Exploring Non-Exhaust Emission Factors

Vehicles also emit particles from processes other than combustion: tyre, brake and road surface wear. An emissions inventory for particulate matter should include them. gtfs2emis implements the non-exhaust methods of the EMEP/EEA guidebook in emi_europe_emep_wear().

Tyre and brake wear emissions are computed as:

\[ TE(i) = dist \times EF_{tsp}(j) \times mf_s(i) \times SC(speed) \]

where:

  • \(TE(i)\) = total emissions of pollutant \(i\) (g),
  • \(dist\) = distance driven by each vehicle (km),
  • \(EF_{tsp}(j)\) = TSP (Total Suspended Particle) mass emission factor for vehicles of category \(j\) (g/km),
  • \(mf_s(i)\) = mass fraction of TSP that can be attributed to pollutant \(i\),
  • \(SC(speed)\) = correction factor for a mean vehicle travelling at a given speed (-).

Tyre

For heavy-duty vehicles such as buses, the tyre-wear emission factor accounts for vehicle size (number of axles) and load:

\[EFTSP^{hdv}_{tyre} = 0.5 \times N_{axle} \times LCF_{tyre} \times EFTSP^{pc}_{tyre}\]

where:

  • \(EFTSP^{hdv}_{tyre}\) = TSP emission factor for tyre wear from heavy-duty vehicles (g/km),
  • \(N_{axle}\) = number of vehicle axles (-),
  • \(LCF_{tyre}\) = a load correction factor for tyre wear (-),
  • \(EFTSP^{pc}_{tyre}\) = TSP emission factor for tyre wear from passenger car vehicles, 0.0107 g/km.

and \[LCF_{tyre} = 1.41 + (1.38 \times LF)\]

where:

  • \(LF\) = load factor (-), ranging from 0 for an empty bus to 1 for a fully laden one.

The function uses the following number of axles for each vehicle class:

vehicle class (j) number of axles
Ubus Midi <=15 t 2
Ubus Std 15 - 18 t 2
Ubus Artic >18 t 3
Coaches Std <=18 t 2
Coaches Artic >18 t 3

The size distribution of tyre wear particles is:

particle size class (i) mass fraction of TSP
TSP 1.000
PM10 0.600
PM2.5 0.420
PM1.0 0.060
PM0.1 0.048

Finally, the speed correction is:

  • \(SC_{tyre}(speed) = 1.39\), when \(speed < 40\) km/h;
  • \(SC_{tyre}(speed) = -0.00974 \times speed + 1.78\), when \(40 \le speed \le 90\) km/h;
  • \(SC_{tyre}(speed) = 0.902\), when \(speed > 90\) km/h.

In emi_europe_emep_wear(), this process is selected with process = "tyre":

library(gtfs2emis)

emi_europe_emep_wear(dist = units::set_units(1, "km"),
                     speed = units::set_units(30, "km/h"),
                     pollutant = c("PM10", "TSP", "PM2.5"),
                     veh_type = "Ubus Std 15 - 18 t",
                     fleet_composition = 1,
                     load = 0.5,
                     process = "tyre",
                     as_list = TRUE)
#> $pollutant
#> [1] "PM10"  "TSP"   "PM2.5"
#> 
#> $veh_type
#> [1] "Ubus Std 15 - 18 t"
#> 
#> $fleet_composition
#> [1] 1
#> 
#> $speed
#> 30 [km/h]
#> 
#> $dist
#> 1 [km]
#> 
#> $emi
#>    PM10_tyre_veh_1 TSP_tyre_veh_1 PM2.5_tyre_veh_1
#>            <units>        <units>          <units>
#> 1:  0.01873998 [g]  0.0312333 [g]   0.01311799 [g]
#> 
#> $process
#> [1] "tyre"

Brake

The heavy-duty vehicle emission factor is the passenger car factor, adjusted to match measurements from heavy-duty vehicles:

\[EFTSP^{hdv}_{brake} = 1.956 \times LCF_{brake} \times EFTSP^{pc}_{brake}\]

where:

  • \(EFTSP^{hdv}_{brake}\) = heavy-duty vehicle emission factor for TSP,
  • \(LCF_{brake}\) = load correction factor for brake wear,
  • \(EFTSP^{pc}_{brake}\) = passenger car emission factor for TSP, 0.0122 g/km.

\[LCF_{brake} = 1 + (0.79 \times LF),\]

where:

  • \(LF\) = load factor (-), ranging from 0 for an empty bus to 1 for a fully laden one.

The size distribution of brake wear particles is:

particle size class (i) mass fraction of TSP
TSP 1.000
PM10 0.980
PM2.5 0.390
PM1.0 0.100
PM0.1 0.080

Finally, the speed correction is:

  • \(SC_{brake}(speed) = 1.67\), when \(speed < 40\) km/h;
  • \(SC_{brake}(speed) = -0.0270 \times speed + 2.75\), when \(40 \le speed \le 95\) km/h;
  • \(SC_{brake}(speed) = 0.185\), when \(speed > 95\) km/h.
emi_europe_emep_wear(dist = units::set_units(1, "km"),
                     speed = units::set_units(30, "km/h"),
                     pollutant = c("PM10", "TSP", "PM2.5"),
                     veh_type = "Ubus Std 15 - 18 t",
                     fleet_composition = 1,
                     load = 0.5,
                     process = "brake",
                     as_list = TRUE)
#> $pollutant
#> [1] "PM10"  "TSP"   "PM2.5"
#> 
#> $veh_type
#> [1] "Ubus Std 15 - 18 t"
#> 
#> $fleet_composition
#> [1] 1
#> 
#> $speed
#> 30 [km/h]
#> 
#> $dist
#> 1 [km]
#> 
#> $emi
#>    PM10_brake_veh_1 TSP_brake_veh_1 PM2.5_brake_veh_1
#>             <units>         <units>           <units>
#> 1:   0.05448105 [g]   0.0555929 [g]    0.02168123 [g]
#> 
#> $process
#> [1] "brake"

Road Wear

Road wear emissions do not depend on speed:

\[TE(i) = dist \times EF^{road}_{tsp}(j) \times mf_{road}\]

where:

  • \(TE(i)\) = total emissions of pollutant i (g),
  • \(dist\) = total distance driven by vehicles in category j (km),
  • \(EF^{road}_{tsp}\) = TSP mass emission factor from road wear for vehicles j (0.0760 g/km),
  • \(mf_{road}\) = mass fraction of TSP that can be attributed to particle size class i (-).

The size distribution of road surface wear particles is:

particle size class (i) mass fraction of TSP
TSP 1.00
PM10 0.50
PM2.5 0.27
emi_europe_emep_wear(dist = units::set_units(1, "km"),
                     speed = units::set_units(30, "km/h"),
                     pollutant = c("PM10", "TSP", "PM2.5"),
                     veh_type = "Ubus Std 15 - 18 t",
                     fleet_composition = 1,
                     load = 0.5,
                     process = "road",
                     as_list = TRUE)
#> $pollutant
#> [1] "PM10"  "TSP"   "PM2.5"
#> 
#> $veh_type
#> [1] "Ubus Std 15 - 18 t"
#> 
#> $fleet_composition
#> [1] 1
#> 
#> $speed
#> 30 [km/h]
#> 
#> $dist
#> 1 [km]
#> 
#> $emi
#>    PM10_road_veh_1 TSP_road_veh_1 PM2.5_road_veh_1
#>            <units>        <units>          <units>
#> 1:       0.038 [g]      0.076 [g]      0.02052 [g]
#> 
#> $process
#> [1] "road"

Viewing emissions

A single call can estimate several processes at once. Here we compute tyre, brake and road wear emissions over 1 km at speeds from 1 to 100 km/h:

library(units)
library(ggplot2)

emis_list <- emi_europe_emep_wear(dist = units::set_units(rep(1, 100), "km"),
                                  speed = units::set_units(1:100, "km/h"),
                                  pollutant = c("PM10", "TSP", "PM2.5"),
                                  veh_type = "Ubus Std 15 - 18 t",
                                  fleet_composition = 1,
                                  load = 0.5,
                                  process = c("brake", "tyre", "road"),
                                  as_list = TRUE)

ef_dt <- emis_to_dt(emis_list,
                    emi_vars = "emi",
                    segment_vars = "speed")

ggplot(ef_dt) +
  geom_line(aes(x = as.numeric(speed), y = as.numeric(emi), color = pollutant)) +
  facet_wrap(facets = vars(process)) +
  labs(x = "Speed (km/h)", y = "Emissions (g)") +
  theme_minimal()

With the output of transport_model(), you can compare hot-exhaust and non-exhaust emissions on a real network, in three steps: a) estimate both types of emissions, b) aggregate them on a spatial grid, and c) plot them.

a) Estimating hot-exhaust and non-exhaust emissions

library(gtfstools)
library(sf)

# read GTFS
gtfs_file <- system.file("extdata/bra_cur_gtfs.zip", package = "gtfs2emis")
gtfs <- gtfstools::read_gtfs(gtfs_file)

# keep a single trip_id to speed up this example
gtfs_small <- gtfstools::filter_by_trip_id(gtfs, trip_id = "4451136")

# run transport model
tp_model <- transport_model(gtfs_data = gtfs_small,
                            spatial_resolution = 100)

# fleet data for the EMEP/EEA model (a single bus type)
fleet_data_ef_emep <- data.frame(veh_type = "Ubus Std 15 - 18 t",
                                 fleet_composition = 1,
                                 euro = "V",   # for hot-exhaust emissions
                                 fuel = "D",   # for hot-exhaust emissions
                                 tech = "SCR") # for hot-exhaust emissions

# emission model (hot-exhaust)
emi_list_he <- emission_model(tp_model = tp_model,
                              ef_model = "ef_europe_emep",
                              fleet_data = fleet_data_ef_emep,
                              pollutant = "PM10")

# emission model (non-exhaust)
emi_list_ne <- emi_europe_emep_wear(dist = tp_model$dist,
                                    speed = tp_model$speed,
                                    pollutant = "PM10",
                                    veh_type = "Ubus Std 15 - 18 t",
                                    fleet_composition = 1,
                                    load = 0.5,
                                    process = c("brake", "tyre", "road"),
                                    as_list = TRUE)

# attach the transport model so emis_grid() and emis_summary() can use it
emi_list_ne$tp_model <- tp_model

b) Aggregating emissions on a spatial grid

# create spatial grid
grid <- sf::st_make_grid(x = sf::st_make_valid(tp_model),
                         cellsize = 0.25 / 200,
                         crs = 4326,
                         what = "polygons",
                         square = FALSE)

# grid (hot-exhaust)
emi_grid_he <- emis_grid(emi_list_he,
                         grid = grid,
                         time_resolution = "day",
                         aggregate = TRUE)
data.table::setDT(emi_grid_he)
pol_names <- setdiff(names(emi_grid_he), "geometry")
emi_grid_he_dt <- data.table::melt(emi_grid_he,
                                   measure.vars = pol_names,
                                   id.vars = "geometry")
emi_grid_he_dt <- sf::st_as_sf(emi_grid_he_dt)

# grid (non-exhaust)
emi_grid_ne <- emis_grid(emi_list_ne,
                         grid = grid,
                         time_resolution = "day",
                         aggregate = TRUE)
data.table::setDT(emi_grid_ne)
pol_names <- setdiff(names(emi_grid_ne), "geometry")
emi_grid_ne_dt <- data.table::melt(emi_grid_ne,
                                   measure.vars = pol_names,
                                   id.vars = "geometry")
emi_grid_ne_dt <- sf::st_as_sf(emi_grid_ne_dt)

# bind grids
emi_grid_dt <- data.table::rbindlist(l = list(emi_grid_he_dt, emi_grid_ne_dt))
emi_grid_sf <- sf::st_as_sf(emi_grid_dt)

c) Plotting the emissions

The map shows daily PM10 emissions in each grid cell, by process:

ggplot(emi_grid_sf) +
  geom_sf(aes(fill = as.numeric(value)), color = NA) +
  geom_sf(data = tp_model$geometry, color = "black") +
  scale_fill_continuous(type = "viridis") +
  labs(fill = "PM10 (g)") +
  facet_wrap(facets = vars(variable), nrow = 1) +
  theme_void()

And the bar chart compares the total emissions of each process:

# emissions by time
emi_time_he <- emis_summary(emi_list_he, by = "time")
emi_time_ne <- emis_summary(emi_list_ne, by = "time")

emi_time <- data.table::rbindlist(l = list(emi_time_he, emi_time_ne))

ggplot(emi_time) +
  geom_col(aes(x = process, y = as.numeric(emi), fill = as.numeric(emi))) +
  scale_fill_continuous(type = "viridis") +
  labs(fill = "PM10 (g)", y = "Emissions (g)") +
  theme_minimal()

References

EMEP/EEA data and reports can be accessed in the following links:

Report a bug

If you have any suggestions or want to report an error, please visit the package GitHub page.