供水管网爆管动态风险评估

摘要

This paper proposed a new dynamic pipe burst assessment model,which took dynamic data of pipes (e.g.velocity,pressure,etc.) into account and was able to evaluate burst risk of the pipes varying with time during a day.A back propagation (BP) neural network was built to reveal the relationship between risk level and the above time-varying factors.In the end,this new model was applied in a real water supply network,and its results showed that the model was able to expose the potential hazard of the pipes with seemingly low risk level,which had a significant advantage compared with the traditional pipe burst assessment model.%文中提出一种新的爆管动态风险评估模型,该模型考虑管段的流速、压力等动态数据,应用BP神经网络揭示管段的爆管风险等级与上述时变因素的关系,从而对管段不同时刻可能出现问题的安全性进行评估.最后,把爆管动态风险评估模型应用到某实际管网中,结果表明该模型能够暴露出一些看似安全管段的潜在爆管风险,相比于传统的爆管评估模型具有较大的优势.