针对当前复杂供水管网泄水事故频发的现象,提出以管网发生故障时的瞬态水力数据信息为样本,采用Hilbert-Huang变换将故障样本信息从水压—时间的时域信号转换为压力幅值—频率的频域信号。分析选取频域信号中的主要特征值坐标作为支持向量机(SVM)的输入数据,并采用粒子群算法(PSO)对SVM参数进行优化,建立了基于HHT和SVM的供水管网故障诊断模型。实例应用表明,该模型在供水管网泄漏诊断中效果较好,从而验证了模型的可行性。 In view of the frequent discharge accidents of complex water supply pipelines, this paper proposes a transient Hilbert-Huang transform to transform fault information from time-domain hydrostatic pressure-time to Pressure amplitude - the frequency of the frequency domain signal. The main eigenvalue coordinates in the frequency domain signal are selected as the input data of Support Vector Machine (SVM), and Particle Swarm Optimization (PSO) is used to optimize SVM parameters. The fault diagnosis model of water supply network based on HHT and SVM is established. The example application shows that the model is effective in the diagnosis of water supply network leakage, which verifies the feasibility of the model.