AFT Product Tips
Applying Communication Time: Let’s Talk About Your Impulse Result/Event Relationships
I have received questions from clients using AFT Impulse where they ask something like
"I closed this valve. Why is the maximum pressure spike way over there and not at the valve?"
This comes from years of developing a "common sense" adapted to steady-state flow. With steady flow, a change between states of equilibrium has a local effect that then spreads across the system in a straightforward path. However, with transients, the location and magnitude of the max pressure spike is based heavily on the geometry of the piping network that the pressure wave reaches during the simulation.
Models of piping systems often have more than one transient event in them that is either set to occur or can be triggered. For example, a relief valve has a transient event that can be triggered by pressure changes in the system. When pressure waves are generated by multiple transient events, these waves and their reflections will spread throughout the system and interact with each other. Some of these wave reflections will have constructive wave interference with each-other. If pressure waves (or their reflections) originated from two or more different transient events and interact constructively, they can make pressure wave spikes that may be larger than the separate pressure waves they originated from. This can happen at a location far from where the transient events that created the pressure waves were located.
Another thing to consider when validating pressure spike results is cavitation, which forms vapor cavities that can collapse and cause significant pressure spikes. When transient event induced pressure waves reflect for their second time, they cause negative pressure waves as part of their communication process. The negative waves can bring a location's pressure below vapor pressure, forming a vapor cavity. When this vapor cavity collapses, it can cause a large pressure wave that will, in turn, interact with other pressure waves or even create other vapor cavity collapses elsewhere.
Therefore, the "common sense" of transient flow is the understanding of how these pressure waves move. The question becomes "When are they going to hit what and what is going to happen when they hit it?"
A key to answering this question is pressure wave communication time.
The communication time formula is t = 2L/a, where t = communication time, a = the wave speed, and L = the distance from the wave-creating transient event to the point it is reflecting from. Communication time is the time it takes the pressure wave from a transient event to travel to the other end of the system (or a reflection point) and back in order to communicate the effects of the rest of the system back to the location of the transient event. This is synonymous to echo-location. An animal that uses echolocation can only react to what it knows is happening, which it figures out by creating a sound wave that then reflects off its surroundings and returns to the animal.
There are very useful insights that can be gathered from an understanding of communication time:
- 1)The effects of a transient event do not reach a location of your piping system until half the communication time between the two points has passed: t_affect = L/a
- 2)The location of the transient event is not affected by what it affects until the communication time (between it and what it affected) has passed: t_communication = 2L/a
Here is how you can use Impulse to take advantage of this insight:
In Impulse's Output window, it can display Communication Time for each pipe. By adding the communication times of the pipes between a transient event's location and a location of interest (like a relief valve) and dividing by 2, you get how much time must pass before the pressure wave of the transient event can start affecting that location (or triggering an event at that location). This is also good for seeing if pressure waves from reflections are going to interact constructively.
Now that you understand the usefulness of communication time data, there are a couple of other insights about transient pressure waves (and simulating them) that I want to share:
- 1) Check valves are known for causing pressure spikes in transient events. However, sometimes adding an instantaneously closing check valve in your model can unrealistically reduce the size of a pressure wave in your model causing cavitation to unrealistically disappear from the results.
- a. When a check valve junction is set to close instantly upon flow velocity equaling zero (right when the flow is reversing), this causes the check valve to stop the flow without having to push against reverse flow. This results in no pressure wave being generated by the check valve. However, a real check valve is going to take some time to close, causing it to have to push against the reverse flow and produce a pressure wave. You can get this realistic effect by adding inertia data to your check valve.
- 2) If pressure waves are still being generated after the wave cycle of your last transient event has happened, there may be a hidden pump internal check valve that is still creating transient events. When the flow velocity oscillates close to zero, it can cause these check valves to chatter, causing odd results.
I hope this provided some useful insight for you in validating transient flow/waterhammer simulations with Impulse. For a detailed explanation of other counterintuitive effects of transient flow on the pressures of your system, you can also read the technical paper "When the Joukowsky Equation Does Not Predict Maximum Water Hammer Pressures", authored by AFT's CEO Trey Walters and Dr. Robert A. Leishear.