Matching electric buses with optimal routes – A new tool that helps make good decisions

Zero-emission vehicles
Clean air

The Bangalore Metropolitan Transport Corporation (BMTC), the largest public transit bus operator in India, has big plans for battery electric buses. It could purchase as many as 1,800 of them in the next several years, and the Karnataka state government aims to transition all 6,500 of BMTC’s buses to zero-emission vehicles by 2030. Some support for this will come from the Government of India’s Faster Adoption and Manufacturing of Electric Vehicles (FAME-II) program.

Bangalore is far from alone. Electric buses are gaining in popularity across the world. China mainstreamed battery electric buses, and up to 30% of new urban bus sales are electric. Santiago, Chile successfully introduced 400 battery electric buses in its fleet, and Bogota, Colombia ordered almost 600 battery electric buses, including articulated, for its TransMilenio bus rapid transit system. The City of Los Angeles is also planning to electrify 155 transit buses and has an ambitious goal to make its entire fleet electric by 2028.

Countries and cities see the urgent need to reduce greenhouse gas emissions and air pollution from transportation activities, and electrifying urban bus fleets could do both. Electric buses produce zero tailpipe emissions of pollutants like particulate matter, nitrogen oxides, and black carbon, and are able to achieve up to 300% energy efficiency improvements compared to diesel-powered buses.

At the same time, these buses are a large investment. While they cost less over the entire service lifetime when using a total cost of ownership (TCO) model, purchase prices can be as much as 150% higher than diesel buses in some markets. Moreover, the new infrastructure needed to support an electric bus fleet is an additional expense. Without proper planning, the charging and range limitations of battery electric buses will also become an issue. Decisions made without good data could lead to problems such as high replacement ratios—i.e., a higher number of electric buses are required to deliver the same service as a single conventional diesel or compressed natural gas bus—and undermined fleet operations. These issues can, in turn, mean higher costs and lower rider satisfaction.

So, BMTC needs answers to questions that many other bus operators around the world have found or will find themselves asking: What are the minimum technical specifications for battery electric buses for a particular route? Which route(s) should get electrified first? What should be the charging strategy for the fleet? TCO models based on empirical data can shed light on these questions, and ICCT is developing a route-level TCO model to support BMTC’s efforts. Our recent working paper describes a new methodology that we developed to identify representative drive cycles based on real-world bus performance data in Bangalore.

What is a drive cycle and why is it worth learning about? In this context, a drive cycle is a representative profile of driving conditions along a single bus route. Such profiles of real-world driving can serve as model inputs to estimate route-level emissions of air pollutants and greenhouse gases, as well as to calculate the costs of alternative technologies at the individual route level. For further modeling analyses, real-world drive cycles provide realistic assumptions and inputs, and are representative of a route’s range of operating conditions, instead of one day’s worth of operations, which could be an outlier. These are the foundation from which to assess how electric buses would perform under realistic conditions for each transit system route.

The ICCT developed the Bangalore route-level drive cycle tool in three steps. First, we collected operations data from buses traveling along four routes; this data included the date of operation, time of day, speed, and elevation. In addition, route-level operations data collected included the total number of buses serving each route, the total length of each route, GPS-based location data of the latitude and longitude of buses on the route, identifying information for each route and bus, schedule information, and the latitude and longitude of all stops. This raw data then went through preparation algorithms to create consistency in format, frequency, and geographic resolution. Quality control and validation removed erroneous data entries and we processed the cleaned GPS data to estimate road grade and vehicle speed at fixed time intervals.

Second, we broke down the cleaned and formatted data into “microtrips” that started and ended when the vehicle had a speed of zero kilometers (km) per hour (h). Multiple microtrips adding up to a predefined time limit were sequenced together to form a candidate cycle, and the candidate cycles were then evaluated based on how closely they resembled the actual full drive cycle. The cycle most representative of the empirical data was chosen as the final drive cycle.

Figure 1. Illustration of a drive cycle for a route developed from real-world operations data.
Figure 1. Illustration of a drive cycle for a route developed from real-world operations data.

Finally, to validate each of the drive cycles we developed, we compared them with the real-world GPS drive cycle data of a bus over the entire route on a randomly selected day. In 10 pairs of Autonomie energy consumption simulation results comparing real-life operations data with drive cycle model output, our drive cycles predicted energy consumption with an average 5% absolute margin of error. The validation results proved that our models are indeed robust and a good proxy to approximate actual energy consumption.

How does this help BMTC? For one, our analysis showed that each of the four studied routes in Bangalore has distinct drive cycle characteristics. Some have average speeds of 27 km/h and maximum speeds of 34 km/h, while others have more stops and are slower, with average speeds of 16 km/h. These characteristics have strong impacts on bus energy needed and fuel consumption, and revealing key differences like these informs decisions that contribute to efficient bus performance along each route, including where to apply bus technologies, how to design schedules, and how to implement charging strategies. For example, low-speed urban routes are more fuel and energy intensive. Refueling or charging strategies for these routes will be different than those for a higher-speed suburban or highway bus route.

These drive cycles are also an important first analysis in the broader TCO model framework, and as a tool they could help transportation policymakers and planners anywhere to identify the best routes for electric buses. Building on this work, the ICCT has already begun modelling route-level energy consumption and conducting a TCO for BMTC’s planned routes for electrification. These analyses will help BMTC make the most efficient decisions regarding routes, bus depots, and fleet procurement. The larger picture developed from the TCO model framework and fleet-wide strategy will be instrumental in formulating plans to navigate the unique challenges of electric buses and achieve successful technology transition at the fleet level.

Sharing advances in best practices regarding research and implementation of soot-free and electric buses is supported by the Climate and Clean Air Coalition’s Heavy-Duty Vehicles Initiative. In this blog, Lingzhi Jin highlights the importance of fleet-wide planning.