Brake Tests

154 Coefficient of Rail Friction. 344. In using the machine, described in Par. 65, to determine the relative condition of the rail surface during different tests, the same section of rail was used at all times. The coefficient of friction, or the ratio of the force required to keep the weights moving and the pressure of these weights upon the rail, is recorded on tlie log as the .kinetic coefficient of rail friction. 345. The values determined range between 12 per cent, and 35 per cent., with the great majority of readings ranging between 22 per cent, and 30 per cent. The records, when taken in connection with simultaneous readings of air temperature and relative humidity, show that the coefficient of rail friction decreases with an increase in the relative humidity for temperatures below the freezing point, whereas for high humidities the coefficient of rail friction is not greatly affected, as long as the temperature is high, but begins to fall as the temperature approaches the freezing point. In other words, the tests confirm common experience, which leads us to expect a bad rail condition with a combination of low temperature and high humidity. 346. As an interesting study of the effects of these various factors, consecutive observations were taken of the coefficient of rail friction, air temperature, relative humidity and barometric pressure for a period of twenty-four hours. These records are shown in Fig. 85 and indicate clearly the tendency of the coefficient of rail friction to increase with increasing temperature and decreasing relative humidity, and to de- • ocino relative humidity and decreasing temperature, crease with increasing relative 1 . © 1. It is significant to note that the coefficient of rail friction obtained on the first reading in the early morning was practically the same as tliat found for a well greased rail later in the day. 347 While the absolute value of the data obtained may be open to question on account of the difference between the condition of the c - 4: thecase of the rolling wheel and the rail, and those surfaces in contact in the case oi LC 105 which existed between the sliding weight and rail when taking the 1 . it is safe to assume that the values obtained observations referred to, it is sale to assit 1 at different times are relative. It is apparent that the average rail condition changes with the seasons of the year (and, to a less degree, with the time of the day), and advantage can be taken of this fact by using a higher braking power in the summer than could be used in the •.9 ,1:1-11ood of a material, if any, increase in wheel winter without the likelihood a n sliding. The above conclusion logically follows from a knowledge of rail conditions in general and their effect irrespective of the readings of the apparatus used during the tests. As a matter of fact there was no -real consistency between the readings obtained for coefficient of rail friction and the amount of wheel sliding experienced. While as