where          r                    and           r             , while in the process varying the sliding velocity

linearly  from  a  user  defined  negative  limit  to  a  positive    one。 As  a  result,  the corresponding dimensionless slide-to-roll ratio, which is defined as

was varied within the range R C [—      ,      ]。  Within the limits of the test machine,  this

range was accomplished for three rolling speed values of            ,       and 15 m/s。

Prior to actual tests, several trial tests were performed to determine the optimal ramp-up and ramp-down rates for the traction tests。 Prior tests done on an earlier version of the same test machine [17] did not have the capability to implement programmed speed profiles, and hence were reported to be impacted by the changes taking places to the contact surface roughnesses throughout a traction test。 Therefore, optimizing (minimizing) the length of test with sufficient data points without any tangible changes to surface roughness profiles was critical。   A number of trial runs    as

specified in Table 2。1 were performed。   The corresponding Rq    values at the end of

each of these trial traction tests are compared in Figure 2。9。 A  5-minute  test was

adapted for the rest of the traction tests as a sufficiently long test with secondary impact on the surface roughnesses。

Each traction test was performed at a user-defined constant normal load ( ) value。  The torque provided to the disk by its drive, which is equal the total traction

torque       experienced by the disk-side driveline, was measured by the torque sensor

Table 2。1 Preliminary tests performed to define a traction test duration with minimum surface roughness changes。

Duration (min) Change in R

Run In 10 -1 (constant)

Test 1 2 -1 to +1

Test 2 2 +1 to -1

Test 3 5 -1 to 1

Test 4 5 +1 to -1

Figure  2。9 Measured  Rq     values  of  roller  and  disk  specimens  throughout       the

preliminary study according to Table 2。1。

shown in Figure 2。2。  Considering that    consists of the contact friction torque     and the disk-side drive shaft bearing friction torque     :

          +     , (2。4)

determining contact traction required determining and removing      from the  measured

    signal。  For this, first       was measured by simply running an unloaded traction   test

(       such that       to yield      =       )。  Next, the traction test was repeated at   the desired       value and               —        was determined as a result of these two tests。  The