Wheel rail contact forces prediction and validation with field tests

Field testing of the wagon health card was completed using the RailCorp AK Track recording car. The wagon health card device was installed on the AK Car to allow the device to be tested and compared with the measured track data and AK car vehicle responses. As wheel rail contact forces cannot be measured directly these were calculated using VAMPIRE simulation. An inverse model with parameters matching the AK Car was embedded in the wagon health card device. The inverse model had been developed in earlier research to predict wheel rail contact forces using only measurements of car body responses; lateral and vertical accelerations, roll, pitch and yaw angular velocities. The device tested was also a proof of concept test for an integrated device with all transducers mounted within the one small 'black box'. Testing was completed in January 2007 on the Parks-Cobar section in NSW. The following paper presents the results and discusses the difficulties in validation without direct wheel-rail contact forces measurements. The results for estimating vertical contact forces were quite encouraging and acceptable for a proof of concept device. There was significant difference in lateral predictions and hence L/V ratios probably due to differences between assumed and actual wheel profiles