View Full Version : Pump operation rules

May 31, 2017, 07:38 PM
Is it possible to define specific operation times for a pump?
How to define the moment when pump begins operating, and the moment when pump stops operating?

I have a system that works fine, except for peak hours (7AM to 9AM). There is a storage tank, but with no elevation. Thus, How can i simulate a pump added to the tank so that it will operate during peak hours.

The easiest way is to use a constant power input pump. However, when I go to extended period, and try other pump, I don't know how to control the operation times of the pump. For instance, the constant power and steady state suggest to pump X flow with Y head gain. If I use those X Y values as design point, I get other results (it always uses the maximum flow which is 2*X).

How could I define the operation rules of the pump?

Patrick Moore
June 1, 2017, 09:17 AM

It sounds like you have two primary questions

1) How do I control a pump to operate at a certain time
2) How do I get that pump to have a specific head and flow when running.

Question 1 How do I control a pump to operate at a certain time --> Answer: There are several ways to control pumps. The easiest way to have a pump run at a specific time is to use a simple control which are on the element itself. If you click on a pipe and select the "Control" icon in the model explorer (looks like a white hand on a blue pipe) you can define controls. See the steps below to add the controls:

Step 1: Identify the action to be done
Step 2: Identify the Method (clock time recommended for operation at a specific time each day)
Step 3: Identify the operator required (in this case the time of day)
Step 4: Identify the clock style for the timer control
Step 5: Select Insert to add a control to the list at the bottom as defined
Step 6: Select Create or Update to create or update the control with the changes made

(click if need larger image)

This type of control will turn the pump on each day at 7 am and turn it off at 9 am if this is the type of control desired.

The most common type of control is more often a level based control that turns a pump on when a tank is low and off at a higher level in the tank and can be done as well if desired by using the Pressure/head option, selecting the tank used for control and defining what level to use for each pump action of on or off.

Now Question 2: How do I get that pump to have a specific head and flow when running. Answer: If when you pump is running the pump is not flowing at the head and flow expected, this essentially means that the system conditions would not allow your pump to run as expected given the way it was defined and the head difference between the suction and discharge sides of the pump. A pump will operate only where the pump curve meets the system curve and if your design point is operating at 2X the design flow then it means that the Design head you used is way too high. It may be that you have the suction head conditions much higher than expected or that whatever is controlling the HGL on the discharge side has the HGL depressed lower than expected, but if the flow is that high it generally means you have a head issue such that the pump head gain is very low which causes the higher than expected flow. This could also occur if this pump is the only water source and it is required to pull way more water to satisfy demand than expected by the design flow of the pump. You may need to investigate that as well. All demand has to be satisfied, so it is either the head conditions or the required demand that is often governing where the pump will flow on its pump curve. A design point curve assumes shutoff head is 2x the design head and the high flow point is at 0 head and 2x the design flow.

This blog post contains a downloadable presentation may help you better understand system curves and pump curves, but this is a concept also well explained in many fluid dynamic texts as well. Reviewing how pumps work goes a long way in understanding what they do in the model as it is the system conditions that govern where a pump will operate by dictating the system curve.

Blog Post Link: (specifically pages 8,10,14, 17 may be helpful to you) Introduction to System Curves – Basic Hydraulic
Considerations (http://blog.innovyze.com/2017/01/24/introduction-to-system-curves-basic-hydraulic-considerations/)

If you ran an EPS you can also ask the model to Plot the pump curve with the operating point on it using the Report Manager and asking it to make a new graph report -> Pump Curve. This will plot the pump curve and the specific operating point at the time selected.

(click if need larger image)

Both of these tools can help you see quickly where the pump is operating. If in the pump output where the flow is high, you look at the pump head gain this will tell you at what point the pump is operating. Knowing these two values you can generally infer why the pump operated where it did based on the suction and discharge heads and/or in some cases the required system demand. If you adjust your system to address the issue (either an error in the suction or discharge head somewhere) or if need be you would need to adjust your pump design head and flow you should easily be able to get the pump to run where you expect it to.


June 1, 2017, 09:48 PM
Thanks, the control option works fine.

Sorry, two more questions.
How could I connect the pump to the system, so that it pumps at high demand, and stores water during low demand. I was making a connection similar to the attached file. But pipe A returns some of the pumped water.

The maximum hourly demand factor is 3.5. But the demand subdaily pattern curve (for instance the one from AWWA 1999) has a maximum factor of 1.8.
Which demand curve should be used to perform EPS for analyzing pumps?

AWWA 1999, Residential end uses of water, AWWA research foundation, Denver-USA, 352 pp

Patrick Moore
June 2, 2017, 08:04 AM

Generally when you model a system you first find out how the actual system is running and then work to get your model to as close as needed replicated that operation. If you are designing a new system then you generally need to look at typical operation and local design standards plus good engineering judgment to lay out a system. There are unfortunately no specific "guidelines' that will always fit every situation. You will need to do a bit of research of the area being modeled to really determine what to use.

Will the tank be a ground tank not at the system HGL? This is the most likely scenario for a system where you are pumping during peak demands and filling during low demand periods. In this case because the tank is at a lower HGL that the discharge of the pump station there has to be some type of control device to prevent the tank from filling which is usually some type of altitude valve. This is necessary due to the head difference in the system and the HGL of the tank. If you simply open a pipe to the tank the rate of flow is generally too high and will negatively impact the pressures on the system side. I suspect this is what you have in "pipe A" you note but did not apparently label. This connection needs to be closed when the pumps are running to prevent recirculation of the pumped water.

In that situation it is relatively easy to do what you say, the pumps are set to run during peak demand periods and are off the rest of the day. The tank is then filled during the non pumping portion of the day based on a Flowrate the system can handle. The fill line should only open when the pumps are not running to avoid recirculation.

As far as the peaking factor goes, use great caution with the AWWA curve. This was based on a single system just to give an example and was not meant to be used in other models when no curve was available. Peaking factors and diurnal curves are very system specific and are always best developed form actual system data from your system SCADA data using a mass balance of flows in and out of a region plus the change in storage. In addition, the smaller the system the higher the peaking factor generally is. The AWWA curve was form a large Midwestern city and has a peaking factor in the ~1.5-2 range which is somewhat typical for a large city. A small subdivision though could easily have a peaking factor of nearly 4. You should look at local usage to best determine what design peaking factors to use for your system as there are no good "guidelines" developed that perfectly fit everywhere. For design a municipality usually has peaking factors and customer usage data they feel comfortable using for design purposes so check with the local municipality. Otherwise you may need to do a sensitivity analysis in case your peaking factor is too low to determine if your design is robust enough to still be ok if your peaking factor is too low.

Hope this helps.