Study On Modeling Optimal Location Of Water Quality Monitoring Sites In Dynamic Water Distribution Network With Exogenous Contamination

To ensure the safety of water quality while the network is invading by the exogenous contamination is a multi-objective complex problem. The water distribution network with complex structure of topology has the characteristic of hydraulic transition and fluid inertia. However, most of the optimal location of water quality monitoring are based on static or quasi-stable state hydraulic conditions at present, which are large different from the operation conditions in real-life network. Based on static state hydraulic condition, a study on the element of flow hydraulic information with time steps changing was done, which was defined as dynamic hydraulic condition. An optimal location model of water quality monitoring sites in dynamic water distribution network, based on the real-life water supply pipe network in the town of TX, GuangDong province, was built to solve the problem of locating the monitoring sites in dynamic hydraulic condition. Also the solving mode studied. The hydraulic model was built by collecting the attribute data, including the network topology structure and operating data. The optimal location model, based on output results of the hydraulic model, was built after analyzing the hydraulic flow in the dynamic hydraulic transition condition and solved by adaptive particle swarm optimization combined k-medoids cluster algorithm. The monitoring point location scheme, calculated by the optimal location model, will be very helpful for the decision-makers to make a decision, which is significant for the management of water distribution network.Firstly, the calculation of hydraulic flow was improved to make out the monitoring information in dynamic hydraulic condition, including the hydraulic flow path and travel time, the quantity of polluted water while the contaminant was monitored, the amount-of-substance and the concentration of contaminant at the monitoring point. According to the comparison with unimproved calculation, the hydraulic flow in dynamic hydraulic condition is more complex, shown as the follows:a) The possibility of hydraulic flow between every two points changes as the time schedule changing. It is influent by the network topology and the flow direction in the pipe during the different time steps.b) The five targets also change between the every two points, which has the possibility of hydraulic flow.c) The total number of the feasible hydraulic flow changes, which percent of change is large.Thus, the static state hydraulic condition can not take the place of the dynamic hydraulic condition in modeling the optimal location of water quality monitoring sites. The improved calculation is adapt for calculating the data in dynamic hydraulic condition and more consistent with real-life network.Secondly, in view of the characteristic of the dynamic hydraulic condition, a general mathematical model was built by summarizing the model with different optimal goals. A standard three-dimensional matrix was built to store the monitoring data of every two network points. The problem of the dynamic hydraulic monitoring data saving was solved by establishing the standard 3-D matrix. It has strong universality and practicability because the monitoring data of different optimal goals can be saved by the matrix, which the general mathematical model can be applicated.Thirdly, the solving mode based on the adaptive particle swarm optimization combined k-medoids cluster algorithm was put forward. It was specific adjusted to the points and monitoring sites on the water distribution network, also the number of the center of cluster and the control value of the small world field were confirmed. The optimal location mathematical model of water quality monitoring sites was solved by the adjusted algorithm and the solving mode was formed. The calculation function of the solving mode was compiled based on the Matlab. Combining the optimal location model and the calculation function, the others network in city can be solved. It can be helpful for the management to carry out the scheme of the monitoring point with different optimal goals.Lastly, five optimal goals were used to build the optimal location model and carry out the schemes for the water distribution network in the city, which are the average monitoring time, the average quantity of polluted water, the average polluted area, the average injected amount-of-substance and concentration of contaminant. The relationship between the five goals and the number of the monitoring points was studied, also the effectivity of the monitoring rate. Special schemes were carried out for each of the five goals. A multiple optimal target scheme of the monitoring sites location, considering the first three goals, was confirmed. The network points and the monitoring effect of the monitoring sites were ensured in the schemes.