Towards Scalable Electric Vehicle (EV) Charging Infrastructure: Think Outside the "Gas Station" Box

Disclaimer: The views expressed here and all articles written by the author on LinkedIN are personal, and does not purport to represent the views of his employer at the time.

This is part of a series of public articles on e-mobility: Electric Vehicles & Ride Sharing Economics, Commercial Fleet EVs - Stealth Revolution, Towards Ubiquitous, Affordable, Ultrafast Charging, Electric & Autonomous: Synergies for a TaaS future, Towards Scalable EV Charging Infrastructure (Think Outside the Gas Station Box), Towards Scalable EV Charging: Hidden Costs of Level 2 Charging at Scale.

Quick ! Will EV charging of the future resemble gas stations ? Or like the supercharger site in Kettleman, CA ?

When you speak to folks who are procuring larger commercial electric vehicle (EV) fleets (bus, truck, van ... what have you), one of the common refrains you will hear is that "procuring the EV is the easiest part... electrification infrastructure is our biggest challenge especially at scale". A couple of articles about it: here (Denver Post) and here (Wired). Why is scalable EV infrastructure a challenge? Before we get there, lets realize two things:

The first thing to realize is a point regarding the rates of battery charging relative to the size of batteries and the time it takes to fully charge any sized battery (which is 30-60 minutes even with fast charging vs 5 mins to fill up a gas tank). These fundamental battery related issues have been extensively discussed in my prior article. Note that these observations hold irrespective of the size of battery pack (small or large), and charging speed (fast or slow - its worse with slower charging)!

Because of these charging-time issues, there is a simple outcome of queuing theory: when service times are longer (30-60 mins EV fast charging vs 5 min gas refueling), you need more servers deployed (i.e. EV fast charging points vs gas stations) to allow convenience of top up and other long-dwell charging behaviors for a larger number of EVs on the road (especially once you start having tens or hundreds of millions of them). What this means is that we will need a larger number of EV charging locations compared to the number of gas station locations.

Note a corollary: if the service times (i.e. charging times) are longer (eg: slower Level 2 AC charging instead of fast Level 3+ DC charging), then the EV assets are under-utilized longer, and you need even more locations for the servers (i.e. EV charging stations) proportional to the ratio of service times (vs fast charging or gas stations), and more widespread. True that a vast number will be at homes or overnight charging spots. And this is OK for consumer charging where the asset is expected to idle overnight. But commercial fleets will always seek productivity and lower idle time of assets (eg: if you could rent out your Tesla overnight, or if you are a Uber fleet owner with EVs with multiple drivers in three shifts, you may also consider it!). Therefore, there will also be a lot of scope for EV ultrafast charging super-hubs (similar to Buc-ee's or Love's large travel stops) strategically located at metro locations (eg: airports, downtown short term parking garages etc) and/or large parts of retail parking lots (eg: malls etc) electrified.

The second thing (a subtle point that will become evident with the growth of EV commercial fleets, especially with autonomous features) to realize is that EV fleets will tend to be more heavily utilized than Internal Combustion Engine (ICE) vehicle fleets since their cost-per-mile economics are much superior in terms of fuel, maintenance and insurance costs (see prior article: here or here). Tony Seba points out that the shift to EV fleets will mean more EV-miles (zero emission miles) per vehicle and in aggregate driven even if the regular ICE fleet is not fully electrified. Recharging facilities will be a function of EV commercial fleet demand. For example, if the cost-per-mile to take a EV ride share approaches that of owning your own vehicle, then there will be a large shift towards EV-ride share miles; similarly logistics fleets will rapidly shift over to electrification of their fleets.

Sam Korus @ ARK Invest points out that Model 3 total cost of ownership (TCO) is better than Toyota Camry (on a 3 year basis; which will become better with EV ride sharing). You can use Model 3 as a proxy for Light Commercial Vehicles (LCVs) as well. In another older piece, they point that 2019 is a cross-over year for LCV (light commercial vehicle) TCO economics for EVs vs ICE vehicles. What this means is that commercial EV fleets in the future will tend to have lower dwell time to get recharged, and their need to keep their assets and drivers productive will need faster EV fast charging turnaround times than we commonly assume today.

For instance today we assume that last mile vans operate only in the day time - but even autonomous or semi-autonomous electric vans of the near future can potentially operate 24x7 or with shorter down times (eg: similar to how airlines minimize the downtime of their planes, and shorten turnaround time to keep them in the air). Similarly think of a Uber or Lyft EV fleet owner who will want to keep their vehicles in service for as long as possible since the value-per-mile earned is much higher than the cost-per-mile and this raw economics will force commercial fleet EV behaviors to drive up asset utilization.

So, as we scale up EVs, we will need FAR MORE EV charging infrastructure than we imagine for ICE vehicles (i.e. gas stations), and especially so, as these EVs are operated in commercial fleets and driven to maximize asset utilization and value. However, most EV charging infrastructure deployed today resembles a "gas station" model (like the first image above) today. Is this a tenable way forward? I think not.

Why Does the Gas Station Model of EV Charging Not Scale Up?

What exactly do we mean by a "gas station" model for Public or Private EV charging?

1. Typically this means that there are a "few" chargers per site or depot (4-6 or a few tens at the outlier sites).
2. Most of them are either slow level 2 (L2) AC chargers at 7-10 kW or a few fast Level 3 (L3) DC chargers (50+KW, dispensing DC power instead of AC power), which each look-and-feel like a dedicated gas pump at a gas station, complete with a dispenser, payment kiosk and dedicated power conversion equipment.

This "gas station" EV charging model is OK for a population of a smaller number of EVs on the road, and especially consumer EVs where they can be charged "at home" with a Level 2 charger. However it is not tenable to scale up for a world of tens of millions or hundreds of millions of EVs. Why?

First, since you have dedicated power conversion at each "EV charging pump" which is 70-90% of the cost of the "pump", when a specific "pump" is not being used, the power conversion investment cannot be leveraged to charge other vehicles. In other words, when a charging bay is idle, you have significant portion of capex that is sitting idle. In other words, the capacity utilization factor per-pump could be modest, and costly capital sitting idle with modest utilization factors is not good economics.

Second, in this "gas pump" model, as you scale up the number of EV charging pumps, you need to linearly scale up (1:1 basis) the costly power conversion as well, i.e. you do not have any significant economies-of-scale. We already saw earlier that at scale we will need a lot more EV charging pumps than gas refueling pumps.

Third, (and this is a subtle one), every commercial location has a electric grid connection with a local utility called a "connected load" which determines the site's commercial tariff class, and may be a constraint based upon available substation capacity of the local utility. The way a utility computes the "connected load" is to sum up all the loads based upon their RATED capacity scaled up by a safety markup, e.g. 125% to determine the total current disconnect capacity. In a gas pump like model of EV charging, this means summing up the rated capacity of the EVSE (EV supply equipment) AC capacity and comparing it with the connected load allowance at the site. For example, if you want to scale up a site to have 100 active L2 chargers (at 11 kW rating each), this means you need 1.1 MVA (Mega watt) of site level connected load capacity BEYOND all other commercial site uses. In fact due to NEC norms, this could be at least 25% higher, i.e. 1.375MVA. This could be a problem since utilities tend to bump up commercial sites to higher levels of tariffs at thresholds such as 250 kVA or 500 kVA. Alternatively, the utility may just have a local constraint and will not approve a higher connected load, even if the customer may be willing to pay for it, i.e. it may be a site-specific constraint to consider.

Fourth, commercial sites incur tariff plans that are a combination of power charges (i.e. $ / kW / month, also called "demand charges") and energy charges (i.e. $ / kWh / month). In addition they may also have monthly charges for power factor (which is a function of power quality of your power conversion equipment for EV charging) and total connected load. As consumers, we tend to pay only energy charges, but commercial sites may have power charges that are as high as (or higher than) energy charges, which can double their monthly bills. The way power charges are computed is by observing the worst-case 15 min energy consumption (in kWh) and divide that by 1/4 (15 min = 1/4 of an hour). In other words, the indiscretion of the "peak" or worst 15 min (compared to "average" power consumption) can cause your monthly bill to be significantly higher than it other wise would be. Peak-to-average ratios of power is often far greater than 2:1! A "gas station" model of EV charging means that several tens of EV chargers (and larger with scale) have to be closely coordinated via software schemes to minimize demand charges, and often not done in practice. Fleet turnaround times may pose additional constraints as well; and these constraints will evolve with the size of the fleet, battery pack sizes, transportation use cases and autonomous capabilities of the fleets.

Fifth, level 2 (L2) AC chargers not only have slow charging speeds, but impose a 10-15% efficiency penalty in both energy (kWh drawn) and power (kW drawn, and hence demand charges) compared to DC charging. While this penalty may be OK for a small site (eg: 5-10 chargers), for a 100 charger site at 11 kW per L2 charger, not only will it raise the connected load requirement to 1.375 MVA as discussed earlier, this power draw penalty can range from 110 - 165 kW, and since the power is drawn for a long period, the energy penalty applies over 6-8 hour periods per charger at this "leakage" rate.

In several markets in US and Europe, the price of power (demand charge, and incremental connected load charge) can range from $10-20/kW/month; and the energy charge can range from 10-25 c/kWh (and possibly higher in higher usage slabs, or based upon time-of-day, or season). If you just do the math on this, you realize that L2 chargers have a significant TCO (total-cost-of-ownership) penalty at scale that can surpass the additional capital cost of DC fast chargers within a few years.

Sixth, fleet owners who need their fleets to be productive for more hours of the day (rather than the 8 hours+ overnight charging model) or drive longer distances than a few tens of km will either have to invest in fast charging infrastructure, or will have to purchase MORE EVs to compensate for the EVs stuck in slow charging. Eg: think of a Uber fleet owner who invests in an EV - will necessarily have to idle their vehicle; or their drivers during charging periods -- they would prefer their fleets to be in operation 24x7. The same is true for port / airport vehicles that go electric; or in a number of other logistics use cases. In other words the replacement of EVs vs ICE fleet vehicles could be greater than 1:1. This additional fleet procurement especially when battery packs are costlier in the early phase of EV growth is a huge inefficiency which can be overcome with more economical fast charging depot models. As discussed earlier, the cost-per-mile economics and increasing levels of autonomy will drive up the demand for fleets to be more utilized.

Finally, power quality matters and is related to bidirectional power capability. The more the power electronics (eg: multiple cheap chargers, or on-board AC charging by vehicles) you put in that handles power in one direction, the more power quality issues it throws up (eg: poor power factor, harmonic pollution, voltage and current distortion etc). This can add up in destructive ways for the grid; and/or throw up penalties for the commercial owner / transit provider. Specifically, a utility may either require compliance to a target power factor with associated penalties/deterrent incentives; or charge for the power factor (eg: via KVA charge which captures both kW and kVar delivered by the grid to the enterprise). Either way, it directly translates into economic impact.

Hopefully these points above begin to articulate why scalable EV infrastructure is a challenge for commercial and fleet owners. They thought that they wanted to buy (electric) vehicles - now they have to deal with the headaches of a fueling station (capex, opex, operations etc) and thinking about specialized electrical infrastructure issues!

But happily, these are solvable problems. The key thing to do is to think outside the box, specifically the "Gas Station" box or mindset.

Scalable EV Charging Infrastructure: Time to Think Outside the "Gas Station" Box !

What does it mean to think outside the "Gas Station" box ?

First, as articulated in my prior article, given the minimum charging times for EVs and more EV chargers, we need to aim for more ubiquitous, affordable, ultra-fast EV charging vs a more limited few-pumps-per-site and few-sites model inspired by the "gas station" model of EV charging.

Second, the affordability point above means that we need to achieve economies of scale in capital expenses, even though we increase the number of EV charging locations or "dispensers" (or "hoses"). We simply cannot allow EV charging capital expenses to linearly increase with number of hoses.

Third, provisioning low speed and inefficient (L2) chargers not only means longer dwell times of charging per charger today for users, but these chargers may need to be ripped and replaced with faster DC charging in the future {not even counting the hidden costs of L2 charging mentioned above}. It is strategically better to provision cabling, trenching, wiring to allow peak-charging capacities to be 200A dc or above (i.e. 80-150kW-dc) one time rather than having to rip / replace in a few years at double the installation cost. This will be especially so with the growth of commercial EV fleet use cases that require the EV asset to be sweated as much as possible, in part, enabled by new autonomous technologies.

Fourth, the connected load constraint on the AC side needs to be decoupled from the peak-capacities provisioned on the DC side per hose/dispenser, and the number of dispensers we provision. This requires a shift away from power conversion inefficiently done in a decentralized fashion, to a more internet style "power routing" model where the aggregate AC power constraint can be distributed to individual hoses/dispensers dynamically, in a software-defined manner. This will also allow the system to evolve flexibly to support a range of fleet EV charging use cases despite short- or medium term constraints on connected load or substation capacity on the utility side. Another viewpoint we discussed in a prior article is that of "virtualization", i.e. virtualization of chargers.

Fifth, granular control of all the operational costs (demand charges per kW, or energy charges per-kWh, power factor costs (ideally unity power factor), and connected load charges). This means software defined control of AC power drawn on a granular level, and subsequent balancing of DC power delivered to individual hoses/dispensers. It also means flexible integration of local renewables and local energy storage modularly over time, and managing such resources to optimize the opex and total cost of ownership). Integration of renewables and accounting will also enable certifiable zero-carbon EV charging, which in turn aligns well with emissions mandates, and government incentives (eg: the Low Carbon Fuel Standards/LCFS program in California, or similar programs in Europe and elsewhere).

Sixth, we need to have excellent control over power quality, and ideally be able to provide grid ancilliary services on demand in real-time. This means sophisticated control of real power and reactive power elements independently, ability to deliver unity power factor (or tunable range of power factors on demand), strict control of harmonic distortion while delivering very high efficiency.

Seventh, the footprint of the power conversion and dispenser/hoses has to be tightly managed to minimize both the installation costs, and meet the site-specific constraints and aesthetics. The footprint is also affected by how the electrical equipment is integrated to enable low installation costs, and potential portability/relocation of the equipment. This aspect is very important for financiers, in addition to the long-term reliability and servicing costs (all part of the total cost of ownership equation).

Seventh, the software defined dynamic management capabilities can extend way beyond just cost-management to value- & yield-management. In other words: we need to go beyond total COST of ownership to maximizing the total VALUE of ownership. For example, on-site software can integrate with cloud-based software for fleet management & dynamic per-vehicle turnaround time & SOC (state of charge) requirements, dynamic retail pricing (eg: Uber-style surge prices), preferred customer loyalty schemes (Platinum / Gold / Silver priorities) in large scale workplace/retail parking, or multi-fleet subscriptions for private charging. It also extends to software-defined ways of capturing smart grid value streams (eg: two way power / energy flows including demand response, ancilliary services, Vehicle to Grid (V2G) participation etc). Some of these points have been also articulated by @azeezmohammed in his article "Charging Fleets, Not Individual Vehicles: Future of EV Charging".

Summary

As you can plainly see, this list above is far more than what a simple "gas station" or "gas pump" model of EV charging allows. Is this vision far away or intractible? Not really! Industrial companies have been delivering such super-flexible power conversion systems in other "industrial-class" markets, and we are bringing it to market today -- literally as you are reading this article. As commercial fleet vehicles start getting electrified, we foresee a significant demand for such "out-of-the-gas-station-box" thinking flexible platforms and solutions.

The future indeed is exciting, and may start happening faster than we think as the fleet of EVs (and subsequently with autonomy) starts increasing on the road in major markets.

LinkedIn: Shivkumar Kalyanaraman 

Disclaimer: The views expressed here are personal.

Twitter: @shivkuma_k

If you like this article, please check out these articles: Towards Scalable EV Charging: Hidden Costs of Level 2 Charging at Scale, "Towards Scalable EV Charging Infrastructure (Think Outside the Gas Station Box)", "Commercial Electric Vehicles (EV) Fleets: The Stealth Growth", "Towards Affordable, Ubiquitous, Ultra-Fast EV Charging: Part 1: Need & Battery Issues", "EV Taxi Fleets & Ride Sharing: Poised for Huge Growth", "Shared EV Transportation in India", "Understanding the Rs. 3/kWh bids in India in 2017", "Distributed / Rooftop Solar in India: A Gentle Introduction: Part 1","Rooftop Solar in India: Part 2 {Shadowing, Soiling, Diesel Offset}", "Rooftop Solar in India: Part 3: Policy Tools... Net Metering etc..." "Solar Economics 101: Introduction to LCOE and Grid Parity" , "Solar will get cheaper than coal power much faster than you think..", "Understanding Recent Solar Tariffs in India", "How Electric Scooters,... can spur adoption of Distributed Solar in India," "Solar + Ola! = Sola! ... The Coming Energy-Transportation Nexus in India", "UDAY: Quietly Disentangling India's Power Distribution Sector", "Understanding Solar Finance in India: Part 1", "Back to the Future: The Coming Internet of Energy Networks...", "Tesla Model 3: More than Yet-Another-Car: Ushering in the Energy-Transportation Nexus", "Understanding Solar Finance in India: Part 2 (Project Finance)", "Ola! e-Rickshaws: the dawn of electric mobility in India", "Understanding Solar Finance in India: Part 3 (Solar Business Models)" , "Meet Olli: Fusion of Autonomous Electric Transport, Watson IoT and 3D Printing".

All LinkedIn Articles/Posts.