monitoring equipment to measure both energy use and the 15-minute peak demand for each of the apartments and the “house”. For some apartments, the monitoring equipment also provided the 15-minute peak demand for individual residential appliances like induction ranges, heat pump water heaters, heat pump clothes dryers and heat pump space heaters. See Table 3 for these data.
Without diving too deep into the details, Table 1 summarizes the top-level results. In general, the NEC estimated loads were around three to six times greater than the measured peak demand. The building has five electric meters, one for each apartment and one for the “house”. The “house” is just the garage and some storage/work rooms. Each apartment and the “house” are shown as columns in Table 1. The “Total” column is the sum of the loads from all five meters (without factoring in diversity) The NEC estimated total is 483 amperes. The monitored sum of the loads (also ignoring diversity) was 111 amperes for the month of November 2025. The NEC oversizing factor is an average of 438%. Data for each of the apartments and the “house” is shown in the table. FYI: Unit #2 has an EV charger; the others do not.
Table 1 – Comparison of NEC Load Calculations to Monitored Loads
The NEC Load Calculations
Table 2 is a summary of the NEC load calculations. These calculations are based on the recommended procedure from several years ago, that I believe is still commonly used by the industry. The following is a summary of the procedure:
- Lines 1-7. The first group of loads include (line 1) general lighting and plugs at 3 W/ft2, (line 2) small appliances like coffee pots and hair dryers at 1,500 W each, and (line 3) laundry equipment at 1,500 W (the clothes dryer is accounted for later). Line 4 is the total if lines 1, 2, and 3. To account for diversity, the first 3,000 Watts are counted at 100% and the remainder is counted at 35%. These calculations are made in lines 5, 6, and 7. Line 7 shows the power to be assumed for each of the apartments and the house.
- Line 8 accounts for fixed appliances like refrigerators, water heaters, microwaves, garbage disposals, dishwashers, etc. These are listed with the estimated power for each They are summed and to account for diversity, multiplied times 75%.
- Lines 9-16. These lines include equipment that is counted at 100%, e.g. all of this equipment is assumed to be on at the same time. Cooking equipment, clothes dryers, heating/cooling equipment and EV chargers fall in this category.
The NEC load calculations estimates are volt-amperes. In this analysis a power factor of 1.0 is assumed such that Watts and volt-amperes are equal. The estimated service loads for Units #1, #3 and #4 are all in the range of 25.8 kW, which would require about 107 amperes, however, this does not include the EV chargers planned for future installation. Unit #2 is larger and has more equipment, including an EV charger. The estimated service load for this apartment is 49.3 kW which would draw 205 amperes. Units #1, #3 and #4 have 125 ampere breaker boxes and Unit #2 has a 200 ampere breaker box.
Table 2 – National Electric Code Load Calculations
These calculations are based on NEC Standard Electrical Load Calculations (Rev 3/23/2016)
Monitored Data
After completing the electrification project previously described in the “saga”, energy monitors were installed at the main service panel for each of the apartments and the house. Unit #2 has a subpanel that feeds the majority of circuits in that unit, so an additional energy monitor was installed in this subpanel and nested under the monitor at the main service panel. Emporia Vue 3 monitors were used in all cases. Each monitor is capable of recording the total load as well as the load of major appliances and/or equipment. The monitors record energy consumption, power and a ratchet keeps track of the 15-minute monthly peak demand. Table 3. Has 15-minute peak demand for the month of November 2025.
Table 3 – 15-Minute Peak Demand for Typical Residential Appliances and Service Panel Total
Summary
As noted previously the NEC estimates of electrical load are three to six times greater than the monitored data. Granted, the monitoring results are just for a single month, but except for heating, little seasonal variation is expected. Data will be analyzed for future months to verify the findings. Also, it is worth noting that the NEC load calculations are intended to provide safe electrical systems. As result, they tend to be conservative. However, some of the assumptions seem a bit extreme. Here are some examples:
- Water heaters are assumed to draw 4500 W, but the monitoring data indicates 394 W for HPWH.
- Electric stoves are assumed to draw 8000 W, but the monitoring data shows an average of 3387 W for induction ranges.
- Clothes dryers are assumed to draw 5000 W, but the monitoring data shows an average of 894 W for heat pump clothes dryers.
- Clothes washers are assumed to draw 1500 W, but the monitoring data shows 678 W.
- Lights and plugs are assumed to draw 3.0 W/ft2, but monitoring data indicates considerably less power.
In Table 3, the difference between “Sum of Above” and “Service Panel Demand” accounts for diversity and the realization that not everything is likely to be on at the same time.