While it’s nice to make good time, when in remote areas, sometimes any route will do.
For EV road trip planning in trippler, I first found the direct route, then chose chargers along the route. This approach can fail when there aren’t enough chargers close to the route, as in remote areas.

Charger hopping is alternative or complementary approach I wanted to explore, where we plan to “hop” from charger to charger within the vehicle’s range to travel from origin to destination. In an extreme cases, this might take us a long way from the direct route, but as long as each charger is in range of two others, or the origin or destination, we’ll complete the trip.
As well as remote areas, this could be useful if typical routes are unavailable (as was the case recently when bushfires closed the Indian Ocean Drive, the preferred route that has all the chargers), or as a mechanism for discovering new routes, when it’s more about the journey than destination!
Charger hopping may be used to planning around charger congestion or charger faults, but trippler is the better option here for now. True resilience draws on multiple strategies.
Data exploration
Let’s explore the OpenChargeMap (OCM) data. With 1127 of a reported 2079 chargers nationwide, I know it’s missing a number of chargers, particularly in the Northern Territory, but it’s what I have.
For charger hopping, I’m interested in which chargers are near other chargers. As expected, chargers are clustered around major population centres on the southern east and west coasts, and strung out like pearls along more remote highways.

While I’ll be working with spatial networks, I’m also interested in how the chargers are distributed within business networks (the operator identified in OCM data) as connectivity/coverage of their networks may be a selling point. Chargefox, the biggest network, appears most widely distributed, while others may service a particular region/state, though there is overlap with other providers. It’s great to see local businesses represented too.

Finally, I was just plain curious about Australia’s highest chargers by elevation. It turns out there are two higher than 1,000m in NSW, in Glen Innes in the Northern Tablelands and Medlow Bath in the Blue Mountains. The visualisation highlights our coastal population and the Great Dividing Range in the south-east.

Data preparation
Back to charger hopping. From each charger we want to know the distance to every other, in order to figure out the viable and best hops for any trip. Openrouteservice (ORS) matrix API gives this data, though we need to process chargers in chunks of fewer than 50 x 50 and reassemble the whole matrix.

Most of these charger pairs are separated by more than 1,000km (and up to 6,000km) so we won’t be hopping between these. However, using these distances, we can build a weighted graph that connects every pair of chargers within some range threshold.

From this graph of connections, we can get groups of chargers that are in mutual reach or “hoppable” across all of Australia – shown below for a range of 200km.

The longer the range, the more chargers will be connected. NB. Tasmania should be part of the main east cost group because, although the ferry route is 415km, there’s minimal driving between chargers at either end. Kangaroo Island chargers are included in the main group as the ferry route is less than 20km.
A hop, skip and jump away
Now, given start and end locations, we can determine if they are in the same “hoppable” group. If not, we need to find a solution to bridge between groups (such as trickle charging on the bridge route). If start and end are the same group, we can find the shortest distance path between them possible by hopping from charger to charger. Hopping from Ceduna SA to Cairns QLD in this manner shows a large detour from the direct route, but it gets us there!

This route was based on minimising the distance travelled. However, we could use the travel time instead (also provided by ORS matrix API). Given the vehicle’s capacity and the power of the charger at the end of each hop, we could we add a time penalty for the charge that would be required at the end of each leg, leading to different routes.

Practically, we’d have to weigh up a major detour such as above, which could involve an extra day (or more!) driving, against the option of trickle charging overnight (or more!) on the direct route.
Trickle charging suggestions are included in the trippler beta version, shown as charging above 100% at a stop before a long leg, though the actual charging location and time remains to be determined by the driver.

It’s also worth noting that we could in theory get a similar result to charger hopping from trippler as-is simply by widening the corridor around the route (currently limited to 5km) to any arbitrary amount, though some techincal limitations remain to be addressed.
Keep an eye out for these features in future versions of trippler!
Leave a Reply