The drive towards better EV DCFC charging reliability - part 2

I first covered this topic in May in the post: Heat is On - Reliability of DCFC stations where I covered this paper and this post will focus on one aspect which is: How do we define reliability. From the paper:

"From the EV driver perspective, a reliable EVSE is one that charges the EV, for the expected duration, after using an appropriate payment method, at the expected rate (i.e., kW)......However, the use of uptime as the reliability metric is controversial since there is no standard definition nor is there a standard calculation methodology.?Given the complexity of the EVSE ecosystem and technology stack, from hardware to software, ensuring a high uptime and assigning “uptime ownership” of each EVSE may be difficult and may require standardization across different jurisdictions."

In interviews with representatives from Charge Point Operators (CPO) like Electrify America and EVgo I hear them say all the right things to explain why reliability is hard. And it is hard. But I also hear a lot of "corporate speak" and in my opinion, it's clearly in the CPO's best interests if they obfuscate defining reliability because it gives them more leeway to downplay the issue and manipulate metrics to make it appear that actual reliability is higher than reality.

So let's start a process to develop a mutually agreeable metric that the industry follows and that government entities who will be handing out Billions of taxpayer dollars can utilize to hold CPOs accountable for maintaining reliability.

The famous Peter Drucker management quote (and used often in energy management) is

"You can't manage what you don't measure".

We can't improve reliability if we don't measure what the different failure modes are. And in order to measure the reasons for failure, we have to define the reasons for failure. From the image at the top of this post I have grouped them into 3 areas that I believe make the most sense:

1) Hardware/Infrastructure

2) Software/Firmware

3) Cell networking

Hardware/Infrastructure issues:

Simple Definition: Things you can touch that are physically broken and repaired by an electrician/technician who can use tools (no spares needed) or replace the equipment with spare parts.

In reading comments around stations that aren't working, commenters say something like "there must be a shortage of technicians" or I've seen CPOs blame lead times on spare parts. Both of those are kind of true, but fairly easily mitigated with a warehouse full of spare parts and service agreement with EV service providers like ABM or ChargerHelp . If good hardware is chosen and a robust Maintenance, Repair, Operations (MRO) plan is in place, this should not be a major impediment to ensuring high reliability.

Software/Firmware:

Simple Definition: Software bugs, errors or glitches located on the DCFC stations on site, in the network system in the cloud, or on the EV.

My personal belief is that this bucket represents most of the issues impacting charging reliability. It's also an issue that can be the hardest to fix once CPOs start rolling out more stations if an underlying software issue is not addressed. It gets patched and patched and bandaided and eventually cascades to a larger failure.

Unlike hardware/infrastructure, this bucket may not always be the fault of the CPO, as the vehicle OEM's software may be the responsible party for a reliability issue. This is ultimately why Tesla is so successful with Superchargers. Tesla controls the EV, the charging station and the network. They control it all so when there's a problem, they can easily address it. It's similar to an Apple laptop running Apple software vs. another laptop using Window's software.

As seen in the above image, there are a lot of points of failure and moving pieces controlled by multiple parties in this bucket. But because there are so many moving parts and software/firmware issues are so hard to pinpoint, it makes it that much more vital to define what the issues are and address them.

If a government entity (States in the case of NEVI) handing out hundreds of millions of dollars for DCFC is holding a CPO to a metric, that entity should be fully informed as to why reliability issues are occurring with no corporate talk obfuscating the issue.

Cell Networking:

Simple Definition: Loss of communication with the cell network service the DCFC installation.

This should be the easiest to solve. I think this one gets blamed a lot because it's an easy scapegoat. But in reality, if the DCFC installation is in an area with good cell coverage, this shouldn't be an issue. If it's in an area of bad cell coverage, the designers should plan to mitigate this during installation and commissioning. I have a hunch that by NOT defining reliability groups/buckets, it allows CPO to blame many software issues on cell networking.

This is the advantage of defining failure mechanisms so we can actually measure what we want to manage. Sunlight is the best disinfectant.

Closing:

Let's all be clear here. The goal of defining the different failure modes of DCFC is not to simply come up with a better metric. The goal is to improve the reliability! And you can't manage (improve reliability) if you aren't measuring (defining and tracking the actual causes/issues impacting reliability).