Document Type : Original Article
Authors
Department of Water Engineering, College of Agriculture, Razi University, Kermanshah, Iran
Abstract
Introduction:
The concept of resilience is particularly important in water distribution networks, which are important urban infrastructures. Estimating and evaluating resilience in each network at the time of design will reduce damage to subscribers and the network. In this research, relationships and functions related to resilience (GRA) in water supply systems and solutions for how to increase resilience using two scenarios of pipe failure and additional needs for the Kangavar network were investigated and implemented.
Material and Method:
By modeling the initial failure modes by increasing the stress intensity and estimating the consequences that arise, the resilience of a system can be evaluated, which includes the following steps (Diao & et al. 2016):
Step 1. Identify the failure mode to evaluate (eg structural failure, excessive demand).
Step 2. Determining the system stress associated with the failure mode and its simulation method (for example, WDS simulation with an additional load on a node for a certain period).
Step 3. Identify the appropriate system and how to measure it (eg ratio of unsatisfied demand to total required demand during the failure period).
Step 4: Simulate the consequences of the failure mode at increasing stress intensity (0%-100% of maximum stress). While stress intensities up to 100% may be highly undesirable, they are theoretically possible and should be considered if a wide range of potential effects are identified. For each given stress value, the appropriate number of failure scenarios is determined.
Step 5. Create a stress-resilience curve that shows the average, maximum, and minimum results produced by the simulation for each given stress value.
Result:
The worst situation for Kangavar network starts at 89% failure and remains until 100% pipe failure. In large networks like the Kangavar network, the graph of the strain duration and the stress duration have a steep slope just like the supply shortage graph. For example, for the value of five percent pipe failure, all three values of minimum, average and maximum strain duration are equal to five. That is, when only five percent of Kangavar's pipes fail, the duration of the strain reaches its maximum.
Among the prevention ways to reduce the lack of supply in case of pipe failure, adding parallel pipes for pipes with a more important position compared to other pipes or looping the network in different areas of the network. For the Kangavar network, 38 pipes were added to the network. These pipes are mostly parallel to the pipes coming out of the tank. After adding only 38 pipes to the network, the amount of supply shortfall is greatly reduced.
The state of excess demand is actually a state in which a number of certain nodes have a need or demand more than their defined capacity in a certain period of time. This situation is actually very similar to when a fire occurs in the network, except that in the case of a fire, the normal and usual demand of the network may decrease a little. But in this simulation, in addition to the normal need and demand, additional needs are also considered.
Figure 12 shows the duration of the fire for hours 18 to 21. In this period of time, in the worst case, i.e. in 100% of the nodes with higher demand, the network faces only 10.82% supply shortage. This means that if all eleven selected points in the network have additional needs at the same time, the network will have an approximately 11% supply shortage. In this situation, it can be seen that the stress applied to the network occurs at the same moment and the duration of the strain is six hours.
Conclusions:
The results showed that the duration of the strain increased with the increase in the percentage of pipe failure. But the graph related to the start of strain had a downward trend and approached zero with the increase in the percentage of pipe failure. The level of resilience for a network is different in different scenarios, in fact it is possible in with equal failure percentage, and the network has more resilience in one scenario than in another scenario. With a slight increase in the percentage of pipe failure in some water supply networks, the amount of strain increases greatly. While in other networks, even with a large increase in failure percentage, the size of the slope strain has started to increase slightly. One of the most important reasons for this depends on the type of network design. For example, if the water supply network is defined as a loop, in case of failure of one of the pipes, part of the lack of demand will be supplied by other pipes. Adding new pipes to the network was one of the solutions considered to increase the resilience of the network in this research.
Keywords
Main Subjects
)