A friend complained that their single-family residence electric bill had gotten much higher in the past year. Since we have the same electricity provider, I looked back at my utility records for the past 40 years. Whereas electric utilities may advertise low energy rates for which you need to read the fine print to see what is really included, I prefer to take the total dollars on the monthly bill and divide it by my monthly electrical energy consumption. Simply, dollars divided by kilowatt hours (KWh).
Figure 1 shows the yearly energy rate since 1980. The rates climbed significantly 15 years ago when “deregulation” became “reregulation” and the reduced rates became burdened with other charges. But in the past year, cost per kilowatt-hour only rose about 1 cent. That meant the customer’s bill increase was either caused by a significant uptick in energy usage, or the electric revenue meter was reading incorrectly. Since the customer didn’t have ready access to their past bills, I placed a power monitor on their electric service to see if the cause could be identified.
I planned on monitoring for a week to compare kilowatt-hour numbers between the utility readings and the power monitor. If the numbers agree over a week, the odds are they would agree over a month. The power monitor was placed at the breaker panel and measured the two voltage conductors to the neutral conductor and the currents in each voltage conductor. Since I was hedging that the problem wasn’t with the revenue meter, I set up the power monitor in energy audit mode, which is similar to a power quality audit with a slight twist.
When monitoring for power quality, the variations in the voltage are usually the areas of concern, whether it is rms variations (sags or swells), transients, harmonics, or voltage fluctuations that result in light flicker. The changes in the voltage waveforms and rms plots are compared against the list of usual suspects to try to determine what caused the disturbance.
In this application, the variations in the current were the main triggers. While it is possible to use the slight change in the voltage that results when loads turn on and off, the current signature of various loads is more likely to be the way to identify which loads are turning on or off at various times in the day. Baseline data showed that about 500 watts were consumed continuously. One of the phases had a familiar current waveform from PQ investigations, shown in Figure 2. It is one often found in commercial facilities where there is a significant amount of IT or electronic equipment. The pure sine wave of a resistive load has an extra bump on the waveform at the positive and negative peaks. This is the result of the superposition of load current of rectified input-switching power supplies, where the current is only drawn in the middle of the waveform at the peaks and when the input voltage is larger in magnitude than the DC link capacitor voltage.
Even in a residence, equipment in “sleep mode” can still draw current with this wave shape. TVs, printers, computers, routers, LCD monitors, etc., usually have such a waveform associated with them in that mode. Other equipment in the residence will have a different waveform and have difference in their cycles of its operation. A microwave, washing machine, air conditioner and dryer have their own unique waveform or current signature. A hair dryer will have a similar signature as a toaster, nearly a purely resistive load with negligible phase shift between the voltage and the current.
The periodicity of the changes in rms current amplitude and wave shape are also clues as to what the equipment is. The cycle time of air conditioners is dependent on ambient temperature, so it is a good practice to record or have that information available when looking at the data. Similar patterns can occur with refrigerators, though if the house is air conditioned at a steady temperature, the pattern should not vary much except during times when it is opened frequently for cooking.
Being able to correlate the data to the appliance or equipment can allow for a comparison against what the specs say for the equipment. As equipment ages, it often draws more current than the rated or specified numbers. A possible clue in this residence is that the bedroom air conditioning units have recently been known to trip the breaker, indicating higher current draw than in the past.
Next month, I will share the signatures of various appliances, and show the offenders that have resulted in the higher energy consumption. Or, it may be that the meter is reading wrong. Time, and the data, will tell.