Knowing what contributes to a household’s energy bill has become more important than ever with rising energy costs, variable electricity rates, and general inflation. Many public utility companies provide customers with an approximate breakdown of how electricity is consumed based on common home profiles, such as size, location, etc.
You may have seen a graph like this on your bill or your utility’s website:
However, charts like this may not accurately reflect an individual household. For example, a new appliance such as a dehumidifier or a failing central air conditioning system may cause an energy bill to suddenly go up.
For most consumers, it is difficult to pinpoint what might be driving up their utility bills based on generic “bill contributors.”
Artificial Intelligence (AI) is becoming more widespread every day, including image recognition, speech and text processing, self-driving vehicles, and smart home devices. Fortunately for consumers, AI is now available to help people better understand how electricity is being used in their homes. Most utilities now provide smart meters where devices like Powerley’s Energy Bridge Hub and Powerlync™ Plug can read energy consumption in real time.
Every appliance or device generates a unique power pattern. The problem in most homes is several appliances are consuming electricity simultaneously, and the combined patterns look like noise. The process of sifting through the noisy energy signal generated by a household to find individual devices is called disaggregation. By using disaggregation, companies like Powerley can arm consumers with better information about how they are spending their precious energy dollars.
Disaggregation relies on a specific form of AI known as Machine Learning. Data scientists like the team within Powerley use data they know represents a specific pattern such as an air conditioner, a heat pump, a refrigerator, or many other common devices. These known power profiles are used to “train” systems that can recognize those devices in a home energy signal. The output of this machine learning model is something called a classifier that can give consumers a detailed breakdown of how much power different devices are consuming at a given point of time.
A typical home power consumption graph may look like this chart from Stanford University:
Disaggregation takes away the guesswork for consumers in determining where their electricity is going.
Variable electricity rates based on time of day increase the cost of electricity during periods of heavy use. Many utility companies across the country are using variable rates to manage demand and to encourage efficiency. Consumers under these rate plans will usually pay more for electricity during the day and less during the night or on weekends. To save money with these variable rate plans, consumers need to know how and when electricity is being used.
The following table is a sample of what a consumer might see with disaggregation and a variable rate plan:
|Off Peak Cost/KW
|Off Peak Cost
This chart shows consumers how they could make several changes to save money:
- Shift laundry to the evening to reduce costs associated with the water heater and washing machine
- Lower thermostat settings during the day when people are at work or school during the winter and raise them during the summer
- Ensure space heaters are only used when necessary
- Upgrade large appliances to more efficient models such as stoves and refrigerators
- Make use of energy efficient lighting such as LED bulbs
- Look for devices that are consuming energy while plugged in but not in operation, such as chargers, small appliances, etc.
Utility companies frequently provide incentives to help consumers with energy savings, such as rebates on new appliances, incentives for recycling older appliances, free LED lightbulbs, free or subsidized smart or programmable thermostats, etc.
Powerley sees disaggregation as a critical feature we currently provide to our customers to help them save money on their utility bills.
We have invested seven years’ worth of research and development toward disaggregation, and we are continuously improving and updating our AI/Machine learning models. While there is a great deal of hype around AI, energy disaggregation provides real value to utility customers.
CHIEF TECHNOLOGY OFFICER AT POWERLEY