Noise can have various negative effects on health, which are important to consider when one engages in dangerous activities, such as driving. Cabin noise can distract the driver, cause stress and lower the cognitive abilities and reaction time. Excessive noise can inhibit the general comfort of passengers. These factors explain why car manufacturers invest financially in NVH measurements and lowering the cabin noise.
A continuous development of the acoustic behaviour of vehicles resulted in the discovery of new noise sources. Modern steering systems utilise an electric servo motor which acts as the origin of the emitted noise. By analysing the spectral structure of the motor NVH measurements we can find the sources of the noise. This information can later be used when designing steering systems to further optimise the vibration and sound emission.
In this paper I examine the most common signal processing methods, such as the spectrogram, Campbell diagram, and order spectrum, and show my implementation in the form of a MATLAB program. This application is capable of reading multiple files and exporting the generated figures, making it possible to process large number of measurements automatically.
Through correlation analysis between motors with different parameters, I will prove that there is no connection between the motor torque ripple and the house vibrations. In order to enhance the robustness and accuracy of the implemented signal processing methods, I propose use of non-linear rpm curves for motor measurements, which can counter the effect of the spectrum smearing and the bias caused by the resonance frequencies.