Calibration Help

Hello! 
I'm tring to calibrate my model with my field data, i already put a VSP for my pump and i got the error "Pump cannot deliver flow or head", when i change the values of the pattern sometimes i got the error and sometimes not, is there any form to find out the adequate pattern form my pump without field data on the downstream of the pump? The best way it's to test various speed factor on the pattern? 

  • Hello Fabio,

    Information on that user notification can be found here: User Notification: Pump cannot deliver head or flow . Generally it means that the pump is operating on the curve, but outside of the values in you table. This could occur with variable speed pumps as well. See this link for more information modeling variable speed pumps: Using Variable Speed Pumps in WaterGEMS and WaterCAD:

    Pattern-based VSPs are typical when you know how the relative speed factor will change with time. If you are using the pattern as a way to calibrate the model, this is not a method I have seen used before. Typically, you would make adjustments to roughness values in the pipes to make changes to headloss. 

    I would recommend reviewing the following link related to model calibration: Water model calibration tips. If you are using WaterGEMS, you could also explore using Darwin Calibrator to help with the model calibration.

    Regards,

    Scott

    Answer Verified By: Fabio Lobo Araujo 

  • Thanks Scott! Usually i run calibrations making adjustments on roughness values, but in this case my Hazen-Williams coefficient it's already really low, the adjust results in values of 100 or 105 for the coefficient, and the pressue values still aren't equivallent to field data, i think that these Hazen Williams values are not coerent with the field, the system of the city has like 64 years, the material it's PVC, it's there any possibly that my coefficient goes under those values? 





      

  • You should not assume that C-factor is the only adjustment to your model to improve calibration. There are many sources of discrepancies between the model and the field data. Here is a list of some. You need to be an engineer and a detective to calibrate a model.

    Table 1-1. Sources of Discrepancies

    Physical

    Pipe size/location

    Pipe connectivity

    Pipe roughness

    Pressure zone boundary

    Pump curves

    Pipe material/age in GIS

    System changes since model built

    Elevation data

     

    Operational

    Valve open/closed/throttled status

    Control valve operation/settings

    Transient events

    Actual operations not matching control rules

    Unusual operations when data were collected

    Tank water levels

    Pump status/speed

    Lack of sufficient sensors/gages

    Water quality reaction rates

    Air blockage in pipe

    Demands

    Spatial allocation

    Model does not reflect conditions when data collected

    Large customers with atypical demand patterns

    Not accounting for seasonal changes in demand

     

    Data

    Inaccurate/uncalibrated gages/meters

    “Latched” data from SCADA

    Understanding SCADA data – average vs. instantaneous

     

     

    Answer Verified By: Fabio Lobo Araujo 

  • Thanks Tom!! 

    It's reasonable to use the item "Pump Status/Speed" to calibrate? Do you already see any calibration by changing speed factors of the pump? 

    I'm comparing pressures at a point some meters away from the pump. I also don't have field data immediately after the pump. So after checking my valves, tank levels and setting a reasonable C-factor, i'm tring to reach field data by changing speed factor of the pump.


  • The key to calibration is understanding WHY the model and field data differ. If it is likely that pump speed is in error, that is something you can adjust. Why would you suspect the speed is incorrect? Are you adjusting speed based on a set point or time controls?

    Answer Verified By: Fabio Lobo Araujo