Through "Technology Lead Design" our team built several prototypes to test our hypotheses to support the idea of a distributed battery network. We quickly learned the disadvantages of targeting consumers and shifted our focus towards commercial fleets.
In conclusion our prototypes received with great positivity from our corporate members. Our evidence appears to support the business viability and technical feasibility of creating a distributed battery network to arbitrage energy between a utility provider and commercial vehicle fleet. Results and requirements for this to be possible are explained below.
1. Problem Statement (HMW)
"How might we" predict the available capacity in vehicles?
e.g. Battery, fuel, and/or energy availability.
What we built
We built a machine learning model that can predict the location and energy utilization of a vehicle. We took commuting data and tried to identify patterns that we might learn from. We learned that most people have very predictable schedules weekly, monthly, and annually.
2. Problem Statement (HMW)
"How might we" measure a vehicle's current utilization?
e.g. Duty Cycle, Idle, Dormant, and Transit Time
What we built
We built an iOS App that can record location data feeding the predictive model. We also used MileIQ to track mobility.
3. Problem Statement (HMW)
"How might we" convince a fleet manager to electrify their fleet?
e.g. What would they have to know make the switch?
What we built
We designed a fleet manager dashboard & predictive arbitrage calculator. We created UI/UX mockups for fleet managers that would measure current fleet use and show potential savings by switching to EV's.
4. Problem Statement (HMW)
"How might we" visualize the available capacity with real data in realtime?
e.g. If we could visualize an entire bus fleet could we measure it's available capacity and downtime?
What we built
We built a V2 of our React web app + V2 of the prediction model. We're about to collect user feedback and learn if this format would be accessible for future users.
5. Problem Statement (HMW)
"How might we" display available storage capacity to the utility?
e.g. Storage Visualizer, Energy Portal, and Energy Marketplace
What we built
Utility Dashboard / Predictive arbitrage calculator product. By creating a dashboard that displays the available capacity to the utility they may be proactive in energy generation.
A system to learn and predict when an electric vehicle owner will be using their electric vehicle.
With commuting data from CoLab and ourselves we're able to learn the following:
HMW measure current vehicle utilization?
Data was collected via onboard device GPS and feed to firebase (Database) which we'd later used to create an api. We recorded the users: starting point, ending point, timestamp, velocity threshold that limits recording only when moving faster then a walk.
Interviewed utility partners and SF municipal fleet manager
Utility partners and Fleet managers agreed this would be a huge determining factors if we could use their data and show the viability of our distributed network.
By treating EV’s as a battery we're able to optimize supply and demand, reducing waste / by turning EV’s into a giant battery.
By allowing fleet managers to have a dashboard view of their fleet vehicles they could take advantage of energy arbitrage selling power back to the grid during peak hours.
Bidirectional alternator infrastructure will be to become as availability to allow EV's to pull and push power back to the grid.
On average 24,000 EV’s discharging 50 kWhs simultaneously would take a 100 MW power plant offline. For utility companies to benefit this system must compensate for at least an entire power plant.
Today there are just over 500,000 EV's on the road in the USA, and sales half roughly increased by 25% from last year. Growing EV sales are vital to motivate utilities to install charging infrastructure.
