An energy provider approached SFL about disaggregating overall household energy consumption data into constituent appliance energy expenditures because the energy provider wished to expand its service offerings to customers. Disaggregating energy data simply means converting raw energy usage data into a list outlining when each individual appliance (i.e. a toaster, refrigerator, heater, etc) was turned on/off.
Disaggregating the energy data allows the energy provider to provide two main, additional services:
- Specific appliance energy consumption data per household
- Personalized, appliance-specific advice and a quantified savings amount based on household energy consumption
Disaggregating the data also provides the capability to build a recommender system to inform households of the savings that can be achieved.
Many companies across all industries ingest and use data that contains information that it is not practical for a human to extract. Using machine learning to extract this information allows companies to better understand customer behavior and has significant value.
SFL developed algorithms that trained combinatorial Hidden Markov Models (HMMs) known as combinatorial HMMs and Deep Learning (CNNs, RNNs) with the raw energy usage data provided (time-series data for households). This allowed for a model to be built that separated, per customer, the underlying behavior of appliances. The client leverage the results to identify faulty devices using too much power and design more targeted programs aimed at reducing energy usage.
SFL’s resulting model was able to successfully disaggregate the raw energy usage data and meet all of the energy provider’s expectations for the project.
Maximizing the value of existing data is always important for a company. Knowing more about customer behavior lets companies make more accurate and responsive decisions, and reach out to specific customers in targeted ways.
In this specific use case, the data was used to identify faulty devices using too much power and design more targeted programs aimed at reducing energy usage. It can provide further feedback to households to encourage cleaner appliances by stating expected savings and also targeted advertisements for new energy-efficient appliances.
This kind of analysis can be taken further with automated feedback, diagnostics, and repair servicing, that can be provided to the customer for greater visibility into the energy expenditure. On a larger scale, these types of data allow for more accurate modeling of energy usage across the country and can allow better energy generation and storage optimizations.