Darwin designer violates velocity constraints

Hello Adithya,

This may be a case of needing to review the setup of the design study. This link has some information that may be useful. I would also highly recommend review WaterGEMS Help. There is a topic called "Advanced Darwin Designer Tips" that may be useful. In that Help topic, there is a second about things to look at when the results do not look right. This may have some steps you can try to improve the results.

Otherwise we will need some additional information about the setup of the design study. It may also be useful to see a copy of the model files. There are two options for sharing your model on Communities, whichever you choose please be sure to zip your files first. The first option is to attach the zip file containing your model to your reply on the forum using the Advanced Reply Editor (you'll find the link below and to the right of the reply box). If your data is confidential please use the instructions in the link below to send it via Bentley Sharefile. Files uploaded to Sharefile can only be viewed by Bentley employees. Please be sure to reply on this thread with the name of the file after it has been uploaded.

Hi Scott,

Yeah I will give a try with your suggestion about advanced darwin designer tips. 

Regarding sharing of files, I have no problem. I am attaching the files herewith . It would be good to have a professional insight into the problem.  

7065.Kadtal_P.rar

My main objective is to try pumped design network in watergems. Instead of taking big network i wanted to try darwin designer for a small sample network then later on go for the big network. But  when i tried DD i have given both pressure and velocity constraints ( cause both are required in my design). As seen in the image attached, DD clearly violates the velocity but takes pressure constraints into consideration. The files attached are just for reference (small network).

Thanks 

Adithya

Hello Adithya,

I did some testing on the model you provided. Included in that was changing the information in the Options tab in the design study so that more trials would be attempted. I also did some manual testing by changing the pipe sizes within the model itself. Increasing some of the information in the Options tab did not yield better results. In fact, the same set of results that you initially see appear to be returned.

I also looked at the model as it currently looks, then tried increasing the size of some of the pipes manually to see how the results differ. The pump flow in your model was 33 L/s with 30.5 meters of head added. Taking P-2 and P-9 as representative pipes, the velocity at P-2 is 5.99 m/s with a flow of 38 L/s. The velocity at P-9 is 7.58 m/s with 33 L/s of flow.

Clearly a larger pipe diameter pipe is needed to decrease the velocity. I manually changed the diameter of P-2, P-4, P-8, and P-9 to 150 mm and recomputed the model. You will notice that the velocities are all around 7.5 m/s and the flows in the pipes are higher. This is expected based on the hydraulics of the system. The pump only needs to add about 7.75 meters of head to the system. This is becasue with the larger pipe diameters the headlosses in the pipes are smaller, even with more flow.

When running Darwin Designer, the design engine is seeing the same sort of thing. Increasing the sizes of the pipes does not necessarily decrease the velocity, since the hydraulics of the system dictate that the pump will provide less head to the system and therefore more flow. More flow means a higher velocity. So Darwin Designer is unable to satisfy the velocity constraint. It therefore tries to minimize the cost by using the smallest possible pipe diameters along with the lowest possible violation.

There are a few things that you will want to look at. First, make sure the tank elevations are accurate for what you are using. If the tanks in the field are not at the same level as the model, this will impact the results. For instance, if the downstream tank was actually had a higher elevation, the head needed from the pump will be higher. The flow would than be less and the velocity smaller. I would also make sure that you elevations are all filled in; the pump and the tanks have a ground elevation set to zero. While this will not impact the hydraulics, it will give you better results in terms of pressure, at least at the pump.

Lastly is the pump itself. This is currently set to us a constant power as the pump definition. I would recommend using a three-point curve or multiple-point curve instead. You should be able to get this information from the pump manufacturer. These are more accurate pump types to use and will therefore yield more accurate results. For more information on the different pump types, please see the Help documentation in the section titled "Pump Theory."

Regards,
Scott

Hello Adithya,

I did some testing on the model you provided. Included in that was changing the information in the Options tab in the design study so that more trials would be attempted. I also did some manual testing by changing the pipe sizes within the model itself. Increasing some of the information in the Options tab did not yield better results. In fact, the same set of results that you initially see appear to be returned.

I also looked at the model as it currently looks, then tried increasing the size of some of the pipes manually to see how the results differ. The pump flow in your model was 33 L/s with 30.5 meters of head added. Taking P-2 and P-9 as representative pipes, the velocity at P-2 is 5.99 m/s with a flow of 38 L/s. The velocity at P-9 is 7.58 m/s with 33 L/s of flow.

Clearly a larger pipe diameter pipe is needed to decrease the velocity. I manually changed the diameter of P-2, P-4, P-8, and P-9 to 150 mm and recomputed the model. You will notice that the velocities are all around 7.5 m/s and the flows in the pipes are higher. This is expected based on the hydraulics of the system. The pump only needs to add about 7.75 meters of head to the system. This is becasue with the larger pipe diameters the headlosses in the pipes are smaller, even with more flow.

When running Darwin Designer, the design engine is seeing the same sort of thing. Increasing the sizes of the pipes does not necessarily decrease the velocity, since the hydraulics of the system dictate that the pump will provide less head to the system and therefore more flow. More flow means a higher velocity. So Darwin Designer is unable to satisfy the velocity constraint. It therefore tries to minimize the cost by using the smallest possible pipe diameters along with the lowest possible violation.

There are a few things that you will want to look at. First, make sure the tank elevations are accurate for what you are using. If the tanks in the field are not at the same level as the model, this will impact the results. For instance, if the downstream tank was actually had a higher elevation, the head needed from the pump will be higher. The flow would than be less and the velocity smaller. I would also make sure that you elevations are all filled in; the pump and the tanks have a ground elevation set to zero. While this will not impact the hydraulics, it will give you better results in terms of pressure, at least at the pump.

Lastly is the pump itself. This is currently set to us a constant power as the pump definition. I would recommend using a three-point curve or multiple-point curve instead. You should be able to get this information from the pump manufacturer. These are more accurate pump types to use and will therefore yield more accurate results. For more information on the different pump types, please see the Help documentation in the section titled "Pump Theory."

Regards,
Scott