Identifying the relationship between tyre performance, fuel consumption and maintenance costs in operating urban bus services: A case study in Sydney, Australia using telematics and fitted sensors

Abstract

Tyres are an expensive input into the cost of a bus business, yet very little is known about what impact an improvement in the performance of tyres, linked in part to improved maintenance practices, might have not only in reducing the costs of tyres but its impact on fuel consumption and overall maintenance costs. This paper uses data collected over a 12 month period from a bus operator in Sydney where we fitted sensors on all tyres of 35 buses monitored on a monthly data capture plan, which together with data obtained from a telematics system installed in each bus and other data describing the operating environment, enabled us to develop a system of equations to identify endogenous and exogenous influences on tread depth loss of each tyre and fuel consumption, and to map this into implications on maintenance costs. The estimated model system is built into a decision support system to investigate ways in which a bus business can improve its cost efficiency through a changed tyre-related maintenance and operating regime. The findings suggest significant opportunities to improve cost efficiency by a careful review of the decisions related to the practices associated with tyres.

Keywords:

• Bus operation
• cost efficiency
• fuel consumption
• maintenance costs
• tyre maintenance

Acknowledgments

This project was undertaken in collaboration with two industry partners: Forrest Coach Lines (ComfortDelGo) and Bridgestone. We acknowledge the contribution of Ben Gao of CDC, Mark Tilley and the maintenance staff at Forest Coach Lines, Helen Lin (to December 2018), David Royle (to November 2018), as well as Bridgestone staff in Tokyo, especially Geert Schoors, Kohei Kumagai, Takeshi Kubota, and Koji Ishida. The project was funded by Bridgestone and the University of Sydney Business School Engaged Activities (EnAct) Support Scheme. Permission to publish this paper by Bridgestone and CDC is acknowledged and we especially thank Tomomoto, Takashi, Senior Vice President, Assistant to Vice President and Senior Officer, CMO, Marketing Solution Strategy, Bridgestone Corporation. We thank two referees for their comments and suggestions to improve the paper.

Disclosure statement

There is no conflict of interest.

Notes

1 In this study we only test power consumption in diesel buses.

2 There are a number of engineering studies focussing only on the tyres which have a tangential relationship with the current paper; however, they do not embrace the larger set of issues that we focus on. See, for example, (Wang et al., 2017) and Klüppel (2014).

3 Australian bus operators rigorously comply with the Vehicle Standard (Australian Design Rule 95/00 – Installation of Tyres) 2018 as detailed in https://www.legislation.gov.au/Details/F2018L01516

4 We excluded the dead running distances because they represent less than 2% of the total kilometres driven, and we did not have enough information on them. For instance, pressure and temperature were not measured when the bus was out of service. After advice from the bus operator and Bridgestone, it was decided to exclude them.

5 An individual bus run refers to every time that a bus leaves the depot and arrives back at the depot. This corresponds with a driver’s shift since there is no hot seat scheduling within FCL operations.

6 Tyre-run level is for all the tyres on the 35 selected buses for each bus-run.

7 Only the front axle is steerable in the buses included in this study.

8 The rigid buses included in this study are 12.5 metres long.

9 The large variation in the RTD periods was mainly because some buses were in maintenance when the tread depth was being measured.

10 We spent a lot of time investigating the potential role of brand but could not find any statistically significant evidence to support any differential impact in the models presented below.

11 This is a dummy variable equal to 1 if the tyre pressure is at its optimal level (between 105 and 115 psi).

12 These numbers represent the percentage of front/back tyres for each type of bus (or all fleet) that are at their optimal level of pressure.

13 These numbers represent the percentage of all the tyres for each bus type (or all fleet) that are in the back outer/inner tyre position.

14 We used the Farrer-Glauber test (Farrar & Glauber, 1967) to test for multicollinearity and did not find any evidence of violation of the test under the F-test and the Chi-square test. The majority of the partial correlation coefficients between the right-hand side variable are less than 0.2.

15 The base level for these dummy variables (1, 0) is the rigid high floor and bendy/articulated bus type.

16 The base level for these dummy variables (1, 0) is the front tyre position.

17 The base level for these dummy variables (1, 0) is the rigid high floor bus type.

18 A number of candidate explanatory variables were tested in this model, such as tyre brand (premium, medium, budget), tyre pressure, weather temperature, and average age of bus drivers. However, their parameters were not statistically different to zero. Equation (6) presents only those variables that were statistically significant.

19 The maintenance cost per km is commercially confidential, so in this paper we include the different items considered as a percentage of the total maintenance costs per km. The average maintenance cost is in the range of AUD$0.50-0.75 per kilometre.

20 The constant of this model provides information on the total maintenance costs per km, which is commercially confidential.

21 This is not the value we have applied for FCL and CDC, which is not available to report.

22 The number of drivers per 1,000 kms is considered independent of the number of drivers per route per vehicle (included in the maintenance cost model). If the user wishes to change the number of drivers per route per vehicle, they can do so using the DSS for the maintenance cost model, as this variable will not have an effect on the dependent variables of the 3SLS model.

23 It is important to recognise that just because the fuel cost savings are relatively small in this illustrative scenario, they are certainly not marginal and indeed the bus operator has indicated that such savings when converted to a total cost per km can be the difference between winning and losing a competitive tender when the time comes to have the contract re-tendered.