A number of passengers were injured and five were killed at the train collision accident that took place in 2000 March due to climbing derailment on the transition curve at the exit from a sharp curve of a Tokyo subway line. According to the investigation by the accident cause investigation committee, it was again recognized that the friction coefficient between wheel and rail was substantially influential for the occurrence of climbing derailment. In running tests, it has been often the case that the friction coefficient reaches nearly twice the values (0.3 to 0.4) observed in the past. We are promoting researches on this theme through simulation, derailment tests and creep force experiments supported by the government. It is an important subject how to evaluate the friction coefficient between wheel and rail. In this contest, a number of researches have been performed on this issue from the viewpoint of adherence and creep force. However, these researches haven't sufficiently clarified or identified the distribution and changes in the value of friction coefficient on main lines for revenue service, parameters that affect the friction coefficient, or the relation between the measurable and actual values of dynamic friction coefficient between rail and wheel.
    Despite the improved reliability of rolling stock, damages on truck parts, mostly due to fatigue, haven't been eliminated yet. We cannot observe most of the damages on axles and wheels in detail, since rapture planes are rubbed. In the case of truck parts recovered with rupture planes under good conditions, we can collect data in quantities by observing rupture planes with stereo and electronic microscopes, examining the metallographic structure and analyzing impurities, inclusions and adhersions, unless there are no secondary damages, based on which we can identify the cause of damage. Even when the fatigue rupture plane remains intact, however, we normally cannot determine the cause of damage. This means that we cannot know the cause of damage by observing rupture planes, implementing laboratory tests, fatigue tests or collecting data in running tests. The metallic fatigue has long been known and investigated in wide ranges. Nevertheless, there are a number of phenomena that are not clarified yet.
    The existence of these phenomena has been known for 100 years to pose an old but new problem. Mankind now travels in the deep space and the IT technology has made rapid strides. Nevertheless, there are a number of facts we cannot clarify. They are not the natural phenomena out of control by mankind, but the kinetics and damages of man-made machines. More than 100 years have passed since the iron wheel-rail railway system was established. The reliability of railway system has progressed, in that signal and protection systems have developed to utilize electronic technologies and trains run at 300km/h or over. An accident will cause unhappiness or damage passengers. In Japan, there are phrases "to know what is unknown" and "to know what has been unknown." It is important to be conscious of the facts we don't know and recognize that nobody has explanations thereof. We, engineers, are required not to forget these phrases, and make ceaseless efforts to know even one of the facts that we haven't known, in order to improve the safety and reliability of railways.