基于贝叶斯理论和遗传算法的管网漏损探测定位

Leakage Detection and Location in Piping Network Based on Bayesian Theory and Genetic Algorithm

针对舰艇消防供水管网结构复杂、管网中用户的用水量与漏损损耗量难以区分等特点,通过将管网中的漏损点等效为具有用水量的虚拟节点,提出了一种基于贝叶斯理论与遗传算法相结合的管网漏损探测定位模型。通过对所有单个虚拟节点模拟管网漏损量,将其剩余压力平方和作为目标函数,通过遗传算法寻找能使目标函数取得最小值确定管网的漏损情况。利用遗传算法搜索在不同漏损管段个数下能使得贝叶斯概率取得最小值的最优管段组合,从而实现漏损定位。仿真结果表明,通过结合贝叶斯理论与遗传算法进行漏损探测与定位具有探测速度快、定位结果准确的优点,可用于解决多点漏损管网的漏损探测和定位问题。

Considering the complexity of the fire fighting system in warship and the difficulty of distinguishing the demand and the leakage at the nodes,by equivalence of the leakage point to virtual node with certain water demand,a leakage detection and localization model based on Bayesian theory and genetic algorithm（ GA） is proposed. The residual square pressure of the leakage point located at all the nodes is calculated as the objective function,total amount of the leakage can be determined by GA,which can make the objective function obtain a minimum value. On the basis above,the leakage pipe and its leakage amount are determined by searching the maximum Bayesian probability with different number of leakage nodes. Optimization programme is made, numerical simulation results demonstrate high-speed detection and high-accuracy localization can be achieved by combining the Bayesian theory and GA. The model can be used to solve the problem of leakage detection and localization with multiple leakage points.