KYPipe Analysis of Water Distribution Network

[meghna77]

Hello everyone,

I am carrying out an analysis on KYPipe 2000 of the existing water distribution network a town to determine its adequacy to meet the present demand.

The supply comes from a Clear Water Tank at the Water Treatment Plant and is gravitated into the distribution network, which comprises of 9 storage tanks of different volumes within the network.

Having drawn out a schematic layout on KYPipe 2000 based on actual lengths, elevations and diameters,demands have been allocated at junction nodes throughout the network and storage tanks have been represented as Tanks (without inflow; No Feedpipe) and the main Clear Water Tank has been represented as a Reservoir. (Total demand = 0.11m3/s)

However, after running the analysis, the following results are obtained:
1. Inflow = 0.16m3/s
2. Outflow = 0.05m3/s
3. Demand = 0.11m3/s

The outflow is representing water flow into the tanks, and the KYPipe Analysis determines the inflow by combining the demand and the flow into the tanks. However, in real sense, the treatment plant is only capable of producing approximately 9,000m3/day. To have a total inflow of 0.16m3/s (13,800m3/day) does not give me a correct picture.

It seems that the system, despite having a total demand of 0.11m3/s provides the tanks with 0.05m3/s, creating a total supply of 0.16m3/s.

I would really appreciate any feedback from anyone who has experience with KYPipe and water distribution networks.

Thanks!

 

[coloeng]

I believe you are modeling under static conditions, i.e. you are looking at only one moment in time. The flows/pressures are controlled by the parameters that you have put into the model such as water elevation in the tanks and reservoir, demands, etc. The model assumes you have unlimited flow into the reservoir for static conditions and then tries to develop the hydraulic gradeline in the rest of the system based on the water elevation at the reservoir. If the elevation of the tanks is below the HGL, they will fill, if the elevation is above the HGL, they will drain. Remember this is only showing a moment in time. The Reservoir is acting like a clearwell at the treatment plant. In real life you try to maintain a constant flow through your treatment plant during daily operations and let your clearwell and storage tanks make up for demands higher than the treatment plant flow.

That being said, if all you are looking at is the capacity of the pipes, you need to run the model under a high stress situation, say max day with a fire flow at some point; peak hour demand; etc. Then you look at the pressures at the nodes in your model and determine if they are adequate, to low, or to high at any of them. You also need to look at the velocities and headlosses in the pipelines to determine if any velocities or headlosses are higher than acceptable. That is typically the purpose of running a static model.

If you want to determine if you have sufficient storage and pumping and if it is operating properly, you need to run an Extended Period Simulation (EPS). This models the system over a given period of time (say 24 hours at 30 minute intervals) with the demands changing as they do in the real system. Again, you should run this under max day conditions to evaluate adequacy of storage, pumping, etc.

Hope this helps.

 

[cvg]

in addition, if you're plant discharges at a certain rate which you say is 9,000 m3/day (0.104 m3/s), than that should be modeled as a constant pumping rate into the clear water reservoir. given that your plant flow rate is slightly less than the daily demand, your simulation should show drawdown in the reservoir and tanks during the day and replenishment at night when demand is less

 

[meghna77]

Thank you for your valued feedback.

I am , indeed, modelling the system at only one moment in time. It just represents the demands in general, and not at any particular time of the day.

I have not tried the EPS yet, but I will definitely give it a shot.

I tried to model a constant pumping rate into the reservoir (the pumping rate being 9,000m3/day) but I am still getting similar results.

The thing that is really confusing me is why I am getting an outflow (flow into the tanks) added to the demands. This is, in turn increasing my inflow to the entire system, and not giving me a correct picture.

What I basically want to model is that if 9,000m3/day is flowing into the system from the tanks at the treatment plant, what are my heads at each of the demand nodes.

Have I missed something here? Is a component being overlooked?

I can attach the .p2k file here if my explanation seems somewhat confusing.

I am new to the analysis of distribution systems on KYPipe so any feedback is very welcome!


Thanks!

 

[cvg]

it seems clear that your plant has the highest head and your tanks are lower. at some point they will fill and altitude valves will shut and then all flow will be from the plant. hold off on eps, you need to fully understand your model first and work the bugs out before you try a simulation.

 

[meghna77]

Yes, the tanks at the treatment plant are at an elevation of 2129 masl, and the 9 tanks are all at elevations lower than this.

The results I am getting from the analysis show that there is flow into one tank, contributing to the demand.

I will try to see what happens if I augment the pipes to the junction nodes that have negative heads.

 

[coloeng]

I agree with cvg, if you are relatively new to KYPipe, hold off on trying an EPS.

When you are modeling a static condition, the inflow into the reservoir is irrelevant. All the model looks at is what is the elevation in the reservoir. It will then calculate the hydraulic grade line (HGL) to all the points in the system based on the elevations of those points and the headlosses between the reservoir and those points. The HGL will equal the water surface of the reservoir minus the sum of the headlosses between the reservoir and whatever point you are looking at in the system. Those headlosses are a function of velocity in the pipelines and any minor losses that you have identified. It will develop the HGL and if they water surface elevation in the tank (you set this elevation as one of the parameters in the model) is below the HGL the tank will fill, if the water surface of the tank is above the HGL the tank will drain and if the water surface in the tank is the same as the HGL there will be no flow into or out of the tank.

To get a true representation of the system you need to know what the demands are at that instant in time (or at least a good approximation)and what the water surface elevations in the tanks and reservoir are. Just because you only have 9,000 cmd going into the reservoir doesn't mean that is all that is going to flow out at any given time. A water distribution system is a dynamic system that is ever changing. In the morning when everyone is getting ready for work, taking showers, starting dishwashers, washing cloths, etc. there is a demand that is higher than the average day demand. As everyone goes to work/school the demand goes down some but usually stays a little above average daily. In the evening when everyone gets home from work demand goes up again and is usually the highest demand of the day. During these periods, when demand is above average, the reservoir at the plant and the tanks in the system are usually draining to provide the additional flow necessary. At night when people are sleeping, your demand drops below average and the reservoir and tanks fill back up. This continues on a regular basis with some days having higher demands than other days because of hotter temperatures. The seasons also play a role in this with winter usually being a time of lower demand and lower peaking factors than summer. Your treatment plant is typically operated at the average daily flow for whatever season you are in with the tanks making up for the demands in the system that are higher than average daily demands.

Until you have an understanding of how a distribution system works, trying to model it is meaningless.

 

[meghna77]

Thanks for your response.

That makes better sense in helping me understand the model.

I understand that it is the water level at the reservoir that governs the HGL in the system. Out of the 9 tanks in my stsyem, 4 are at about the same level and it is only one of them that has water filling it up, while the other 3 are draining to supply the rest of the system. I understand now that this is due to the HGL which may be lower or higher than the water level in the main reservoir, and thus either resulting to flow into the tank or out of the tank.

I have augmented certain sections in the system to ensure I have no negative pressures in the system. The residual heads at each of the junction nodes are positive. I was only worried about the flow into one of these tanks, because I need to make sure that does not affect the capacity of the system.

Again, the point you have raised about demands being different at different times of the day is quite valid. I will look into fine-tuning the system a little later on, once I get the hang of the basics.

 

[cvg]

your lowest tank does not deliver water under daily flow conditions, it fills up. that's ok, but at some point it will be completely full. I would suspect that most of the time it is full. at that point, you should have a control valve that will shut off the flow into the tank to prevent overfilling. once that closes, than the system pressures will increase and the total outflow will be lower. you may want to investigate this a bit further since it is somewhat unusual to have several tanks "floating" on the system at different elevations. normally they would all be about the same elevation. are you sure you don't have a lower pressure zone that this one tank serves?

 

[meghna77]

I think I was not very detailed in my explanation.

This is what happens:

There is a treatment plant with a reservoir which is the source of supply for the entire system (at an elevation of 2129 masl).

The entire system comprises of 9 tanks within the network. Once the water leaves the reservoir, it gravitates to 4 tanks at an elevation of approximately 2060 masl. These 4 tanks store water and also supply water to the rest of the system comprising of 5 more tanks further downstream.

Out of the 4 tanks mentioned above, it is one of them that shows an outflow (meaning water flowing into it, while the rest show a positive flow, insinuating they are draining water into the network).

I hope this makes it easier to understand the system.