由于运行中管道实际摩阻因数与设计采用值的不一致会影响供水系统的运行质量和安全，为得到准确的摩阻因数以反馈设计，以及对运行中的供水系统进行系统诊断和优化调度，提出一种基于混合粒子群优化算法（hybrid particle swarm optimization,HPSO）和管网水力计算模型的管段摩阻因数智能反分析方法，该方法以监测点处的水压监测值与水力计算模拟水压值的二乘误差最小为目标，通过HPSO的强大全局寻优能力计算不同管段的摩阻因数。对室内试验管网模型的反演结果表明：采用正常工况下反演得到的摩阻因数模拟水压监测点处水压与实际测量值之间误差最大为2.87%；在爆管工况下，水压模拟值与实际测量值最大相对误差为2.63%，说明通过该方法进行摩阻因数的反演具有较强的稳定性。

The water supply network is the lifeline project of the city.Under the background of the implementation of the rural revitalization strategy,the security of rural drinking water and the accelerated construction of the urban-rural water supply integration project,the water supply network has also become the lifeline project of the vast rural areas.In order to ensure the quality and quantity of water supply and complete the water supply task economically and reasonably,it is necessary to carry out scientific planning and design and precise optimization and scheduling of the water supply network.The hydraulic simulation calculation of the water supply network is the basis for the planning,design,operation scheduling and fault diagnosis of the pipeline network.Among the factors affecting the accuracy of the hydraulic calculation model of the pipeline network,the influence of the friction coefficient is particularly prominent.In order to narrow the gap between friction factor theory and practice,a calculation method based on hybrid particle swarm optimization algorithm and nodal water pressure method is proposed. Hybrid particle swarm optimization model is obtained by combining the selection mechanism of basic particle swarm optimization and genetic algorithm.The difference between hybrid particle swarm optimization and particle swarm optimization lies in that the particle swarm has to cross operate after updating the velocity and position,and replace the parent particle with the offspring particle.The crossover operation makes the offspring inherit the advantages of the parent particles and theoretically strengthens the ability to search the region between the particles.In the algorithm,the initial friction resistance factor is randomly selected and put into the Hazen-William formula.The square root of the square error of the square error between the water pressures is calculated based on the actual water pressure value of the nodes monitored by the pipeline network model as the fitness value,and the fitness is evaluated.When the accuracy is met or the maximum number of iterations is reached,the frictional resistance factor is output. In order to verify the feasibility of the method,a 2.5 m×2.5 m square flat pipeline network model was established in the laboratory experiment,which contained 9 basic rings.Four water pressure monitoring points are set in the model pipeline network node.Based on the friction resistance factor inversion under normal working conditions,the hydraulic simulation model of the pipeline network is established on this basis.In the case of pipe bursting,virtual nodes are introduced into the pipeline network model to simulate the pipe bursting point,and the simulation results are compared with the measured values to verify the accuracy of the model. The results show that:(1) HPSO has a strong global optimization ability.The maximum relative error of the Hazen-Williams factor of each pipe section obtained by inversion under normal and special conditions (pipe burst) is 7%,with high accuracy and better adaptability.(2) Calculate the outlet water pressure of the monitoring point under the normal condition and the pipe burst condition with the inversion value of the Hazen-Williams factor under the normal condition.The maximum relative error between the result and the actual water pressure monitoring value is only 2.87%.The inverse problem solved by the method in this paper is well-posed.