De-carbonization of the electric grid, particularly with renewable energy sources such as wind and solar, requires shifting demand where possible to periods when energy is plentiful and cheap. Electric Vehicle (EV) charging lends itself well to “control”, especially in environments such as homes, offices and malls – unlike at supercharger stations where you just want to fill up fast and be on your way. Developing the concept and code for such control in the open allows for cross industry collaboration and the evolution of standards through code-first. With this mindset, engineers from VMware and Camus Energy, a start-up in the energy space, launched the open source project Kinney. While Kinney is still in its infancy, the plan is to factor conditions such as weather, upstream grid load, a vehicle’s typical daily travel needs and available time to charge to intelligently control charging. Kinney has the potential to be useful in a variety of Internet of Things (IoT) situations that support the sustainable generation and use of energy and, as such, can help us understand how we might achieve that goal.
VMware: sustainability and open source
VMware’s virtualization software makes data centers more energy efficient and we continue to innovate in that space. We also have a Microgrid proof-of-concept underway that will include two roof-top solar panel arrays, two large batteries, multiple Chargepoint EV charging stations, a smart office building and a small onsite data center. Our goals are to gain knowledge on how to operate as an electric island, to provide, if necessary, power to the city of Palo Alto in the case of grid load shedding (recent California heat waves and wildfires stressing the urgency of such capabilities), in addition to sharing our testbed with academic research partners such as Stanford and Vanderbilt.
Further, VMware recognizes the value of open source in developing industry standards through code and fostering cross-industry collaboration. Across VMware, we are making contributions to multiple open source projects, including Kubernetes, The Linux Kernel, EdgeX Foundry and the Spring framework.
Building Project Kinney
VMware has access to multiple ChargePoint charge stations. We needed, however, to upgrade our license to access monitoring and management capabilities such as reading and shedding load (Chargepoint API). Soon after we upgraded our license, we were “talking” SOAP and had a Python implementation to programmatically start and stop vehicle charging. Camus Energy developed Go support, a tedious manual task because XML namespace handling is not as mature in Go as in Python.
By mid-March 2020 we had collected three solid days worth of charge data when the COVID-19 lockdown went into effect. With all of us working from home, the on-campus EV charging flatlined!
While historical data is great for developing insights and auxiliary code, it does burst the bubble on cool demos and engineering trade-off algorithms. A month into the situation, we “adapted” by focusing our efforts on developing a simulator – station and vehicle.
More recently, the Camus team successfully demo-ed their EV charge controller to a potential client.
Our next step is to integrate EV monitoring and management functionality into the VMware Microgrid proof-of-concept. We will be using the EdgeX Foundry framework to collect data, analyze locally and effect control operations. We will also develop algorithms to glean vehicle driving insights to more effectively balance energy sharing in office environments and demand shifting in homes, to better meet energy availability. We shall be presenting on this at Open Source Summit Europe 2020.
Help us push forward microgrids and demand shifting
Kinney currently only supports ChargePoint charge stations. Might you have access to and be interested in developing support for charge stations from other vendors?
Might you be interested in defining and registering a common way to determine what a vehicle needs in terms of minimum charge? This is valuable regardless of where you are charging, at home or the office or at the mall. Under grid constraints, this would be a way for the conscientious customer to provide insights to the system to more effectively distribute energy.
Together we could build a common control plane – one that will prove more and more valuable as renewable energy sources proliferate and EV adoption rises.
A little further down, broader in scope, we would like to associate with each load its context and importance. For example a light bulb in a stairwell is essential when someone is using the stairs. Heating and cooling are critical on extreme temperature days. Using such criteria we can more intelligently control energy in smart building environments. This is analogous to scheduling compute workloads in data centers where batch workloads such a report generation can run in the background on a best effort basis.
If you’d like to join a sustainability project in its early stages, this is your chance! Explore the code base at Kinney Github and let us know how you’d like to help.