We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
基于支持向量机的综合管廊火灾纵向温度实时预测.
- Authors
陆梓萍; 王 克; 周晓冬; 王 董; 姜 楠; 贾欣苗; 杨立中
- Abstract
Underground utility tunnel fires cause huge economic losses and damage to the city. Given the rapid fire development and unique confined spatial structure of utility tunnels, an accurate and real-time fire temperature prediction system is needed for firefighting making decisions and guiding fire operations. In this study, numerical simulation models of utility tunnel cable fires with five fire locations were established, and a database containing the location of the fire source, heat release rate, fire duration, temperature, as well as the spatial relationship between thermocouples and fire sources was set up. Combined with support vector machine(SVM), a data-driven real-time temperature prediction model was proposed, which realized the temperature forecast in utility tunnel fire scenarios. Moreover, this paper proposed a method to improve the dataset structure around fire sources of artificial intelligence. The method has excellent performance in prediction accuracy and time cost, showing great potential in smart firefighting application of underground utility tunnel.
- Publication
Fire Safety Science, 2023, Vol 32, Issue 2, p94
- ISSN
1004-5309
- Publication type
Article
- DOI
10.3969/j.issn.1004-5309.2023.02.04