Residential Live Load Modeling Based on Panoramic VR Viewing and Network Big Data

Abstract:

Reliable live load value is the prerequisite for reliability design of civil engineering buildings， and the collection of live load samples is the key data foundation for load modeling. Aiming at the problems of difficulty in entering households， weighing， low efficiency， and small samples that exist in the traditional research methods of sample collection and weighing in households， and based on the research thinking of big data， a new research method of residential persistent live load based on panoramic VR（virtual reality） house viewing system is proposed. Through the VR system of open network platform， multi-source heterogeneous data such as pictures and videos， combined with Internet resources， comprehensive network crawler， and other means， live load investigation is realized by non-weighing method， and practical research is conducted by taking residential buildings as examples. The characteristic value of residential live load was obtained by statistical analysis and load combination of 4 676 rooms in 9 cities （with a total area of nearly 75 000 square meters）. The results show that the standard deviation of sustained live loads significantly increases at any given time， the time of live load variations is different among different cities， which， overall， presents a decreasing trend. The combination of VR system and internet big data can achieve efficient and convenient residential live load surveys， forming a new research method for live loads.

Key words:live load;VR(virtual reality) system;big data;load code;structural engineering