Sergey Nivens - Fotolia
When temperatures go up, people blast their air conditioners, and when people blast their air conditioners, the electrical grid struggles.
In order to address this persistent problem, the City of Palo Alto Utilities (CPAU) department launched a customer behavior analytics program to help it anticipate demand and manage excessive use. This has helped the department deliver a more stable supply of power to the many large industrial users in the municipality.
"The point of the program is to get a reliable supply," said Karla Dailey, senior resource planner at CPAU.
Since putting the program in place, CPAU has been able to reduce power consumption during extreme heat events by an average of 1.2 megawatts, which could mean the difference between keeping air conditioners going and experiencing a loss of power.
The utility first partnered with cloud-based software and services vendor AutoGrid in 2011 to develop its Demand Response Optimization and Management System, an opt-in program where participants agree to have their energy consumption throttled back by the utility when demand is expected to be high. In return, participants receive credits for meeting pre-set targets for power use reduction. Participants include SAP's Palo Alto office, Xerox subsidiary PARC and the local Veteran's Hospital.
The program works by the utility feeding an AutoGrid database with data collected from meters throughout the city. AutoGrid then runs predictive models against this data to determine when the grid is likely to be taxed beyond capacity. CPAU can then limit the flow of power to participants. The limit is typically so small that it would just require the user to put their air conditioner from 68 degrees to 70, for example. But this small reduction across a broad participant base allows the utility to guarantee a smooth supply to all its users, avoiding brown-outs.
Public utilities generally cover large geographic regions, but CPAU is unique in that it serves only the city of Palo Alto. Being smaller means it has fewer resources to apply to a large-scale tech project like the Demand Response program. Dailey said this was the main reason they chose to go with a managed service analytics provider rather than do all the work in-house.
There have been some challenges. For example, midway through the summer, a few participants wanted to change their power reduction targets. This change altered all of the back-end analytics that determine how much power consumption users must cut back during a heat wave. If the program were run entirely in-house, this change would have been relatively easy, but Dailey said it required some back and forth with their vendor to get the changes made properly.
Still, given the limited resources of the utility, Dailey said it makes sense to go with a managed service. It made for a fast implementation -- deployment took less than 30 days -- and opens the door to rapid scaling should CPAU find more users who want to participate in the program. All of this would have been difficult to manage with only in-house resources.
"On balance, it's better to use something that was developed outside," Dailey said.
Next StepsFind out your customer behavior analytics project
Learn what data you need to do customer analytics
See how to develop a customer analytics strategy