Results 1 to 4 of 4

Thread: System Curves for EPS run

  1. #1

    Join Date
    Mar 2016
    Posts
    6

    System Curves for EPS run

    Hi

    I need to produce some system curves for a specific pump in H2OMap model of a bulk water supply network. The model is currently set up as a 5 day EPS run.
    The pump in question pumps from Reservoir A to Reservoir B and the suction side of downstream pump.

    When I produce a system curve for the pump in question at time step 0, the curve is as I expect (very minor static lift and shallow curve). However when I produce system curves at other time steps the static lift is much higher than I expect - the static should be minor given there is little variation in head between the two reservoirs (max of ~2.4m but the model system curve output produces ~8-12m) and the system curve is almost a straight line and an inverted curve. As a sanity check I've gone back to the EPS run at these particular time steps to check the static and I can't replicate the static head that the system curves are giving me.

    The time steps I picked to run the system curves at were time steps where the pump should be operating as per the EPS run.

    Can you see where I may be going wrong? Are there any specific tricks to using the system curve tool? Any guidance would be appreciated.

    Thanks
    Gina

  2. #2
    Forum Moderator

    Innovyze Employee



    Innovyze Employee



    Join Date
    May 2015
    Posts
    435
    Gina,

    One thing to keep in mind about system curves is that there is a family of possible curves for most systems that will vary as the system operates. The trick is usually to identify conditions for the boundary upper and lower system curves as then all other curves will fall in between the boundaries.

    If you want to quickly test what is happening, shut off the pumps and place a positive demand on a suction node and a negative demand of the same value on the discharge side. The head point of the system curve at that flow is the discharge head - Suction head. You can then see if anything unusual is happening in the system to explain the large system curve head. Just select a flow value where they system curve seems odd and use that flow in the manual run.

    If you have a pumped only zone on the suction side though, what can occur is that as the model varies the demand to generate the system curve, there may not be enough pump supply to support the higher flows and this depresses the HGL on the suction side. This may explain what is occurring in your system.

    If this does not resolve your issue please consider forwarding your question on to support@innovyze.com and we can help you investigate the possible cause in more depth.

    While it sounds like you have a pretty good grip on the hydraulics that impact system curves here is a bit of information that may help you and others better understand key points that govern system curves.

    When a system curve analysis is run, this shows the basic procedure used by the model and or in a manual run.
    System Curve Analysis (click for larger image if necessary)
    System Curve Analysis.jpg

    For most systems the maximum and minimum static lift would be calculated as follows:
    Min and Max static Lift (click for larger image if necessary)
    max and min static lift.jpg

    But on the system curve the actual static lift for the minimum layer can also be slightly lower than shown in the previous curve when the demands are high. This occurs, because the receiving tank is supplying water to the system and the headloss between the tank and the discharge of the station can depress the HGL on the discharge of the station which in essence lowers the static lift the station would actually "see" once it started running. In the curve below this is why the Red minimum system curve is lower than the min static lift with zero demand curve.

    See below for an example of possible system curves for a given system
    Example Family of System Curves (click for larger image if necessary)
    Family of system curves-simple.jpg

    To understand to model tool, there are a few things the modeler should always keep in mind.
    1. The model will assume that all flow on the system curve goes through the specific pump that was selected for the analysis. If the pipes connected to the pump are on the smaller side, they may not have enough capacity to push large flows through and can experience higher than expected headloss at higher flows. Also the model max flow for the system curve is calculated form the pump curve and is the flow point where the projected curve would have zero head. Since that flow is generally too small for the entire pump station combined flow of all pumps some modelers create a "dummy" pump to run the system curve analysis which has larger connecting pipes that can allow for larger flows and run the system curve using a combined pump curve for all the pumps at the station This allows the system curve tool to calculate the full pump station flow range using the dummy pump with a modified curve is used. (NOTE: if you have a suite license you can easily run a system curve with multiple pumps or pumps in parallel as well using the Pump System Analyst extension as well and this can often be even easier)
    2. The system curve will be run either as a steady state or as an EPS run where hours are selected. To make sure your system boundary curves are identified one often has to set up each conditions for the maximum or minimum curve and then run the analysis. The model will assume the boundary conditions at the time specified govern when running the analysis. The modeler must make sure these conditions represent conditions to get the boundary system curves or you will just calculate intermediate curves. This is often most easily done by using two separate steady state runs with the conditions set to create the maximum and minimum boundary system curve.


    Patrick Moore

  3. #3

    Join Date
    Mar 2016
    Posts
    6
    Hi Patrick
    Thanks for your response, I have done the checks as you suggested. I should have mentioned in my post that for this system there are 2 pumps which operate in parallel. As per your point 1 above, I have now modelled a 'dummy' pump, which represents the combined 2 duties, and the resulting system curves are as I expect.

    I would like to understand more about why my initial approach did not work. The two parallel pumps are set up with the same controls so they run at the same time. When I produce the system curves for a single pump for the EPS varying times is it the second pump operation which is interfering with the system curve, causing the unusual shape? Or does the model know to turn the other parallel pump off? (I've included below the system curve outputs for my initial approach and revised approach).

    The suite extension that you mention "Pump System Analyst", is this available in H2OMap? Please let me know if I should email the support team or whether this can be addressed here.

    Initial approach (single pump selected for system curve output) System Curve Results:
    Pump curve 1 - inital approach.jpg

    Revised approach
    (combined pump curves) System Curve Results
    Combined pump curve.jpg

    Thanks,

    Gina

  4. #4
    Forum Moderator

    Innovyze Employee



    Innovyze Employee



    Join Date
    May 2015
    Posts
    435
    Gina,

    The model tool will calculate the system curve under the assumptions you specify. However, while you can run the system curve tool when another pump is running, generally people evaluate system curves as a "pump station" property rather than a "pump" property. This means the evaluation is based on the energy requirements to make 1, 2, 3, up to the maximum number of pumps run and evaluate the pump operation under the range of possible conditions so that one can choose the best pump to fit that pump station.

    In your initial approach you are only evaluating the system curve for a single pump at a time under different operating conditions (caused by different EPS run times) which usually changes the static lift and the system demand so that the three curves don't overlap. In addition the maximum flow for 1 pump is not large enough to get the full station flow. Since you likely ran the three analyses with system demand there is a more sinusoidal shape to the curve in the initial portion of the curve as initially not all of the pump station supply goes to the receiving tank. Once the pump station supply exceeds the zone demand all of the remaining water has to go to the tank and you start to see a more expected shape of the system curve for flows above 800-1000 l/s.

    If you recall my previous post your curves look more like the lower red line which is the expected shape of the system curve when there are system demands occurring.

    Once you used the dummy pump the system curve tool can run at higher flows and you see the full shape of the curve that is relevant for multiple pumps running.

    We highly recommend you evaluate the upper and lower bounds of your pump stations system curves though as all other possible curves for the pump station will fall within the boundary system curves. Often, the maximum system curve is found at maximum static lift and no system demand. This forces all water to the receiving tank and generally has the highest headloss. The minimum system curve is often associated with maximum system demand and the minimum static lift. But each engineer should always evaluate his system to confirm what the boundary system curves are for that particular system. To run the model at the boundary conditions it is often necessary to set up the boundary conditions as desired and to run the system curve analysis at time = 0 (or as a steady state run) when you can easily specify the conditions the model will see. This often requires multiple runs of the system curve tool as the boundary conditions are generally quite different for the maximum and minimum curves.

    Note: It is recommended to turn off all other pumps at a pump station when running a system curve as the model will only turn off the pump selected to run the system curve on. The modeler should have the pumps off and disable controls that could turn the pumps on as well. This may be easiest to simply use the facility manager to inactivate the pump elements to ensure they don't run. If allowed to run, the other pumps will interfere with the system curve analysis and could lead to confusion when interpreting the results.

    Also, the Pump Station Analyst (PSA) tool is included with our InfoWater and H2ONET suite license platforms, but is unfortunately not currently included with H2OMAP Water suite tool at this time. When the PSA tool is available, it can serve as an additional resource for evaluating pump stations and calculating system curves that can supplement the system curve tool in the run manager.

    Hope the helps,

    Patrick Moore
    Quote Originally Posted by Gina_L View Post
    Hi Patrick
    Thanks for your response, I have done the checks as you suggested. I should have mentioned in my post that for this system there are 2 pumps which operate in parallel. As per your point 1 above, I have now modelled a 'dummy' pump, which represents the combined 2 duties, and the resulting system curves are as I expect.

    I would like to understand more about why my initial approach did not work. The two parallel pumps are set up with the same controls so they run at the same time. When I produce the system curves for a single pump for the EPS varying times is it the second pump operation which is interfering with the system curve, causing the unusual shape? Or does the model know to turn the other parallel pump off? (I've included below the system curve outputs for my initial approach and revised approach).

    The suite extension that you mention "Pump System Analyst", is this available in H2OMap? Please let me know if I should email the support team or whether this can be addressed here.

    Initial approach (single pump selected for system curve output) System Curve Results:
    Pump curve 1 - inital approach.jpg

    Revised approach
    (combined pump curves) System Curve Results
    Combined pump curve.jpg

    Thanks,

    Gina

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •