It has been observed that the co-efficient of friction increased and reached a higher point and then gradually decreased. How can I explain the reason for this trend? The experiment was carried out in spare lubrication condition.
I believe Mamoun is right. It is very hard to speculate about what is causing this behaviour without knowing more about the test conditions. Obviously, the x-axis is the rotational speed.
I would be interested in knowing the geometry of the contact. Ball-on-disc, pin on disc etc. Also, is the contact pure sliding, or rolling and sliding? Load and corresponding contact pressure would also be interesting. Finally, what kind of lubricant is used, and what do you mean with "spare lubrication condition"?
The relatively low frition coefficients suggests to me that we are talking about a more than marginallt lubricated contact as long as no exotic materials are involved.
This is a counterintuitive trend. If you look at a typical Stribeck Curve, which describes the relationship between COF and sliding speed of two lubeicated surfaces, you would find that the trend is opposite to the shown one. Is this result repeatable?
Following your results, as you increased the speed you have got time variable temperature over the contact, and also a continuously running-in.
I obtained the same results on ball bearings lubricated by oil mist. In my opinion, you have got a running-in procedure. To avoid this, I increased gradually the speed, and I kept it constant for few minutes at the desired speed. So, I obtained a point for 500 rpm in the first experiment, the second point for 800 rpm in the second experiment, and so on. I considered the friction coefficient the mean for a steady value of few minutes. Also, it is very important to keep the temperature constant. My advice is to heat-up the testing device at a higher temperature than desired, and let the device to cool free up to the desired temperature. Choosing a higher start temperature than expected during the tests, you could eliminate the influence of the temperature.
Another solution is to develop this tests very quickly at the room temperature, and to measure the temperature during the tests (with an X ray apparatus). Try after a theoretical modeling, taking into account the temperature during the test and the instant friction coefficient.
You can carry-out a running-in of the tribological couples of the tested materials, and repeat your tests as indicated above. Good luck!