(Unedited) Podcast Transcript 318: Open Source Electric Bus

This week we’re joined by Erica Eggleton, a PhD Candidate at the University of Washington to chat about her work on Route Dynamics, an open source program that estimates the energy demand for electric bus routes run by King County Metro.

Below is a full (unedited for the moment) transcript:

Jeff Wood (1m 27s):
Erica Eagleton, welcome to the talking head waist podcast.

Erica Eggleton (1m 33s):
Thank you so much for having me well, thanks

Jeff Wood (1m 35s):
For being here. So before we get started, can you tell us a little?

Erica Eggleton (1m 38s):
Yeah, so I’m a fourth year PhD student at the university of Washington, or UDaB as I might call it sometimes. And in the chemical engineering department, I’m also a university of Washington clean energy graduate fellow. I work in a research group led by professor Dan Schwartz. It’s traditionally an electrochemistry group, but we’ve been getting more into lithium ion batteries the past few years. And so my research works on modeling energy storage and conversion systems. We’ll talk about one of those projects a bit more today and yeah, lastly, I’m just a part of our graduate student Senate here at UDaB where one of my roles is to be on the university transportation committee.

Erica Eggleton (2m 29s):
So really passionate about transit and our community

Jeff Wood (2m 33s):
That you start to become passionate about transit.

Erica Eggleton (2m 36s):
Yeah. Great question. So it really started a long time ago when I was in middle school and my family moved to Germany and I lived there for seven years. So actually like all of middle and high school or, you know, those really formative years

Jeff Wood (2m 54s):
And I became a major

Erica Eggleton (2m 57s):
Transit user. It was just so right. I mean, if you’ve ever been to Germany or Europe in general, there are such dense transit networks from my little village on the outside of Stuttgart. I could get to the city, I could go to other cities. I could even go to other countries in Europe, you know, I just needed the right ticket. And that just gave me so much independence led to so many great experiences and just really loved it at the time. And yeah. Then realized that I started to kind of take it for granted when I moved back to the us and for college, I went to Montana state university.

Erica Eggleton (3m 42s):
So it kind of went from one extreme to the other where like I got my first car and needed to drive just to like go get groceries and stuff like that. And that just made me way more aware of like my carbon footprint and that I really missed it. I was really excited to come to Seattle for grad school and get more access to that again.

Jeff Wood (4m 5s):
Did you have a favorite trip in Europe that you had to take a train or transit to?

Erica Eggleton (4m 9s):
Oh man. That’s so hard. So hard to pick. I mean, we always tried to take train trips every year for my mom’s birthday. Cause she’s also like a fellow transit geek I would say. So that took us like Prague to Austria. Yeah. All over. So it’s really hard to pick.

Jeff Wood (4m 34s):
So when I was younger, my parents lived in Rotterdam for a year and I went to go visit for Christmas and we went to Heidelberg in Germany for Christmas. And that was fun. The Christmas markets and stuff are amazing. Your tree in the background reminds me of the good times.

Erica Eggleton (4m 48s):
Yeah. I mean, honestly my time in Germany probably contributed to my love for Christmas, love the Christmas markets and all that jazz for sure.

Jeff Wood (4m 58s):
And so an exciting part of, of your work, I imagine is that whole battery discussion and how it’s kind of taking off now in terms of people are getting really excited to see what’s happening. I mean, what’s it like in the office right now, talking about this transition that’s happening.

Erica Eggleton (5m 13s):
I mean, there’s so many aspects we talk about, like in my group, there are people that are like looking at diagnostics for batteries that use for anything from like your personal electronics to electric vehicles. So that’s just always on our brain, but as far as like the transition for transit agencies in general. Yeah. I mean, we’re just excited to see that lots of cities are doing this. We are seeing like lots of policy changes in Washington and then seeing like policy changes on the national level to where more and more cities are going to be making this transition to electrifying their transit in the future.

Erica Eggleton (5m 58s):
So just know that this is going to be something that a lot of people need more information about. So it’s really motivating for me to try to give more answers.

Jeff Wood (6m 10s):
So what are you all working on generally in terms of battery technology at the moment in the clean energy Institute?

Erica Eggleton (6m 17s):
Yeah, so it really varies. So we, the clean energy Institute itself is made up of research groups in various departments from like chemistry, material science, chemical engineering, electrical engineering, civil engineering, and I’m probably missing some others too. So many groups are working on batteries in different forms. Some are looking at the materials itself. So like finding the safest and most energy dense materials you can use for the batteries. And then there are modeling groups working on optimizing the grid to include energy storage systems groups working on yeah.

Erica Eggleton (7m 0s):
More like systems level approaches. Like my group is looking at specific applications with transit agencies like King County Metro and how we can understand the batteries a bit more. So yeah. All kinds, all kinds of things. It’s hard to just talk about a few,

Jeff Wood (7m 21s):
Well, that’s a hot topic as best as transitioning to electric. What are some of the things we need to think about from your perspective in terms of that transition to electric buses?

Erica Eggleton (7m 29s):
So some of the things that we’re interested in is just how to plan for this transition better and also implement data driven approaches. So I’m sure we’ll talk a bit more about this software we’ve developed called route dynamics. And the goal of this software is to help create a clear pathway for transit agencies that are making these electrification shifts and understanding what characteristics of routes and vehicle specifications are important to get the most out of your batteries and infrastructure planning.

Erica Eggleton (8m 15s):
And so things that these agencies can be considering are things like the geographic location of the route, like how hilly the routes are, how the ridership of the routes changes like how full the buses are also considering acceleration as well. So like how fast you’re stopping and starting where like bus stop locations are and other events like that. Yeah. So essentially taking those type of things that might be important for logistics management in general for the transit agencies, but they’re also valuable for what you need for your battery electric buses as well.

Erica Eggleton (8m 58s):
So trying to merge those together

Jeff Wood (9m 1s):
A bit about the tool, I mean, you just mentioned a number of things that are integrated into the tool, but I’m wondering where did the idea for the tool came from and you know, how it was developed and what is

Erica Eggleton (9m 11s):
Yeah. So the idea of the tool came from this motivation of trying to get more information with our only relying on like onboard vehicle data. And we were also reading more about like what type of open source data is available. And then in the literature found models such as vehicle dynamics models that essentially are considering all the forces acting on the bus and then estimating the power requirements. So essentially what the software entails is you taken this data and you get power and energy requirements out that you can use to analyze your different routes

Jeff Wood (9m 57s):
And you use mapping tools to use GIS to calculate some of that stuff. It sounded like some of the processing was pretty intense as well.

Erica Eggleton (10m 5s):
Yeah. Yeah. Great point. So to revisit some of the different data first that goes into it, it’s the geographic information systems or GIS data. That’s essentially the location of where the bus route is, and also its elevation. So with that, you can determine the elevation profile for the route and determine how steep it is and stuff like that. But yeah, you’re right. It was a little hard to figure that point out because it needed to be filtered a certain amount. So that way we can get accurate estimates. So part of the software itself, one part is just to even clean that data to give us smooth elevation profile.

Erica Eggleton (10m 49s):
So that’s one of the inputs that we can use.

Jeff Wood (10m 54s):
You have acceleration. I’m curious. One of the things that came into my mind when I was reading the paper was, you know, what about those drivers that just shoved the shove, their foot on the gas or on the, on the pedal, I guess I should say now that the gas is going to disappear at some point,

Erica Eggleton (11m 7s):
We think about these like little scenarios all the time. And cause, I mean, I guess real quick, that’s something that’s kind of fun about this, right? Like this is something we’ve all experienced. Like you’ve experienced being on a bus and feel like jerk out of nowhere or you slamming on the brakes and holding onto the bar. And so we can kind of like feel these different, really fast accelerations or fast decelerations and breaking from our, our lives. And so yeah, that’s something we want to include. And so we’re making it, so that’s something that the user can input. They can input what acceleration profiles you want to follow.

Erica Eggleton (11m 48s):
But yeah, that is something that would then have to be at least right now, pretty consistent. So like one thing looking at is from King County Metro, we were given an ideal acceleration profiles. Like when you’re buying a bus, you want to make sure that it can meet this acceleration profile that they view is ideal. And so that is what we’re following when the buses accelerate your way from a bus stop or something like that. But yeah, one thing you could do to analyze these routes further is play around with those. Like what if it’s just a constant acceleration? What if it’s some obscene fast acceleration, like you might’ve experienced? And so that is something you can play around with, with this tool.

Jeff Wood (12m 31s):
You also have the ridership numbers. I’m curious if those are just imputed from the data that the agency gives you or is that like real-time information. I know that real-time information is really popular right now because of the pandemic. And people want to know, you know, how many people on the bus so that they can practice physical distancing, but I’m curious how that plays into it and how the ridership numbers are computed as well.

Erica Eggleton (12m 52s):
Yeah. So right now it’s also just a user input. We, while developing this, Reedus given like ridership numbers for like a typical month for certain routes and that’s what we’re inputting, but that is something you could change. Like right now it seems like the easiest to be like, what if you’re only seeing half of those numbers or a certain percentage of those numbers, then you can start to see how that would change your output. And yeah, so no like real time input right now, but that would definitely be interesting to look at further, just we require a few adjustments.

Jeff Wood (13m 30s):
So you take all these measures, the elevation, the acceleration and the writers, and then you kind of calculate a, you know, a battery usage. What’s the benefit to the agency to have that kind of calculation.

Erica Eggleton (13m 40s):
So multiple, multiple uses. So first you can think of it, we get these power and energy requirements and the energy is really great for like understanding the range. I feel like that’s one of the first problems you might think of when you’re thinking of putting in a electric vehicle anywhere, can it actually complete the trip you want it to make. And so based off of these route characteristics and vehicles, specifications, and other data input into the model, you can get an estimate of how much energy would be required to complete that trip and see if the range or the battery capacity on this vehicle can meet that.

Erica Eggleton (14m 21s):
So that’s one thing you can do. Another is, so we talked about the power requirements. So understanding what parts of the trip might be requiring higher loads, and that can be used for understanding that battery degradation component. There are certain events that can be more stressful on the battery and those normally happen at those higher charge and discharge rates. An interesting thing we can look at with the load profiles is when it’s too, because when a bus is hitting the brakes, many of them are, have regenerative braking as well. So they’re actually recharging the batteries at different times.

Erica Eggleton (15m 2s):
And so what are the rates of that? So looking at those, the rates at which the battery is charging and discharging, we can start to find different routes that have more stressful events for the batteries and use that information for trying to optimize its lifetime. When you’re thinking about the different forces on the bus that are affecting it, or that make it harder from get to point a to point B, like it’s really a perfect analogy to think about you riding your bike on a route, trying to go from point a to point B. And what makes it harder or easier for you? So like going up a Hill is a lot harder, or if you’re wearing a backpack that’s full of textbooks, that’s a lot harder than just being free.

Erica Eggleton (15m 45s):
And it’s also like a funny analogy. Cause like if you’re really into biking, like my dad, for example, he even has like a power meter that tells him the wattage that he’s actually putting out as he’s writing. And so it’s like just a perfect comparison because that’s exactly what we’re measuring like based off of all those things that makes it harder, easier to get from point a to point B, we can estimate the load.

Jeff Wood (16m 10s):
One of the things that I thought about was I know Seattle is famous for the times when the streets ice over a little bit. And so there’s some streets, I think in Capitol Hill or other places where you have these videos outside of people’s windows and you see the bus sliding down the Hill trying to get up the Hill and can’t get up the Hill. And so, you know, that made me think about how much energy usage or not energy usage if you’re just sliding down the Hill is necessary. And I, I think that goes to that same point too. I know that, you know, Montreal has been trying out the electric recharging buses. And I think one of the problems they had was in the winter is trying to keep the bus heated. Right. And so haven’t used all that electricity for the heating. I imagine the same as in Phoenix, when in the summertime, when you’re having to cool the bus down, thinking about how much energy is going to be expended, doing all those extra things that you don’t think about, you know, energy usage, you just think about the bus driving, but there’s all these other little details that need to be fixed out before you can determine how much battery you need for any given trip or any given day.

Erica Eggleton (17m 7s):
Yeah. That’s another huge part of it that auxiliary power, right? How much is just going to keep the bus comfortable. And then that’s also interesting to look at too because different temperatures can also be stressful for the battery as well. I mean, we always like joke, I guess in our group that batteries are just like us, that like, they like to be at a certain comfortable temperature and anything else. And you start to get all these like weird degradation things happening. And so yeah, considering those type of weather parameters would not only help the model for that extra power, but also help understand like what other stresses the battery might be.

Jeff Wood (17m 51s):
That sounds like a really good new Yorker cartoon, the temperamental battery and have like, you know, the battery walking into a room with like a coat on it’s like it’s kinda cold out. It’s kinda cold in here, you know, bad in New York, a joke. There’s the big discussion about electrification and a move towards battery electric buses and you’re you’re in Seattle. So you know that there are trolley buses. I’m wondering why we wouldn’t just move to trolley buses. Why, why batteries? Why do we need them if we can do this other thing that is already proven for over a hundred years?

Erica Eggleton (18m 19s):
Yeah. Great question. And I, one of the biggest things is that infrastructure component, right? I mean like looking at King County Metro, for example, like there’s so many bus lines that the infrastructure for having overhead wires for all of those routes would be enormous. And so, yeah, this seems like a great option for agencies that already have like diesel buses. They don’t want to put overhead wires across their whole city and instead just put a couple charging stations throughout the route, but they can keep going. So that, that infrastructures can be really hard, especially when you have a lot of routes

Jeff Wood (19m 4s):
And the tool makes it easier to measure the output of the battery. I’m wondering why it doesn’t just have a meter on it. Why isn’t there a meter on the, on the battery? It tells me gas tanks.

Erica Eggleton (19m 15s):
Yeah. So any electric vehicle or these battery electric buses, do you have a battery management system on them that is like measuring the voltage and the current and keeping track of that. And that’s actually data that we’re going to be using to kind of like ground truth or model or see how well it’s doing at predicting these things, the keyword, their prediction. So this type of model can kind of compliment those things and it can help predict for future routes. So say you’re just like looking into making this transition and you’re wanting to make like, have questions like which buses to buy and in what routes you don’t have that meter data yet.

Erica Eggleton (19m 56s):
And instead can use this to help make those decisions. Or you want to play around with moving the buses around, like maybe this bus we know is on a stressful route. What if we, a couple of days a week put it on this flatter route and that’s not as stressful. You can start to kind of play around with those things, using this, instead of just looking on the data on board itself,

Jeff Wood (20m 22s):
What’s been the response from King County, Metro so far on using the tool.

Erica Eggleton (20m 26s):
They’re excited about it. They’ve been working with us to like, see what kind of things we can learn from it. So for example, like me just showing them that with all this data that we have available in King County, this is the analysis we can have and then helping them make these decisions. Yeah. There’s still like early on in this transition and they’ve already been doing analysis on their own on like where out should go, but to also include other factors that like aren’t in my model, like climate justice parameters, such as like what communities would benefit more from emotionless vehicles.

Erica Eggleton (21m 9s):
And so one thing I’ve been looking at is based off those routes that they’re interested in on that, which ones are like best candidates from an energy efficiency standpoint as well.

Jeff Wood (21m 22s):
If you had unlimited resources, what else would you add to the model? Oh, wow.

2 (21m 28s):
I mean, I really

Erica Eggleton (21m 29s):
Interested in something you’ve brought up now the real time data component. I know there’s so much real time data you could add, like you mentioned ridership, there’s also, at least in Seattle, there’s an app called one bus away where you can follow where your buses are in real time. And so like the real time geographic location is available as well. So yeah, like with unlimited time and resources would love to get all those integrated into the model as well. And then you could have these like real time power estimates being made for each one. And so something to look forward to.

Jeff Wood (22m 11s):
So this is kind of a, I don’t know if it’s a question as a comment and it’s kind of funny and I hope you enjoy it as well. But I was reading the paper last night and I know that planning jargon is really bad. And I wanted to read my favorite sentence from the paper so we can kind of have a bit of a laugh together. The Skippy Savitsky Goulay filter was used, which fits the data, using a user-defined polynomial over a selected window. I just read the sentence. And I was like, wait, what?

2 (22m 39s):
And I

Jeff Wood (22m 39s):
Enjoyed that because of all the jargon that goes into our industry generally is no, we’re always saying this MPO said this about the RTP and it’s just, I just had a laugh at that and I hope you do too.

Erica Eggleton (22m 53s):
It’s definitely easy to go into the weeds.

2 (22m 58s):
Yeah. I,

Erica Eggleton (22m 59s):
I honestly forgot that that was in the part you were reading and kind of goes with what you were saying earlier. There was a lot of filtering needed for that geographic data. I mean, and I guess just going on like a side tangent here. So my department also has a data science program that I’ve been involved with. So like data management, software development, and that’s why like this project was extra cool to me, you know, combining the chemistry and chemical engineering side with this cool open source software. But yeah, the first, like almost six months, nine months of it was just trying to make this one data curve smooth, making the data clean and it felt like a true data scientist just cleaning data forever until I got some actually cool results.

Jeff Wood (23m 56s):
What’s the difference in energy necessarily between like say the hilliest root and the flat is true. Is it such a huge shift that it makes such a big difference?

Erica Eggleton (24m 5s):
So elevation is one of the ones where we’re seeing a considerable effect. And that’s also, I guess yeah. Why we’re interested in this because a transit agency in Kansas might not really care about these kinds of planning tools as much as like Seattle, where it’s really hilly. And I am reminded of that every time I ride my bike. And so, yeah, we’re definitely seeing considerable differences that, and the acceleration profile as well. So like that acceleration has a decent impact. And so then if you combine them accelerating up a really steep Hill, that route might not be really great for your battery because it’s like a double whammy at that point.

Jeff Wood (24m 54s):
Are there any other use cases that are kind of fantastical that you’ve thought of for using this data set and the tool

Erica Eggleton (25m 1s):
Besides like transit, planning and optimization? I could see this being used for any electric vehicle, right? I mean, this type of information could also be useful when planning your trip and your Evie. And I know that there have been some people that have looked into that before, but trying to, not only like, you know, think of Google maps right now, when you like, are trying to find the most efficient route to the grocery store, time-wise a lot, if you’re also considering, well, what’s the most efficient route for my battery of my car. So that could be really useful. Or also, I mean, like right now, this tool’s looking at transit fleets, but as we continue to move towards electrification of other things, like what about fleets of ups, FedEx, they’re moving large quantities of things.

Erica Eggleton (25m 56s):
And if they’re moving towards electric vehicles, they might want to optimize their routes as well. So yeah, I could see this being applied to many other forms of transportation.

Jeff Wood (26m 8s):
Yeah. We had some folks on recently talking about autonomous delivery vehicles and it seems like, you know, without passengers it’s just freight. Right. And so thinking about all these in-town examples, I mean, moving people on a bus is to a certain extent, similar to moving goods inside of a city. It’s not like a long haul thing or a long distance thing necessarily, but it would be interesting, especially in a city like Seattle to see how like an Amazon or even some of these new companies would be able to, you know, think about, cause they’re all trying to make electric vehicles. So think about how, you know, you would distribute your goods and services and use, you know, the ability to measure your battery outputs. I think that’s really fascinating. So it seems like it has a lot of different applications outside of transit, which I like the transit one of the best myself, but, but it seems like very useful.

Jeff Wood (26m 55s):
So where can folks find the module online if they want to go and check it out?

Erica Eggleton (26m 59s):
Yeah. So this can be found on get hub and maybe we can include the link here and essentially get hub is a platform for people to share open source software. So I guess that’s something I’ve not really touched on much yet, but we have been developing this with the idea that we want this to be open access, something that other transit agencies could look at and offer feedback. So yeah, like one of the benefits of get hub is you can leave comments or issues they’re called, if you have like certain ideas or opinions, and there’s also the availability for you to download the code and play around with it.

Erica Eggleton (27m 41s):
And so that way it can be developed by a large community instead of just our small team. And so that’s a really fun aspect of it.

Jeff Wood (27m 52s):
Awesome. What’s next for you? What’s the next project you’re working on? Are you just going to be improving this project?

Erica Eggleton (27m 57s):
Yeah. Well, this is definitely something that I’ll be working on for a bit still improving it. Ground-truthing it a bit like I was saying before with other data from Metro, but yeah, I’m also involved in other open access data projects in the battery space, just really motivated and making this field like super interdisciplinary and collaborative and sharing data sharing software. So that way we can try to solve these problems faster. Nice.

Jeff Wood (28m 31s):
And Erica, where can folks find you if you wish to be found?

Erica Eggleton (28m 35s):
Sure. Yeah. Well, one way I can be found is on Twitter. My handle is just at Eric Angleton. You can also follow this project and other like clean energy Institute related projects on Twitter as well. There’s is at U w underscore CEI. So yeah, that’s where you can be in the know what’s going on there.

Jeff Wood (28m 58s):
Well, Erica, we thank you so much for joining us. We really appreciate it. Thanks for your time. And thanks for joining us to talk to me. Headways podcast is your project of the overhead wire on the [email protected]. Sign up for a free trial of the overhead wire daily or 14 year old daily city’s news list by clicking the link at the top, right of the overhead wire.com. And please, please, please put the pod in a pitch on.com/the overhead wire many thanks to our current patrons for their ongoing support. And as always, you can subscribe to this podcast on iTunes, Stitcher, SoundCloud, overclass Spotify, and wherever you get your podcasts. And you can always find the traditional home at USA dot Street’s blog.org. See you next time at talking headways.