Measurements of masked thresholds for warning signals during work on the permanent way

Status:

completed 09/2010

Aims:

During work on the permanent way, acoustic warning signals are used to warn workers to clear the tracks in the event of danger, i.e. when a train is approaching. Automatic warning systems are increasingly used for this purpose. They are activated by rail contacts fitted to the track on which trains are running, and generally take the form of a warning signal chain adjacent to the track. Remote-controlled automatic warning systems can also be fitted to tracklaying machinery itselves.
Systems from two manufacturers are in use in Germany. Their signals differ in their frequency characteristics and temproral structure. Since the workers must adjust to the relevant signal, the German Social Accident Insurance Institutions for the building trade (BG BAU) and for the railway services (EUK) and Deutsche Bahn AG are in favour of only one signal being used in the future. The study is therefore to ascertain which of the two signals currently in use can be heard more easily under the given working conditions.

Activities/Methods:

The masked threshold is the level at which a signal is detected with a probability of 50% (i.e. in half of all cases) over a given background noise. Standardized methods exist for calculation of this threshold. However, they consider only the spectral properties of the signal and the background noise. Application of these methods therefore yields meaningful results only for constant sounds. Other influencing variables such as harshness, tonality and sharpness are not considered.
For this reason, tests were conducted in the laboratory on test subjects in order to determine the masked thresholds for the two signals by experimentation. For the purpose of the test, the background noise (operating noise of a tracklayin machine) was emitted at a fixed level, and the warning signal issued over it at a range of levels. The test subjects indicated in each case when they had heard the signal. The discrete values enabled the level to be calculated at which 50% of the signals presented to the test subjects were detected. These series of tests were conducted for both warning signals, and for a number of background noises and directions of incidence of the warning signal.
In a further part of the test, the pair comparison test, the two warning signals were played in turn to the test subjects over the background noise, and the subjects required to state which of the two had been easier to detect. This test was performed for a number of signal levels, all of which however lay above the masked threshold for the signals (above threshold hearing tests).
Supplementary experiments examined whether the results could be confirmed when hearing protection suitable for the hearing of signals during track work was worn. The suitability of hearing protection for track work was determined at the IFA ? Institute for Occupational Safety and Health of the German Social Accident Insurance by means of an calculation based upon the sound-attenuation values. A hearing test on the individual worker at the workplace is also required before the system is used in practice.

Results:

When allowance was made for the measurement uncertainty, no significant difference was measured between the masked thresholds of the two signals in all situations studied (with variation of the background noise and the direction of incidence).
By contrast, the above threshold measurements (pair comparison) delivered a clear result. This test, which resembles conditions in the field, showed that the Minimel signal used by Schweizer was more suitable as a warning signal on track work than the Autoprowa signal used by Zöllner. This also holds true when hearing protection suitable for track work is used.
With the agreement of the project initiators, the results were made available to Deutsche Bahn AG.

Last Update: