A novel fusion modeling method for the acoustic fields of diesel engine laboratories

The complex geometric parameters of the diesel engine, the ancillary equipment, and the pipe system make it difficult to establish an accurate acoustic field model of a reference diesel engine laboratory using 3D geometric modelling. A novel fusion modeling method based on machine vision and 3D geometric modeling is proposed to achieve efficient and accurate modeling for the acoustic field of the diesel engine laboratory. Firstly, an enhanced scale-invariant feature transform algorithm and an enhanced random sample consensus algorithm are proposed for the machine vision-based modeling of the diesel engine, the ancillary equipment, and the pipe system. Then, the laboratory walls and the engine bench with simple geometric parameters are modeled by CATIA software. Finally, the two types of models are fused according to the location relationships and are converted into an acoustic field model. Experimental results show the geometric dimensions and sound pressure levels errors of the reconstructed model are less than 0.5% and 2.5 dB, respectively, this indicates that the proposed method can accurately model the acoustic field of the reference diesel engine laboratory without complex geometric parameters.