LSTM模型在海绵城市建设中的应用

针对海绵城市给排水规划设计中排水性路面建设难、流体管网充满度预测不准确等问题,提出了一种渗透性路面来解决城市内涝问题,并对其建设方式和使用材料进行了说明,同时为了对规划流体管网的排水能力进行预测,将LSTM神经网络和卷积神经网络结合,提出一种改进的LSTM模型对流体管网充满度进行预测。最后,对文中的透水性路面性能和流体管网充满度预测算法进行试验。试验结果表明,渗透性路面的排水能力和降噪能力效果显著,改进后的LSTM模型预测结果准确度提高了2%。预测时间缩短了3倍,大大提高了预测效率。

Application of LSTM Model in Sponge City Construction

Aiming at the problems of difficulty in constructing drainage pavement and inaccurate prediction of fluid filling degree of fluid pipe network in the planning and design of sponge city water supply and drainage,a permeable pavement was proposed to solve the problem of urban waterlogging,and explained its construction methods and materials used.In addition,in order to predict the drainage capacity of planned fluid pipe network,an improved LSTM model was proposed by combining LSTM neural network and the convolutional neural network to predict fluid filling degree of fluid pipe network.Finally,the prediction algorithm of permeable pavement performance and the fluid fullness of fluid pipelines were tested in this paper.Test results showed that the drainage capacity and noise reduction ability of permeable pavement were significant,and the accuracy of the improved LSTM model was improved 2%.The forecast time has been shortened by 3 times.The prediction efficiency is greatly improved.