Disk model effect for road surface roughness using convolution method

Der-Shen Yang, Qianhui Zhang, Hao Xu, Kwesi Sagoe-Crentsil, Wenhui Duan

Research output: Contribution to journalArticleResearchpeer-review

1 Citation (Scopus)


In most related studies on road surface roughness, the vehicle's wheel is often using a contact point model rather than a disk model. This results in neglecting the wheel's size and interaction with the road. Consequently, the vehicle's response may not be genuinely reflected, especially for the massive topic of noise, vibration, and harshness (NVH). Unlike the existing approach targeting the power spectrum, this paper proposes a new convolution method to tackle the disk effect and operates directly on the spatial domain, i.e. road surface roughness. By using a designed periphery function, it can simulate the wheel geometry passing through road surface roughness. The periphery function acts as a filter to the road surface roughness that can filter out smaller oscillations. Some examples involving roughness from ISO 8608 standards were tested. It is shown herein that the proposed method can match the theoretical result (using the geometry method (GM)) not only in the spatial domain but also in the power spectral density (PSD). Since the convolution is performed under the spatial domain, the proposed method can directly apply the disk model to any existing road surface roughness with different spectral compositions in practice. Understanding the disk effect reduces the higher frequency of the vehicle's response depending on roughness severity, which may significantly impact the vehicle design for ride comfort, road surface roughness extraction, bridge health monitoring using the drive-by method, etc.

Original languageEnglish
Article number2340033
Number of pages20
JournalInternational Journal of Structural Stability and Dynamics
Issue number16
Publication statusPublished - 22 Sept 2023


  • Convolution
  • disk model
  • point model
  • power spectral density
  • road roughness

Cite this