Prediction of Rolling Contact Fatigue Using Instrumented Wheelsets

The measured wheel/rail forces from four wheels in the leading truck of a coal hopper car during one revenue service roundtrip were used to by the Wheel Defect Prevention Research Consortium (WDPRC) to predict rolling contact fatigue (RCF) damage. The data was recorded in March 2005 by TTCI for an unrelated Strategic Research Initiatives project funded by the Association of American Railroads (AAR). RCF damage was predicted in only a small portion of the approximately 4,000 km (2,500 miles) for which data was analyzed. The locations where RCF damage was predicted to occur were examined carefully by matching recorded GPS and train speed/distance data with track charts. RCF is one way in which wheels can develop tread defects. Thermal mechanical shelling (TMS) is a subset of wheel shelling in which the heat from tread braking reduces a wheel’s fatigue resistance. RCF and TMS together are estimated to account for approximately half of the total wheel tread damage problem [1]. Other types of tread damage can result from wheel slides. The work described in this paper is concerning pure RCF, without regard to temperature effects or wheel slide events. It is important that the limitations of the analysis in this paper are recognized. The use of pre-existing data that was recorded two years prior to the analysis ruled out the possibility of determining the conditions of the track when the data was recorded (rail profile, friction, precise track geometry). Accordingly, the wheel/rail contact stress was calculated with an assumed rail crown profile radius of 356-mm (14 inches). RCF was predicted using shakedown theory, which does not account for wear and is the subject of some continuing debate regarding the exact conditions required for fatigue damage. The data set analyzed represents the wheel/rail forces from two wheelsets in a single, reasonably well maintained car. Wheelsets in other cars may produce different results. With this understanding, the following conclusions are made. - RCF damage is predicted to accumulate only at a small percentage of the total distance traveled. - RCF damage is predicted to accumulate on almost every curve 4 degrees or greater. - RCF damage is primarily predicted to accumulate while the car is loaded. - RCF damage is predicted to accumulate more heavily on the wheelset in the leading position of the truck than the trailing wheelset. - No RCF damage was predicted while the test car was on mine property. - Four unique curves (8 degrees, 7 degrees, 6 degrees, and 4 degrees) accounted for nearly half of the predicted RCF damage of the loaded trip. In each case, the RCF damage was predicted to accumulate on the low-rail wheel of the leading wheelset. - Wayside flange lubricators are located near many of the locations where RCF damage was predicted to accumulate, indicating that simply adding wayside lubricators will not solve the RCF problem. - The train was typically being operated below the balance speed of the curve when RCF damage was predicted to occur. - The worst track locations for wheel RCF tend to be on curves of 4 degrees or higher. For the route analyzed in this work, the worst locations for wheel RCF tended to be bunched in urban areas, where tight curvature generally prevails.