基于支持向量机的综合管廊火灾纵向温度实时预测

Real-time longitudinal temperature prediction of utility tunnelfires based on support vector machine

城市地下综合管廊一旦发生火灾,会对城市造成很大的经济损失和社会影响。考虑到火灾的快速发展和综合管廊狭长受限的特殊空间结构,迫切需要一套准确、实时的火灾温度预测系统辅助消防救援人员制定决策和指导消防行动。建立了5种不同火源位置的地下综合管廊电缆火灾数值模型,结合支持向量机(SVM),根据火源位置、热释放速率、火灾发生时间以及待测点与火源之间的相对位置关系开发了一种数据驱动的温度实时预测模型,实现了地下管廊火灾场景内的纵向温度预测,提出了在火源附近数据结构的优化方案,提高了火源附近的预测准确度。该方法在预测性能和预测时间方面取得了优异的性能,展示了人工智能在火灾预测应用中的优越表现和发展前景。

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.