By Miguel Simão, Lead Data Scientist at Stratio Automotive 

The UN’s COP 27 conference on climate change this November has once again put carbon footprint reduction pledges in the spotlight. Ambitious plans to reduce greenhouse gas emissions by 55% by 2030 and to ban ICE cars and vans from 2035 are already in place, but with transport considered to be the highest-emitting sector of the UK economy, accounting for 22% of total GHG emissions (113 MtCO2e) in 2019, it’s clear that public transport companies will have a major role to play in reaching these targets. With very little choice in the matter, providers of all sizes will have to make the change within the next few years to ensure business continuity.  

However, it’s not simply a matter of switching to electric vehicles (EV). The technology is still quite new and presents a range of issues that need to be understood and factored into fleet managers’ plans. High upfront cost, insufficient charging infrastructure and limited driving range remain significant barriers to adoption, particularly for public transport providers concerned with protecting the bottom line. While plans to phase out ICE vehicles won’t go away, public transportation providers need a medium-term solution if they are to help achieve a 55% reduction in greenhouse gas emissions by 2030.  

For the deployment of both an interim solution and the successful ultimate transition to EVs, the integration of artificial intelligence and data analytics will be a key tool in public transport’s belt. Modern vehicles and especially EVs have a lot of sensors used to monitor their operation that can be used by artificial intelligence (AI) to provide real-time, actionable insight into the internal health of the buses’ systems. This is the basis for a predictive fleet maintenance solution that can detect and even prevent public transport vehicle breakdown. This improves fleet availability and reliability, and therefore traffic conditions in cities with more efficient bus services – reducing emissions in the process. Moreover, the same approach to harnessing vehicle data can enable dedicated fuel consumption strategies with lower emissions across ICE public transport fleets; helping drivers to improve their driving efficiency so that they spend as little time and fuel as possible on each journey.  

The creation of this level of visibility over a fleet, combined with data analysis and artificial intelligence, is also a great stepping stone on the road to EV transition. Because of the high upfront costs of investing in an electric fleet, and the higher cost of fixing or replacing parts, fleet managers must be able to quickly establish a strategic understanding of how to run an EV fleet in a cost effective way that allows them to make a return on the investment as quickly as possible. Utilising AI and data analysis will be a critical element to achieving this.  

For example, in order to take advantage of an electric bus’ cheaper running costs and lower maintenance requirements, vehicles will need to be operated for longer periods and more intensively compared to traditional ICE fleets. This means keeping vehicles out of the workshop and on the road, as well as extending the life of the more expensive EV components to save on maintenance costs. Vehicle maintenance therefore has to be transitioned from an inefficient preventive approach which involves the replacement of parts before their expected end-of-life to a data-led predictive method. By collecting vehicle data from regular operation to predict the true remaining useful life (RUL) of components, fleet operators are able to safely increase the life of components, thus making vehicle maintenance more predictable and less expensive. By using AI to provide real-time, actionable insight into the internal health of electric buses, maintenance managers can diagnose malfunctions remotely, without having to recall a vehicle off the road and physically look into it. An entire fleet can be monitored remotely, at any time and from anywhere. Vehicle servicing can be optimally scheduled, maximising time on the road, avoiding unexpected breakdowns and extending the operational life of parts.   

These data-driven insights can also help public transport fleet operators overcome problems relating to the limited range of EV vehicles and the need for more frequent charging as the battery degrades. An electric bus that has a maximum range of 300km when new and only needs one charge to complete its daily route may require an additional charging session to complete the same distance after a few years due to battery degradation, creating route planning complications. Moreover, the range of one charge may vary due to uncontrollable factors such as weather, traffic, route and vehicle load. Using advanced data analytics and machine learning to combine battery data with other variables affecting range allows fleet managers to accurately predict a vehicle’s remaining range, and what the expected battery capacity loss will be in the next few years.  

This can be integrated within a predictive maintenance solution in order to understand when the battery must be replaced or when the vehicle has to be operated with shorter routes. By being able to monitor and visualise metrics such as the State of Charge (SoC) optimality over time (the percentage of time spent at a healthy SoC range) and the Depth of Discharge (DoD – how much the battery is discharged between consecutive charges) a fleet operator can understand if the operation profile can be changed to maximise battery life, reducing the total cost of ownership of electric buses.  

This type of data analytics supported by artificial intelligence will be fundamental to a more carbon-friendly and, ultimately, electric approach to public transport. Developing a predictive fleet maintenance solution which analyses vehicle data to enhance efficiency and bring down costs will be central to transport companies’ successful contribution to climate goals everywhere.