چكيده به لاتين
One of the underwater noise sources is hydrodynamic noises, which are caused by various
factors such as the movement of an object in the fluid and the formation of a boundary layer
around the body (flow noise), the interaction of the vortices of the turbulence flow with the
elastic shell of the body (flow- induced noise), the separation of the fluid by body movement,
etc. This noise can be considered as both self-noise and propagation noise in the environment.
Hydrodynamic noise reduction is crucial due to the development of sonar equipment for
engineering, military, and navigational purposes. The identifying features of this noise's source
should be understood to decrease it. Today, most of the self-noises of the moving underwater
vehicle are well known. For example, the machinery noise of a vehicle such as the combustion
chamber, generators, pumps, valves, etc., has been identified and the pattern of each of these is
known for a sonar system. A sonar system can identify the sources of these noises by using its
library and by comparison. However, the detection of arbitrary underwater body movement by
a sonar system is difficult. In other words, the movement of the body leads to the formation of
hydrodynamic noise, and it is difficult to identify this noise due to its complexity. Despite many
efforts to reduce the noises caused by the presence of marine structures and underwater vehicle
hulls (hydrodynamic noises), these noises have received little attention, due to the complex
nature of their generation and propagation, the time-varying environment of the sea and the
complex mathematical equations. Therefore, the purpose of this research is to explore and
identify the characteristics of flow noise sources in passive sonar systems by analyzing the
mechanism of acoustic noise generation and its propagation by feature extraction from acoustic
pressures. To extract the useful feature, first, this noise must be generated around the source at
different locations. The simulation method was used for this purpose. The numerical simulation
method occupies an important part of this research because feature extraction in the next step
is completely dependent on the accuracy of the numerical results in this step. Considering the
computational cost and simulation problems for long submersible bodies, this research has tried
to identify the location of the source first by using the minimum information recorded in the
receivers around the noise source. Then, using the same information, the flow noise
characteristics around the source were calculated. In the next step, using the characteristics of
the flow around the source, the features of this noise were investigated and extracted. In the
following, experimental data has also been used. At this step, hydrodynamic noise must be
separated from background noises. Therefore, the de-noising algorithms were explored and the
flow noise signals are de-noised with two new approaches. Finally, the feature of this noise was
extracted in the time domain, frequency domain, and a new domain (graph theory).
