Before and After Data Used to Assess Performance of a Predictive Maintenance System Enabled by Intelligent Transportation System Technologies in Ravenna, Italy.
Ravenna, Italy
Testing an innovative predictive management system for bus fleets: outcomes from the Ravenna case study
Summary Information
Maintenance of transit assets can be both costly and time consuming for transit agencies. Predictive maintenance systems may offer a way for transit agencies to reduce maintenance costs while not compromising vehicle availability or safety. Predictive maintenance systems continuously collect data about transit assets and feed this data into an algorithm which, in combination with other data such as weather data, make predictions about when an asset will need repair. The European Bus System of the Future project conducted a pilot of a predictive maintenance system for a transit bus fleet in Ravenna, Italy, and evaluated the results in 2017. A team of researchers installed a series of sensors on transit buses which monitored oil quality in near real-time. The team specifically chose to only monitor oil quality because most engine failures are related to oil contamination. This data on oil quality, along with additional on-board data collected through a vehicle telematics device, were then sent for use in a predictive maintenance algorithm running in a back-office system. The algorithm was then able to make predictions about potential engine failures and transmit alerts to maintenance staff for corrective action.
Methodology
Researchers compared key performance indicators from before and after implementation of the predictive maintenance system, including measuring outcomes in operations and maintenance.
Findings
- Oil required per vehicle per 10,000 kilometers (6,214 miles) was reduced by 31.8 percent when compared to the period before the predictive maintenance system was implemented.
- Breakdowns per vehicle per 10,000 kilometers (6,214 miles) were reduced by 8.08 percent compared to the before period.
- Overall average maintenance time per vehicle per 10,000 kilometers (6,214 miles) did not change. Researchers noted that the type and duration of maintenance activities can vary.
- While maintenance time per vehicle did not change, the labor effort required for data management, measured by monthly workload in full-time equivalent staff, increased by 17 percent. Most of this effort was attributed to tuning the supporting information technology system during implementation, and was expected to diminish during ongoing operations.
