Data-Based Probabilistic Damage Estimation for Asphalt Shingle Roofing

Guoqing Huang, Hua He, Kishor C. Mehta, Xiaobo Liu

Research output: Contribution to journalArticlepeer-review

28 Scopus citations


Asphalt shingles on residential building roofs are susceptible to damage, and often blow off, during windstorms. The loss of shingles can also result in damage to the content in the interior of a residence by allowing the penetration of rain. This paper presents the data-based probabilistic damage estimation procedure to predict wind-induced damage on asphalt shingle roofing, using wind pressure data from wind tunnel testing. First, the probability distribution of peak wind pressure over a certain period for pressure data associated with each measurement tap is estimated. Then, the failure probability of the shingle associated with each tap and the damage ratio for the entire roofing shall be determined. Finally, a neural network is adopted to predict the wind-induced damage ratio for asphalt roof shingles considering multiple contributing factors such as wind speed, wind angle of attack, building sizes, roof slope, and terrain roughness.

Original languageEnglish
Article number04015065
JournalJournal of Structural Engineering (United States)
Issue number12
StatePublished - Dec 1 2015


  • Asphalt shingle
  • Damage ratio
  • Hermite polynomial model
  • Neural network
  • Translation process model
  • Wind effects
  • Wind-induced damage


Dive into the research topics of 'Data-Based Probabilistic Damage Estimation for Asphalt Shingle Roofing'. Together they form a unique fingerprint.

Cite this