’Wisdom of the crowd’ saves time and money in the bus traffic
Unplanned stops for heavy-duty vehicles are problematic. Maintaining vehicles proactively can save service costs and improve transportation efficiency. How to early on detect faults and predict component failures for a fleet of city buses has been investigated in new research from Halmstad University.
Buses constantly move around in the city traffic. Sometimes they break down, or stop working in a satisfactory way. Unplanned downtime caused by machine failure costs money, both for the repair itself, but also because of the loss of time, not least for the passengers.
”The purpose of my study is to predict the need of maintenance, to see whenever equipment needs inspection and repairs. And this can be done to prevent severe consequences due to machine failures”, says Yuantao Fan, PhD student in Signals and Systems Engineering who has written the licentiate thesis “A Self-Organized Fault Detection Method for Vehicle Fleets”.
Huge amount of data
The case study within his work focuses on vehicle air system problems of city buses. Moreover, the study shows that the method he has used works for detecting other, various faults.
With a highly digitised electronic system and hundreds of sensors mounted on-board a modern bus, a huge amount of data is generated from daily operations. The data is transferred to a back-office service, a database that is processing the information.
”The data collected reflects the status of the vehicle. The data is analysed and, when the statuses of the buses are compared, you find out which bus is mostly unlike the others. It is assumed that most buses work fine. And if a vehicle behaves abnormally we can judge if it is in a bad condition based on data from a group of buses.”
“Wisdom of the crowd"
This concept is called ’Wisdom of the crowd’ and Yuantao Fan shows that it works equally well as an expert system, which specifically looks for one type of fault that has been predicted by domain expert. Instead, his method can detect many different faults and signals that stand out from the group, also the more ’unthinkable’ ones.
The method, which contains an algorithm that is used on this bus fleet, can also be used in other areas. Examples are machines in factory production lines or power grids where it is important to detect tiny metal cracks. Maintenance that can be automatically arranged well before failure, saves resources for many different actors in society.
Text: Kristina Rörström