Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Investigation into the improvement of tire management practices Zhou, Jie 2007

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2006-0743.pdf [ 10.59MB ]
Metadata
JSON: 831-1.0081066.json
JSON-LD: 831-1.0081066-ld.json
RDF/XML (Pretty): 831-1.0081066-rdf.xml
RDF/JSON: 831-1.0081066-rdf.json
Turtle: 831-1.0081066-turtle.txt
N-Triples: 831-1.0081066-rdf-ntriples.txt
Original Record: 831-1.0081066-source.json
Full Text
831-1.0081066-fulltext.txt
Citation
831-1.0081066.ris

Full Text

INVESTIGATION INTO THE IMPROVEMENT OF TIRE MANAGEMENT PRACTICES by JIE ZHOU B.Eng. Northeastern University, China, 1992 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES (Mining Engineering) THE UNIVERSITY OF BRITISH COLUMBIA February 2007 © Jie Zhou, 2007 Abstract Mining mobile equipment is becoming larger in order to lower operation cost and increase productivity; however, their performance characteristics are often limited by the tire capabilities. Although considerable progress has been made in the field of tire research, the lack of quantitative research on tire performance characteristics impedes the advancement of off-the-road tires. This is due to the complexity and severity of tire interactions with the surrounding environment and the difficulty in assessing the impact of these interactions on tire life. Longer haul roads, and a severe operating environment make tires more vulnerable than other components of a vehicle. Furthermore, due to a thriving mining industry and a lack of tire production capacity, there is a world wide shortage of tires. As such mines and tire service companies are interested in finding suitable tire shortage coping strategies to improve tire performance, extend tire life and hence lower operating costs. This thesis presents the work done in a collaborative research project between the University of British Columbia and Fountain Tire Mine Service Ltd. This thesis reviews the relevant tire performance research, analyzes tire interactions with the surrounding environment, tire maintenance practices such as managing air pressure and rotation, and operational issues using case studies based on data collected from two mines. The research resulted in enhanced understanding of the impact of some factors and practices on off-the-road tires in mining application, as well the research provide a basis for improved tire management. Table of Contents Abstract 1 1 Table of Contents iii List of Tables vii List of Figures xiii List of Symbols xv Glossary of Terms xix Acknowledgements xxi 1. Introduction 1 1.1 Background 1 1.2 Research Objectives 3 1.3 Thesis Outline 4 2. Literature Review 5 2.1 Off-The-Road Tire Fundamentals 5 2.1.1 Tire basics 5 2.1.2 Tire removal 10 2.1.3 Influencing factors 13 2.1.4 Tire selection 15 2.2 Off-The-Road Tire Care and Maintenance 16 2.3 Tire Performance Research 21 2.3.1 Highway tire research 21 2.3.2 Off-the-road tire research 25 2.3.2.1 Tire-truck interactions 25 2.3.2.2 Tire-ground interactions 27 2.3.2.3 Tire-maintenance interactions 31 2.3.2.4 Tire-environment interactions 33 2.4 Real-Time Tire Monitoring and Control 34 2.4.1 Highway tire sensing 34 2.4.2 Automatic inflation of tires 36 2.4.3 Off-the-road tire monitoring and control 38 3. Fountain Tire Mine Service Limited Research Project 43 3.1 Maintenance Practices 43 3.2 Selected Mines , 47 3.2.1 The Kemess Mine 47 3.2.2 The Gibraltar Mine 48 3.3 Case-specific Objectives and Approach 50 4. Data Analysis 52 4.1 Data Collection 52 4.2 Data Treatment and Assumptions 54 4.3 Methodologies 56 4.3.1 SAS software package 56 4.3.2 Analysis methods 57 5. Results and Discussion 66 5.1 The Kemess Mine 66 5.1.1 Tire Rotation 66 5.1.1.1 Relation between tire life and rotation frequency 66 5.1.1.2 Relation between tire wear rate and rotation frequency 68 5.1.1.3 Tire rotation sequences 70 5.1.2 Rotation Time 74 5.1.2.1 Regression analysis 74 5.1.2.2 Correlation analysis 78 5.1.3 Tire Air Pressure Setting 81 5.1.3.1 M A N O V A analysis 82 5.1.3.2 Tire air pressure spikes analysis 85 5.1.4 Tire Failure Analysis 87 5.2 The Gibraltar Mine 91 5.2.1 Tire Rotation 91 5.2.1.1 Relation between tire life and rotation frequency 91 5.2.1.2 Relation between tire wear rate and rotation frequency 94 5.2.1.3 Tire rotation sequences 98 5.2.2 Rotation Time 104 5.2.2.1 Regression analysis 104 5.2.2.2 Correlation analysis 108 5.2.3 Tire Air Pressure Setting 112 5.2.3.1 M A N O V A analysis 113 5.2.3.2 Tire air pressure spikes analysis 118 5.2.4 Tire Failure Analysis 121 5.3 Comparison of Results between the Two Mines 128 6. Conclusions 134 7. Future Work 137 References 139 Appendix A Fountain Tire Mine Service Limited tire air pressure and tread depth check sheet (Fowler and Huntingford, 2005) 148 Appendix B Examples of IID graphical tests (Kemess Mine) 150 Appendix C SAS output of canonical correlation between the rotation interval hours and the tire life at the Kemess Mine 153 Appendix D SAS output of canonical correlation between the three rotation interval hours and the corresponding three age-specific wear rates at the Kemess Mine.. 163 Appendix E SAS output of tire air pressure M A N O V A analysis at the Gibraltar Mine (40.00R57) 174 Appendix F Equations to obtain results 189 List of Tables Table 2-1 Tire removal summary 11 Table 2-2 Tire tread area removals (Goodyear, 2005) 12 Table 2-3 Estimates of linear regression with statistically significant covariates (15" radial tire) (Krivsov et al, 2002) 23 Table 4-1 Example of M A N O V A data structure 64 Table 5-1 Tire life vs. tire rotation frequency (40.00R57) (Kemess) 67 Table 5-2 Average rotation hours (40.00R57) (Kemess) 67 Table 5-3 Tire wear rate vs. tire rotation frequency (40.00R57) (Kemess) 69 Table 5-4 Age-specific tire wear rate for each rotation period (40.00R57) (Kemess) 69 Table 5-5 Age-specific tire wear rate (mm/hour) according to wheel positions (40.00R57) (Kemess) 69 Table 5-6 Tire life vs. rotation sequence for rotation frequency 1 (40.00R57) (Kemess) 70 Table 5-7 Tire life vs. rotation sequence for rotation frequency 2 (40.00R57) (Kemess) :...71 Table 5-8 Tire life vs. rotation sequence for rotation frequency 3 (40.00R57) (Kemess) 72 Table 5-9 Regression statistics for installation to first rotation (40.00R57) (Kemess)....74 Table 5-10 Regression coefficients for installation to first rotation (40.00R57) (Kemess) 74 Table 5-11 Regression statistics for first to second rotation (40.00R57) (Kemess) 76 Table 5-12 Regression coefficients for first to second rotation (40.00R57) (Kemess)....76 Table 5-13 Correlations between the rotation interval hours between each rotation (R l , R2 and R3) and the tire life (L) (40.00R57) (Kemess) 79 Table 5-14 Correlations between the age-specific wear rates between each rotation ( A l , A2 and A3) and the tire wear rate (W) (40.00R57) (Kemess) 79 Table 5-15 Correlations between the three rotation interval hours and the three age-specific wear rates (40.00R57) (Kemess) 80 Table 5-16 Effect of tire installation air pressure on tire life and wear rate (40.00R57) (Kemess) 81 Table 5-17 Effect of changing pressure on tire life (40.00R57) (Kemess) 82 Table 5-18 Effect of changing pressure on tire wear rate (40.00R57) (Kemess) 82 Table 5-19 Least squares means . comparison of different air pressure (40.00R57) (Kemess) 84 Table 5-20 Least squares means comparison of different air pressure (40.00R57) (Kemess) 85 Table 5-21 Summary of tire air pressure measurement (40.00R57) (Kemess) 86 Table 5-22 Summary of tire air pressure measurement with over 130% rating (40.00R57) (Kemess) 86 Table 5-23 Summary of tire air pressure measurement with over 140% rating (40.00R57) (Kemess) 86 Table 5-24 Tire life comparison according to various positions (40.00R57) (Kemess).. .88 Table 5-25 Failure number comparison according to wheel positions (40.00R57) (Kemess). 89 Vlll Table 5-26 Tire life vs. tire rotation frequency (37.00R57) (Gibraltar) 92 Table 5-27 Tire life vs. tire rotation frequency (40.00R57) (Gibraltar) 93 Table 5-28 Average rotation hours (37.00R57) (Gibraltar) 93 Table 5-29 Average rotation hours (40.00R57) (Gibraltar) 94 Table 5-30 Tire wear rate vs. tire rotation frequency (37.00R57) (Gibraltar) 95 Table 5-31 Tire wear rate vs. tire rotation frequency (40.00R57) (Gibraltar) 95 Table 5-32 Age-specific tire wear rate for each rotation period (37.00R57) (Gibraltar) 95 Table 5-33 Age-specific tire wear rate for each rotation period (40.00R57) (Gibraltar) 96 Table 5-34 Age-specific tire wear rate (mm/hour) according to wheel positions (37.00R57) (Gibraltar) 96 Table 5-35 Age-specific tire wear rate (mm/hour) according to wheel positions (40.00R57) (Gibraltar) 97 Table 5-36 Tire life vs. rotation sequence for rotation frequency 1 (37.00R57) (Gibraltar) 98 Table 5-37 Tire life vs. rotation sequence for rotation frequency 1 (40.00R57) (Gibraltar) 99 Table 5-38 Tire life vs. rotation sequence for rotation frequency 2 (37.00R57) (Gibraltar) 100 Table 5-39 Tire life vs. rotation sequence for rotation frequency 2 (40.00R57) (Gibraltar) 101 Table 5-40 Regression statistics for installation to first rotation (37.00R57) (Gibraltar) 104 Table 5-41 Regression statistics for installation to first rotation (40.00R57) (Gibraltar).. 105 Table 5-42 Regression coefficients for installation to first rotation (37.00R57) (Gibraltar) 105 Table 5-43 Regression coefficients for installation to first rotation (40.00R57) (Gibraltar) 105 Table 5-44 Regression statistics for first to second rotation (37.00R57) (Gibraltar).... 107 Table 5-45 Regression statistics for first to second rotation (40.00R57) (Gibraltar).. ..107 Table 5-46 Regression coefficients for first to second rotation (37.00R57) (Gibraltar).. 107 Table 5 -47 Regression coefficients for first to second rotation (40.00R5 7) (Gibraltar) ..108 Table 5-48 Correlations between the rotation interval hours between each rotation ( R l , R2) and the tire life (L) (37.00R57) (Gibraltar) 109 Table 5-49 Correlations between the rotation interval hours between each rotation (R l , R2) and the tire life (L) (40.00R57) (Gibraltar) 109 Table 5-50 Correlations between the age-specific wear rates between each rotation ( A l , A2) and the tire wear rate (W) (37.00R57) (Gibraltar) 110 Table 5-51 Correlations between the age-specific wear rates between each rotation ( A l , A2) and the tire wear rate (W) (40.00R57) (Gibraltar) 110 Table 5-52 Correlations between the two rotation interval hours and the two age-specific wear rates (37.00R57) (Gibraltar) 110 Table 5-53 Correlations between the two rotation interval hours and the two age-specific wear rates (40.00R57) (Gibraltar) - I l l Table 5-54 Effect of tire installation air pressure on tire life and wear rate (37.00R57) (Gibraltar) 112 Table 5-55 Effect of tire installation air pressure on tire life and wear rate (40.00R57) (Gibraltar) 113 Table 5-56 Effect of changing pressure on tire life (37.00R57) (Gibraltar) 114 Table 5-57 Effect of changing pressure on tire life (40.00R57) (Gibraltar) 114 Table 5-58 Effect of changing pressure on tire wear rate (37.00R57) (Gibraltar) 114 Table 5-59 Effect of changing pressure on tire wear rate (40.00R57) (Gibraltar) 115 Table 5-60 Least squares means comparison of different air pressure (37.00R57) (Gibraltar) 116 Table 5-61 Least squares means comparison of different air pressure (37.00R57) (Gibraltar) 116 Table 5-62 Least squares means comparison of different air pressure (40.00R57) (Gibraltar) 117 Table 5-63 Least squares means comparison of different air pressure (40.00R57) (Gibraltar) 117 Table 5-64 Summary of tire air pressure measurement (37.00R57) (Gibraltar) 119 Table 5-65 Summary of tire air pressure measurement (40.00R57) (Gibraltar) 119 Table 5-66 Summary of tire air pressure measurement with over 120% rating (37.00R57) (Gibraltar) 119 Table 5-67 Summary of tire air pressure measurement with over 120% rating (40.00R57) (Gibraltar) 120 Table 5-68 Summary of tire air pressure measurement with over 125% rating (40.00R57) (Gibraltar) 120 Table 5-69 Tire life comparison according to various positions (37.00R57) (Gibraltar) 122 Table 5-70 Tire life comparison according to various positions (40.00R57) (Gibraltar) 123 Table 5-71 Failure number comparison according to wheel positions (37.00R57) (Gibraltar) 124 Table 5-72 Failure number comparison according to wheel positions (40.00R57) (Gibraltar) 125 List of Figures Figure 2-1 Anatomy of earthmover bias ply tire (Goodyear, 2005) 7 Figure 2-2 Anatomy of earthmover radial ply tire (Goodyear, 2005) 7 Figure 2-3 Tire dimensions (Goodyear, 2005) 8 Figure 2-4 Trends in earthmover tires (Goodyear, 2005) 9 Figure 2-5 Two Piece Assembly (O'Neil, 2003) 10 Figure 2-6 Cause and effect (Elkview Mine et al, 2005) 14 Figure 2-7 Formation of the fore and aft interfacial shears (Lippmann, 1986) 22 Figure 2-8 Relation between load and tire life (25 ton truck, 18x25" tire) (De and Mukhopadhyay, 1989) 26 Figure 2-9 Relation between tire life and loaded speed (100 ton truck, 27.00R49 Tire) (Carter, 1998) 27 Figure 2-10 Wheel locations (Knights and Boerner, 2001) 28 Figure 2-11 Impact No. of failures at different wheel locations (Caterpillar 793C truck, 40.00 R57 tire) (Knights and Boerner, 2001) 29 Figure 2-12 Elements of the tire in 2-D (Fervers, 2004) 30 Figure 2-13 Load deflection characteristic from test and simulation (14.00R20 tire) (Fervers, 2004) 31 Figure 2-14 Relation between tire life and inflation pressure (25 ton truck, 18x25" tire) (De and Mukhopadhyay, 1989) 32 Figure 2-15 Continuous remote real-time tire monitoring (Werner and Barrowman, 2002) 40 Figure 2-16 Integration of tire monitoring information at a mine site (Werner and Barrowman, 2002) 41 Figure 3-1 Tire maintenance services provided by Fountain Tire Mine Service Limited (Fowler and Huntingford, 2005) 44 Figure 3-2 Tire wheel position acronyms 45 Figure 4-1 Example of rotation process acronyms for rotation frequency 1 58 Figure 4-2 Example of rotation process acronyms for rotation frequency 2 or 3 59 Figure 4-3 Canonical Correlation relationships at the Kemess Mine 61 Figure 4-4 Canonical Correlation relationships at the Gibraltar Mine 61 Figure 5-1 Suggested rotation sequence (40.00R57) (Kemess) 73 Figure 5-2 Pareto analysis (40.00R57) (Kemess) 90 Figure 5-3 Suggested rotation sequence for 37.00R57 (Gibraltar) 102 Figure 5-4 Suggested rotation sequence for 40.00R57 (Gibraltar) 103 Figure 5-5 Pareto analysis (37.00R57) (Gibraltar) 126 Figure 5-6 Pareto analysis (40.00R57) (Gibraltar) 127 Appendix A Fountain Tire Mine Service Limited tire air pressure and tread depth check sheet (Fowler and Huntingford, 2005) 148 Appendix B Examples of IID graphical tests (Kemess Mine) 150 List of Symbols A l : Age-specific wear rate of each tire between installation and 1 s t rotation. A2 : Age-specific wear rate of each tire between 1s t rotation and 2 n d rotation. A3 : Age-specific wear rate of each tire between 2 n d rotation and 3 r d rotation. H: Tire running hours. L : Tire life. L a : Average tire life for each rotation acronym. Lai: Tire life of each individual tire under each rotation acronym. L c : Average tire life for each position criterion. L C j : Tire life of each individual tire under each position criterion. Lf: Average tire life for each rotation frequency. L f i : Tire life of each individual tire under each rotation frequency. L p : Average tire life for each installation air pressure. L P i : Tire life of each individual tire under each installation air pressure. N a : Number of tires under each rotation acronym. N b : Number of tires before each rotation. N c : Number of tires under each position criterion. Nf: Number of tires under each rotation frequency. Nj: Number of pressure measurements under each air pressure rating. N m : Number of matching. N m a : Number of spike pressure measurements with all installation air pressure setting N m i : Number of spike pressure measurements with the lowest installation air pressure setting. N m o : Total number of pressure measurements for all wheel positions under each air pressure rating. N m w : Number of pressure measurements for each wheel position under each air pressure rating. N p : Number of tires under each installation air pressure. N p a : Number of pairs of installation air pressure settings. N p s : Number of installation air pressure settings. N r : Number of tires under each rotation period. N r r : Number of rear wheel position rotations. N t 0 : Total number of tires which had experienced spike pressures. N t p : Total number of pressure measurements under all air pressure ratings. N u : Number of tires with uneven remaining tread under each rotation acronym. N w : Number of tires under each wheel position. N w o : Number of worn out tires which had experienced spike pressures. Pi: Percentage of pressure measurements for each air pressure rating. Pi: Percentage of spike pressure measurements with the lowest installation air pressure setting. P m : Percentage of matching. P u : Percentage of tires with uneven remaining tread for each rotation acronym. P w : Percentage of worn out tires which had experienced spike pressures. P w 0 : Percentage of pressure measurements for each wheel position under each air pressure rating. Rl : Rotation interval hours of each tire between installation and 1s t rotation. R2 : Rotation interval hours of each tire between 1s t rotation and 2 n d rotation. R3 : Rotation interval hours of each tire between 2 n d rotation and 3 r d rotation. Rb: Average rotation hours before each rotation. Rbi: Rotation hours of each individual tire before each rotation. T : Tire tread consumption. T a : Average remaining tread depth for each rotation acronym. T a j : Remaining tread depth of each individual tire under each rotation acronym. Tj : Inboard remaining tread depth measurement of an unevenly worn out tire. T 0 : Outboard remaining tread depth measurement of an unevenly worn out tire. T„ : Remaining tread depth of an unevenly worn out tire. W : Tire wear rate. W a : Average tire wear rate for each rotation acronym. W a j : Tire wear rate of each individual tire under each rotation acronym., Wf: Average tire wear rate for each rotation frequency. Wfi: Wear rate of each individual tire under each rotation frequency. W p : Average tire wear rate for each installation air pressure. W P i : Wear rate of each individual tire under each installation air pressure. W r : Average age-specific tire wear rate for each rotation period (interval). W r j : Age-specific wear rate of each individual tire under each rotation period. W w : Average age-specific tire wear rate for each wheel position. W W j : Age-specific tire wear rate of each individual tire under each wheel position. Glossary of Terms Age-specific tire wear rate (mm/hour): A measurement of the speed of a tire tread consumption related to a specific period. It is equal to the tire tread consumption during that period divided by the tire running hours during that period. Even remaining tread (evenly worn out): The outboard and inboard measurements of the remaining tread depth of a tire are the same. Rotation frequency: How many times a tire has been rotated until scrapped, e.g., rotation frequency 0 means a tire has no rotation at all until scrapped while rotation frequency 1 means that a tire has been rotated once until scrapped, etc. Rotation hours: The hours that a tire has run since its installation before each rotation. Rotation interval: The running hours of a tire between two rotations. Rotation sequence: The order of wheel positions according to which a tire is rotated for each rotation since its installation. Rotation time: The accumulated running hours since its installation when a tire is rotated. Scrapped tire: A tire that has finished its service and will not be used anymore. Tire matching: A maintenance practice to keep the tire tread depth difference between rear pairs of tires on a vehicle to less than 10 mm. Tire positions Front positions: These include two wheel positions: LF and RF. Rear positions: These include four wheel positions: LRO, LRI, RRI and RRO. Left positions: These include three wheel positions: LF, LRO and LRI. Right positions: These include three wheel positions: RF, RRO and RRI. Rear inside positions: These include two wheel positions: LRI and RRI. Rear outside positions: These include two wheel positions: LRO and RRO. Rear left side positions: These include two wheel positions: LRO and LRI. Rear right side positions: These include two wheel positions: RRO and RRI. Tire rotation: A tire maintenance practice that moves a tire from one wheel position to another wheel position of the same or a different vehicle. The purpose of tire rotation is detailed in section 4.1. Tire wear rate (mm/hour): A measurement of the speed of a tire tread consumption. It equals to the consumed tire tread depth divided by its running hours when it is scrapped. Uneven remaining tread (unevenly worn out): The outboard and inboard measurements of the remaining tread depth of a tire are not the same. Worn out tire: A tire that has finished its service due to lack of remaining tread depth. The remaining tread depth for worn out varies depending on maintenance practices. Acknowledgements The author is deeply thankful to his thesis supervisor Dr. Robert Hall for his excellent guidance and constant encouragement, which made this thesis possible. The author would like to thank Fountain Tire Mine Service Limited for the financial support and the training for the work presented in this thesis. In particular, special thanks go to Mr. Greg Fowler for his invaluable input and advice. My sincere gratitude also goes to Mr. Ken Huntingford for his vision, guidance and patience on this research project. The author would also like to acknowledge the Kemess Mine and the Gibraltar Mine for their support during this research project. The help from Mr. David Shearer of Fountain Tire Mine Service Limited and Mr. Dave Wright of The Goodyear Tire & Rubber Company is highly appreciated. The love and support from my wife, my son and my parents, even so far away have given me the confidence and motivation to complete this thesis. 1. Introduction 1.1 Background Currently there is a serious world wide large tire shortage due to global economic growth. The rapid development and massive construction "boom" in some Asian countries such as China and India have boosted the demand for raw materials and the prices for most commodities are at very high levels. To take advantage of this trend, mines are reopening, expanding and more new mines are being developed, hence more larger mobile equipment are being purchased. However, mobile equipment performance characteristics are often limited by tire capabilities. Due to the cyclic nature of the mining industry, large tire manufacturers were reluctant to invest to expand their production capacities. As a result, many purchasers are receiving new equipment without tires, and it is not unusual to see some equipment idle awaiting replacement tires in some mines. It is anticipated that the current situation will continue into 2007 (O'Neil, 2006). With low or virtually no inventory of large tires, mines are trying their best to avoid operational losses due to parked equipment without spare tires. In addition to looking at used tires to do recapping or using bias-ply instead of radial ply tires, many mines are paying more attention to tire care programs. Furthermore, tires are in top three along with operator salaries and fuel in open pit mine haulage costs (Knights, et al, 2001). Therefore, it is necessary for mines and tire service companies to modify their tire management approaches in order to adapt to larger equipment, severer operating environments and a need for lower operating costs. One company that has recognized the importance of better understanding how tire management decisions affect tire life is Fountain Tire Mine Service Ltd. As a result of this they have funded a collaborative research project with the University of British Columbia to investigate tire management practices from an engineering perspective in efforts to develop strategies to further improve tire life. This thesis presents the work based on data collected from the Kemess Mine and the Gibraltar Mine. Analyses on tire rotational practices such as rotation sequence and rotation time, tire failures and tire air pressure setting at installation are presented and discussed. 1.2 Research Objectives The primary objective of this research is to improve tire management through the application of engineering research and analyses techniques. The focuses of the research are as following: • Enhance the understanding of tire maintenance practices, relevant mine operation and their impact on tire performance by analyzing tire maintenance practices and relevant operational issues. • Expand the knowledge Of off-the-road tire management for further improvement by determining the optimum set points and control limits for major variables. • Establish an analytical approach to analyze tire performance data. • Develop the application of analytical methods in the field of off-the-road tire research. • Understand and quantify the influences of major maintenance factors on tire life. • Examine and modify current tire management approaches if needed. 1.3 Thesis Outline This thesis consists of eight chapters and six appendices. A brief introduction of these chapters is as following: Chapter 1 introduces the research background and research objectives. Chapter 2 presents basic knowledge on off-the-road tires and relevant tire research, monitoring and control. Chapter 3 describes current Fountain Tire Mine Service Ltd. tire maintenance practices, introduces the tire research project, the selected mine sites, case-specific objectives and research approach. Chapter 4 explains data collection, treatment, assumptions and adopted methodologies. Chapter 5 shows analytical results for the two mines and presents a discussion and comparison on four parts: tire rotation, rotation time, tire air pressure setting and tire failures. Chapter 6 presents conclusions. Chapter 7 describes the future work. 2. Literature Review 2.1 Off-The-Road Tire Fundamentals " 2.1.1 Tire basics Different from highway tires, today's off-the-road tires are generally large, heavy tires used for various types of operations such as in mining, forestry and agricultural industries. Off-the-road tires have a variety of applications. They are used for graders, loaders, scrapers, rigid frame haul trucks and articulated haul trucks. Other applications include mobile cranes, airport emergency equipment, port handling equipment, logging vehicles, special steel mill slab/slag vehicles, and military vehicles. Diversified applications demand various designs in terms of tread pattern and depth, compounding, structure etc. Because of the complexity and severity of their interactions with the working environment, off-the-road tires often utilize the most advanced technology available in the tire industry. There are a vast range of off-the-road tire sizes; this research is mainly about tires used on rigid frame haul trucks with payloads up to 300+ tons. Two Goodyear tire sizes are involved in this research: 40.00R57 and 37.00R57. The tire types for these two sizes are RL-4B and RL-4M+ respectively. The number before dot (40 or 37) is the overall width of the tire in inches; 00 is the 96% aspect ratio (section height/section width); R means radial construction and 57 is the rim diameter in inches. RL means rock lug or radial lug; 4 means E-4 industry code and B and M+ represent internal design sequence (Goodyear, 2005). There are two types of tire construction: bias ply and radial ply. Historically, bias ply tires have been used in most new haulage trucks and usually their services are good in various working environments. However, in many cases operation costs can be lowered through utilization of radial tires. Figure 2-1 and Figure 2-2 are anatomies of earthmover bias ply and radial ply tires respectively. Because of radial ply tire's distinctive structure and compounding, it provides many advantages over bias ply. They include: • "Longer tread wear for reduced operating costs. • Cooler running for higher speeds, increased fleet output. • Excellent traction. • Improved penetration resistance for reduced downtime. • High flotation (tire road resistance capabilities in soft underfoot conditions) for superior mobility. • Improved vehicle cushioning, resulting in lower maintenance costs. • Superior operating economy through a 5-8% reduction in fuel consumption" (Goodyear, 2003). Flippers Figure 2-1 Anatomy of earthmover bias ply tire (Goodyear, 2005) Sidewall Figure 2-2 Anatomy of earthmover radial ply tire (Goodyear, 2005) Figure 2-3 shows tire dimensions in both unloaded and loaded conditions. As it indicates, when a load is applied, tire radius changes which results in sidewall deflection. Tires are designed and allowed to operate with a certain percentage of deflection. Section W i d t h T Outside Diameter Rim Flange Height Section Height I f . Rim Width Loaded Radius Dual Spacing Loaded Sectio Width Total Footprint Length Width Figure 2-3 Tire dimensions (Goodyear, 2005) With the increase in equipment size tire manufacturers have been under pressure to develop larger and more robust products utilizing advanced technology available. Figure 2-4 shows the trends in earthmover tires. % B D .:3 :0 Cotton Cord Rayon Cord First Tube!ess EM Tires R^dia! fiWTires; Nylon Cord 15° Tire 30;00-33 13.00-24 21.00-35 First Wide 1 8 0 0 _ 3 3 Base Tire 29.5-25 40.00-57 0 0 4 9 ^ 37.00R5? 36.00-51 0 M35? 1940: 1945 1950 1955 1960 1965: 1970 1975, 1980 19.85 1990 1995 2000 2005 2010 Year Figure 2-4 Trends in earthmover tires (Goodyear, 2005) Because the tire treadbelt usually has shorter life than the casing, Two Piece Assembly was designed by Goodyear in that the treadbelt can be separated from the casing as shown in Figure 2-5. Aiming at longer tire life, Two Piece Assembly offers flexible tread designs and larger footprints. The following are advantages of Two Piece Assembly: • No cracks can migrate from the treadbelt to the casing. • Higher temperature performance ability. • Air migration-resistant from the air chamber to the treadbelt. • Less replacement time (3 hours in contrast to 8 hours for conventional tires) (O'Neil, 2003). Figure 2-5 Two Piece Assembly (O'Neil, 2003) Reprinted with permission of Mining Engineering Magazine and the Society for Mining, Metallurgy and Exploration Inc. 2.1.2 Tire removal Due to the complexity of tire interactions with the surrounding environment, there are various tire removal reasons. Basically these removals can be summarized into five categories (Goodyear, 2005) as shown in Table 2-1. It is apparent that understanding various tire removal (failure) reasons and their characteristics helps make a tire management program more efficient. Goodyear (2005) summarized tire removals for tread area, shoulder area, sidewall area and bead area. Table 2-2 is a summary of tire tread area removals. Table 2-1 Tire removal summary Removal category Removal reasons in each category Tread area Cuts Separations Shoulder area Cuts Separations Sidewall area Cuts Separations Radial splits Bead area Separations Flange erosion Cracking Liner area Splits Lifting liner Wrinkled liner Loose fabric Table 2-2 Tire tread area removals (Goodyear, 2005) Distinguishing characteristics Rock cut Separated Liner damage Belts exposed Heat evident Rubber reversion Worn out No No No Yes No No Cut separation Yes Yes No Yes No No Tread impact Yes No Yes Yes Yes No Tread cut through Yes No Yes No No No Heat separation No Yes No No Yes Yes T K P H related separation No Yes No No Yes No Tread chunking Yes No No Yes No No Lug base cracking No No No No No No Stone drilling Yes No No No No No Tongue out Yes Yes Yes Yes Yes No Delaminating No Yes No No No No 2.1.3 Influencing variables Because of off-the-road tire operation complexity, their performance depends on more than just one influencing factor. Considering the complete surroundings, influencing factors on tires can be divided into the following four categories (Zhou et al, 2006): • Tire-truck interactions. These include truck payload, truck size, truck mechanical structure, truck speed, wheel position, tire make, etc. • Tire-ground interactions such as tire friction, haul road profile, etc. • Tire-maintenance interactions. These include tire alignment, tire rotation, tire air pressure setting, preventative maintenance, etc. and • Tire-environment interactions such as climate, snow, rain, etc. Figure 2-6 is a detailed description of major variables from Elkview Mine et al (2005). Figure 2-6 Cause and effect (Elkview Mine et al, 2005) 2.1.4 Tire selection It can be seen from 2.1.3 that tire selection involves a great deal of complexity. The selection of a proper tire includes evaluating the following parameters: TKPH (Ton-Kilometer per Hour, a rating system that rates a tire's ability to perform work from a heat perspective), ply rating, tire size, tread pattern and depth, inflation pressure, vehicle speed, haul road profile, payload, weather conditions, tire life and tire cost (De and Mukhopadhyay, 1989). Goodyear recommended that three basic tire requirements be considered: "suitable physical size to well fit within the wheel; be able to carry the proper load, say, at least G V W (Gross Vehicle Weight); and meet minimum traction requirements" (Goodyear, 2004). As well any tire selection should follow the Tire and Rim Association (TRA) standards. 2.2 Off-The-Road Tire Care and Maintenance Currently, a lack of quantitative research on off-the-road tires impedes the development of tire care and maintenance management programs. According to Michelin's Earthmover Group (1999), "Tires represent the largest portion of a haul truck's hourly operating expenses for the surface mining business. In some cases the cost can be as high as $50 per hour, which is twice the cost of fuel". Therefore, it is necessary for mines and tire service companies to modify their tire management approaches in order to adapt to larger equipment, severer operation environments and demand for lower operation costs. A sound tire management program includes a correct and regular job site survey, the right tire replacement and maintenance, and most importantly the understanding of tire performance. It is a comprehensive management system from a perspective of tire, vehicle mechanical system and operation. The following is a brief review of major tire management components. -Job site survey With regards to job site survey, haul roads play a key role in reducing tire wear and tire impact failures. The road design variables such as curve, crown, slope angle and road course materials etc. have significant influences on tire performance. On the other hand, even though the haul road is correctly maintained, poor loading and dumping areas can destroy tires. Knights et al (2001) pointed out that loading or dumping zones caused frequent tire impact failures at a mine. They also made some recommendations such as "superelevate the surface of the dumping zones to reduce the ground pressure on the left tires when turning". To systemize and optimize job site survey, many companies have designed distinctive job site survey forms such as those by the Goodyear Tire & Rubber Company (Goodyear, 2005). -Correct maintenance Correct tire maintenance will extend tires life and strengthen their performance. These include correct mounting and matching procedures, suitable air pressure setting, etc. Some of the maintenance operation may seem trivial but important. For example, truck suspension systems influence tire alignment. Even a 1/10" misalignment of tandem axles will have a significant impact on highway tire costs because of its connection with tire wear (Kal Tire, 2005). A master gauge should be used monthly to adjust the accuracy of air gauges (Fowler and Huntingford, 2005). Tire air filling is an important aspect of correct tire maintenance. Recently, nitrogen tire inflation systems have been of interest in mines. Nitrogen inflation has the following advantages: • The tire air pressure can be maintained constant longer due to nitrogen's inert characteristics. • Nitrogen runs cooler which results in less heat buildup and therefore improves tire performance. • Nitrogen is a dry gas with no moisture which has a reduced oxidation on the rim and tire inner liner. • Because nitrogen is non-combustible and non-corrosive, nitrogen inflation is more environmentally safe (Sympson, 2006). Preventive maintenance is far more important than run-to-failure policy in the case of tires. Daily air pressure checks, regular tread depth checks, cycle time records, weight studies and T K P H (Ton-Kilometer per Hour) analyses provide mines with valuable information in assessing a tire's life and developing a corresponding replacement policy. Unfortunately, productivity is always the first priority in mines and makes preventive maintenance hard to run efficiently at times. The right equipment and tools are conducive to tire maintenance. For re-treading in the factory, one interesting electronic equipment for checking tire casing is the NDI Casing Analyzer. This equipment detects whether there are any air leaks or casing separations using ultrasonic sound waves (Kal Tire, 2005). -Record keeping It is obvious that tire management software facilitates the tire management process. One of these software packages is Goodyear's E M T R A C K III, a tire and wheel tracking software which records tire types, air pressure, tread depth, disposition, removal reason etc. and generates tire performance analysis and cost reports. TyreArm™ is a tire application risk management software package developed by Klinge & Co. Pty Ltd. It is used to record and analyze tire performance information. It can also generate tire and rim maintenance alerts automatically. The exports from this software are compatible with some reliability software such as Weibull R (Rasche et al, 2003). -Human factors There is no doubt that truck drivers, shovel and loader operators play important roles in tire care. Truck over speeds, frequent braking, overloading, over watering etc. result in reduced tire performance. "Tread wear and casing life differed by as much as 30 percent among drivers operating the same equipment while hauling similar loads over identical routes" (Pit & Quarry University, 2001). -Understanding tire performance One important aspect of a tire management program is the understanding of tire performance. Valuable information can be gained by doing analyses on tire interactions, maintenance practices such as rotation and air pressure setting and operational issues. This information can be used to assist mines or tire service companies in the formulation of improved tire management. Martin Fester (Chadwick, 2005) mentioned that there are three categories of factors that affect equipment profitability and mobility: tire life-cycle cost factors such as wear resistance and reparability; equipment productivity increase factors such as reduced downtime and load capacity; and operator's job comport and safety. He also pointed out that understanding these basics are conducive to improving modern earthmover tire management. 2.3 Tire Performance Research 2.3.1 Highway tire research Although off-the-road tires have evolved for more than half a century, limited published research is available. In contrast much research can be found in the field of highway tires. This research includes tire friction analysis, tire skid resistance, tire traction test and traction measurement, braking test, fatigue test, hydroplaning and viscoelasticity, stress distribution, etc. (Ludema, et al, 1973; Janowski, 1983; Lippmann, 1986). Some of this research is quite detailed such as the modeling of airplane tires. In this research, eight independent variables: depth of macro-texture, depth & density of contaminant, speed, tire inflation pressure, vertical loading, and nominal tire width & diameter are used to model the friction (Balkwill, 2003). Skid resistance is an area with lots of research. Generally weather has a very significant effect on skid performance. It is noted that the skid resistance is the minimum on wet ice. Road skid resistance changes with ambient temperature that influences tire tread rubber characteristics. It is also observed that the tire-road combination friction-temperature gradient depends on both the road surface texture and the tire tread compound (Ludema andGujrati, 1973). Stress distribution between tire tread and road is another highway tire research area. Figure 2-7 shows the formation of the fore and aft interfacial shears. As is shown, there exists a speed difference between the two adjoining and parallel surfaces. As a result a linearly increasing shear of the intervening material is generated (Lippmann, 1986). Figure 2-7 Formation of the fore and aft interfacial shears (Lippmann, 1986) Reprinted, with permission, from STP 929-The Tire Pavement Interface, copyright ASTM International, 100 Barr Harbor Drive, West Conshohocken, PA 19428 With regards to traditional reliability modeling, most mobile equipment reliability analyses are on the complete machine or some important components such as an engine. Tires are only treated as a part of the reliability analysis. Krivsov et al (2002) did tire reliability analyses using a regression approach. Their research focused on tread and belt separation (TBS). By using tire geometry and physical properties as potential tire life influencing factors, a Cox survival regression model was established through laboratory test data. Because the t values are negative and the P-values for peel force and tread radii in Table 2-3 are smaller than 0.05, the authors concluded that the tread radii and the peel force are significant factors that affect TBS failures inversely. This means the TBS failure rate will decrease when the tread radii and the peel force increase. On the other hand, the TBS failure rate will increase when the operating time increases (Positive t value and less than 0.05 P-value). Table 2-3 Estimates of linear regression with statistically significant covariates (15" radial tire) (Krivtsov et al, 2002) Reprinted from Reliability Engineering & System Safety, Vol. 78, Krivtsov et al, Regression Approach to Tire Reliability Analysis, Page 272, Copyright (2002), with permission from Elsevier Beta Standard error t value P Value Constant 126.173 17.304 7.292 0.0000 Peel force -1.213 0.184 -6.597 0.0000 Tread radius 1 -0.073 0.016 -4.729 0.0000 Tread radius 2 -0.166 0.071 -2.327 0.022 Operating time 0.163 0.015 10.948 0.0000 Tire tread wear is a major factor of ultimate tire removal. Veith (1986) classified wear mechanism into five categories: adhesive wear, abrasive wear, erosive wear, corrosive wear and fatigue wear. He also defined the rubber friction as F=yaFa+YdFd+YwFw where F a , Fd, and F w are the frictional forces caused by adhesion, deformation, and wear or abrasion respectively; ya, jd and y w are the corresponding coefficients. Tire tread wear rate is one important tire performance index. Brenner et al (1975) did some tests and concluded that the intrinsic wear rate of tires was constant and there was no sign of a significant difference between new and used highway tires. The authors adopted a linear regression method to find out the relationship between tire tread depth and tire running mileage. This method was suggested to avoid problems resulting from the use of arithmetic or geometric mean. Various factors such as road profile, climatic conditions, vehicle mechanical system, and drivers etc. affect tire tread wear rate. For highway tires, wear rate of cutting abrasion increases linearly with load while wear rate of frictional or fatigue abrasion increases exponentially with the load. Ambient temperature's effect on tire wear rate is different: cutting abrasion wear rate sharply decreases from high to a minimum at subfreezing temperatures, and then gradually increases with rising temperature. Fatigue or frictional abrasion wear rate keeps decreasing when temperature rises (Ambelang, 1973). Veith (1995) pointed out that the wear rate of tire tread compounds is a complex physical-chemical process. It depends not only on the tire tread compounds properties but also on certain tire application, operational and seasonal factors. He also demonstrated the effect of varying temperature and seasonal conditions on tire tread wear performance characteristics. Different tire air pressure would cause uneven load distribution among tires, and therefore influence tire tread wear rate. Huhtala et al (1989) pointed out that as the tire inflation pressure increased, the road strain values (damage to the road) increased and the truck tire centre had the highest contact pressure between the tire and the pavement. 2.3.2 Off-the-road tire research Due to different characteristics between highway and off-highway tires, the different environments they operate in, and more specifically off-the-road tire operation complexity, highway tire research only provides some basic concepts and methodologies for off-the-road tire research. The following off-the-road tire research is presented according to four tire interactions as explained in section 2.1.3. 2.3.2.1 Tire-truck interactions Load is a very important factor in tire-truck interactions. Overloading is often encountered in mines due to maximized production, uneven load distribution, varying material density, etc. Exceeding T R A (Tire & Rim Association) and tire manufacturers' recommendations, overloading results in more tire heat build up and shortened tire life. Figure 2-8 shows the relation between load and tire life as determined by De and Mukhopadhyay (1989). The figure has been removed because of copyright restrictions. The X axis represents tire life in hours and the Y axis represents percentage of tire load. The tire life increases from approximately 1750 hours to 8250 hours in a concave shape when the percentage of load decreases from approximately 140% to 70%. Figure 2-8 Relation between load and tire life (25 ton truck, 18x25" tire) (De and Mukhopadhyay, 1989) Obviously tire life reduces when speed increases. Higher speed, longer distance combined with overload lead to tire overheats and eventually premature failures. TKPH (Ton-Kilometer per Hour) is one indicator of such a combination. Figure 2-9 shows the relation between tire life and loaded speed as determined by Carter (1998). 0 10 20 30 40 LOADED SPEED {MILE HOUR) Figure 2-9 Relation between tire life and loaded speed (100 ton truck, 27.00R49 Tire) (Carter, 1998) 2.3.2.2 Tire-ground interactions For tire-ground interactions, research has been carried out on the relationship between haul roads and tire failures. By using statistical correlation between off-the-road tire failures and open pit haulage routes Knights et al (2001) compared the number of failures before and after haulage routes change for a large copper and molybdenum mine in Chile. Here, the authors assumed the failure data distribution was Poisson and F tests were conducted to find out whether correlation existed between haulage roads and these failures at certain confidence levels. An F test was also carried out to test the impact failures with respect to different wheel locations (Figure 2-10 and Figure 2-11). Furthermore, a distribution was used to quantify the effects of the traffic change on tire life. After identifying these relations, corrective action was recommended to modify truck and shovel operator practices, adopt a suitable corresponding maintenance care program, and redesign and modify the dumping area (Knights and Boerner, 2001). Figure 2-10 Wheel locations (Knights and Boerner, 2001) Reprinted with permission of Mining Engineering Magazine and the Society for Mining, Metallurgy and Exploration Inc. 15 14 | 1 2 § 10 I 8 6 « z 4 2 0 5 a 4 Tire positron 6 Figure 2-11 Impact No. of failures at different wheel locations (Caterpillar 793C truck, 40.00R57 tire) (Knights and Boerner, 2001) Reprinted with permission of Mining Engineering Magazine and the Society for Mining, Metallurgy and Exploration Inc. There is a lot of research on mine haul road design. They provide mines with basic guidelines in designing and maintaining haul roads (Kaufman and Ault, 2001; Tannant, D. D., and Regensburg B., 2001; Thompson and Visser, 2002). The importance of haul road design and maintenance was explained by De and Mukhopadhyay (1989), "Properly designed and maintained haul roads lead to reduced fuel consumption, higher vehicle speed, longer tire life, and more comfortable and safer riding. Haul road surface materials determine the rolling resistance and tractive coefficient which in turn affects vehicle speed and gradeability. For each addition of 1% grade the vehicle has to overcome lOkg/ton of gross weight". Another interesting approach for modeling tire-soil interaction is F E M (Finite Element Method) simulation. Fervers (2004) developed a 2-D (two dimensional) F E M model that reproduced mechanically the basic components of a tire. In this model, the tire elements are tread, belt, carcass, air filled volume and rim (Figure 2-12). Load deflection characteristics from test and simulation were compared (Figure 2-13). The relation between soil deformation, compaction and tire pressure on different soil types were also covered. The relation between soil deformation, compaction and tire pressure on different soil types were also covered. The figure has been removed because of copyright restrictions. This figure is a two dimension drawing of a tire. The elements in this drawing is tread, belt, carcass, air filled volume and rim. Figure 2-12 Elements of the tire in 2-D (14.00R20 tire) (Fervers, 2004) The figure has been removed because of copyright restrictions. The X axis represents tire deflection in mm and the Y axis represents the wheel load in kN. There are three dashed concave lines and three solid concave lines with three pressure settings (7.5 bar, 2.5 bar and 1.5 bar). The tire deflection increases in a concave shape for all three pressure settings when the wheel load increases. For the same wheel load, the tire deflection increases when the tire pressure decreases. For the same tire deflection, tires with higher pressure have the higher wheel load. Also three solid lines from simulation almost coincide with the three dashed lines from test. Figure 2-13 Load deflection characteristic from test and simulation (14.00R20 tire) (Fervers, 2004) 2.3.2.3 Tire-maintenance interactions Tire air pressure setting at installation is usually recommended by the tire manufacturer. As discussed in 2.3.1, tire air pressure settings influence tire performance such as tire life and tread wear rate. However, there is a lack of relevant published research on off-the-road tires. Since tire air pressure actually supports the vehicle weight, it is believed to be one of the most important indexes in terms of tire performance. Proper tire air pressure setting according to the specific working environment is one of the keys to longer tire life. Both over-inflation and under-inflation can cause premature tire failures. A n over-inflated tire has a narrower contact with the road which results in premature wear on one tread area; on the other hand, an under-inflated tire leads to excessive sidewall flexing and heat buildup (De and Mukhopadhyay, 1989). Pit & Quarry University (2001) pointed out that "the operation of pneumatic tires with too much or too little air is similar to the operation of an engine with too much or too little oil". Figure 2-14 shows the effect of under-inflation on off-the-road tire life. The figure has been removed because of copyright restrictions. The X axis represents the percentage of inflation and the Y axis represents the tire life in hours. The tire life increases linearly from approximately 1000 hours to 5000 hours when the percentage of inflation increases from 60% to 100%. Figure 2-14 Relation between tire life and inflation pressure (25 ton truck, 18x25" tire) (De and Mukhopadhyay, 1989) Tire rotation is one important tire maintenance practice. It includes rotation sequence and rotation time. Currently no published research on off-the-road tire rotation sequence and rotation time is available. It is the author's belief that both tire tread wear and tire failures (including damage) decide the rotation time, therefore both factors should be considered when determining the rotation time. In order to obtain longer tire life, more worn out and less scrapped tires due to failures are desired. This means a lower wear rate is conducive to longer tire life. 2.3.2.4 Tire-environment interactions Environment factors influence tire performance. This can be in the form of effects on tire air pressure, tire inside temperature, etc. As well Blackwell (1996) demonstrated that tires were subject to the most damage with the presence of water, typically occurring in May, June and July at one mine. However, nothing else could be found in the published literature on environmental effect on off-the-road tire life. 2.4 Real-Time Tire Monitoring and Control 2.4.1 Highway tire sensing Highway tire sensing is a field with a lot of research and advancements. A major purpose of this technology is the continuous monitoring of tire air pressure or tire inside temperature. Other applications include the measurement of tire deformation, tire friction, etc. Continuous tire air pressure monitoring leads to many benefits such as increased fuel efficiency, more comfortable ride, reduced tire wear and reduced number of accidents, etc. Conventional wireless tire sensing systems use an active sensor to send encoded signals to an outside receiving system through antenna. However, an active sensor needs a battery or a separate electromagnetic coupling to transmit energy. The battery has to be durable or the additional bulky energy transmission needs large equipment (Grossmann, 1999). To solve the aforementioned problem, in the last decade, SAW (Surface Acoustic Wave) sensor has been developed as one M E M S (Micro-Electro-Mechanical Systems) device. The difference between SAW and other M E M S devices is that it does not have any moving parts. Since the SAW sensor is completely passive, it does not need an additional battery (Varadan, et al, 2000). A SAW sensor consists of a piezoelectric substrate with two IDTs (inter-digital transducers) on its surface. These two IDTs are used to convert signals between voltage variations and elastic mechanical waves. The sensor is mounted on the tire valve stem through which it is interrogated by an antenna (Stelzer, et al, 2001). The SAW sensor can be made of different materials. Grossmann (1999) suggested the use of quartz crystals as passive sensors. The theory behind this is the tire air pressure change would influence the natural frequency of a quartz crystal. The sensor is fixed at the rim of the wheel and is connected to the outside antenna at the axis. According to the author, this system is simple, reliable and compact. Varadan, et al (2001) suggested the use of a polymer based SAW sensor which is compatible to the silicon housing of the tire valve stem because of their similar material properties. Another approach to monitoring tire air pressure is the measurement of vibration. Craighead (1997) proposed measuring wheel/axle subsystem vertical vibration when a vehicle is moving to detect tire air pressure change, and to determine the degree of wheel unbalance through transducers installed on the wheel. The tire pressure change is sensed by the change of tire stiffness, and the wheel unbalance can be determined by the fundamental wheel rotational frequency and its harmonics. Craighead (1997) also mentioned sensing tire air pressure decrease based on the change of tire rolling radius. Because the signals are obtained from A B S (Advanced Braking System) to reduce additional cost of installing sensors, its application is limited to the vehicles with A B S . Currently, the majority of tire air pressure monitoring systems sense internal tire inflation pressure. This may cause inappropriate warnings due to the decrease of internal tire temperature (Barbanti, et al, 2004). In their paper, the authors presented a new concept: inflation state which is the difference between tire internal absolute pressure and external environment absolute pressure. Because the tire inside temperature and outside temperature influence these two pressures respectively, warning threshold air pressures were established based on outside temperature variations. For tire stress or deformation measurement, a novel approach was proposed by Sergio et al (2003). This approach continuously measured a mechanical deformation of the tire using the tire itself as a sensor and the steel wires inside a tire as impedance electrodes. In this way, the tire deformation can be detected by an impedance change (the space between the layers of steel wires). According to the authors, this method avoided the disturbance caused by embedded SAW sensors. 2.4.2 Automatic inflation of tires The purpose of automatic inflation of tires is to adapt the moving vehicle to various road conditions by changing its tire air pressure through the use of a central tire inflation system (CTIS). Originally developed for military applications before World War II, CTIS has been used in other industries. One area that has widely applied CTIS is the forestry industry (Mills, 2004). The use of CTIS has resulted in many benefits such as increased vehicle slope climbing capability, less damage to the roads, reduced rolling resistance, etc. CTIS has five main components. These components are air pressure unit, air lines to transmit air to the tires, air control valves, rotary union hardware and computerized control interface to select suitable tire air pressures according to specific operating situations (Owende, et al, 2001). Traction is an important index of vehicle mobility. By conducting experiment on sandy and loess soil surfaces, Pytka, et al (2005) concluded that reducing tire inflation pressure on a military truck from 390KPa to 200KPa influenced the off-road traction by increasing DBP (drawbar pull) at 20% and 31% for sand and loess respectively. Bradley (1993) also tested CTIS in western Canadian log-hauling conditions. The test found a 39% traction gain on loose gravel and a reduced tire wear rate as well. The effectiveness of variable tire pressure (VTP) on thin pavements and gravel roads were investigated by Shalaby, et al (2002). The authors pointed out that lower tire pressure reduced rut depth. This minimizes the damage of vehicle on soil structure. Because reducing tire inflation pressure leads to a larger contact area and increased sidewall deflection, it mitigates the tire sinkage and minimizes the tire slip (Trenn, 1997). Reducing tire inflation pressure also results in reduced resonant frequency and therefore a smoother ride. Adams, et al (2004) wrote that CTIS improved the ride of an agricultural vehicle by 99% on average based on ISO 2631 standard. Douglas, et al (2000) did testing on road dynamic contact stresses under tires with CTIS. They concluded the vertical contact stresses were non-uniform under heavy load and low pressure. The authors thought the differences in transverse and longitudinal contact stresses of a tire subject to low and high inflation pressures may be a reason to use CTIS. 2.4.3 Off-the-road tire monitoring and control As mentioned before, regular (daily preferred) tire air pressure, tread depth and tire inside temperature checks are an indispensable part of sound tire management. These checks are usually conducted when vehicles come in for refuel; however sometimes a truck has to stop halfway for these checks at the cost of lost production. It is estimated that in a typical operation, about half a man-year or about 30,000 US dollars in labor costs is spent on checking those tires manually (Pit & Quarry University, 2001). Being such a labor intensive and limited method, manual checking can not reflect tire performance in a "continuous" way. Since the late 1990s, several OEMs (Original Equipment Manufacturer) and a few independent companies have introduced tire-monitoring systems. Among these are: Michelin's Earthmover Management System (MEMS), Goodyear's Intelligent Off-The-Road tire monitoring system (IOTR), Rimex's TyreSense and Fuller Brothers' Tire Analysis System (TAS). To date, TAS has been one of the most successful (Werner and Barrowman, 2002). One common thing about these tire monitoring systems is that they all involve pressure and temperature sensors. M E M S , IOTR and TAS place sensors inside the tire (Werner and Barrowman, 2002) while Tyresense screws the sensor directly into the wheel or rim(Rimex, 2006). They can then be linked to intelligent mining systems through on-board receiver and wireless communication network. Compared with tire data downloading by a handheld unit or at a fixed location, this is a big step forward. Figure 2-15 is a depiction of remote tire monitoring. Figure 2-15 Continuous remote real-time tire monitoring (Werner and Barrowman, 2002) Published in the proceedings of the CIM Annual General Meeting—Vancouver 2002. Reprinted with permission of the Canadian Institute of Mining, Metallurgy and Petroleum There are some differences among these system interfaces: " M E M S data is polled by Modular Mining System on a regular basis; while IOTR data is sent to Modular Mining System at a fixed frequency; TAS interface sends event driven messages and supports the functionality to poll for tire data" (Werner and Barrowman, 2002); and TyreSense's minimum data logging is 15 minutes (Rimex, 2006). It is obvious that many benefits will be obtained from an integration of tire sensing into a mine management system as shown in Figure 2-16. Through this system, variables such as pay loads, vehicle speed, haul distance and tire temperature, etc. can be displayed together and their relationship will be better understood. This provides insight into tire life and performance control such as load distribution, tire failure location, tire failure detection and air temperature threshold setting, dispatch decision, etc. As a result, quick subsequent actions can be taken to improve tire life. P OPS Off Highway Haiti Truck COtitecSm -tk era A H^f Sftttm fiqrMr Tift Srsitm M i l Efitsm mm Sitt Centra! Computer Operational information 4 EBP System Figure 2-16 Integration of tire monitoring information at a mine site (Werner and Barrowman, 2002) Published in the proceedings of the CIM Annual General Meeting—Vancouver 2002. Reprinted with permission of the Canadian Institute of Mining, Metallurgy and Petroleum One attractive aspect of real-time tire monitoring is that T K P H (Ton-Kilometer per Hour) and other tire-specific algorithms are used in the tire monitoring systems or mine management systems to help make more informed decisions (Ednie, 2002; Werner and Barrowman, 2002). Some mines and companies testing and using real-time tire monitoring and control systems are: Fording River (TAS), Syncrude (MEMS) (Ednie, 2002), Kemess Mine (TyreSense) (Fowler and Huntingford, 2005) and Quebec Cartier Mining (TyreSense) (Michel, 2005). Fording River and Syncrude reported positive results (Ednie, 2002). With all the potential advantages, real-time tire monitoring and control technology still has some challenges. Existing problems include short sensor life, data inaccuracy, event time lagging and lack of GPS system, etc. (Michel, 2005; Fowler and Huntingford, 2005). Hopefully, these problems will be solved in the near future. 3. Fountain Tire Mine Service Limited Research Project 3.1 Maintenance Practices Fountain Tire Mine Service Limited is a primary supplier and maintenance contractor of mining haul truck tires to mines in western Canada. As such, they have recognized the importance of enhanced understanding of tire performance. Specifically they believe their competitiveness can be improved through enhancement of the maintenance services they provide. As a result this research is funded to analyze tire management practices from an analytical approach. Figure 3-1 is a summary of tire maintenance services provided by Fountain Tire Mine Service Limited. Currently, two concerns regarding improving the company's tire maintenance service to increase tire performance are tire rotation and tire air pressure setting at installation. Tire rotation, one area that has not been addressed in the current literature, is to move a tire from one wheel location to another wheel location of the same or a different vehicle. The purpose of tire rotation is (Fowler and Huntingford, 2005): • For safety; front tires support the drivers cab. • For tire matching; 10mm tread depth difference need to be rematched. • For even tire load and air pressure. • For reducing tire failures and extending tire life. Tire Maintenance Services Tire replacement due to worn out Tire air pressure setting at installation Tire replacement due to damages or failures Tire air pressure setting at installation Tire rotation (matching) Tire air pressure setting at installation Regular check (tire air pressure and tread depth) Scrap tire inspection Tire repair (retread) Cycle time record, weight studies Site audits Figure 3-1 Tire maintenance services provided by Fountain Tire Mine Service Limited (Fowler and Huntingford, 2005) Usually new tires are installed in the front positions, and then rotated to rear positions. Different rotation practices exist: some mines do rotations based on tread wear percentage; some do it based on a certain tread depth while some do it based on tire running hours. Improvements in these practices can be made by deciding appropriate rotation frequency and sequence to maximize tire life. The following tire wheel position acronyms are used by Fountain Tire Mine Service Limited and are adopted in this thesis: LF—left front; RF—right front; LRO—left rear outside; LRI—left rear inside; RRI—right rear inside; RRO—right rear outside. Figure 3-2 shows the tire wheel locations. Figure 3-2 Tire wheel position acronyms Regular tire air pressure and tread depth checks and scrap tire inspections provide valuable information as the basis for data input into the E M T R A C K III -a tire data management software. In order to continuously improve tire performance and obtain relevant information for further analyses, Fountain Tire Mine Service Limited designed a new air and tread depth check sheet as shown in Appendix A. 3.2 Selected Mines Two mines have been selected by Fountain Tire Mine Service Limited for this research: the Kemess Mine and the Gibraltar Mine. Currently Fountain Tire Mine Service Limited is responsible for the tire maintenance services of these two mines. 3.2.1 The Kemess Mine As Northgate's principal operation, the Kemess South gold and copper open pit mine and mill are located in the northern central British Columbia, about 430 kilometers northwest of Prince George. Temperatures vary significantly throughout the year: from as low as about -35 degrees Centigrade in January to over 30 degrees Centigrade in July. There is a lot of snowfall in winter around the mine and it is common to have up to two meters snow cover on the ground (Northgate Minerals Corporation, 2006; InfoMine Inc., 2006). The Kemess South open pit mine is designed with 15 meter high benches and 45 degree slope angles. The current nominal production rate is 52000 tons of ore and 85000 tons of waste per day. Conventional crushing, grinding and floatation techniques are used to process and produce gold-copper concentrates (Northgate Minerals Corporation, 2006). The mine has two electric cable shovels, a hydraulic shovel, and a loader which supply ore and waste to the fleet of 15 Euclid R260 End Dump haulage trucks with 260 tons payload. The trucks move ore to the primary gyratory crusher located adjacent to the open pit and waste rock to storage dumps (Northgate Minerals Corporation, 2006). The tire size for these trucks was 40.00R57, type was RL-4B for the period when data was collected. The manufacturer of these tires is Goodyear. Recently the tire size was switched to 46/90R57 (Fowler and Huntingford, 2005). The Kemess Mine is actively involved in seeking ways to improve tire life. By working with Fountain Tire Mine Service Limited and the Goodyear Tire & Rubber Company, they installed advanced electronic pressure and temperature sensors in each tire and implemented a state of the art tire management program. As a result, a 40 % increase in tire life and a $600,000 decrease in annual tire costs have been gained. (Northgate Minerals Corporation, 2006). The Kemess Mine is designed with left side driving. This will have a different influence on tire performance than right side driving at other mines. 3.2.2 The Gibraltar Mine The Gibraltar copper and molybdenum mine is located in southern central British Columbia, approximately 68 km north of Williams Lake. The mine began development in 1972. After more than 20 years operation, the mine was shut down in 1998 due to low copper prices. In October 2004, the mine recommenced its operation amid a positive and strong metal market conditions. The average annual temperature is approximately 4.2 degrees centigrade around the mine area (Taseko Mines Limited, 2006; Environment Canada, 2006). Currently the nominal production rate at the Gibraltar Mine is 36000 tons of ore and waste per day, among which copper accounts for 98.6%. In March 2006 Taseko Mines Limited approved a 62 million Canadian dollars investment for the Gibraltar Mine's concentrator facility expansion and upgrading. It is expected that annual copper production will reach 100 million pounds by the year of 2008 while annual molybdenum production will arrive at 1.35 million pounds due to this investment (Taseko Mines Limited, 2006). The Gibraltar Mine has two loaders and a fleet of 11 End Dump haulage trucks. Among these trucks, five are MT-4000 Unit Rig with 240 tons payload; six are MT3700B Terex with 180 tons payload. For the period when data was collected, the tire size for MT-4000 Unit Rig was 40.00R57, type was RL-4B; while for MT3700B Terex the tire size was 37.00R57, type was RL-4M+. The manufacturer of these tires is Goodyear (Fowler and Huntingford, 2006). The Gibraltar Mine is right side driving. 3.3 Case Study Objectives and Approach The case study objectives of this research project are detailed as follows: • Determine appropriate tire rotation sequences for the Kemess Mine and the Gibraltar Mine. • Suggest first tire rotation time ranges for the two mines based on tread consumption. • Establish tire tread consumption prediction models at the two mines. • Analyze the influence of tire installation air pressure settings on tire life and tire tread wear rate and clarify the relationship between wheel positions and air pressure spikes, at the two mines. • Examine tire failures with different wheel positions to determine dominant tire damages or failures at the two mines. The following activities were undertaken to achieve these objectives: Firstly, the author attended an off-the-road tire training course at the Gibraltar Mine. This training provided the author with basic knowledge of off-the-road tires. Secondly, visits to the Kemess Mine and the Gibraltar Mine were made. Through these visits, a better understanding of Fountain Tire Mine Service's tire maintenance practices and relevant mining operational practices was gained. Relevant data and information were also collected during these visits. Thirdly, an analytical approach was established and data analysis using methods such as statistical correlation, Pareto analysis, etc. was performed. A detailed description of these methods is in Chapter 5. Lastly, the analytical results were presented to Fountain Tire Mine Service Ltd. and mine sites for implementation. 4. Data Analysis 4.1 Data Collection The relevant data for this research was mainly extracted from the E M T R A C K III software. Additional data and information were collected during mine visits and through communication with Fountain Tire Mine Service Limited and mine sites. At the Kemess Mine, the data for tires installed and eventually scrapped between May 2003 and December 2005 is used for analyses. Because the Gibraltar Mine reopened in October 2004, the data for tires installed and eventually scrapped between its opening and April 2006 is used for analyses. The data extracted from E M T R A C K III software include the following reports: • Tire history report: this report includes tire brand number, tire disposition such as installation and removal, vehicle ID and wheel positions, event date, tire hours, tread depth and air pressure check values, etc. This data is used for tire rotation practice and tire installation air pressure setting analyses. • Scrap tire report: this report includes tire brand number, remaining tread depth, date tires were installed or removed, removal reasons, tire hours and adjusted cost etc. This report is the source data for tire failure analyses. • Scrap analysis report: used for Pareto analysis, this report includes tire removal reasons, number of tires removed for different removal reasons and dollars lost, etc. Other additional data being collected from the Kemess Mine and Fountain Tire Mine Service Limited are truck maintenance records, mine site ambient temperature record, haul road profile, tire change sheets, cycle time record and weight studies, etc. These data provide supplementary information and will be used in the future to quantify the impact of major operational and environmental factors on tire performance. 4.2 Data Treatment and Assumptions The following data treatment principles apply in this research: • Only tires that have been installed and scrapped during the specific periods (the Kemess Mine: between May 2003 and December 2005; the Gibraltar Mine: between October 2004 and April 2006) are used for analyses. This provides an approach to compare tire life, rotation practices and various aspects regarding tire performance. • Tires that were initially installed at the front positions are included in the analysis. These tires account for the majority of all tires. This complies with the current common tire installation practices: tires are usually installed at the front positions and then rotated to the rear positions. • Cross-check different reports to confirm data validity. In general the following assumptions are made for data analyses in this thesis: • Assume data is independent and identically distributed. Two scatter plots of the data can be used to test this assumption. One is to plot the ith data (tire life or tire tread depth consumption) versus the (i-l)th data. Here i means the order of data occurrence. If the points lie along a straight line, the data is dependent. The other is to plot the cumulative occurrence numbers of tire life or tread depth consumption data versus the cumulative tire life or tread depth consumption data. A straight line means the data have an identical distribution (Hall, 2005). Examples of IID graphical tests are shown in Appendix B. Assume the source data are correct. Assume all the other influencing factors stay the same when doing analyses on specified factors. When there is a difference between the tread depth records of "on hold for inspection" and "scrap inspection" in the tire history report, assume the value of "scrap inspection" is the correct one. Because the tire hours in the tire history report are engine hours, and they include the hours that tires are not moving either loaded or empty such as when a truck is stopped for a coffee break, refueling, etc.; also differences in truck speeds makes tire hours inaccurate to some extent, it is assumed tire hours are the same as the hours that tires have actually run in this thesis. 4.3 Methodologies 4.3.1 SAS software package SAS (Statistical Analysis System) is an integrated large scale application software system. It has many powerful functions such as data analysis, report writing and graphics, operations research, business planning and forecasting, etc. Currently, SAS is extensively used in the fields of medicine, science, financial and social science, etc. that involve data management and application. There are three major windows in SAS software: log, output and editor windows. The log window is used to check procedures by looking at whether there is any error message; the output window is used for viewing the results while the editor window is used to write programming codes. The SAS codes are comparatively easy and a large amount of application codes can be found on line through searching for specific procedures (PROCs). This allows the user to focus more on the application area rather than on programming itself. The basic procedures to run SAS are: import data into SAS software (the data may be in different formats including Microsoft Excel); writing corresponding analysis codes in the editor window; run and then export the output into Microsoft Word or Excel. 4.3.2 Analysis methods In order to achieve the research objectives, an analytical approach was established and different data analysis methods were used. In addition to the general statistical analyses, the following methods were used in this thesis: Tire rotation sequences At both the Kemess Mine and the Gibraltar Mine, for rotation frequency 1, the following approach and rotation process acronyms are used to represent different rotation sequences (Figure 4-1): • Si-same side, rotated to inside position of the same or a different truck, e.g. from LF to LRI. • SO-same side, rotated to outside position of the same or a different truck, e.g. from LF to LRO. • Ol-opposite side, rotated to inside position of the same or a different truck, e.g. from LF to RRI. • OO-opposite side, rotated to outside position of the same or a different truck, e.g. from LF to RRO. Figure 4-1 Example of rotation process acronyms for rotation frequency 1 For rotation frequencies 2 and 3, the same approach and rotation process acronyms as for rotation frequency 1 are defined (Figure 4-2): Sl-e.g. from LRO to LRI; SO-e.g. from LRI to LRO; OI-e.g. from LRO to RRI; OO-e.g. from LRI to RRO Regression analysis Regression is a statistical technique used to find relationships between variables for the purpose of predicting future values. R Squared is the square of the correlation coefficient between the response (Y) values and the predicted response values, and it measures how successful the regression fit is in explaining the variation of the data. An R Squared value of 1 means a perfect fit. Standard error is the standard deviation (positive square-root of the variance) of the errors associated with regression fit. After an estimation of a coefficient, the t-statistic for that coefficient is the ratio of the coefficient to its standard error. That coefficient can be tested against a t distribution to determine how probable it is that the true value of the coefficient is really zero (hypothesis). The P-value is associated with the t Statistic. The bigger the t Statistic value, the smaller the P-value (Kutner, et al, 2005). In this analysis, the confidence level is set at 95% (a=0.05). Thus, if P<0.05, reject the hypothesis. Regression method is used in this thesis to find relationship between the tire tread consumption and the corresponding tire running hours. It is assumed that the tire wear rate is constant within each rotation period, i.e. the tire tread consumption is linear with tire running hours for a given position. The tire tread depth consumption (T) is used as one Y variable and the corresponding running hour data (H) is used as one X variable. Using Microsoft Excel's regression function a regression analysis is performed on these data. At both mines, for installation to first rotation, the intercept coefficient is set to zero because the tread depth consumption is zero at zero hours. At the Kemess Mine, for installation to first rotation, 213 tires are used while 82 tires are used for first to.second rotation. At the Gibraltar Mine, 18 and 27 tires are used for 37.00 R57 and 40.00R57 for installation to first rotation respectively while 10 and 15 tires are used for first to second rotation for 37.00 R57 and 40.00R57 respectively. Correlation analysis Canonical Correlation analysis studies the relationship between two sets of variables (Dillon and Goldstein, 1984). For this reason, Canonical Correlation is carried out using the SAS software package at both mines to determine the relationships as shown in Figure 4-3 and Figure 4-4. Tire life (L) Correlation Rotation interval hours ( R l R2 R3) ^ w i Correlation Correlation 1 Tire wear rate (W) Age specific wear ^ w rates ( A l A 2 A3) Figure 4-3 Canonical Correlation relationships at the Kemess Mine Tire life (L) Correlation Rotation interval hours ( R l R2) ^ W i Correlation Correlation 1 Tire wear rate (W) Age specific wear rates ( A l A2) Figure 4-4 Canonical Correlation relationships at the Gibraltar Mine In the SAS output, positive correlation values mean that i f one variable is increased or decreased, the other variable will be increased or decreased too; negative correlation values mean the opposite. A correlation value of 1 or -1 means perfect correlation. Examples of SAS output are shown in the appendices C and D. M A N O V A analysis In order to understand the influence of tire installation air pressure on tire life and wear rate, a multivariate analysis of variance (MANOVA) is conducted using the SAS software package. M A N O V A is an extension of analysis of variance (ANOVA) to include more than one dependent (Y) variable in the analysis, and is used for testing the equality of mean values of Y variables over several treatments. The alpha level is the P value (probability value) which is a predetermined acceptance level. Usually it is set at 0.1 or 0.05. The alpha level is set at a=0.05 in this thesis (a=0.1 for M A N O V A analysis on tire life at the Kemess Mine). Also Type III SS (Sum of Squared errors) is used because it examines all of the effects simultaneously. F value is a measurement of ^ MeanSquare distance between individual distributions: F = , Mean Square of Error Type HISS . „ . . Mean Square = - — . As F goes up, the P value goes down (i.e., DF (degree of freedom) more confidence in there being a difference between two means) (LeMay, 2006). The hypothesis for this analysis is the vectors of means of Y variables are the same, i.e. changing air pressure (class variable) has no effect on Y variables. If p>0.05, accept the hypothesis; i f p<0.05, reject the hypothesis. At the Kemess Mine, the X class variable includes four different tire installation air pressure settings (100 psi, 105 psi, 110 psi and 95 psi); the two Y dependent (response) variables are tire life and tire wear rate. At the Gibraltar Mine, the X class variable consists of three different tire installation air pressure settings (100 psi, 102 psi and 104 psi for 37.00R57; 100 psi, 102 psi and 105 psi for 40.00R57); the two Y dependent (response) variables are tire life and tire wear rate. Table 4-1 is an example of the data structure. This data is then input into the SAS software. A n example of SAS output is attached in Appendix E. Table 4-1 Example of MANOVA data structure Air pressure setting (psi) Tire life L (hours) Wear rate W (mm/hour) 100 (C) 5109 0.0127 100 (C) 2596 0.0287 105 (C) 3791 0.0206 105 (C) 1710 0.0222 110(C) 3485 .0.0221 110(C) 2467 0.0199 95 (C) 4471 0.0157 95 (C) 3627 0.0185 To understand which air pressure settings contribute more to the difference on tire life and tire wear rate, t tests for comparison of means between pairs are carried out. For this (Lemay, 2006). NPs! purpose, the a value is calculated as: a = , while N p a = O05 Npa 2!(N Ps-2)! 4! 0 05 At the Kemess Mine, N P . = _ = 6 , a= = 0.0083. If p<0.0083, the 2!(4-2)! 3! difference between means is significant. At the Gibraltar Mine, N p a = — = 3, 2!(3-2)! a = 0.05 = 0.0167. If p<0.0167, the difference between means is significant. Note this comparison is quite conservative because the a value is divided by number of pairs to avoid statistical error (Lemay, 2006). The assumptions for M A N O V A are as following: • The variances of all response variables are equal among treatment groups (installation air pressure settings). This can be checked by looking at the residual plots. Evenly distributed residuals around zero indicate an equal variance (Lemay, 2006). • Response variables are normally distributed. This can be checked by looking at normality tests. If p>0.05, the normality assumption stands (Lemay, 2006). • Assume the influence of other factors such as hauling distance, pit depth and ambient temperature are the same for different initial air pressure settings. Pareto analysis Pareto analysis is a method to prioritize. maintenance efforts by classifying failures according to costs. Firstly failures are arranged in a descending order based on their corresponding costs; then the cumulative cost of failures is plotted versus the cumulative number of failures (Hall, 2005). In this thesis, the cumulative lost tire cost (original + retread + repair - adjustment) is used as Y axis and the cumulative number of tires is used as X axis. 5. Results and Discussion 5.1 The Kemess Mine 5.1.1 Tire Rotation 5.1.1.1 Relation between tire life and rotation frequency As can be seen from Table 5-1, the tire life increased when the tire rotation frequency increased at the Kemess Mine. The only inconsistency was for rotation frequency 4 which had a lower average tire life than frequency 3. By comparing the average tire life for each rotation frequency (Table 5-1) with the average rotation hours of each rotation (Table 5-2) it is seen that the average tire lives of four frequencies are longer than the average rotation hours of four rotations: average tire life hours 2254 for frequency 0 vs. average rotation hours 1490 for the first rotation; average tire life hours 3474 for frequency 1 vs. average rotation hours 2833 for the second rotation; average tire life hours 4026 for frequency 2 vs. average rotation hours 3319 for the third rotation; and average tire life hours 4228 for frequency 3 vs. average rotation hours 3380 for the fourth rotation. This-means that at least some tires had the opportunities to go to the next rotation. Because there is not enough data for rotation frequencies 4 and 5, the statistics on these two are omitted. Table 5-1 Tire life vs. tire rotation frequency (40.00R57) (Kemess) Rotation frequency Number of tires Average tire life (hours) Standard deviation (hours) 0 25 2254 1305 1 131 3474 983 2 57 4026 673 3 22 4228 622 4 2 3483 939 5 1 5264 N / A Sum 238 Average 3555 1066 Table 5-2 Average rotation hours (40.00R57) (Kemess) Number of tires Average rotation hours First rotation 213' 1490 Second rotation 82 2833 Third rotation 25 3319 Fourth rotation 3 3380 Fifth rotation 1 4949 This ascending tire life trend with ascending tire rotation frequency is partly explained by the purposes of tire rotation as discussed in section 3.1. These quantitative results suggest that tire life can be improved by adjusting the rotation interval/ frequency. 5.1.1.2 Relation between tire wear rate and rotation frequency It is known that tire tread wear rate is one of the important tire performance indices. As can be seen from Table 5-3 and Table 5-4, the tire wear rate decreased when! the tire rotation frequency increased and age-specific tire wear rate also decreased with rotation period. Further looking at Table 5-5, it is obvious that front tires (or tires in their early lives) had higher wear rate, and rear right tires had higher wear rate than rear left tires for 1s t to 2 n d rotation. Also generally rear outside tires wore faster than rear inside tires. Table 5-3 Tire wear rate vs. tire rotation frequency (40.00R57) (Kemess) Rotation frequency 0 Rotation frequency 1 Rotation frequency 2 Rotation frequency 3 Tire wear rate (mm/hour) 0.0180 0.0172 0.0170 0.0165 Table 5-4 Age-specific tire wear rate for each rotation period (40.00R57) (Kemess) Installation-1st rotation 1st -2nd rotation 2nd-3rd rotation 3rd rotation -Scrap Age-specific wear rate (mm/hour) 0.0187 0.0178 0.0153 0.0151 Table 5-5 Age-specific tire wear rate (mm/hour) according to wheel positions (40.00R57) (Kemess) Wheel position LF RF LRI LRO RRI RRO Installation-1st rotation 0.0176 0.0201 1 st -2nd rotation 0.0161 0.0162 0.0196 0.0208 2nd-3rd rotation 0.0146 0.0132 0.0094 0.0186 5.1.1.3 Tire rotation sequences Tire rotation is important, so is the rotation sequence. It is believed the sequence of the rotation will influence tire life; this is due to various reasons such as tread wear pattern, load distribution, superelevation, cooling, etc. From Table 5-6 it can be seen for the rotation frequency 1 tires that were rotated to 01 had the longest average lives while tires that were rotated to SI had the shortest average lives. Thus, based on this analysis tires should be rotated to 01 on their first rotation to maximize life. Table 5-6 Tire life vs. rotation sequence for rotation frequency 1 (40.00R57) (Kemess) SI SO OI OO Number of tires 34 27 42 27 Average tire life (hours) 3244 3483 3702 3374 Average remaining tread depth (mm) 31.1 27.9 23.9 27.5 Tire wear rate (mm/hour) 0.0166 0.0164 0.0165 0.0170 Percentage of tires with uneven remaining tread 32.4% 40.7% 28.6% 40.7% The results (Table 5-7 and Table 5-8) show that the rotation priority considerations for frequencies 2 and 3 were SO-OO-OI-SI and OO-SI-OI-SO respectively based on average tire lives, i.e. SO and 0 0 would be the first considerations for the second and the third rotation respectively. Here it should be noted there is no real tire life difference between 0 0 and SI (Table 5-8: 4534 vs. 4528 hours) for rotation frequency 3. For this analysis, 0 0 is adopted. A caution also needs to be taken for rotation frequency 3 due to its comparatively low number of tires. It is observed from Table 5-6, Table 5-7 and Table 5-8 that the best position for each rotation frequency generally had the lowest wear rate for their respective frequency. Table 5-7 Tire life vs. rotation sequence for rotation frequency 2 (40.00R57) (Kemess) SI SO 01 0 0 Number of tires 11 14 20 12 Average tire life (hours) 3821 4395 3913 3973 Average remaining tread depth (mm) 20.3 14.7 19.6 15.6 Tire wear rate (mm/hour) 0.0169 0.0160 0.0167 0.0175 Percentage of tires with uneven remaining tread 45.5% 35.7% 35% 41.7% Table 5-8 Tire life vs. rotation sequence for rotation frequency 3 (40.00R57) (Kemess) SI SO OI OO Number of tires 7 3 7 5 Average tire life (hours) 4528 3728 3922 4534 Average remaining tread depth (mm) 14.1 16 18.6 17.5 Tire wear rate (mm/hour) 0.0157 0.0185 0.0169 0.0149 Percentage of tires with uneven remaining tread 14.3% 0% 42.9% 20% Based on the above analyses, it is suggested that the rotation sequence from the first rotation to the third rotation be carried out according to OI-SO-00 at the Kemess Mine (Figure 5-1). There was no tire that had actually gone through this rotation sequence. Three tires went through OI-SO sequence with an average tire life of 4280 hours (the average tire life at the mine is 3555 hours). Figure 5-1 Suggested rotation sequence (40.00R57) (Kemess 5.1.2 Rotation Time 5.1.2.1 Regression analysis It can be seen from Table 5-9 that regression gives a reasonable result with an R Squared value of 0.9289. According to Table 5-10, the equation for tread consumption between installation and first rotation is: T = 0.0166*H (1) Where T represents tread depth consumption (mm) and H represents tire running hours. Table 5-9 Regression statistics for installation to first rotation (40.00R57) (Kemess) R Squared 0.9289 Standard Error 4.5 Observations 1262 Table 5-10 Regression coefficients for installation to first rotation (40.00R57) (Kemess) Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Intercept 0 N / A N / A N / A N / A N / A X Variable 1 0.0166 0.0001 128.432 0 0.0164 0.0169 According to Fowler and Huntingford (2006), 55 mm remaining tread depth (30mm tire tread consumption) is the optimum removal tread depth for front positions and is also the optimum tread depth for safety reasons at the Kemess Mine. This is the general guideline. A tire with less tread to void ratio has more chances of rock cut through. On the other hand, if replaced earlier many new inventory tires would have to be used. Using equation (1), the replacement hours for the first rotation is: H = T/0.0166 = 30/0.0166 = 1807hours (2) Note this rotation time is different from 1100 hours at the Kemess Mine as in section 5.1.4. The 1100 hour rotation time for left front tires is for the purpose of reducing heat separation failures while the above 1807 hours is based on 30 mm tire tread consumption. A further analysis that combines various influencing factors is needed. Due to various demands both in terms of tire inventory, matching and truck availability for tire changes it may not be feasible to change at exactly 1800 hours. Therefore, it is suggested that the first rotation hour be set at between 1750 and 1800 hours. This is longer than the current rotation practice (Table 5-2: 1490 hours for the first rotation on average) at the Kemess Mine. Similarly, regression analyses are carried out on the first to second rotation. The results are shown in Table 5-11 and Table 5-12. The R squared value is lower than the R squared value between installation and the first rotation. There is not enough data for regression analysis for second to third rotation and third rotation to scrap. Table 5-11 Regression statistics for first to second rotation (40.00R57) (Kemess) R Squared 0.7036 Standard Error 7.3 Observations 507 Table 5-12 Regression coefficients for first to second rotation (40.00R57) (Kemess) Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Intercept 8.7 0.9 9.5470 0 6.9 10.5 X Variable 1 0.0130 0.0004 34.6209 • 0 0.0123 0.0137 There are various factors that influence the rotation time at rear positions. These factors include: • Wear rate, lower wear rate means longer running hours between rotations. • Wear pattern, compensating the wear pattern difference requires similar running hours at latter wheel position to the previous wheel position. • Failure mechanism, which requires the rotation interval to be shorter at some wheel positions to avoid certain failures. Tire matching is an important part of tire rotation. Checking the data, it is found that matching accounted for a large percentage (48%) of rear rotations. According to the data, the reasons for matching are: • One tire of the pair (LRI-LRO or RRI-RRO) is scrapped or damaged. If a suitable tire to match it is available then it is installed. Otherwise the good tire is removed and saved to be put into service when the next match tire is needed. This accounts for the majority of matching. When choosing tires to match an attempt is made to maintain the tread difference between rear pairs of tire to less than 10mm. • When the remaining tread depth difference between the pair of tires is more than 10mm, the tire with less remaining tread depth needs to be removed if there is a suitable tire available. In reality the tires chosen for matching do not have exactly the same tread depth-often with more than 5 mm difference. Although the wear rate differences between rear pairs for the first to the second rotation was only about 0.0001 mm/hour for left side and about O.OOlmm/hour for right side (Table 5-5), the tread wear rate difference between some individual tires can be bigger than that (sometimes half of the average wear rate). This is because of the comparatively high standard deviation of wear rate for each wheel position. This means for two perfectly matched tires, the hours needed to arrive at 10mm tread difference vary significantly. Due to the variability of rotation practice and complexity of involving factors, further analyses and possibly some tests are needed in order to determine the suitable rotation time for rear rotations. 5.1.2.2 Correlation analysis From Table 5-13 and Table 5-14, it can be seen that the first rotation interval hours influenced the tire life most while it is the second age-specific wear rate that influenced the tire wear rate most. It is also noticed from Table 5-15 that only the correlation between age-specific wear rate A l and first rotation interval hours R l was comparatively high (-0.5963), this means that age-specific wear rates are not determined solely by the rotation interval hours between each rotation. Because there are positive correlations between R l , R2 and tire life L (Table 5-13), and negative correlations between R l , R2 and A2 (Table 5-15), it seems extending both R l and R2 moderately will extend tire life. Table 5-13 Correlations between the rotation interval hours between each rotation (Rl, R2 and R3) and the tire life (L) (40.00R57) (Kemess) L R l 0.5481 R2 0.5239 R3 0.2653 Table 5-14 Correlations between the age-specific wear rates between each rotation (Al, A2 and A3) and the tire wear rate (W) (40.00R57) (Kemess) W A l 0.3534 A2 0.5685 A3 0.0672 Table 5-15 Correlations between the three rotation interval hours and the three age-specific wear rates (40.00R57) (Kemess) R l R2 R3 A l -0.5963 0.0242 0.1688 A2 -0.3477 -0.3196 -0.3908 A3 0.0039 -0.0239 -0.0744 It should be noted that due to the limit of data (only 25 tires for the Kemess Mine), the results from the analysis should be used with caution until additional data becomes available. 5.1.3 Tire Air Pressure Setting The tire installation air pressure at both mines is recommended by the tire manufacturer. Due to various factors this value has been increasing over the years at the Kemess Mine. In December 2005 the tire installation pressure was increased from 110 psi to 115 psi for front positions. The effectiveness of this set pressure will be evaluated after December of 2006 when relevant data such as tire life are available. From Table 5-16, it can be seen that generally tire life decreased with increasing tire installation air pressure; as well the tire wear rate increased when the tire installation air pressure increased. Table 5-16 Effect of tire installation air pressure on tire life and wear rate (40.00R57) (Kemess) 95 psi 100 psi 105 psi 110 psi Number of tires 7 21 171 39 Average life (hours) 3932.6 4032 3502.1 3467.1 Tire wear rate (mm/hour) 0.0160 0.0154 0.0172 0.0196 5.1.3.1 MANOVA analysis As can be seen from Table 5-17 the P value for tire life is 0.0483<0.05, so the effect of changing air pressure on tire life is significant at 95% confidence. The P value for overall tire wear rate is 0.0001O.05 (Table 5-18), therefore the hypothesis of no effect is rejected and the air pressure change has an influence on the tire wear rate. Table 5-17 Effect of changing pressure on tire life (40.00R57) (Kemess) Dependent variable: L (tire life) Source DF F Value P r > F Pressure 3 3.29 0.0483 Table 5-18 Effect of changing pressure on tire wear rate (40.00R57) (Kemess) Dependent variable: W (tire wear rate) Source DF F Value P r > F Pressure 3 17.95 0.0001 Comparing the corresponding P values in Table 5-19, it is found there are significant differences between 3 and 1 (110-100), 3 and 2 (110-105) with P values less than 0.0083 (0.05/6). On the other hand, the differences between means are not significant between 1 and 2, 1 and 4, 2 and 4, 3 and 4. This means 110 psi air pressure setting plays a significant role in changing the tire tread wear rate. There are no P values less than 0.0083 (0.05/6) in Table 5-20. This means that there are no differences between means of tire life among the four pressures at 95% confidence. As mentioned in section 4.3.2, this result is conservative compared to that of Table 5-17. At 90% confidence level, it can be seen the p value between 3 and 1 (110-100) is less than 0.0167 (0.1/6). A l l the other p values are larger than 0.0167. This means at 90% confidence level, the effect of changing air pressure on tire life is gradual and it starts to become significant at 110 psi setting. Table 5-19 Least squares means comparison of different air pressure (40.00R57) (Kemess) Least squares means for effect of pressure Pr > |t| for HO: LSMean(i)=LSMeanG) Dependent variable: W (tire wear rate) i/j 1 (100 psi) 2(105 psi) 3 (11.0 psi) 4 (95 psi) 1 (100 psi) 0.0473 0.0001 0.4313 2(105 psi) 0.0473 0.0001 0.5604 3 (110 psi) 0.0001 0.0001 0.0177 4 (95 psi) 0.4313 0.5604 0.0177 Table 5-20 Least squares means comparison of different air pressure (40.00R57) (Kemess) Least squares means for effect of pressure Pr > |t| for HO: LSMean(i)=LSMeanG) Dependent variable: L (tire life) i/j 1 (100 psi) 2(105 psi) 3 (110 psi) 4 (95 psi) 1(100 psi) 0.0386 0.0100 0.5204 2(105 psi) 0.0386 0.1748 0.2830 3(110 psi) 0.0100 0.1748 0.1248 4(95 psi) 0.5204 0.2830 0.1248 5.1.3.2. Tire air pressure spikes analysis Table 5-21 shows that most of the tire air pressure measurements are in the range of 100%-130% of initial set pressure at the Kemess Mine. It is noticed that most (16 out of 17) spike pressures (over 140%) happened with 95. psi cold installation. It is also noticed that about 51% of tires (38 tires) which had experienced spikes (over 130%) were worn out with comparatively long hours. As well from Table 5-22 and Table 5-23 it can be seen that rear inside positions had more air pressure spikes than rear outside positions and left front had higher spike percentage than right front. The reason for the latter may be the left side driving at the Kemess Mine. Table 5-21 Summary of tire air pressure measurement (40.00R57) (Kemess) Air pressure rating <100% 100%-120% 120%-130% 130%-140% >140% Number of measurements 333 2327 1116 129 17 Percentage 8.49% 59.33% 28.45% 3.29% 0.43% Table 5-22 Summary of tire air pressure measurement with over 130% rating (40.00R57) (Kemess) Wheel position LF RF LRI LRO RRI RRO Number of measurements 11 4 50 25 37 19 Percentage 7.53% 2.74% 34.25% 17.12% 25.34% 13.01% Table 5-23 Summary of tire air pressure measurement with over 140% rating (40.00R57) (Kemess) Wheel position LF RF LRI LRO RRI RRO Number of measurements 1 0 4 0 9 3 Percentage 5.88% 0% 23.53% 0% 52.94% 17.65% 5.1.4 Tire Failure Analysis Table 5-24 is a summary of tire life according to various positions at the Kemess Mine: • It is noted that rear outside and rear right positions had higher average scrap lives than rear inside and rear left positions respectively. • In most cases, evenly worn tires had higher average tire life than unevenly worn tires. The exception to this was right outside positions (including right front positions), which might have something to do with the left side driving at the Kemess Mine. The unevenly worn tread acted as a kind of super elevation. According to the data showing failure modes for different wheel positions in Table 5-25 the following may be concluded: • Cut separation, cut tread and impact break mainly accounted for shorter lives of rear inside tires; • Heat separation mainly accounted for shorter tire lives of left wheel—especially left front wheel positions. Table 5-24 Tire life comparison according to various positions (40.00R57) (Kemess) Number of Number of tires with Average tire tires with Average tire uneven scrap life (hours) even scrap life (hours) tread tread Rear inside 42 3481.7 85 3729.7 Rear outside 35 3789.3 61 3723.4 Rear left side 33 3415.1 68 3628.5 Rear right side 44 3776.3 78 3813 Front 3 2490.3 12 1306.8 Rear 77 3621.5 146 3727.1 Left 34 3371.5 74 3411.6 Right 46 3732.6 84 3659.2 Table 5-25 Failure number comparison according to wheel positions (40.00R57) (Kemess) 1L 1R LRO LRI RRI RRO Cut separation 4 7 8 5 Cut sidewall 1 1 7 9 7 8 Cut tread 1 2 4 7. 8 4 Heat separation 4 1 1 1 Impact break 1 3 7 12 6 Worn out 22 23 36 29 Turn up separation 1 1 1 1 Separation 1 1 Repair failure 1 Chipper separation 1 Ply ending separation 1 In order to identify the major problem areas where efforts can be directed, a Pareto Analysis (Hall, 2005) was conducted (Figure 5-2). From this graph it can be seen that cut tread, impact break, cut separation and cut sidewall accounted for about 80% of lost tire cost (original + retread + repair - adjustment). Based on the average tire lives for cut tread (2744 hours) and impact break (2969 hours) it can be concluded that better maintenance of haul roads and clean up of loading and dumping areas will minimize cut tread and impact break failures. As well the four heat separation failures that happened at left front positions had an average tire life of 1104.3 hours; it seems that left front tires should be rotated at some value less than this to minimize heat separation. 0% -I 1 1 1 ! 1 0% 20% 40% 60% 80% 100% 120% Cumulative failures (%) Figure 5-2 Pareto analysis (40.00R57) (Kemess) 5.2 The Gibraltar Mine To further demonstrate the efficacy of the analytical methods and the applicability of the results at the Kemess Mine, the same data analyses are performed at the Gibraltar Mine. Due to a low number of tires and a limited data at the Gibraltar Mine, caution should be taken when applying the results in this section. 5.2.1 Tire Rotation 5.2.1.1 Relation between tire life and rotation frequency At the Gibraltar Mine, the tire life also generally increased when the tire rotation frequency increased (Table 5-26 and Table 5-27). Comparing the average tire life for each rotation frequency (Table 5-26 and Table 5-27) with the average rotation hours (Table 5-28 and Table 5-29) it is seen that the previous are longer than the latter for two frequencies of 37.00R57 and 40.00R57 tires: average tire life hours 4393 and 3667.8 for frequency 1 vs. average rotation hours 4224 and 3104 for the second rotation; average tire life hours 5223.5 and 3985.3 for frequency 2 vs. average rotation hours 4479 and 3763 for the third rotation. This also means that at least some tires had the opportunities to go to the next rotation (second and third rotation). Because there is not enough data for rotation frequencies 0, 3 and 4, the statistics on these three are omitted. As 40.00R57 tires are supporting heavier payload (240 tons) than that (180 tons) of 37.00R57 tires, their average life was shorter (3771.3 vs. 4666.2 hours). The average rotation hours of 40.00R57 tires (Table 5-29) were also shorter than that of 37.00R57 tires (Table 5-28) for each rotation. These quantitative results at the Gibraltar Mine confirm that tire life can be improved by adjusting the rotation interval/ frequency as demonstrated at the Kemess Mine. Table 5-26 Tire life vs. tire rotation frequency (37.00R57) (Gibraltar) Rotation frequency Number of tires Average tire life (hours) Standard deviation (hours) 0 1 2506 N / A 1 8 4393 1029 2 8 5224 563 3 2 4610 352 Sum 19 Average 4666 986 Table 5-27 Tire life vs. tire rotation frequency (40.00R57) (Gibraltar) Rotation frequency Number of tires Average tire life (hours) Standard deviation (hours) 0 2 1240 1075 1 12 3668 1007 2 10 3985 694 3 1 4942 N / A 4 4 4520 475 Sum 29 Average 3771 1107 Table 5-28 Average rotation hours (37.00R57) (Gibraltar) Number of tires Average rotation hours First rotation 18 2565 Second rotation 10 4224 Third rotation 2 4479 Table 5-29 Average rotation hours (40.00R57) (Gibraltar) Number of tires Average rotation hours First rotation 27 1533 Second rotation 15 3104 Third rotation 5 3763 Fourth rotation 4 4121 5.2.1.2 Relation between tire wear rate and rotation frequency At the Gibraltar Mine, as can be seen from Table 5-30 and Table 5-32, the trend of tire wear rate was not obvious while the age-specific tire wear rate increased with the rotation period for 37.00R57 tires. For 40.00R57 tires the tire wear rate generally decreased when the tire rotation frequency increased (Table 5-31) and age-specific tire wear rate generally decreased too with the rotation period (Table 5-33). Checking Table 5-34 and Table 5-35, it is obvious that rear left tires had higher wear rate than rear right tires for 37.00R57 and rear left inside position had higher wear rate than the other rear positions for 40.00R57; rear outside tires wore faster than rear inside tires for 37.00R57 while rear inside tires wore faster than rear outside tires for 40.00R57. The wear rate difference between rear pair tires (Table 5-34 and Table 5-35) was bigger than that at the Kemess Mine (Table 5-5). The numbers in parentheses in Table 5-34 and Table 5-35 are number of tires. Table 5-30 Tire wear rate vs. tire rotation frequency (37.00R57) (Gibraltar) Rotation frequency 0 Rotation frequency 1 Rotation frequency 2 Rotation frequency 3 Tire wear rate (mm/hour) 0.0184 0.0152 0.0155 0.0172 Table 5-31 Tire wear rate vs. tire rotation frequency (40.00R57) (Gibraltar) Rotation frequency 0 Rotation frequency 1 Rotation frequency 2 Rotation frequency 3 Rotation frequency 4 Tire wear rate (mm/hour) 0.0173 0.0164 0.0145 0.0146 0.0160 Table 5-32 Age-specific tire wear rate for each rotation period (37.00R57) (Gibraltar) Installation-1st rotation 1 st -2nd rotation 2nd-3rd rotation Age-specific wear rate (mm/hour) 0.0142 0.0190 0.0212 Table 5-33 Age-specific tire wear rate for each rotation period (40.00R57) (Gibraltar) Installation-1st rotation 1st -2nd rotation 2nd-3rd rotation 3rd -4th rotation Age-specific wear rate (mm/hour) 0.0154 0.0151 0.0162 0.0131 Table 5-34 Age-specific tire wear rate (mm/hour) according to wheel positions (37.00R57) (Gibraltar) Wheel position LF RF LRI LRO RRI RRO Installation-1st rotation 0.0140 (8) 0.0144 (10) 1 st -2nd rotation 0.0145 (1) 0.0224 (5) 0.0144 (2) 0.0175 (2) 2nd-3rd rotation 0.0212 (2) Table 5-35 Age-specific tire wear rate (mm/hour) according to wheel positions (40.00R57) (Gibraltar) Wheel position LF RF LRI LRO RRI RRO Installation-1st rotation 0.0146 (15) 0.0165 (12) 1 st -2nd rotation 0.0167 (3) 0.0123 (3) 0.0159 (3) 0.0153 (6) 2nd-3rd rotation 0.0218 (1) 0.0146 (2) 0.0181 (1) 0.0119 (1) 3rd-4th rotation 0.0210 (2) 0(1) 0.0104 (1) 5.2.1.3 Tire rotation sequences From Table 5-36 and Table 5-37 it can be seen for the rotation frequency 1 tires that were rotated to OI had the longest average lives for both 30.00R57 and 40.00R57 while tires that were rotated to SI and OO had the shortest average lives respectively. Therefore, based on this analysis both type tires should be rotated to OI on their first rotation to maximize life. It is also noted that there was almost no evenly worn out tires as is evident in Table 5-36, Table 5-37, Table 5-38 and Table 5-39. Table 5-36 Tire life vs. rotation sequence for rotation frequency 1 (37.00R57) (Gibraltar) SI SO OI OO Number of tires 5 3 4 5 Average tire life (hours) 4256 5528 5656 4906 Average remaining tread depth (mm) 33.9 25.7 17.8 21.8 Tire wear rate (mm/hour) 0.0151 0.0131 0.0142 0.0155 Percentage of tires with uneven remaining tread 100% 100% 100% 80% Table 5-37 Tire life vs. rotation sequence for rotation frequency 1 (40.00R57) (Gibraltar) SI SO OI 0 0 Number of tires 6 1 3 6 Average tire life (hours) 3613 3635 3989 3257 Average remaining tread depth (mm) 25.4 28 19.8 34 Tire wear rate (mm/hour) 0.0165 0.0157 0.0163 0.0157 Percentage of tires with uneven remaining tread 100% 100% 100% 83.3% The results of Table 5-38 and Table 5-39 show that the rotation priority considerations for frequency 2 are SI-SO-OO-OI for 37.00R57 and OI-OO-SI-SO for 40.00R57 based on average tire lives, i.e. SI and OI would be the first considerations for the second rotation for the two types of tires respectively. According to the above analyses, it is suggested that the rotation sequence from the first rotation to the second rotation be carried out according to OI-SI for 37.00R57 tires (Figure 5-3) and OI-OI for 40.00R57 tires at the Gibraltar Mine (Figure 5-4). There was one 37.00R57 tire that actually went through the suggested OI-SI sequence with a tire life of 5778 hours (average tire life for 37.00R57 is 4666 hours). For 40.00R57, two tires have actually gone through the suggested OI-OI sequence with an average tire life of 4351 hours (the average tire life for 40.00R57 is 3771 hours). Table 5-38 Tire life vs. rotation sequence for rotation frequency 2 (37.00R57) (Gibraltar) SI SO OI OO Number of tires 3 1 4 2 Average tire life (hours) 5766 5413 5026 5137 Average remaining tread depth (mm) 9.7 19.5 21.3 18 Tire wear rate (mm/hour) 0.0153 0.0145 0.0153 0.0156 Percentage of tires with uneven remaining tread 100% 100% 100% 100% 5778 hours (average tire life for 37.00R57 is 4666 hours). For 40.00R57, two tires have actually gone through the suggested OI-OI sequence with an average tire life of 4351 hours (the average tire life for 40.00R57 is 3771 hours). Table 5-38 Tire life vs. rotation sequence for rotation frequency 2 (37.00R57) (Gibraltar) SI SO OI OO Number of tires 3 1 4 2 Average tire life (hours) 5766 5413 5026 5137 Average remaining tread depth (mm) 9.7 19.5 21.3 18 Tire wear rate (mm/hour) 0.0153 0.0145 0.0153 0.0156 Percentage of tires with uneven remaining tread 100% 100% 100% 100% Table 5-39 Tire life vs. rotation sequence for rotation frequency 2 (40.00R57) (Gibraltar) SI SO OI 0 0 Number of tires 2 1 5 2 Average tire life (hours) 3937 3746 4055 3980 Average remaining tread depth (mm) 33.8 18.5 24.7 28.3 Tire wear rate (mm/hour) 0.0130 0.0178 0.0149 0.0142 Percentage of tires with uneven remaining tread 100% 100% 100% 100% 2:SI To another truck r . ._ .T__ LRI 2:SI To another truck RRI Figure 5-3 Suggested rotation sequence for 37.00R57 (Gibraltar) 5.2.2 Rotation Time 5.2.2.1 Regression analysis As seen in Table 5-40 and Table 5-41 regression has reasonable results with an R Squared value of 0.9826 and 0.9356 respectively. According to Table 5-42 and Table 5-43, the tread consumption equations between installation and first rotation for 37.00R57 and 40.00R57 tires are: T = 0.0147 * H (3) T = 0.015 * H (4) Where T represents tread depth consumption (mm) and H represents tire running hours. Table 5-40 Regression statistics for installation to first rotation (37.00R57) (Gibraltar) R Squared 0.9826 Standard Error 3.4 Observations 134 Table 5-41 Regression statistics for installation to first rotation (40.00R57) (Gibraltar) R Squared 0.9356 Standard Error 4.1 Observations 166 Table 5-42 Regression coefficients for installation to first rotation (37.00R57) (Gibraltar) Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Intercept 0 N / A N / A N / A N / A N / A X Variable 1 0.0147 0.0002 86.5763 0 0.0143 0.0150 Table 5-43 Regression coefficients for installation to first rotation (40.00R57) (Gibraltar) Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Intercept 0 N / A N / A N / A N / A N / A X Variable 1 0.0150 0.0003 48.9412 0 0.0144 0.0156 Because there is no available information on optimum removal tread depth for the Gibraltar Mine the same value (30mm tire tread consumption) is assumed and used. Using equation (3) and (4), the replacement hours for the first rotation of 37.00R57 and 40.00R57 are: H = T/0.0147 = 30/0.0147 = 2047hours (5) H = T/0.0150 = 30/0.0150 = 2000hours (6) For the same reasons as at the Kemess Mine, it is suggested that the first rotation time be set at between 2000 and 2050 hours for 37.00R57 and between 1950 and 2000 hours for 40.00R57 at the Gibraltar Mine based on 30mm tread consumption. This is different from the current practice (2565 hours on average for 37.00R57 and 1533 hours on average for 40.00R57) at the Gibraltar Mine. Similarly, regression analyses are carried out on the first to second rotation. The results are shown in Table 5-44, Table 5-45, Table 5-46 and Table 5-47. There is not enough representative data for regression analysis for second to third rotation and third rotation to scrap. Table 5-44 Regression statistics for first to second rotation (37.00R57) (Gibraltar) R Squared 0.9919 Standard Error 5.1 Observations 59 Table 5-45 Regression statistics for first to second rotation (40.00R57) (Gibraltar) R Squared 0.6322 Standard Error 8.7 Observations 90 Table 5-46 Regression coefficients for first to second rotation (37.00R57) (Gibraltar) Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Intercept 0 N / A N / A N / A N / A N / A X Variable 1 0.0148 0.0002 84.063 0 0.0145 0.0152 Table 5-47 Regression coefficients for first to second rotation (40.00R57) (Gibraltar) Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Intercept 5.5 2.6 2.0629 0.0421 0.2 10.7 X Variable 1 0.0124 0.0010 12.3 0 0.0104 0.0144 Checking the data, it is found that matching accounted for a large percentage (92% for 37.00R57 and 52% for 40.00R57) of rear rotations at the Gibraltar Mine. The matching percentage for 37.00R57 tires was higher than that for 40.00R57 tires. 5.2.2.2 Correlation analysis From Table 5-48 to Table 5-51, it can be seen that the second rotation interval hours influenced the tire life most while it is the first age-specific wear rate that influenced the tire wear rate most for both types of tires. As well the correlation between the age-specific wear rate A2 and first rotation interval hours R l was comparatively high for 37.00R57 type tires (Table 5-52); while it is the correlation between the age-specific wear rate A l and second rotation interval hours R2 that was comparatively high for 40.00R57 type tires (Table 5-53). This means that age-specific wear rates are not determined solely by the hours between each rotation. For 37.00R57 tires, because there is a positive correlation between R2 and tire life L (Table 5-48), and negative correlation between R2 and wear rates A l and A2 (Table 5-52), it seems extending R2 and shortening R l moderately will extend tire life. For 40.00R57 tires, because there is a positive correlation between R2 and tire life L (Table 5-49), and negative correlation between R l and wear rates A l , A2 (Table 5-53), it seems extending both R l and R2 moderately will extend tire life. Table 5-48 Correlations between the rotation interval hours between each rotation (Rl, R2) and the tire life (1) (37.00R57) (Gibraltar) L R l 0.1505 R2 0.3137 Table 5-49 Correlations between the rotation interval hours between each rotation (Rl, R2) and the tire life (1) (40.00R57) (Gibraltar) L R l -0.2941 R2 0.7972 Table 5-50 Correlations between the age-specific wear rates between each rotation (Al, A2) and the tire wear rate (W) (37.00R57) (Gibraltar) W A l 0.6912 A2 0.5850 Table 5-51 Correlations between the age-specific wear rates between each rotation (Al, A2) and the tire wear rate (W) (40.00R57) (Gibraltar) W A l 0.8213 A2 0.6674 Table 5-52 Correlations between the two rotation interval hours and the two age-specific wear rates (37.00R57) (Gibraltar) A l A2 R l 0.4239 -0.5236 R2 -0.1934 -0.2654 Table 5-53 Correlations between the two rotation interval hours and the two age-specific wear rates (40.00R57) (Gibraltar) A l A2 R l -0.2363 -0.3778 R2 0.6750 -0.1151 5.2.3 Tire Air Pressure Setting Unlike the Kemess Mine, since the Gibraltar Mine reopened in October 2004, there has not been any significant change in installation air pressure. As shown in Table 5-54, for 37.00R57 tires, the trends for both tire life and tire wear rate are not obvious. For 40.00R57 tires, generally life decreased with the increasing installation air pressure; as well the wear rate increased when the installation air pressure increased (Table 5-55). Table 5-54 Effect of tire installation air pressure on tire life and wear rate (37.00R57) (Gibraltar) 100 psi 102 psi 104 psi Number of tires 13 5 1 Average life (hours) 4651.1 4761.4 4387 Tire wear rate (mm/hour) 0.0154 0.0170 0.0155 Table 5-55 Effect of tire installation air pressure on tire life and wear rate (40.00R57) (Gibraltar) 100 psi 102 psi 105 psi Number of tires 19 8 2 Average life (hours) 3880.6 3644.6 3239.5 Tire wear rate (mm/hour) 0.0146 0.0168 0.0213 5.2.3.1 MANOVA analysis As shown from Table 5-56 and Table 5-57 the P values for tire life are 0.9530>0.05 and 0.4803>0.05 respectively, so the effect of changing air pressure on tire life is not significant for both 37.00R57 and 40.00R57 tires. The P values for tire wear rate is 0.2064>0.05 for 37.00R57 and 0.0253O.05 for 40.00R57 tires (Table 5-58 and Table 5-59), therefore the effect of changing air pressure on tire wear rate is not significant for 37.00R57; on the other hand, the hypothesis of no effect is rejected and the air pressure change has a significant influence on the tire wear rate of 40.00R57 tires. Table 5-56 Effect of changing pressure on tire life (37.00R57) (Gibraltar) Dependent variable: L (tire life) Source DF F Value P r > F Pressure 2 0.05 0.9530 Table 5-57 Effect of changing pressure on tire life (40.00R57) (Gibraltar) Dependent variable: L (tire life) Source DF F Value P r>F Pressure 2 0.75 0.4803 Table 5-58 Effect of changing pressure on tire wear rate (37.00R57) (Gibraltar) Dependent variable: W (tire wear rate) Source DF F Value P r > F Pressure 2 1.74 0.2064 Table 5-59 Effect of changing pressure on tire wear rate (40.00R57) (Gibraltar) Dependent variable: W (tire wear rate) Source DF F Value P r > F Pressure 2 4.25 0.0253 Because there are no P values less than 0.0167 (0.05/3) in Table 5-60, Table 5-61 and Table 5-62, the effect of changing air pressure on tire life for both types and on tire wear rate for 37.00R57 is not significant from a statistical point of view. Looking at the corresponding P values in Table 5-63, it is found that there are significant differences between 3 and 1 (105 psi-100 psi), with the P value less than 0.0167 (0.05/3). On the other hand, the differences between means are not significant between 1, 2 (100 psi-102 psi) and between 2, 3 (102 psi-105 psi). This means the effect of changing air pressure on the tire tread wear rate of 40.00R57 tires is gradual, and it starts to become significant at 105 psi setting. Table 5-60 Least squares means comparison of different air pressure (37.00R57) (Gibraltar) Least squares means for effect of pressure Pr> |t for HO: LSMean(i)=LSMeanG) Dependent variable: L (tire life) i/j 1 (100 psi) 2(102 psi) 3 (104 psi) 1 (100 psi) 0.8734 0.8086 2(102 psi) 0.8734 0.7606 3 (104 psi) 0.8086 0.7606 Table 5-61 Least squares means comparison of different air pressure (37.00R57) (Gibraltar) Least squares means for effect of pressure Pr > |t for HO: LSMean(i)=LSMeanG) Dependent variable: W (tire wear rate) i/j 1 (100 psi) 2(102 psi) 3 (104 psi) 1 (100 psi) 0.0891 0.9630 2(102 psi) 0.0891 0.5103 3 (104 psi) 0.9630 0.5103 Table 5-62 Least squares means comparison of different air pressure (40.00R57) (Gibraltar) Least squares means for effect of pressure Pr> |t for HO: LSMean(i)=LSMean(j) Dependent variable: L (tire life) i/j 1 (100 psi) 2(102 psi) 3 (105 psi) 1 (100 psi) 0.7763 0.2692 2(102 psi) 0.7763 0.2354 3 (105 psi) 0.2692 0.2354 Table 5-63 Least squares means comparison of different air pressure (40.00R57) (Gibraltar) Least squares means for effect of pressure Pr> |t for HO: LSMean(i)=LSMeanG) Dependent variable: W (tire wear rate) i/j 1 (100 psi) 2(102 psi) 3 (105 psi) 1 (100 psi) 0.1250 0.0126 2(102 psi) 0.1250 0.1061 3 (105 psi) 0.0126 0.1061 5.2.3.2. Tire air pressure spikes analysis At the Gibraltar Mine, the tire air pressure spikes analysis results are discussed as following: • Table 5-64 and Table 5-65 show that most of the tire air pressure measurements are in the range of 100%-120% of initial set pressure for both 37.00R57 and 40.00R57 tires. 40.00R57 tires have comparatively higher peak psi rating range than 37.00R57 tires. • It is noticed that most spike pressures (over 120%) happened with 100 psi cold installation: 83% for 37.00R57 and 89% for 40.00R57. • It is also noticed that 57% 37.00R57 (4 tires) and 46.7% 40.00R57 (7 tires) which had experienced spikes (over 120%) were worn out with comparatively long hours. • Comparatively 40.00R57 type has more tires (51.7%-15 tires) which had experienced over 120% air pressure spikes than 37.00R57 type (36.8%-7. tires). • From Table 5-66 it can be seen that front and left rear inside positions had more air pressure spikes than the other positions for 30.00R57 tires; from Table 5-67 and Table 5-68 it can be seen that rear inside positions had more air pressure spikes than rear outside positions for 40.00R57 tires. Table 5-64 Summary of tire air pressure measurement (37.00R57) (Gibraltar) Air pressure rating <100% 100%-110% 110%-120% 120%-125% Peak psi (124%) Number of measurements 48 136 104 18 3 Percentage 15.69% 44.44% 33.99% 5.88% 0.98% Table 5-65 Summary of tire air pressure measurement (40.00R57) (Gibraltar) Air pressure rating <100% 100%-110% 110%-120% 120%-130% Peak psi (130%) Number of measurements 86 263 111 27 1 Percentage 17.66% 54% 22.79% 5.54% 0.21% Table 5-66 Summary of tire air pressure measurement with over 120% rating (37.00R57) (Gibraltar) Wheel position LF RF LRI LRO RRI RRO Number of measurements 9 6 3 0 0 0 Percentage 50% 33.33% 16.67% 0% 0% 0% Table 5-67 Summary of tire air pressure measurement with over 120% rating (40.00R57) (Gibraltar) Wheel position LF RF LRI LRO RRI RRO Number of measurements 4 4 6 2 6 5 Percentage 14.81% 14.81% 22.22% 7.41% 22.22% 18.52% Table 5-68 Summary of tire air pressure measurement with over 125% rating (40.00R57) (Gibraltar) Wheel position LF RF LRI LRO RRI RRO Number of measurements 1 1 1 0 3 0 Percentage 16.67% 16.67% 16.67% 0% 50% 0% 5.2.4 Tire Failure Analysis At the Gibraltar Mine, Table 5-69 and Table 5-70 are summaries of tire life according to various positions: • It is noted for 37.00R57 that rear inside and rear left positions had slightly higher average scrap lives than rear outside and rear right positions respectively. This is opposite to the observations from the Kemess Mine. • For 40.00R57 it is noted that rear inside and rear right positions had higher average scrap lives than rear outside and rear left positions respectively. This is still different from the Kemess Mine. • The difference of driving sides at the two mines may explain the above differences. According to the data showing failure modes for different wheel positions in Table 5-71 and Table 5-72 the following may be concluded: • Cut separation, cut sidewall, cut tread and impact break mainly accounted for shorter lives of rear inside tires for 37.00R57. • It is cut separation, cut sidewall, and impact break that mainly accounted for shorter lives of rear inside tires for 40.00R57. • There were no heat separation failures for both types. • Similar to the Kemess Mine, almost all the rear positions of both types had impact break failures. • It is also noted there were almost no evenly worn out tires at the Gibraltar Mine (only 2 evenly worn out tires). Table 5-69 Tire life comparison according to various positions (37.00R57) (Gibraltar) Number of tires Average tire life (hours) Rear inside 23 5085.3 Rear outside 14 5016.1 Rear left side 17 5074 Rear right side 20 5046.5 Front 1 2506 Rear 37 5059.1 Left 17 5074 Right 21 4925.5 Table 5-70 Tire life comparison according to various positions (40.00R57) (Gibraltar) Number of tires Average tire life (hours) Rear inside 22 3790.3 Rear outside 13 3697.4 Rear left side 14 3684.7 Rear right side 21 3803.2 Front 2 1240 Rear 35 3755.8 Left 16 3379.1 Right 21 3803.2 Table 5-71 Failure number comparison according to wheel positions (37.00R57) (Gibraltar) 1L 1R LRO LRI RRI RRO Cut separation 1 1 Cut sidewall 1 1 Cut tread 1 2 Heat separation Impact break 2 1 2 1 Worn out 6 6 9 3 Turn up separation Separation Repair failure Chipper separation Ply ending separation Chafer separation 1 Table 5-72 Failure number comparison according to wheel positions (40.00R57) (Gibraltar) 1L 1R LRO LRI RRI RRO Cut separation 1 1 3 Cut sidewall 1 1 Cut tread 1 2 Heat separation Impact break 1 2 4 1 Worn out 5 8 5 Turn up separation Separation Repair failure Chipper separation Ply ending separation Chafer separation 1 Figure 5-5 and Figure 5-6 are Pareto analyses for tires at the Gibraltar Mine. As seen from these graphs impact break and cut separation accounted for about 60% of lost tire cost for both type tires. Based on the average tire lives for impact break (4604.4 and 3348.9 hours respectively) and cut separation (2214 and 2303.5 hours respectively) it can be concluded that better maintenance of haul roads and clean up of loading and dumping areas will minimize cut separation failures. From the tire failure analysis results of the two mines it appears that different mines have different failure mechanisms due to different operating environments. 5.3 Comparison of Results between the Two Mines As can be seen in the previous sections there are some similar results for tires at the two mines. However other results are different. This is due to the different operational environments at the two mines. This section is a summary of result comparison for 40.00R57 tires between the two mines. The result comparison between 37.00R57 and 40.00R57 tires at the Gibraltar Mine is covered in the previous sections of this chapter. -Tire life • The tire life increased when the tire rotation frequency increased at both mines (Table 5-1, Table 5-27). • The average tire life at the Gibraltar Mine was 3771.3 hours (Table 5-27); while at the Kemess Mine it was 3555.4 hours (Table 5-1). The truck size difference at the two mines could be one of the reasons for this. -Wear rate • The tire wear rate decreased when the tire rotation frequency increased at both mines (Table 5-3, Table 5-31). • The age-specific tire wear rate also decreased with rotation period at both mines (Table 5-4, Table 5-33). • Both the tire wear rates and age-specific tire wear rates for each rotation period at the Kemess Mine (Table 5-5) were generally higher than that at the Gibraltar Mine (Table 5-35). • Generally rear outside tires wore faster than rear inside tires at the Kemess Mine (Table 5-5). It is the opposite at the Gibraltar Mine (Table 5-35). • Generally rear right tires had higher wear rate than rear left tires at the Kemess Mine (Table 5-5). Rear left inside position had higher wear rate than the other rear positions at the Gibraltar Mine (Table 5-35). • The wear rate difference between rear pair tires at the Gibraltar Mine (Table 5-35) was bigger than that at the Kemess Mine (Table 5-5). -Rotation sequence • The suggested rotation sequence for the first rotation is OI at both mines (Figure 5-1, Figure 5-4). • The suggested rotation sequence for the second rotation is SO at the Kemess Mine (Figure 5-1) and OI at the Gibraltar Mine (Figure 5-4). -Rotation time • The average rotation hours for each rotation were basically the same at the two mines. They were slightly shorter at the Kemess Mine (Table 5-2, Table 5-29). • Linear regression gives lower R squared values for first rotation to second rotation than for installation to first rotation at both mines (Table 5-9, Table 5-11, Table 5-41 and Table 5-45). • The suggested first rotation time is between 1750 and 1800 hours at the Kemess Mine (current hours: 1490) and between 1950 and 2000 hours at the Gibraltar Mine (current hours: 1533) based on the 30mm tread consumption (Section 5.1.2.1, Section 5.2.2.1). • Matching accounted for a similar percentage (48%) of rear rotations at the Kemess Mine (Section 5.1.2.1) to that (52%) at the Gibraltar Mine (Section 5.2.2.1). • Extending both R l (rotation interval hours between installation and first rotation) and R2 (rotation interval hours between first and second rotation) moderately will extend tire life at both mines (Section 5.1.2.2, Section 5.2.2.2). -Air pressure setting • Generally tire life decreased with the increasing tire installation air,pressure; as well the tire wear rate increased when the tire installation air pressure increased at both mines (Table 5-16, Table 5-55). • The air pressure change has a significant influence on the tire wear rate at both mines (Table 5-18, Table 5-59). • The air pressure change has a significant influence on the tire life at the Kemess Mine; on the other hand, the effect of changing air pressure on tire life is not significant at the Gibraltar Mine (Table 5-17, Table 5-57). -Spike air pressure • The tire air pressure variation range at the Kemess Mine (Table 5-21: 100-140%) was wider than that at the Gibraltar Mine (Table 5-65: 100-130%). • Most tire air spike pressures happened with the lowest cold air pressure installation at both mines (95 psi at the Kemess and 100 psi at the Gibraltar). This is because the lower initial setting air pressure allows tire air pressure to rise at a higher percentage (Section 5.1.3.2, Section 5.2.3.2). • The worn out percentages of tires which had experienced spike pressures were similar at the two mines (51% at the Kemess vs. 46.7 % at the Gibraltar) (Section 5.1.3.2, Section 5.2.3.2). • The peak psi (Table 5-21: over 140%) at the Kemess Mine was higher than that (Table 5-65: 130%) at the Gibraltar Mine. • Rear inside positions had more air pressure spikes than rear outside positions at both mines (Table 5-22, Table 5-23, Table 5-67 and Table 5-68). • Left front had higher air pressure spike percentage than right front at the Kemess Mine (Table 5-22 and Table 5-23) while there was no difference between front tires at the Gibraltar Mine (Table 5-67 and Table 5-68). -Tire failure • Rear outside and rear right positions had higher average scrap lives than rear inside and rear left positions respectively at the Kemess Mine (Table 5-24); rear inside and rear right positions had higher average scrap lives than rear outside and rear left positions respectively at the Gibraltar Mine (Table 5-70). • Cut separation, cut tread and impact break mainly accounted for shorter lives of rear inside tires at the Kemess Mine (Table 5-25); while it is cut separation, cut sidewall, and impact break that mainly accounted for shorter lives of rear inside tires at the Gibraltar Mine (Table 5-72). • Unlike the Kemess Mine (Table 5-25), there were no heat separation failures at the Gibraltar Mine (Table 5-72). • Both mines had comparatively high-number of impact failures. Similar to the Kemess Mine (Table 5-25), almost all the rear positions had impact break failures at the Gibraltar Mine (Table 5-72). There were a comparatively high percentage of evenly worn out tires at the Kemess Mine (Table 5-24); on the other hand, there were almost no evenly worn out tires at the Gibraltar Mine. Cut tread, impact break, cut separation and cut sidewall (about 46%) accounted for about 80%) of lost tire cost at the Kemess Mine (Figure 5-2). At the Gibraltar Mine, impact break and cut separation (about 33%) accounted for about 63% of lost tire cost (Figure 5-6). The Pareto analysis curve at the Gibraltar Mine is steeper than that at the Kemess Mine. Dominant failures can be minimized through better maintenance of the haul roads and clean up of the loading and dumping areas at both mines (Section 5.1.4, Section 5.2.4). 6. Conclusions Tires are playing a continuing important role in the mining industry. This thesis presents a brief review of the relevant research and carries out analyses on tire rotation practices such as rotation sequence, rotation time and tire air pressure setting at installation. An approach to applying statistics for the determination of rotation time and installation air pressure setting influence on tire life and tread wear rate has been demonstrated. Also tire rotation sequence acronyms are created to assist in analyzing and determining rotation sequence. As additional data is collected, greater confidence in the estimates obtained from the models will be gained. Currently the suggested rotation sequence is being implemented by Fountain Tire Mine Service Ltd. at the Kemess Mine and computer programs will be developed based on these analytical methods and procedures. In general this thesis has achieved the following: • Enhanced the understanding of tire maintenance practices, relevant mine operation and their impact on tire performance. • Expanded the knowledge of off-the-road tire management. • Established an analytical approach and developed the application of analytical methods in the field of off-the-road tire research. • Quantified the influences of major maintenance factors on tire life. • The knowledge gained from the analyses helped in the evaluation of current tire management approaches. Case-specifically, the following conclusions arise from the research: At the Kemess Mine: • The suggested rotation sequence is OI-SO-00 from the first rotation to the third rotation (note: this sequence is different from the current practice at the Kemess Mine). • The suggested first rotation time is between 1750 and 1800 hours based on 30mm tread consumption. • The tread consumption prediction models are established for two rotation periods. • High tire installation air pressure does influence tire life and tire tread wear rate. • Rear inside and left front positions experienced the most air pressure spikes. • Dominant failures can be minimized through better maintenance of the haul roads and clean up of the loading and dumping areas. At the Gibraltar Mine: • The suggested rotation sequences are OI-SI for 37.00R57 and OI-OI for 40.00R57 from the first rotation to the second rotation. • The suggested first rotation times are between 2000 and 2050 hours for 37.00R57 and between 1950 and 2000 hours for 40.00R57 based on 30mm tread consumption. • The tread consumption prediction models are established for two rotation periods of both type tires. • Higher tire installation air pressure does influence tire tread wear rate of 40.00R57 tires. • Left rear inside and front positions for 37.00R57 and rear inside positions for 40.00R57 experienced the most air pressure spikes. • Dominant failures can be minimized through better maintenance of the haul roads and clean up of the loading and dumping areas. Summarizing the results from the two mines, it can also be concluded: • Previous analyses have shown some similar results for tires at both mines; other results differ due to different operating environments. Therefore independent analyses are necessary for each mine. It should also be noted that some results are based on limited data from the two mines. Thus their general applicability as of yet is not known. 7. Future Work To continue the current analyses on tire maintenance practices, the following specific things are being pursued: • Ongoing data is being collected from the Kemess Mine and the Gibraltar Mine for additional analyses. • Further analyses will be carried out with the objective of determining suitable tire air pressure for different operating situations. • Data will be collected from other mines to perform similar analyses. • Develop optimized tire maintenance practices and procedures specific to Fountain Tire Mine Service Ltd. based on the analytical results and the company's experience. The focus of the ongoing project, however, will be on quantifying the impact of major operational and environmental factors on tire performance. This will provide insight into the tire interactions for improved tire management. Analyses will be conducted on the following relations in two categories: • Loading. This includes the relations between payload and tire life, payload and tire tread wear rate, payload and tire air pressure, etc. Road resistance. This includes three relations: the relation between tire life or tire tread wear rate and haul roads profile such as haul distance, road slope angles and number, road curve radii and number, etc.; the relation between haul roads profile and failure mechanism; the relation between tire life or tire tread wear rate and the climate such as ambient temperature, rain, snow, etc. References Adams, B. T., Reid, J. R, Hummel, J. W., Zhang, Q. and Hoeft, R. G , "Effects of Central Tire Inflation Systems on Ride Quality of Agricultural Vehicles", Journal of Terramechanics, Vol. 41, No. 4, October, pp. 199-207, 2004. Amberlang, J. C , "Testing of Tire Treadwear under Laboratory and under Service Conditions", Tire Science and Technology, TSTCA, Vol 1, No. 1, February, pp. 39-46, 1973. Balkwill, K. J., "Development of A Comprehensive Method for Modeling Performance of Aircraft Tires Rolling or Braking on Dry and Precipitation-contaminated Runways", microform, Montreal Transport Canada, Transportation Development Centre, 2003. Barbanti, G.., Pellicciary, M . , Andrisano, A. , "On Tire Monitoring Systems Temperature Compensation", SAE 2004 World Congress & Exibition, Detroit, MI, USA, March 2004. Blackwell, G., "Estimation of Large Open Pit Haulage Truck Requirements", C I M A G M , Edmonton, Canada, pp. 255-273, 1996. Bradley, A . H. , "Testing A Central Tire Inflation System in Western Canadian Log-hauling Conditions", Roads & Transportation, Technical Note TN-197, Forest Engineering Research Institute of Canada, June, 1993. Brenner, F. C , Scheiner, S. R., and Kondo, A. , "Effect of Tire Wear on Wear Rate", Tire Science and Technology, TSTCA, Vol . 3, No. 4, November, pp. 235-251, 1975. Carter, R. A. , "Where the Rubber Meets the Road", Coal Age, Vol . 103, December, pp. 28-30, 1998. Chadwick, J., " O E M Components", Mining Magazine, pp. 12-14, February, 2005. Craighead, I. A. , "Sensing Tyre Pressure, Damper Condition and Wheel Balance from Vibration Measurements", Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, Vol. 211, November, pp. 257-265, 1997. De, A . and Mukhopadhyay, A . K. , "Selection, Maintenance, and Relations of Various Parameters for Off-Highway Hauling Tires", Off-Highway Haulage in Surface Mines, eds. Golosinski and Srajer, Balkema, Rotterdam, pp. 153-159, 1989. Dillon, W. R. and Goldstein, M . , "Multivariate Analysis, Methods, and Applications", John Wiley and Sons, Toronto, pp. 337-359, 1984. Douglas, R. A. , Wood ward, W. D. H. , and Woodside; A . R., "Road Contact Stresses and Forces Under Tires with Low Inflation Pressure", Canadian Journal of Civil Engineering, Vol. 27, No. 6, December, pp. 1248-1258, 2000. Ednie, H. , "Real-Time Monitoring Avoids Early Failures", C I M Bulletin, Vol . 95, October, pp. 17-19,2002. Elkview Mine, Goodyear and Fountain Tire Mine Service Ltd., "Cause and Effect Table", March, 2005. Environment Canada, "Canadian Climate Normals 1971-2000" via the Internet, 9 August 2006, http://www.climate.weatheroffice.ec.gc.ca/climate normals/index e.html. Fervers, C. W., "Improved F E M Simulation Model for Tire-Soil Interaction", Journal of Terramechanics, http//www.elsevier.com/locate/jterra, Vol . 41, Issues 2-3, April-July, pp. 87-100, 2004. Fowler, G. and Huntingford, K., Personal Communication, Fountain Tire Mine Service Limited, Vancouver, Canada, 2005. Fowler, G. and Huntingford, K. , Personal communication, Fountain Tire Mine Service Limited, Vancouver, Canada, 2006. Goodyear, "Goodyear Off-The-Road Tires Engineering Data Book", 2003. Goodyear, "Off-The-Road Tire Training, Graduate Course", 2004. Goodyear, "Off-The-Road Training, Level II", the Gibraltar Mine, June 13-15, 2005. Grossmann, R., "Wireless Measurement of Tire Pressure with Passive Quartz Sensors", Proceedings of the 1999 Smart Structures and Materials-Sensory Phenomena and Measurement Instrumentation for Smart Structures and Materials, Newport Beach, CA, USA, SPIE Vol. 3670, pp. 214-222, 1999. Hall, R. A. , "Maintenance Engineering", MINE 582 Course Notes, University of British Columbia, Vancouver, Canada, January-April, 2005. Huhtala, M . , Pihlajamaki, J. and Pienimaki, M . , "Effects of Tires and Tire Pressures on Road Pavements", Transportation Research Record, Vol . 0361-1981, pp. 107-114, 1989. InfoMine Inc., "Kemess" via the Internet, 9 August 2006, http://vvfww.infomine.com/minesite/minesite.asp?site=kemess. Janowski, W. R., "Tire Traction Testing in Adverse Environments", Frictional Interaction of Tire and Pavement, A S T M STP 793, eds. Meyer, W.E. and Walter, J. B., American Society for Testing and Materials, pp. 65-78, 1983. Kal Tire, "Tire Management Services Program" via the Internet, 24 January 2005, http//www.kaltire.com/commercial/medium_truck_tires/management.php. Kaufman, W. W. and Ault, J. C , "Design of Surface Mine Haulage Roads-A Manual", National Institute for Occupational Safety and Health, Pittsburgh Research Laboratory Library, 2001. Knights, P. F. and Boerner, A . L. , "Statistical Correlation of Off-Highway Tire Failures with Open Pit Haulage Routes", Mining Engineering, Vol . 53, August, pp. 51-57, 2001. Krivtsov, V . V . , Tananko, D. E. and Davis, T. P., "Regression Approach to Tire Reliability Analysis", Reliability Engineering & System Safety, http//www.elsevier.com/locate/ress, Vol . 78, Issue 3, December, pp. 267-273, 2002. Kutner, M . H. , Nachtsheim, C. J., Neter, J., and L i , W., "Applied Linear Statistical Models", McGraw-Hill/Irwin Series Operations and Decision Sciences, 2004. LeMay, V . , "Applied Multivariate Statistics", FRST 531 Course Notes, University of British Columbia, Vancouver, Canada, January-April, 2006. LeMay, V. , "Problems in Statistical Methods", FRST 533 Course Notes, University of British Columbia, Vancouver, Canada, September-December, 2006. Lippmann, S. A. , " Effects of Tire Structure and Operating Conditions on the Distribution of Stress between the Tread and the Road", The Tire Pavement Interface, A S T M STP 929, eds. Pottinger, M . G . and Yager, T. J., American Society for Testing and Materials, Philadelphia, pp. 91-109, 1986. Ludema, K. C. and Gujrati, B. D., "An Analysis of the Literature on Tire-Road Skid Resistance", A S T M Special Technical Publication 541, American Society for Testing and Materials, Philadelphia, Pa., pp. 11-17, 29-42 and 141-145, 1973. Michel, P., "QCM-Tyresense Project", Large Tire User Group Fall Meeting, Edmonton, Canada, November 16 t h-17 t h, 2005. Michelin's Earthmover Group, "Ten Tips to Guarantee Tire Toughness and Longevity", Coal Age, Vol . 104, August, pp. 51-52, 1999. Mills, R. W., "Central Tire Inflation Systems and the Effects of Tire Pressure on Logging Roads", Graduating essay, the Faculty of Forestry, the University of British Columbia, April, 2004. Northgate Minerals Corporation, "the Kemess Mine" via the Internet, 8 August 2006, http://www.northgateexploration.ca/frame kemess mine.html. O'Neil, T., "Earthmover Tires-Big tires, Big Shortages", Mining Engineering, Vol . 58, February, pp. 29-32, 2006. O'Neil, T., "Goodyear Unveils Two Piece Assembly for Mining Tires", Mining Engineering, Vol . 55, October, pp. 38-40, 2003. Owende, P. M . O., Hartman, A . M . , Ward, S. M . , Gilchrist, M . D. and O'Mahony, M . J., "Minimizing Distress on Flexible Pavements Using Variable Tire Pressure", Journal of Transportation Engineering, Vol. 127, No. 3, May/June, pp. 254-262, 2001. Pit & Quarry University, "Lesson 10 Tires" via the Internet, 26 January 2001, http//www.pitandquarry.com/univLl 0-1 .html; http//www.pitandquarry.com/univLl 0-2.html; http//www.pitandquarry.com/univLl 0-3 .html. Pytka, J., Dabrowski, J., Zajac, M . , and Tarkowski, P., "Effects of Reduced Inflation Pressure and Vehicle Loading on Off-road Traction and Soil Stress and Deformation State", Journal of Terramechanics, Vol. 43, No. 4, October, pp. 469-485, 2006. Rasche, T., Klinge, T., Sugden, N . , Trenberth, D., Field, P. and Southey, R., "Best Practice Tyre Application Risk Management (TYREARM™) at BHP Billiton's B M A Goonyella Riverside Mine", Fifth Large Open Pit Mining Conference, Kalgoorlie, WA, pp. 71-78, 3-5 November, 2003. Rimex, "TyreSense" via the Internet, 4 August 2006, http://www.tyresense.com/. Sergio, M . , Manaresi, N . , Tartagni, M . , Guerrieri, R., and Canegallo, R., "On Road Tire Deformation Measurement System Using A Capacitive-resistive Sensor", Proceedings of IEEE Sensors, Vol. 2, No. 2, pp. 1059-1063, 2003. Shalaby, A . and Reggin, A. , "Surface Rutting of Thin Pavements and Gravel Roads Under Standard and Reduced Tire Inflation Pressures", Canadian Journal of Civil Engineering, Vol. 29, No. 5, October, pp. 679-691, 2002. Stelzer, A. , Schimetta, G.., Reindl, L., Springer, A . and Weigel R., "Wireless SAW Sensors for Surface and Subsurface Sensing Applications", Proceedings of SPIE-The International Society for Optical Engineering, Vol. 4491, pp. 358-366, 2001. Sympson, T., "Pushing Nitrogen" via the Internet, 11 August 2006, http://www.pitandquarrv.com/pitandquarry/article/articleDetail.isp?id=254530. Tannant, D. D., and Regensburg B., "Guidelines for Mine Haul Road Design", School of Mining and Petroleum Engineering, Department of Civil and Environmental Engineering, University of Alberta. 109 p., 2001. Taseko Mines Limited, "Gibraltar" via the Internet, 8 August 2006, http ://www. tasekomines. com/tko/Gibraltar. asp. Thompson, R. J. and Visser, A . , T., "An Integrated Haul Road Design System to Reduce Cost per Tonne Hauled", World Mining Equipment Haulage 2002 Conference, Tucson, A z . , U S A . , 19-22 May 2002. Trenn, J. A. , "Using Variable Tire Pressure Technology with Off-road Forestry Equipment", Graduating essay, the Faculty of Forestry, the University of British Columbia, May, 1997. Varadan, V. K. , Jose, K. A . , and Varadan, V. V , "Design and Development of Passive MEMS-IDT Sensors for Continuous Monitoring of Tire Pressure", Proceedings of SPIE-The International Society for Optical Engineering, Vol. 4236, pp. 242-251, 2001. Varadan, V. V , Tellakula, A . R., Hollinger, R. D., L i , C. T., and Varadan, V. K., "Wireless IDT Microsensors for Subsurface Sensing", Proceedings of SPIE-The International Society for Optical Engineering, Vol. 4129, pp. 352-358, 2000. Veith, A . G., "The Most Complex Tire-Pavement Interaction: Tire Wear", The Tire Pavement Interface, A S T M STP 929, eds. Pottinger, M.G. and Yager, T. J., American Society for Testing and Materials, Philadelphia, pp. 125-158, 1986. Veith, A . G., "Tire Tread Wear-the Joint Influence of Compound Properties and Environmental Factors", Tire Science and Technology. TSTCA, Vol . 23, No. 4, October-December, pp. 212-237, 1995. Werner, J. T. and Barrowman, K. , "Remote Real-Time Tire Monitoring in Open Pit Mines", CIM Conference, Vancouver, 2002. Zhou, J., Hall, R. A . , Fowler, G. and Huntingford, K., "Applications of Engineering Analysis to Improve Tire Management", CIM Conference, Vancouver, May, 2006. Appendix A Fountain Tire Mine Service Limited tire air pressure and tread depth check sheet (Fowler and Huntingford, 2005) A0 "SBUHETBlSL sesaerB a Bf/ST~SE!P TO. MINE SITE to add - weather - wet / dry / snow SERVICEMEN DATE CHECKED A B c D E F H C L.F. L.R.O. L.R.I. R.R.I. R.R.O. R.F. Truck Truck Truck Ambient Time Loading 0 L R.t.d. R.t.d. R.t.d. R.t.d. R.t.d. R.t.d. No. Hours Location Temp. Checked Tool T D (mm) psi (mm) psi (mm) psi (mm) psi (mm) psi (mm) psi Comments Or Observations Appendix B Examples of IID graphical tests (Kemess Mine) Correlation Test 6,000 -| 5,000 4,000 I | 3,000 2,000 1,000 * • • • • 1 • • » •* ^ • * » • — — — • — - — • —•—»--> 1,000 2,000 3,000 4,000 (i-l)th tire life 5,000 6,000 Correlation Test Trend Test Cumulative tread consumption Appendix C SAS output of canonical correlation between the rotation interval hours and the tire life at the Kemess Mine Relations of three rotation interval hours with tire life 08:51 Wednesday, April 26, 2006 The CANCORR Procedure Rotation interval hours 3 Tire life 1 Observations 25 Means and Standard Deviations Standard Variable Mean Deviation Label R l 1317.7600 439.9469 R l R2 1049.4000 627.1576 R2 R3 952.1600 698.6783 R3 L 4209.3600 682.9315 L The CANCORR Procedure Correlations among the Original Variables Correlations Among the rotation interval hours R l R2 R3 R l 1.0000 -0.0383 -0.1862 R2 -0.0383 1.0000 -0.1493 R3 -0.1862 -0.1493 1.0000 Correlations Among the tire life 1.0000 Correlations Between the rotation interval hours and the tire life L R l 0.5481 R2 0.5239 R3 0.2653 The C A N C O R R Procedure Canonical Correlation Analysis Adjusted Approximate Squared Canonical Canonical Standard Canonical Correlation Correlation Error Correlation 1 0.9031 0.8981 0.0376 0.8156 Eigenvalues of Inv(E)*H = CanRsq/(l-CanRsq) Eigenvalue Difference Proportion Cumulative 4.4242 1.0000 1.0000 Test of HO: The canonical correlations in the current row and all that follow are zero Likelihood Approximate Ratio F Value NumDF Den DF Pr>F 0.1844 30.97 3 21 <.0001 NOTE: The F statistic is exact. Multivariate Statistics and Exact F Statistics S=l M=0.5 N=9.5 Statistic Wilks' Lambda Pillai's Trace Value F Value NumDF Den DF Pr>F 0.1844 30.97 0.8156 30.97 Hotelling-Lawley Trace 4.4242 30.97 Roy's Greatest Root 4.4242 30.97 3 3 3 3 21 21 21 21 <.0001 <.0001 <.0001 <.0001 The CANCORR Procedure Canonical Correlation Analysis Raw Canonical Coefficients for the rotation interval hours hours 1 R l R l 0.0017 R2 R2 0.0011 R3 R3 0.0008 . Raw Canonical Coefficients for the tire life lifel L L 0.0015 The CANCORR Procedure Canonical Correlation Analysis Standardized Canonical Coefficients for the rotation interval hours hours 1 R l R l 0.7324 R2 R2 0.6877 R3 R3 0.5329 Standardized Canonical Coefficients for the tire life lifel L L 1.0000 The CANCORR Procedure Canonical Structure Correlations Between the rotation interval hours and Their Canonical. Variables hours 1 R l R l 0.6069 R2 R2 0.5801 R3 R3 0.2938 Correlations Between the tire life and Their Canonical Variables lifel L L 1.0000 Correlations Between the interval hours and the Canonical Variables of the tire life lifel R l R l 0.5481 R2 R2 0.5239 R3 R3 0.2653 Correlations Between the tire life and the Canonical Variables of the interval hours hours 1 L L 0.9031 The CANCORR Procedure Canonical Redundancy Analysis Raw Variance of the rotation interval hours Explained by Their Own The Opposite Canonical Variables Canonical Variables Canonical Variable Cumulative Canonical Cumulative Number Proportion Proportion R-Square Proportion Proportion 1 0.2286 0.2286 0.8156 0.1865 0.1865 Raw Variance of the tire life Explained by Their Own The Opposite Canonical Variables Canonical Variables Canonical Variable Cumulative Canonical Cumulative Number Proportion Proportion R-Square Proportion Proportion 1 1.0000 1.0000 0.8156 0.8156 0.8156 The CANCORR Procedure Canonical Redundancy Analysis Standardized Variance of the rotation interval hours Explained by Their Own The Opposite Canonical Variables Canonical Variables Canonical Variable Cumulative Canonical Cumulative Number Proportion Proportion R-Square Proportion Proportion 1 0.2637 0.2637 0.8156 0.2151 0.2151 Standardized Variance of the tire life Explained by Their Own The Opposite Canonical Variables Canonical Variables Canonical Variable Cumulative Canonical . . Cumulative Number Proportion Proportion R-Square Proportion Proportion 1 1.0000 1.0000 0.8156 0.8156 0.8156 The CANCORR Procedure Canonical Redundancy Analysis Squared Multiple Correlations Between the rotation interval hours and the First M Canonical Variables of the tire life M 1 R l R l 0.3004 R2 R2 0.2745 R3 R3 0.0704 Squared Multiple Correlations Between the tire life and the First M Canonical Variables of the rotation interval hours M 1 L L 0.8156 Appendix D SAS output of canonical correlation between the three rotation interval hours and the corresponding three age-specific wear rates at the Kemess Mine Relations of three wear rates with three rotation 18:06 Tuesday, April 11, 2006 interval hours The CANCORR Procedure wear rates 3 rotation interval hours 3 Means and Standard Deviations Standard Variable Mean Deviation Label A l 0.0207 0.0103 A l A2 0.0152 0.0049 A2 A3 0.0153 0.0100 A3 R l 1317.7600 439.9469 R l R2 1049.4000 627.1576 R2 R3 952.1600 698.6783 R3 The CANCORR Procedure Correlations Among the Original Variables Correlations Among the wear rates A l A2 A3 A l 1.0000 -0.0450 -0.0455 A2 -0.0450 1.0000 -0.0167 A3 -0.0455 -0.0167 1.0000 Correlations Among the rotation interval hours R l R2 R3 R l 1.0000 -0.0383 -0.1862 R2 -0.0383 1.0000 -0.1493 R3 -0.1862 -0.1493 1.0000 Correlations Between the wear rates and the rotation interval hours R l R2 R3 A l -0.5963 0.0242 0.1688 A2 -0.3477 -0.3196 -0.3908 A3 0.0039 -0.0239 -0.0744 The C A N C O R R Procedure Canonical Correlation Analysis Adjusted Approximate Squared Canonical Canonical Standard Canonical Correlation Correlation Error Correlation 0.8101 0.7771 0.0702 0.6562 0.4852 0.4398 0.1561 0.2354 0.0201 0.2040 0.0004 Eigenvalues of Inv(E)*H CanRsq/(l-CanRsq) Eigenvalue Difference Proportion Cumulative 1.9089 1.6011 0.8610 0.8610 0.3078 0.3074 0.1388 0.9998 0.0004 0.0002 •1.0000. Test of HO: The canonical correlations in the current row and all that follow are zero Likelihood Approximate Ratio F Value Num DF Den DF P r>F 1 0.2627 3.77 9 46.392 0.0013 2 0.7643 1.44 4 40 0.2391 3 0.9996 0.01 1 21 0.9274 Multivariate Statistics and F Approximations S=3 M=-0.5 N=8.5 Statistic Value F Value Num DF Den DF Pr>F Wilks'Lambda 0.2627 3.77 9 46.392 0.0013 Pillai's Trace 0.8920 2.96 9 63 0.0053 Hotelling-Lawley Trace 2.2172 4.53 9 26.846 0.0011 Roy's Greatest Root 1.9089 13.36 3 21 <.0001 NOTE: F Statistic for Roy's Greatest Root is an upper bound. . The CANCORR Procedure Canonical Correlation Analysis Raw Canonical Coefficients for the wear rates wearl wear2 wear3 A l A l -56.8897 78.6943 7.7307 A2 A2 -170.1834 -109.0483 -24.8666 A3 A3 -9.6371 -10.9373 98.5868 Raw Canonical Coefficients for the rotation interval hours hours 1 hours2 hours3 R l R l 0.0021 -0.0010 -0.0003 R2 R2 0.0007 0.0008 0.0012 R3 R3 0.0007 0.0010 -0.0008 The CANCORR Procedure Canonical Correlation Analysis • Standardized Canonical Coefficients for the wear rates wearl wear2 wear3 A l A l -0.5852 0.8095 0.0795 A2 A2 -0.8.367 -0.5361 -0.1223 A3 A3 -0.0968 -0.1099 0.9904 Standardized Canonical Coefficients for the rotation interval hours hours 1 hours2 hours3 R l R l 0.9032 -0.4594 -0.1182 R2 R2 0.4281 0.4885 0.7782 R3 R3 0.5227 0.7177 -0.5241 The CANCORR Procedure Canonical Structure Correlations Between the wear rates and Their Canonical Variables wearl wear2 wear3 A l A l -0.5432 0.8386 0.0399 A2 A2 -0.8087 -0.5707 -0.1424 A3 A3 -0.0562 -0.1378 0.9889 Correlations Between the rotation interval hours and Their Canonical Variables hours 1 hours2 hours3 R l R l 0.7895 -0.6117 -0.0504 R2 R2 0.3155 0.3989 0.8610 R3 R3 0.2906 0.7303 -0.6183 Correlations Between the wear rates and the Canonical Variables of the rotation interval hours hours 1 hours2 hours3 A l A l -0.4400 0.4069 0.0008 A2 A2 -0.6551 -0.2769 -0.0029 A3 A3 -0.0455 -0.0668 0.0199 Correlations Between the rotation interval hours and the Canonical Variables of the wear rates wearl wear2 wear3 R l R l 0.6395 -0.2968 -0.0010 R2 R2 0.2556 0.1935 0.0173 R3 R3 0.2354 0.3543 -0.0124 The CANCORR Procedure Canonical Redundancy Analysis Raw Variance of the wear rates Explained by Their Own The Opposite Canonical Variables Canonical Variables Canonical Variable Cumulative Canonical Cumulative Number Proportion Proportion R-Square Proportion Proportion 1 0.2051 0.2051 0.6562 0.1346 0.1346 2 0.3647 0.5698 0.2354 0.0858 0.2204 3 0.4302 1.0000 0.0004 0.0002 0.2206 Raw Variance of the rotation interval hours Explained by Their Own The Opposite Canonical Variables Canonical Variables Canonical Variable Cumulative Canonical Cumulative Number Proportion Proportion R-Square Proportion Proportion 1 0.1870 0.1870 0.6562 0.1227 0.1227 2 0.3678 0.5547 0.2354 0.0866 0.2093 3 0.4453 1.0000 0.0004 0.0002 0.2094 The CANCORR Procedure Canonical Redundancy Analysis Standardized Variance of the wear rates Explained by Their Own The Opposite Canonical Variables Canonical Variables Canonical Variable Cumulative Canonical Cumulative Number Proportion Proportion R-Square Proportion Proportion 1 0.3174 0.3174 0.6562 0.2083 0.2083 2 0.3493 0.6668 0.2354 0.0822 0.2905 3 0.3332 1.0000 0.0004 0.0001 0.2907 Standardized Variance of the rotation interval hours Explained by Their Own The Opposite Canonical Variables Canonical Variables Canonical Variable Cumulative Canonical Cumulative Number Proportion Proportion R-Square Proportion Proportion 1 0.2691 0.2691 0.6562 0.1766 0.1766 2 0.3555 0.6246 0.2354 0.0837 0.2603 3 0.3754 1.0000 0.0004 0.0002 0.2604 The CANCORR Procedure Canonical Redundancy Analysis Squared Multiple Correlations Between the wear rates and the First M Canonical Variables of the rotation interval hours M 1 2 3 A l A l 0.1936 0.3592 0.3592 A2 A2 0.4292 0.5059 0.5059 A3 A3 0.0021 0.0065 0.0069 Squared Multiple Correlations Between the rotation interval hours and the First M Canonical Variables of the wear rates M 1 2 3 R l R l 0.4090 0.4971 0.4971 R2 R2 0.0653 0.1028 0.1031 R3 R3 0.0554 0.1809 0.1811 Appendix E SAS output of tire air pressure MANOVA analysis at the Gibraltar Mine (40.00R57) M U L T I V A R I A T E ANALYSIS OF V A R I A N C E 17:43 Monday, August 28, 2006 The G L M Procedure Class Level Information Class Levels Values Pressure 100 102 105 Source DF Model 2 Error 26 Corrected Total 28 Number of observations 29 Dependent Variable: L L Sum of Squares 111.3421 Mean Square 55.6711 1918.6579 73.7945 2030.0000 F Value Pr > F 0.75 0.4803 R-Square CoeffVar Root M S E L Mean 0.0548 57.2692 8.5904 15.0000 Source DF Type ISS Mean Square F Value Pr> F Pressure 111.3421 55.6711 0.75 0.4803 Source Pressure DF Type HISS Mean Square F Value Pr>F 111.3421 55.6711 0.75 0.4803 Source Model Error DF 26 Dependent Variable: W W Corrected Total 28 Squares 0.0580 0.1772 0.2352 Sum of Mean Square 0.0290 0.0068 F Value Pr > F 4.25 0.0253 R-Square CoeffVar Root M S E WMean 0.2465 -4.5488 0.0826 •1.8150 Source Pressure DF Type I SS Mean Square F Value Pr > F 0.0580 0.0290 4.25 0.0253 Source Pressure DF Type HISS Mean Square F Value Pr>F 0.0580 0.0290 4.25 0.0253 The G L M Procedure Least Squares Means Standard L S M E A N Pressure L L S M E A N Error Pr > |t| Number 100 15.2105 1.9708 <.0001 1 102 16.2500 3.0372 <.0001 2 105 8.0000 6.0743 0.1993 3 Least Squares Means for effect Pressure Pr > |t| for HO: LSMean(i)=LSMean(j) Dependent Variable: L /j 1 2 3 1 0.7763 0.2692 2 0.7763 0.2354 3 0.2692 0.2354 Standard L S M E A N Pressure W L S M E A N Error Pr>|t| Number 100 -1.8415 0.0189 <0001 1 102 -1.7864 0.0292 <.0001 2 105 -1.6771 0.0584 <.0001 3 Least Squares Means for effect Pressure Pr > |t| for HO: LSMean(i)=LSMeanO) Dependent Variable: W i/j 1 2 3 1 0.1250 0.0126 2 0.1250 0.1061 3 0.0126 0.1061 NOTE: To ensure overall protection level, only probabilities associated with pre-planned comparisons should be used. E = Error SSCP Matrix L W L 1918.6579 -1.5169 W -1.5169 0.1772 Partial Correlation Coefficients from the Error SSCP Matrix / Prob > |r| DF = 26 L W L 1.0000 -0.0823 0.6833 W -0.0823 1.0000 0.6833 H = Type III SSCP Matrix for Pressure L W L 111.3421 -1.7503 W -1.7503 0.0580 Characteristic Roots and Vectors of: E Inverse * H , where H = Type III SSCP Matrix for Pressure E = Error SSCP Matrix Characteristic Characteristic Vector V'EV=1 Root Percent L W 0.3427 92.13 -0.0051 2.2716 0.0293 7.87 0.0223 0.7218 M A N O V A Test Criteria and F Approximations for the Hypothesis of No Overall Pressure Effect H = Type III SSCP Matrix for Pressure E = Error SSCP Matrix S=2 M=-0.5 N=11.5 Statistic Value F Value NumDF Den DF Pr>F Wilks' Lambda 0.7236 2.19 4 50 0.0829 Pillai's Trace 0.2837 2.15 4 52 0.0878 Hotelling-Lawley Trace 0.3720 2.30 4 28.992 0.0829 Roy's Greatest Root 0.3427 4.46 2 26 0.0217 NOTE: F Statistic for Roy's Greatest Root is an upper bound. NOTE: F Statistic for Wilks' Lambda is exact. Plot of residl *predictl. Symbol is value of Pressure. 15 -15 ifrfinffsnfnf'fSfnffffffss'nnnffifiirfffnnffnff'iffffunffH'ff 3 10 12 14 16 18 predict 1 Plot of resid2*predict2. Symbol is value of Pressure. resid2 , 0.20 " -0.20 -1.850 -1.825 -1.800 -1.775 -1.750 -1.725 -1.700 -1.675 predict2 NOTE: 6 obs hidden. The UNIVARIATE Procedure Variable: residl Moments N Mean Std Deviation Skewness Uncorrected SS Coeff Variation 29 0 8.2779 -0.1612 1918.6579 Sum Weights Sum Observations Variance Kurtosis Corrected SS Std Error Mean 29 0 68.5235 -0.9656 1918.6579 1.5372 Basic Statistical Measures Location Variability Mean 0.0000 Std Deviation 8.2779 Median 0.7895 Variance 68.5235 Mode . Range 29.0395 Interquartile Range 12.9605 Tests for Location: Mu0=0 Test -Statistic- -p Value-Test Student's t t 0 Pr>|t| 1.0000 Sign M 0.5 Pr>=|M| 1.0000 Signed Rank S 3 Pr>=|S| 0.9496 Variable: residl Tests for Normality —Statistic— -p Value-Shapiro-Wilk Kolmogorov-Smirnov Cramer-von Mises Anderson-Darling W 0.9684 D 0.0685 W-Sq 0.0282 A-Sq 0.2274 P r < W 0.5163 P r > D >0.1500 Pr > W-Sq >0.2500 Pr>A-Sq >0.2500 Quantiles (Definition 5) Quantile Estimate 100% Max 13.7895 99% 13.7895 95% 11.7895 90% 11.7500 75% Q3 6.7500 Quantiles (Definition 5) Quantile Estimate 50% Median 0.7895 25% Q l -6.2105 10% -13.2105 5% -13.2500 1% -15.2500 0 % M i n -15.2500 Extreme Observations Lowest Highest Value Obs Value Obs -15.2500 27 9.7500 23 -13.2500 20 9.7895 . 12 -13.2105 13 11.7500 21 -11.2105 17 11.7895 9 -9.2105 6 13.7895 4 Stem Leaf # Boxplot 1 00224 5 | 0 56779 5 +-—+ 0 12233 5 *--+--* -0 43210 5 | | -0 98765 5 +-—+ -1331 3 | -15 1 | .—+.—+.—+-—+ Multiply Stem.Leaf by 10**+1 Variable: residl Normal Probability Plot 12.5+ *++*++ * _2 5+ - | ~ | _ * h < * * | .)—)-***-(-** | -|_|-*-|—(.** -17.5+++++*+ +....+.—+.—+.—+.—+----+.—+.—+-—+----+ -2 -1 0 +1 +2 The UNIVARIATE Procedure Variable: resid2 Moments N 29 Sum Weights 29 Mean 0 Sum Observations 0 Std Deviation 0.0796 Variance 0.0063 Skewness -0.6080 Kurtosis 1.1265 Uncorrected SS 0.1772 Corrected SS 0.1772 Coeff Variation . Std Error Mean 0.0148 Basic Statistical Measures Location Variability Mean 0.0000 Std Deviation 0.0796 Median 0.0122 Variance 0.0063 Mode . Range 0.3599 Interquartile Range 0.1023 Tests for Location: Mu0=0 Test -Statistic- -p Value-Student's t t 0 Pr>|t| 1.0000 Sign M 3.5 Pr>=|M| 0.2649 SignedRank S 15 Pr>=|S| 0.7519 Variable: resid2 Tests for Normality Test —Statistic— -p Value-Shapiro-Wilk W 0.9513 Kolmogorov-Smirnov D 0.1250 Cramer-von Mises W-Sq 0.0751 Anderson-Darling A-Sq 0.5166 P r < W 0.1981 P r > D >0.1500 Pr>W-Sq 0.2362 Pr>A-Sq 0.1826 Quantiles (Definition 5) Quantile Estimate 100% Max 99% 95% 90% 75% Q3 0.1658 0.1658 0.1193 0.1011 0.0497 Quantiles (Definition 5) Quantile Estimate 50% Median 0.0122 25% Q l -0.0526 10% -0.0820 5% -0.1924 1% -0.1941 0% Min -0.1941 Extreme Observations •Lowest Highest-Value Obs -0.1941 1 -0.1924 20 -0.0820 2 -0.0761 3 -0.0685 28 Value Obs 0.0561 17 0.0685 29 0.1011 18 0.1193 19 0.1658 27 Stem Leaf 1 7 1 02 0 55667 0 0001122444 -0 421 -0 887655 -1 -1 99 ....+.—+-—+.—+ Multiply Stem.Leaf by 10**-1 # Boxplot 1 I 2 I 5 +-—+ 10 *„+--* 3 I I 6 +-—+ 2 I Variable: resid2 Normal Probability Plot 0.175+ +*+++++ -0.175H •++ +-*++*++ _j_ * * * _j—1_ ++++++ l-+*+ * .+.—+.—+-—+-—+-—+-—+-—+-—+-—• -2 -1 0 +1 +2 Appendix F Equations to obtain results Tire life vs. rotation frequency (Table 5-1, Table 5-26 and Table 5-27) Nf Lf = J]Lfi / N f , f=0 to 5 (Kemess), f=0 to 3 (4) (Gibraltar). i=l Average rotation hours (Table 5-2, Table 5-28 and Table 5-29) Nb Rb = ^ R b i / N b , b=l to 5 (Kemess), f=l to 3 (4) (Gibraltar). i=l Tire wear rate vs. tire rotation frequency (Table 5-3, Table 5-30 and Table 5-31) Nf W f = £ Wf, / N f , f=0 to 3 (Kemess), f=0 to 3 (4) (Gibraltar). i=l Age-specific tire wear rate for each rotation period (Table 5-4, Table 5-32 and Table 5-33) Nr Wr = £ Wn / N r , r=l to 4 (Kemess), r=l to 3 (4) (Gibraltar). i=l Age-specific tire wear rate according to wheel positions (Table 5-5, Table 5-34 and Table 5-35) Nw Ww = J] Wwi / Nw , w: L F , R F , L R O , L R I , R R I , R R O . i=l Tire rotation sequences (Table 5-6, Table 5-7, Table 5-8, Table 5-36, Table 5-37, Table 5-38 and Table 5-39) Na Na Na La = ^ L a i / N a , Ta = ^ T a i / N a , Wa = £ Wai / Na , Pu = N u / N a , a: SI, S O , OI , 0 0 . i=l i=l i=l Tu = ( T 0 + T i ) / 2 . Matching percentage Pm = N m / N r r . Tire life or wear rate vs. installation air pressure (Table 5-16, Table 5-54 and Table 5-55) Np Np Lp = ^ L p i / N p , W p = ^ W P i / N P , p : 95psi, lOOpsi, 105psi, HOpsi (Kemess), p: lOOpsi, i=l i=l 102psi, 104 (105)psi (Gibraltar). Summary of tire air pressure measurement (Table 5-21, Table 5-64 and Table 5-65) Pi = N i / N t P , i : different pressure ratings. Summary of tire air pressure measurement with over 130%, 140%, 120% and 125% rating (Table 5-22, Table 5-23, Table 5-66, Table 5-67 and Table 5-68) Pwo = Nmw / Nmo, w: L F , R F , L R O , L R I , R R I , R R O , o: different over percentage ratings. The percentage of spike pressures with the lowest installation air pressure Pi = Nml / Nma . Worn out tire percentage which had experienced spikes (over certain percentage) Pw = Nwo / Nto . Tire life comparison according to various positions (Table 5-24, Table 5-69 and Table 5-70) Nc Lc = ^ L d / N c , c: rear inside, rear outside, rear left side, rear right side, front, rear, left, i=l right. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0081066/manifest

Comment

Related Items