Inclusive utility investment programs such as Pay As You Save® (PAYS®) face a steep challenge of ensuring that all participants will enjoy net bill cost savings after receiving energy efficiency services. This summer, Stephen Bickel, with Ethan Goldman of Resilient Edge, and John Theurer of Zither Labs, co-authored the paper “Towards residential upgrade savings guarantees: An AMI-based diagnostic interface,” which was included in the European Council for an Energy Efficient Economy’s (eceee) peer-reviewed Summer Study proceedings. This paper explores the diagnostic tool used to analyze hourly electric AMI (advanced metering infrastructure) data to automatically flag projects for review and help identify what types of issues could be impacting performance.
The PAYS system is a model for offering energy efficiency upgrades to customers by removing barriers which typically prevent people from participating in traditional upgrades. By using a voluntary tariff, utilities can provide energy efficiency upgrades and recover their costs without requiring the resident to provide steep upfront costs or have prime credit scores and moderate to high income. Under PAYS annual cost of the tariffed charges to the customer cannot exceed 80% of the annual estimated savings.Typically, the customer enjoys reduced annual energy costs, but PAYS programs do not offer a savings guarantee.
There are five primary reasons that expected savings may not be fully realized:
- Changes in usage behavior
- Installation of appliances
- Overestimated savings
- Upgrades were not properly installed
- Unrelated equipment failure.
In order for inclusive utility investment programs to offer savings guarantees, utilities will need to find ways to manage these risks to deliver on such promises.
One electric cooperative is using a diagnostic tool built on top of the open standard calTRACK method for calculating weather normalized energy use using meter data to provide physically meaningful model outputs.
Metrics used by the diagnostic charts include heating balance point, cooling balance point, heating degree-days, cooling degree-days, heating demand, cooling demand, base load, r-squared (indication of how well the model fits the year of training data), and CVRMSE (coefficient of root mean squared error, an indication of how much error exists between the model and the actual energy use).
The diagnostic system can identify early signs of potential deficiencies, determine if a fix is needed, and verify that any such issues are effectively remediated. It is a valuable tool in scaling up whole-home efficiency projects. Its measurements can help promote one of the key value propositions for inclusive utility investment programs – that the participants’ annual bills do not go up, unless the occupants’ lifestyle changes in a way that necessitates increased spending
We recommend that programs should use standard, transparent tools such as these to monitor the success of every project. Even if programs determine that an explicit performance guarantee is impractical or unwise, the evaluative process described in the paper can help with customer satisfaction and identify systemic problems that prevent programs from meeting their goals or maximizing climate impact.
These tools will likely need to be supplemented by automated calculations, visualizations, and decision rules with phone interviews and on occasion on-site investigations to resolve ambiguities. The results of such investigations can be compiled into a knowledge base of ground-truth examples that will help improve the accuracy and saliency of tools in the future.
