基于遗传算法的供水管网需求乘数因子校核

Calibration of Water Demand Multipliers in Water Distribution Systems Using Genetic Algorithms

水力模型已广泛应用于给水管网设计、分析与运行中。在所有水力模型中,需水量是导致模型输出最不确定的参数之一。因用水情况不确定,使得管网中的节点需水量变得异常复杂。在大多数实际管网中,用于校核节点需水量的监测设备数量有限,且小于未知量个数,使得节点需水量校核作为欠定问题,令节点需水量校准产生较大误差,并且传统遗传算法校核节点需水量的方法是假定所有节点的需求乘数因子一致,这也导致校核后的模型无法接近真实运行情况,因此提出在欠定条件下用遗传算法解决需求乘数因子的校核问题。通过对一个实际案例多次运行并取平均值作为结果进行验证,结果表明,遗传算法的校核结果不仅能够与被测位置的实际值相拟合,而且可以得到非测量位置的管道流量和节点水头,其中校核后的节点水头和管道流量误差较小,平均误差分别为1.78%、4.05%。该方法相比于传统校核方法具有更高精度,且更能反映出管网真实运行情况,同时还避免了传统校核方法中因遗传算法产生局部最优解而导致误差偏大的问题,对于大型管网模型校核也具有一定参考价值。

Hydraulic models have been widely used for design,analysis,and operation of water distribution systems.As with all hydraulic models,water demands are one of the main parameters that cause the most uncertainty to the model outputs.However,estimation of the demand parameters is usually complicated due to the stochastic behavior of the water consumptions.This is an underdetermined system where the number of measurements is limited and less than the number of unknows,so the calibration of water demand will produce a large error,and the traditional genetic algorithm to calibrate the water demand of nodes is to assume that the demand multiplier factors of all nodes are the same,which leads to the problem that the calibration of model can not be close to the real operation.This paper presents an approach to calibration of the demand multiplier factors under an ill-posed condition where the number of measurements is less than the number of parameter variables.The problem is solved by using a genetic algorithm (GA).A practical case was run multiple times was taken and the average to validate that not only is the GA able to match the calibrated values at measured locations,but by using multiple runs of the GA model,the flow rates and nodal heads at nonmeasured locations can be estimated.The error of node head and pipeline flow is small after check,and their average error is 1.78% and 4.05%,respectively.Compared with the traditional optimized method,this method has higher accuracy and can reflect the real operation of the pipe network.At the same time,it also avoids the problem of large error caused by the local optimal solution produced by genetic algorithm in the traditional optimized method.It also has a certain reference for the calibration of large pipe network model.