Research areas
Psychoacoustics
Introduction
During the starting procedure of electric railbound vehicles (e.g. light rails
and subways), the motor in connection with the traction converter is a dominant
noise source apart from the gear. Due to an increasing sensibility of operators
and passengers, a more detailed knowledge about the optimisation potential of these sound
sources is wanted. In the development phase, however, acoustic criteria often play a minor
role compared with economic requirements, since it is not sufficiently known how
railbound-vehicle sounds are perceived by man in detail. The project presented
here should contribute to an advanced knowledge about technical sources
and psychoacoustic effects, so that finally an acoustic guideline about
all relevant parameters will be developed in order to optimise sound quality
of starting railway vehicles.
The project comprises three essential steps of work:
1) Where do the sounds come from?
On the basis of a detailed literature survey, an overview is elaborated of all technically
relevant parameters that influence traction noise. Among them are, e.g., the type of motor,
type of power-electronic devices, type of circuit, choice of current signal, choice of gear teeth numbers.
2) How are these sounds perceived?
In order to investigate psychoacoustic aspects, a multitude of vehicle-interior sounds are
recorded and used as stimuli for listening tests, so that relevant parameters for sound
quality (e.g., tonality) can be identified. Subsequently, instrumental quantities are
determined which should correlate as well as possible with perceived effects.
3) How can the sounds be optimised?
As synthesis of the first two steps, it is investigated which technical effects prove to be
critical from a psychoacoustic point of view. Finally, one wants to know how to construct a
railway vehicle with a traction noise as pleasant as possible. In order to avoid extensive
parameter studies on real vehicles, a simplified acoustic model is developed which comprises
synthesised traction-noise components and added real rolling-noise components. The model should
not only yield equal psychoacoustic quantities compared with reality, but it also should be
indistinguishable from real recordings. After successful validation, optimisation studies
are carried out with this model, e.g. with different temporal characteristics of PWM switching frequency.
It is emphasised that traction noise normally cannot be reduced significantly, since it is
unavoidable with respect to modern traction equipment. However, it will be shown that the
sound quality of many vehicles can in fact be slightly optimised by use of several
"adjusting screws". Even if they might not influence specificied quantities, they can
contribute to an increased product quality or competiveness, respectively.
Progress
Future work
- Completion of dissertation, publication presumably in 2005
