OK, apologies to Philip K Dick fans for the title... My 'need more love' comment refers to the fact that it is often said these days that EVs 'need connected services' in a way that conventional vehicles maybe don't. I thought it might be interesting to explore that hypothesis a little bit.
So to get started, I suppose the Dick reference (OK, that just got my blog blocked by any of you using family safe browsers) is more appropriate than it might first seem. While YOU might be busy dreaming of electric sheep (whatever floats your boat) it's more likely that your EV will be slightly busy charging itself, plugged securely into a nice high voltage supply in your garage. As an aside, it might arguably be more useful to be plugged into a lower voltage, higher amperage supply, but that's not the way the interconnect standards are written and from a safety and convenience point of view high current connections are sub-optimal, so the die is already cast there, I believe.
But there is a fundamental question to be answered there. Is it OK that everyone plugs in and trickle charges as soon as they get home, once EVs, EREVs (extended range electric vehicles) and PHEVs (plug in hybrid electric vehicles) start being sold in volume? If you think about that in US terms for a moment, which state is likely to have the highest penetration for the foreseeable future of these vehicles? Full marks for the answer 'California'. And which state of the union has some of the biggest power capacity and distribution challenges, especially in summer months? Right again.
So if nothing else, the means of being able to variably schedule charge cycles so the load can be deferred to non-peak periods and potentially spread across the population of vehicles is fairly important. This is actually not that hard to achieve, but it does require a service that provides a logical interconnection between utility companies and EVs. In an ideal world, all EV charge cycles would fit within a small number of hours to be easily schedulable in this spread model and still be available when their users need them in the morning.
Even though a simple charge scheduling service would help, the model breaks down pretty quickly. How is this enforced - what's to stop users turning 'renegade' for their own convenience and mobility security ('damn everyone else, I want to ensure I have a full battery at 6am')? What is my battery is sufficiently large, or my in-home charger infrastructure so simple (eg. just a straight 110v connection) that I need 12 or 13 hours to charge my battery? What if things become critical despite all our best predictions and the grid needs to disconnect some users to save itself?
Fairly quickly I believe we arrive at a scenario where unmitigated charging is unlikely to be very acceptable for very long. In other words, an EV with a simple adapter cord plugged directly into a standard mains voltage socket. There are two areas of rationale for this - one being the requirements of the power company to manage these charging events, the other being the convenience of the user, who will probably want more than 110v in most cases quite soon, and can probably be sold on the benefits of spreading the charge load over time, perhaps purely on the notion that taking 'surplus' generating capacity at non-peak times (or at least removing the need to throttle down generation where it can be done, which still carries expense) is good for the planet, as it removes stress from peak times and avoids the need for additional capacity through improved utilization.
If that kind of altruism isn't sufficient to carry the day - and it probably won't be (excuse my skepticism) then it's relatively easy to start incentivising users to behave in prescribed ways. For example, offering a home electricity rate discount for people that subscribe to mitigated charging vs. unmitigated - in some ways this is not dissimilar to the interruptable power supply tariff available in certain states for running air conditioning, or the overnight tariffs available in some countries originally for running storage heating as well as domestic appliances on a timer switch. The mitigation device - charging station in the garage - could easily police appropriate behaviour by reporting when the car was connected to it. Between the device, the vehicle, and the home meter, there are many levels of sophistication that could be applied to detect 'rogue' charging, even though I am sure there are many creative people who would try to find ways around those techniques.
Logically, though, we are probably on a roadmap towards a domestic electricity market. This already exists for grid participants and large corporate users - and technology certainly provides what we need to enable extension to consumers. How might this work? In the case of home overnight charging, to stick with that pattern for a while, it's likely that the rate you would pay:
- might vary from one night to the next depending on demand (which as we all know depends on multiple other factors - weather conditions, available capacity, extraordinary events that increase consumption eg. a superbowl, cup final, grand prix or perhaps these days the final of some televised 'talent' show)
- would vary depending on when you are asking for power (more expensive in early evening when people are cooking, less expensive in the small hours)
- would probably vary according to the rate at which you want to draw power from the grid (higher rates would attract a premium except in the case where the utility finds itself with a surplus to sell for some period of time - in some cases they might want to incent high charge at times of very low usage to clear the grid of chargers for the morning peak)
It's almost certainly the case that this application of market forces is the only way to incent the kind of behaviour and coordination between participants that is needed to optimize this ecosystem (to justify that statement further would require a sociopolitical discussion, so I won't). At this stage our simple service - which up to now just scheduled charging, monitored usage, and cut the connection to the grid in extremis - just got rather more complicated, although the principles remain the same.
Now the user reaches home and connects the car to the charger unit. At some stage - whether on a per event basis or more generically via the application of business rules and thresholds - the consumer makes some decisions and in some cases tradeoffs between what he or she is willing to pay and the desire to have a fully charged battery at some time. Certain strategies might carry more risk than others - in general, the more the charge window is deferred and the longer it takes the more likely that you will not have a full battery in the morning even if notionally you should. Equally high charge rates on a hot day may not be achieveable in practice due to battery pack overheating and the need to back off the charge rate commensurately. The consumer can interact with this information and this strategy on a number of devices - the vehicle screen, the charger unit itself, on a smartphone, on a PC, or on an intelligent TV screen (what we sometimes refer to as the '10 foot display' like Windows Media Center or Xbox).
Once you drive away from home, the same energy market serves your needs in enabling a diverse set of charging solutions. From slow/trickle charge points built into public and private car parking spaces, to 'charge while you wait' facilities built around natural gathering points such as branches of Starbucks or popular shopping destinations, to charge facilities added to existing filling station real estate - with a probable bias towards fast charge systems to the extent that the can be capacitized (bad pun there for geeks and EEs) - a range of solutions are likely to become available, all ideally enabled by the same marketplace systems, and findable / reservable by connected services as appropriate.
Reservable?, I hear you say... until all parking places are equipped with charge points, it's likely that smart car park operators will allow EVs to reserve the spaces that actually have this facility, rather than them being 'wasted' by a conventional vehicle. Additionally, it's highly likely that even when battery packs emerge that can handle fast charge, that public capacity to use this capability is likely to be severely limited. Even a relatively modest 1.6kWH battery pack (as fitted to the GM Volt, providing 40m plug in range from 70% of its capacity) would pull about 8000W to charge in 15 minutes. If you assume a 415v connection (about as high as you could practically go) that's 20 amps. Per vehicle. Getting that power from the grid in real time in a distributed model is a long way from being trivial. So to satisfy fast charge, local storage is likely to be required - the equivalent of today's underground fuel tanks, perhaps. That means limited capacity over time, which probably means if you want to guarantee getting that charge when you need it you might want some kind of reservation system (which would, I am sure, be a premium service).
(As an aside, it's a pretty scary thought that those batteries/ultracapacitors might sit right next to the legacy underground fuel tanks. That said, it's got to be better than the Formula 1 KERS implementations that actually put the battery pack IN the fuel tank...)
By the way, the emergence of this domestic energy market is not driven purely by, or to serve, vehicles. Quite the contrary. The increased trend towards having your own generating capability in the search for greener power - whether this be solar, wind, or thermal - is likely to substantially change the energy ecosystem away from a centralized to a decentralized generation model. Don't take my word for it though, our counterparts that look after the utility industry (stand up Jon and Larry!)and the experts in their industry that they talk to say so. And of course there is a bit of interaction here - in that you might want 'your own' power to be used for your own purposes before returning any surplus to the grid. This is clearly more efficient, although if the grid is paying a good price for power today, you might choose to defer some of your own requirements in order to help the community.... or more the point make money.
So now we need a set of services that handle the continuous variability of rates and the contractual context and billing mechanisms that go along with it. As well as access to reasonably sophisticated scheduling algorithms and marketplace systems that offer and sell a certain amount of power in pre-booked blocks from some timeframe ahead of consumption right down to real time. Some of these are much more in the domain of the utility industry to figure out (like I say, it's not just about EVs), but many of these services need at least an EV-related binding, and EVs being charged will be one of the larger and more persistent energy consumers in many households.
Which brings us to the next point, which has something of a 'holy grail' nature to it in the automotive - utilities intersection. VTG - or 'vehicle-to-grid' capabilities offer another level of interaction between energy producers (or marketplaces) and energy consumers. If we sell a few million vehicles all of which have a - in relative terms - thunking great power storage capability on them, how can this be used to help spread, or timeshift, generation from trough to peak load? In other words put in less generation capacity compared to demand, and do a better job of smoothing it over time. This has been a prohibitively expensive in a centralized generation model, but given that EVs need their battery packs anyway, it just became more viable as a concept. So now, EVs take a full charge overnight when power is relatively cheap and plentiful - whether they think they need it or not. The following day, those vehicles that don't require their full charge can return the power to the grid at a time when the grid most needs it.
Again, a marketplace concept is probably the only way to make this work effectively. I've already described how this applies to taking charge at night. In this model, perhaps rates would be set to incentivize full charging in those markets that require help in smoothing load in this way. The reverse model would apply during the day at peak load times... the utility would offer to buy back surplus power from EV users at rates that would vary according to demand. Just like with the mechanism for managing charging, the consumer can be as involved in this as he or she wants to be. Consumers can set thresholds and rules and have this all managed automatically - or can be notified at certain thresholds when they want to be involved in the decision making. The only requirement is that the vehicle is connected to a charging - or in this case discharging - station at the time the power is needed. Since it is expected that a number of lower charge rate charge points will appear in private and public car parks over the next few years (they already exist in some all around the world) this becomes increasingly more practical.
But wait a moment, I hear you say... I might buy into the notion that utility companies can do a good enough job of forecasting demand and capacity to be able to manage this well, but how does my car or the system know how much of the charge I need in my battery to get me through the rest of the day?
This brings us onto another class of more familiar automotive services - optional - and therefore relatively undersubscribed - for conventional vehicles, but in this model at least rather more critical for EVs. In order to enhance predictive abilities around charge requirements - or even effective range - EVs need improved navigation, traffic, and calendar integration capabilities.
Let's explore this a little further. Over time, it's not that hard for an EV to develop a profile of how I drive the vehicle under various circumstances. If we assume that privacy issues can be handled, merging this data with how this type of vehicle performs in the real world as a class can only enhance the ability of the vehicle-cloud ecosystem to predict my range / energy consumption for certain types of driving. Now overlay first class real-time and predictive traffic, probably based on probe data from connected vehicles in the aggregate. Plus weather data that is also enhanced with local probe information, whether this be external temperature, rain sensors, or information from the traction/stability control systems.
Since EVs can be more sensitive to topography as well as traffic and how they are driven, the map information associated with my routes can be integrated into this model, and even used to plot routes that are optimzied for energy consumption as well or in combination with other characteristics.
Add to this, then, information obtained from my calendar about my plans for the day and the locations associated with them. Or even if this is not explicitly loaded, an estimated itinerary based on my movements in prior days, taking account of daily, weekly, monthly and seasonal patterns. For many users, there is a huge convenience factor in having a well defined calendar synchronized with my car in any case - since it means I may never need to type an address or even select a destination in my navigation system again. The car just knows where I am going - and if my plans change I can interact with the itinerary easily to get it right. In the vehicle, or via any other device available to me, since the information lives at least in part in the cloud.
So now the system has a pretty good idea of where I am going, when I am doing it, what traffic and weather are likely to be like, and how I will drive under those circumstances. It also knows the safety margins I like to have on range, the performance of my battery ('fresh' or well conditioned batteries require less of a safety margin), the availability of any 'fast charge' stations in the area I could potentially get stranded in later, and/or whether I have onboard generating facilities as a last resort (EREV - but notice I said last resort. The idea is NOT to return power to the grid only to cause fossil fuel consumption as a result but to operate the vehicles intelligently on plug in power as much as possible).
From all of this, it's possible to automagically determine whether I have surplus power to return to the grid. This can get quite sophisticated. It may be that the grid badly needs power between 2 and 3pm this afternoon, but that the crisis is over by 4pm. My vehicle might 'overprovide' on surplus power knowing that it can recouperate the power at cheaper rates between 4 and 5pm when I plan to leave, or should I leave early, I can stop at a fast charge station and still recharge at rates lower than I was paid for the surplus power.
More cautious users can opt out of these schemes, or reduce their risk by requiring personal authorization for transactions, or set more conservative thresholds in the business rules engine, but even a relatively low level of participation can probably make a positive contribution to the environment and the energy ecosystem, especially if coupled with green charging options at home such as solar, wind, or thermal.
It may not have escaped notice that a lot of the services we talked about apply equally well to non EVs, and that there are substantial gains in convenience and efficiency of vehicle operation to be gained in those scenarios as well. This is entirely true, and was the premise that drove this blog entry - is EV just a note in the margin for a connected car strategy?
My conclusion is a reluctant no. Reluctant, because I'd like to say that the compelling logic of what can be done with connected services and conventional cars has driven an explosion in the uptake of these services, if not with current vehicles, then with the next generations of such vehicles in the pipeline. From my viewpoint, however, this is not the case. Early providers of these services are still struggling to monetize their offerings, there is no industry clarity regarding the consumers willingness to pay for these capabilities, and automakers are being understandably conservative regarding driving adoption and integration of these capabilities. In other words, the conventional vehicle marketplace has yet to prove that it needs these capabilities.
The EV marketplace does need these capabilities, particularly if we want EVs to deliver the benefits to the environment and energy policy that we are all hoping for. The EV divisions and teams of those same automakers we just referenced realize this, and are looking for smart ways of enabling these high capability connected ecosystems to come together, or at least to enable the first steps in the journey I outlined above. EVs will hit the streets and be bought by consumers whether or not de facto or de jure standards are in place, and with or without the planning and cooperation of their utility companies. Coordinated effort is needed by utilities, automakers, and technology providers to ensure that a coherent ecosystem can rapidly emerge to service the needs of its clients/participants.
Coda
That seemed like a great concluding paragraph, but I'm not quite done yet. That line of reasoning was a natural way to hopefully prove that EVs really are going to need at least a basic set of enabling services, coupled with other ecosystems, to survive and thrive in the way their makers and purchasers want them to. However there are yet other reasons why EVs are a more 'natural' market as connected vehicles, and for completeness I will at least mention these in turn below.
At least at introduction and for the foreseeable future, all EV variants command a premium price over their conventional cousins. In some cases this can be as high as a factor of 2. By dint of nothing other than simple math, this increases the margin available to integrate additional technology and capabilities to enable the delivery of advanced user experiences and connected services.
Additionally these platforms are in most cases all- or largely- new. New platforms, unencumbered by evolutionary design and sourcing practices unlock the ability to apply core innovation to vehicle lines. Electrical architectures in general are likely to be much more integrated and software-driven. This lowers the barriers to entry in terms of providing more connected and integrated experiences and capabilities in the vehicle.
And more specifically, all electric vehicle designs that I have seen sport a reasonably advanced head unit that integrates several vehicle functions and provides monitoring capabilities for the vehicle drivetrain. Why is this the case, when we don't typically do that for conventional vehicles, with the exception of a few high performance and specialist vehicles? Partly this is I think a function of custom and practice. Toyota set the standard for this with the first generation Prius, and its continuous display highlighting power management in action. It's notable that new and old Prius drivers alike seem to prefer this screen to the other pages on offer as a default. I don't think it's unfair to suggest that a new idea is rapidly propogated across the industry, and therefore most other products in market or in pipeline in this class have their own version of 'that' screen.
Secondly, building trust in the technology and making the driver feel secure is a consideration here, and conceivably more practical considerations. Dashboards up to the 1970s had an array of minor gauges and dials showing water and oil temperature, oil pressure, alternator volts, battery charge, and in many cases other operating characteristics of the vehicle. Why? To some degree reliability was not what we are used to these days and the ability of the driver to monitor for potential incipient failure was useful, but to some degree this was a function of custom and practice and what the driver thought he needed to keep an eye on, even if many of the gauges barely moved from nominal. From that time onwards - enabled by evolving technology - many gauges were replaced with warning lights or ultimately nothing at all, evolving to todays message consoles and self-checking systems.
While vehicles still have limited range, or potentially limited range depending on how they are driven, or limited range without burning fuel, the need to keep the driver very aware of those circumstances is clearly very real. Smarter systems will compensate for this need over time - as described above - but until these are firmly established, the driver may well need to keep an eye on the trendline of energy consumption in order to plan his journey accordingly. We don't want to see too many EVs stranded at the side of the road! Apart from anything else, roadside assistance with a spare fuel tank can get me on my way with a conventional vehicle in 5 minutes or so (once they reach me). This is not true of a depleted first generation EV, where sufficient charge to get to the next fast charge station might take quite a while to instantiate.
Finally and slightly more intangibly, EVs may be ushering in a new and alternative sense of value into the automotive market. The compelling drivers for vehicle purchase other than practical aspects - those considerations that make buying a car an emotional purchase that defines your lifestyle and not just an appliance - have mostly been performance and appearance (often itself an analog of performance - 'it looks as if it is fast' (even if it isn't)) for the past few decades - at least since the late 1950s. EVs are not optimized for this! They are about being ecologically responsible and low running costs, albeit with a higher investment cost. That mindset seeks alternative differentators to the classic ones, and with a high technology ethos being inherent in the EV marketplace, more digital lifestyle in the car, advanced user experiences, and connected services fit this requirement pretty well. Especially when they can help justify the notionally pragmatic outlook of the EV buyer by making the vehicle itself more practical, or by allowing an EREV buyer to run the car on plug-in power most of the time. It's easy to see how these can rapidly turn into 'musts' for this platform in a way that they haven't - as yet - for most internal combustion vehicles.
There is one more aspect of electric vehicles and their affinity with connected services that I should mention. EVs are the platform of choice - at least conceptually if not necessarily practically - for intelligent city programs and their associated shared-use and dedicated-use fleets. This is a huge and hot topic in and of itself, and therefore will be left for now until a future blog entry.
That's all I have - this may not be an exhaustive list or set of arguments, and I'm interested in any I have missed. In general, i'm certainly interested in comment or debate on this topic, and don't claim in any way to have all the answers, so if you have an opinion or wish to take issue with what I've said, please either comment away or send me an email.
By the way, I want to name check a few people who - knowingly or otherwise - have contributed over many discussions/years to the ideas that have been articulated in this blog entry. I defininitely do not claim to be the sole progenitor of any, or all of them. Contributors include Larry Cochrane, Drew Gude, Norbert Braendli, Francois Richard, Manuel Simas, Tom Phillips, Martin Hall, Mark McNulty, Ed Muth, James Hutchinson, Steve Lewis, Miguel Rodrigues, David White, Chad Stayton, Trevor Stayton, and last-but-definitely-not-least Spyros Sakellariadis.
John
