Results 1 to 2 of 2

Thread: Elevated Tank level EPS trending

  1. #1

    Elevated Tank level EPS trending

    I'm in the process of updating my model but my model tanks are trending opposite of my field/SCADA tank trending. Demands have been updated via mass balance, scaled to match plant production and allocated via geocoded billing data by "closest node". All tank, pump and valve boundary conditions match field/SCADA. One tank trends opposite for the first three hours of the day then trends correctly, one tank is spot on, one tank trends correctly but the levels are way off, too low, then starting at 1500 hrs two tanks trending is off again. 24 hr EPS, 1 hour timestep reporting. Any suggestions would be greatly appreciated.

    Happy 4th!!

  2. #2
    Forum Moderator

    Innovyze Employee

    Innovyze Employee

    Join Date
    May 2015

    I know its been a while since you posted your question, but I had a few suggestions you might consider trying if you are still having questions.

    When a modeler is trying to get tank trending within the model to match SCADA field data the following items are often good to check:

    1) Verify the starting boundary conditions are identical with the model and the SCADA data. (it appears you have done this) Obviously starting at a different level will impact when pumps or valves with controls turn on.
    2) Check to make sure the total system demand for both the model and field data are identical. Even if you load meter data with billing data, you will have to scale the meter data to match actual system consumption. One of the best ways to check system consumption is during diurnal curve creation using a system mass balance methodology. The average system demand calculated by the mass balance for a particular region is the same region that you use to scale the meter demands to match what was calculated in the mass balance. If the total system demands do not match, it will be difficult to get tank trending to match as the water consumed and supplied will be off and this will show up as a difference in the water storage tank levels.
    3) Diurnal Curve Errors part 1: When calculating diurnal curves make sure they are normalized. What this means is if you calculate the demand for each hour of the day for a region and divide that by the average demand for the simulation period you will have a normalized diurnal that is a multiplier from the average demand. This demand will never be greater or less than the base demand (i.e. the demand included in the Demand1 -> Demand 10 fields) over the simulation period. If the demand is not normalized, the overall demand will not match what was observed on the day of comparison and can lead to errors. The best way to check is to sum up all of the individual diurnal points calculated. The sum will equal the number of simulation periods in the pattern. So for a 24 hour pattern it will sum to 24. Watch out for rounding errors as well. If the model uses only 2 decimal places check the rounded values to make sure they add up to the correct total or small errors can be introduced.
    4) Diurnal Curve Errors part 2: If the areas used for a mass balance are large (i.e. cover multiple pressure zones) there is a greater chance that the overall pattern for the region will differ from a specific part of the system. Thus even if you have a correctly calculated diurnal, you may have errors in the model as certain portions of the system use water differently than the calculated pattern says they should. This is the one reason that the more specific and smaller the regions used to calculate the diurnals are generally more accurate and help the model to better replicate the real system.
    5) System controls: Generally if tank trends are not matching SCADA values, the most common culprit is errors in the system controls for pumps (and sometimes valves) that fill and drain the tanks. If a given tank is off, look at the pumps station flows for the stations that fill the tank and at the pump stations that pull water from that tank to feed another zone. Pump control errors will much more strongly impact tank results than errors in the demands or diurnals will simply based on the high flow associated with pumping. You may need to look at the inflows and outflows of the entire zone to get the whole picture. If pumps are not running at the same time in the model as they were in SCADA this will cause major differences in the tank trending. When checking you will need to usually make adjustments to the set point controls that tell a pump when to turn on or off in the model. Zero in on the controls impacting the tank level, and more often than not that will resolve the trending errors.

    Hope this helps you or others having a similar issue.

    Innovyze Support

Tags for this Thread

Posting Permissions

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