1. Background
As a follow-up to our initial look at visualizing data from the FAF5 database using QGIS (link to previous post), we're working to further develop this into a geospatial decision support tool. The aim of the tool is to help trucking fleet owners assess where and how best to decarbonize their fleets by transitioning to alternative fuels and powertrains. The code is being maintained in this GitHub repo.
2. Summary of New Functionality
We've made quite a few additions to the basic FAF5 visualization functionality introduced in the last post. Big shout-out here to Micah Borrero, who's been making awesome contributions to the backend code to help implement different parts of the new functionality while working with the MCSC as an undergraduate researcher with MIT's UROP program.
Visualizing Imports and Exports by FAF5 Region
In the previous post, we used the FAF5 data to visualize freight transported along U.S. highway networks. Since then, we've added functionality to visualize the total freight transported between different FAF5 regions, as shown in Figure 1 below.
This is done by first processing data contained in a csv file produced by FAF5 that quantifies freight carried by mode (trucking, inland shipping, air and combinations thereof), commodity, origin and destination, with one row per origin-destination trip and commodity. The data is aggregated (optionally with filtering applied) to sum over all ton-miles carried to/from each FAF5 region, and the resulting freight flows per region are merged with the shapefile of FAF5 region boundaries shown in the previous post for visualization.
Figure 1: Total annual ton-miles imported and exported by truck to/from each FAF5 region per square mile
This regional visualization is an important complement to the highway network flows because it helps the user to see where the freight transported along the highway networks is actually coming from and going to. I considered a few ways of visualizing the freight transport (tons, ton-miles, etc.), and found that in general the total ton-miles transported divided by the surface area of the FAF5 region gives the most intuitive visualization. Ton-miles is chosen rather than tons because it accounts not only for the amount of freight transported, but also the distance that it's carried over. Dividing by the surface area of the FAF5 region gives a visualization of freight transport density that's independent of the chosen size of the FAF5 region.
Filtering FAF5 Data
Functionality has also been added to help visualize FAF5 data of interest by filtering on various dimensions, described in the following sections.
By Import/Export
First, the freight flow data can be filtered according to whether the freight is being imported into or exported from the region.
Figure 2: Total annual ton-miles imported (left) and exported (right) by truck to/from each FAF5 region
By Mode and Commodity
The FAF5 data can also be filtered according to the mode by which the freight is carried (all maps in this post show freight carried by truck), and by the commodity carried. The commodity falls into one of 43 categories, including animal feed, coal and fertilizer.
Figure 3: Total annual ton-miles of cereal grains imported and exported to/from all FAF5 regions per square mile
By Origin/Destination
It can also be helpful to filter for exports originating from or being exported to a particular origin or destination region.
Figure 4: Total annual ton-miles of exports to all FAF5 regions originating from Dallas-Fort Worth (highlighted in green)
Lifecycle Emissions Associated with Freight Flows
We've also added functionality to evaluate emissions associated with today's freight flows. This is done by combining fuel production, and tailpipe emission intensities (gCO2 / gallon) from the GREET tool developed by Argonne lab, along with vehicle characteristics (fuel type, fuel efficiency) and operating conditions (typical loaded and unloaded weight, annual miles driven, percent miles carrying different commodities) from the 2002 vehicle inventory and use survey (VIUS). The VIUS program was suspended for nearly 20 years after 2002. Following this suspension, a VIUS survey was done more recently in 2021, with the results expected to be available at the end of 2023.
The combination of these three datasets is shown at high level in Figure 5.
Figure 5: High-level diagram showing the data sources (GREET, VIUS, and FAF5) that are combined to evaluate emissions associated with freight flows
Figure 6 shows in some more detail how the information from each dataset is integrated. For each freight flow i in the FAF5 database, the FAF5 data specifies which commodity is being transported, the range over which it's transported, and the total ton-miles transported between the origin and destination. The VIUS data is used to evaluate the average fuel efficiency (mpg) and payload for trips carrying the given commodity over the given range. This information, along with the ton-miles carried from the FAF5 data and the emissions per gallon (epg) of a truck burning a given fuel from GREET is combined using the equation in the central grey diamond in Figure 6.
Figure 6: Detailed diagram showing how emissions data from GREET and vehicle characteristic and operation data from the VIUS are combined with the FAF5 data to evaluate associated emissions
Our analysis of the VIUS data found that the vast majority (>99%) of ton-miles were carried by diesel trucks in 2002. Based on this, we assume that the trucks are burning diesel when evaluating the emissions per gallon with GREET for the present freight flows. The resulting emissions map, shown in Figure 7 for imports and exports of all commodities by truck, ends up looking very similar to the freight flow map (with only some small variation arising from varying fuel efficiencies and payloads for different commodity types).
In the future, we'd like to investigate how the emissions map changes for different electrification and alternative fuel penetration scenarios, in which case we expect regional effects (eg. the regional intensity of the electrical grid) to become more pronounced.
Figure 7: Emissions associated with total annual imports and exports of all commodities using the methodology visualized in Figures 5 and 6
Additional Layers
We've also added some additional layers that we anticipate could offer additional decision-making support, including:
Locations of public EV charging and alternative refueling stations on highway corridors, obtained from the DOE's Alternative Fuels Data Center.
Figure 8: Locations of charging and alternative refueling stations on highway corridors
Regional CO2 intensity of the electrical grid, obtained from the eGRID database.
Commercial electricity price by state, obtained from the EIA electricity database.
Maximum demand charges by utility. Demand charges are assessed based on the maximum power draw that a given user places on the electrical grid over a short time period (typically 15 minutes or so), with the idea of penalizing users for the rapid ramp-up of power plant outputs needed to accommodate high power draws. These are obtained from a compilation of all historical demand charge rates compiled by the National Renewable Energy Lab (NREL).
The last three layers listed above will be illustrated in the next section.
3. Identifying Candidate Electrified Trucking Corridors
This section illustrates a sample scenario analysis to identify promising highway corridors for electrified trucking.
As a first step, all pairs of origin/destination FAF5 regions were listed in order of the total annual ton-miles transported from the origin region to the destination region, as shown in Figure 9.
Figure 9: List of origin-destination pairs, ordered according to annual ton-miles transported between them
Given that the Dallas-Fort Worth (DFW) region appears near the top several times as an origin, we filter the FAF5 data to look at imports into other regions originating from DFW in more detail, as shown in Figure 10. Also highlighted in the cyan are potential corridors connecting DFW to regions with high imports originating from DFW.
Figure 10: Total annual ton-miles of imports into all FAF5 regions originating from DFW (highlighted in green). Major corridors connecting DFW to other FAF5 regions are highlighted with adjacent cyan lines, and each highlighted corridor is labeled with a light blue box.
The corridors are then ranked from 1 (most promising) to 5 (least promising) according to various criteria using available decision support layers, as summarized in Figure 11 and Table 1.
Figure 11: Visual rankings of candidate highway corridors from most promising (1) to least promising (5) for electrified trucking (more details in Table 1)
Fig. 11 label | Relevant Layer | Ranking Approach | Texas Triangle Rank | DFW-Durham Rank | DFW-St Louis Rank | DFW-LA Rank | DFW- El Paso Rank |
A | Highway Flux (annual tons / highway link) | Large highway flux suggests large potential pool of charging station users --> faster payoff of initial charging infrastructure costs. | 1 | 4 | 2 | 3 | 5 |
B | DCFC Charging Stations | Though most existing DCFC stations likely won't be usable by trucks, their presence suggests the route is already compatible with charging. | 1 | 2 | 5 | 3 | 2 |
C | CO2 Intensity of the Grid | Low CO2 intensity is desired to minimize lifecycle emissions associated with EV truck charging. | 1 | 3 | 4 | 2 | 2 |
D | Commercial Electricity Price | Though charging rates will likely differ from the commercial electricity price, they will likely be correlated --> lower is preferred. | 1 | 3 | 2 | 4 | 1 |
E | Maximum Demand Charges | Large power draws needed to charge truck batteries may incur demand charges --> lower is preferred. | 1 | 2 | 2 | 2 | 2 |
All | All | Average rankings over all relevant layers | 1.0 | 3.2 | 3.0 | 2.8 | 2.4 |
Table 1: Ranking of candidate highway corridors shown in Figure 11 from most promising (1) to least promising (5) for electrification
Based on this sample scenario analysis, the Texas triangle connecting DFW to Austin, San Antonio and Houston comes out as a clear winner in terms of suitability as an electrified trucking corridor among freight corridors originating in DFW. This finding is corroborated by a 2020 analysis of high-potential regions for electric truck deployments done by Jessie Lund (Rocky Mountain Institute) and Mike Roeth (North American Council for Freight Efficiency).
4. Next Steps
There are lots of directions that I'd like to take the this tool from here, and I'm always happy for more ideas on what would make it a useful resource to support fleets in navigating the transition to alternative fuels and powertrains.
1. DataHub Integration: I'm currently working with a principal research scientist Emre Gencer at the MIT Energy Initiative and an awesome developer Brilant Kasami to integrate this tool into an emerging web-based platform called the DataHub, which aims to centralize data sources that the MCSC is drawing from, and tools that we're developing.
My ultimate goal for this geospatial mapping tool is that it can become a user-interactive web app where users can easily select which layers to visualize, do basic analyses with the layers, and upload their own geospatial data to overlay on the map.
2. Additional Decision Support Layers: I've gotten lots of great suggestions for other mapping layers that could help with decision support, including:
Current and proposed regional incentives to support fleets in transitioning to alternative fuels and powertrains.
Proposed infrastructure projects (eg. public truck charging stations along highways) to support transitions along particular corridors.
Regional abundance of resources for alternative fuels (eg. hydrogen reservoirs, biofuel production).
Regional capacity of the grid to accommodate increased load from high-powered truck charging.
3. Extend the tool to other modes and countries: While the tool currently focuses on trucking in the U.S., there's no reason why it couldn't be extended to other countries, or to other modes like aviation and shipping, provided that quality freight flow and emissions data can be obtained with the needed coverage.
4. Scenario Analysis: As I alluded to in Section 2, I think a powerful use case for this tool will be to project the evolution of regional transportation emissions under alternative penetration scenarios of alternative fuels and powertrains (eg. X% of trucks transition to electric, Y% to hydrogen fuel cell and Z% to biofuels). Both the FAF5 freight flow data and the GREET module include projections out to 2050. In the case of electrification and potentially hydrogen, the regional emission intensity of the electrical grid will also play an increasingly important in these lifecycle emissions projections.
Comments