In this project, we leverage more than 10 years worth of data from millions of passenger vehicle safety inspections in Pennsylvania to track the deterioration of tire tread thickness. Ensuring that tire tread thickness is greater than 2/32 of an inch for all tires on the vehicle is an explicit criteria mandated by state law during annual safety inspections. When tire tread depth data is tracked over time for an individual vehicle, one can see how tread of tires decreases, and can also make estimates and extrapolations of deterioration relevant to the whole fleet. In this project, we will organize the available data to the vehicle level and attempt to create time series of tire wear for each vehicle over all available records. We will then develop "deterioration curves" for the tire tread values based on the annually updated values. We will then aggregate the vehicle-level deterioration rates to create fleet average deterioration rates for tire tread, and compare these values with those presented in the literature. Using these deterioration rates, we will produce vehicle-level estimates that predict when a car's tread would fall below the specified safety threshold, because if not addressed by purchasing new tires, these cars would be unsafe to drive. More specifically, we care about predicting what percent of cars in a year would be expected to fall below the threshold at some point between the dates of the annual safety inspections.
In this project, we leverage more than 10 years worth of data from millions of passenger vehicle safety inspections in Pennsylvania to track the deterioration of tire tread thickness. Ensuring that tire tread thickness is greater than 2/32 of an inch for all tires on the vehicle is an explicit criteria mandated by state law during annual safety inspections. When viewed over time, one can see how tread of tires decreases. In limited data available from the state on safety inspections (several hundred thousand vehicles per year), the only annual value recorded is the "minimum" value of the 4 tires inspected (which just shows that all other tires are at least as high as that value), while in privately available data from our industry partner CompuSpections (roughly a million records per year), values for all 4 tires are recorded (every year). We will organize the available data to the vehicle level and attempt to create time series of tire wear for each vehicle over all available records. We will then develop "deterioration curves" for the tire tread values based on the annually updated values. We will then aggregate the vehicle-level deterioration rates to create fleet average deterioration rates for tire tread, and compare these values with those presented in the literature. Using these deterioration rates, we will produce vehicle-level estimates that predict when a car's tread would fall below the specified safety threshold, because if not addressed by purchasing new tires, these cars would be unsafe to drive. More specifically, we care about predicting what percent of cars in a year would be expected to fall below the threshold at some point between the dates of the annual safety inspections. Finally, we will consider the "post-inspection" deterioration rate and consider whether slightly higher thresholds should be set in order for a vehicle to pass an inspection. For example, should the inspection value be increased by 1 or 2/32" because many vehicles are otherwise falling below the acceptable level currently, and less would be expected to fall below threshold with higher thresholds. We will communicate the results of our research to interested stakeholders such as the PA Department of Transportation, tire companies, inspection companies, etc.
This project is scheduled to be completed by December 2017. In spring of 2017, I am working with a recently graduated MS student in Civil and Environmental Engineering, Yi (Vickie) Liu, to prepare the database to be used in the analysis, and to perform some preliminary analysis of the data. During the summer I will do some further analysis of the data. In the fall, I have recruited a new PhD student in EPP to assist with the analysis of data and completion of writing tasks.
Deployment is just being considered, but we are interested in speaking with representatives of inspection stations and dealerships to train them to identify opportunities to encourage drivers to get new tires when their vehicles are close to the safety threshold.
Over the course of this one year project, we anticipate developing vehicle-level deterioration rates for vehicles. We want to develop algorithms and data analytics that can be shared with various stakeholders, and to develop messaging for consumers that educates them on the issues of safety as related to tires.
Name | Affiliation | Role | Position | |
---|---|---|---|---|
pacharya@andrew.cmu.edu | Acharya, Prithvi | EPP | Other | Student - PhD |
pf12@andrew.cmu.edu | Fischbeck, Paul | EPP/SDS | Other | Faculty - Tenured |
yi6@cmu.edu | Liu, Vickie | CEE | Other | Student - Masters |
hsm@cmu.edu | Matthews, H. Scott | CEE/EPP | PI | Faculty - Tenured |
No match sources!
No partners!