Below are some responses to frequently asked questions. 

1. There are many easy and simplified surge analyses (e.g., Joukowski relation). Why do we need to use InfoSurge?

2. What is the difference between Method of Characteristic (MOC) and Wave Characteristic Method (WCM)? Why WCM is applied to InfoSurge?

3. What are the pressure and flow units used in InfoSurge?

4. Is it possible to run a large InfoSurge model (e.g., 10,000 links) with a small link InfoWater license (e.g., 1,000 link license)?

5. How to create a transient using pump operation (e.g., pump speed change, pump startup and trip)?

6. How to create a transient using valve operation?

7. How to create a transient using demand change?

8. What should I do if my surge model is not running?

9. Can I simply use the steady analysis model for surge analysis?

10. How to simulate a check valve?

11. What is the "Pipe Segment Length Tolerance?"

12. What is the error message: "invalid pipe connection configuration at surge protection device?

13. What is the error message:"Surge node connections exceed limit (10)?"

14. What is the pressure-sensitive demand in Run Manager?

15. How do you simulate the leakage?

16. What is the Leakage Constant, Leakage Factor and Intrusion Integral? How to use the intrusion modeling?

17. Why intrusion is important in surge analysis?

18. What is the initial air volume in the air valve?

19. I am going to run a surge analysis with a huge water distribution network. To save the computation time, I am planning to use skeletonization techniques to reduce its size. What is the effect of the skeletonization on surge analysis?

20. What data is reported for a surge protection device?

21. How does InfoSurge handle non-TCV valves like pressure reducing valve or flow control valve?

22. How does InfoSurge model vapor cavitation?

23. I am getting very severe pressure spikes in my model. What is happening?

24. Where do I look for cavitation problems?

25. What can I do to determine if cavitation is causing a problem in my model?



  • 1. There are many easy and simplified surge analyses (e.g., Joukowski relation). Why do we need to use InfoSurge?

Water hammer analysis is extremely important but its complexity, perhaps coupled with its mysterious nature and the need for specialized analysis tools, has led to a number of published guidelines for simplifications in the analysis. However, most simple expressions, such as the Joukowski relation, are only applicable under a set of highly restricted and often unrealistic circumstances. When the required conditions are met, the simple relationships are both powerful and accurate. In the case of the Joukowski relation, the two most important restrictions are that there should be only a small head loss resulting from friction and no wave reflections from any hydraulic devices or boundary conditions in the system. If these conditions are not met, the Joukowski expression is no longer valid and the conclusions that are based on this rule may also not be applicable. Moreover, the Joukowski relation does not consider liquid column separation (cavitation). If a negative surge is below the vapor pressure, all gas within the water is gradually released and the collapse of the cavities will result in a large pressure surge spike.

No simplified rules can provide a prediction of the worst-case performance under all transient conditions. The water hammer response in water distribution systems is strongly sensitive to system-specific characteristics, and any careless generalization and simplification could easily lead to incorrect results and inadequate surge protection. Comprehensive water hammer analysis, InfoSurge, is not only needed, but this approach is both justified (by its importance) and practical, owing to the rapid development of fast computers and both powerful and efficient numerical simulation models for water hammer analysis. 


  • 2. What is the difference between Method of Characteristic (MOC) and Wave Characteristic Method (WCM)? Why WCM is applied to InfoSurge?

Both the MOC and WCM method are capable of accurately solving for transient pressures and flows in water distribution networks including the effects of pipe friction. The MOC requires calculations at interior points to handle the wave propagation and the effects of pipe friction. The WCM handles these effects using pressure waves. Therefore, for the same modeling accuracy, the WCM will usually require fewer calculations and provide faster execution times. Because of the difference in calculation requirements and the comparable accuracy of the two techniques, use of the WCM may be more suitable for analyzing large pipe networks. Several examples and paper are presented to compare both MOC and WCM in the following website.

http://www.innovyze.com/products/surge/


  • 3. What are the pressure and flow units used in InfoSurge?

1. Similar to InfoWater Pro, the pressure and flow units are specified in the Simulation Options.

2. Velocity (e.g., wave speed) is in ft/s or m/s based on which flow units are specified in simulation options.

3. Inertia is in lb-ft^2 or N-m^2 based on which flow units are specified in simulation options.


  • 4. Is it possible to run a large InfoSurge model (e.g., 10,000 links) with a small link InfoWater Pro license (e.g., 1,000 link license)?

If the number of the links in a InfoSurge model is more than that of InfoWater Pro license link, InfoSurge will not run. This is because a surge analysis requires its steady analysis for the initial condition. Therefore, InfoWater license will be checked first.


  • 5. How to create a transient using pump operation (e.g., pump speed change, pump startup and trip)?

Please look at "Creating a Transient by Pump".


  • 6. How to create a transient using valve operation?

Please look at "Valve Operation".


  • 7. How to create a transient using demand change?

Please look at "Junction Demand Change".


  • 8. What should I do if my surge model is not running?

When a surge analysis doesn’t run correctly, user can use the Report icon next to the Run icon to open the SURGE.RPT notepad report that will list the problems. Generally, the first thing is to check the “Hydraulic Run Only” check box and re-run the surge analysis. This can give some idea if the error is from a non-converged steady state simulation or from the surge analysis. Please look at more detail information at General Surge Modeling Tips.


  • 9. Can I simply use the steady analysis model for surge analysis?

Most hydraulic elements of steady analysis model can be used directly to surge analysis but a few should be carefully considered. For example, a check valve in pipe (under the Modeling Data of pipe) will be ignored during surge analysis. This is because the location of the check valve is important in surge analysis but is not in steady analysis. Therefore, the check valve should be replaced using active valve in surge analysis. The more detail information regarding the difference between steady and surge analyses can be found in General Surge Modeling Tips.


  • 10. How to simulate a check valve?

There are three possible ways to represent a check valve.

1. In case of check valve in the discharge side of pump, it can be simply represented with the check valve option in Pump Surge Data.


2. A check valve can be separately represented as an Active Valve. In this case, we need to find the corresponding stem change curve to represent the check valve operation (A surge analysis without check valve control may need to be done first to find the time of flow direction change).

3. An active valve can be used as a check valve without using stem change curve. In this case, the valve will work as a simple local loss component and the check valve option in Active Valve Data should be assigned. When the flow direction through the valve is changed, the check valve will be working (same as the Option 1).



  • 11. What is the "Pipe Segment Length Tolerance?"

It is the smallest pipe length that will be considered for Surge analysis. If a pipe has smaller length that the value specified in the “Pipe Segment Length Tolerance”, then automatically the length of the pipe will be consider as equal to the “Pipe Segment Length Tolerance” for the Surge calculations. The smaller this value, the longer time the analysis will take to be completed. Note that pipe lengths in the model must be adjusted so each pipe will be a length such that the pressure wave will traverse the pipe in a time which is an exact multiple of the computational time increment.

  • 12. What is the error message: "invalid pipe connection configuration at surge protection device?”

A surge protection device should be located in a junction connected to two pipes only. If the device is placed on the other junction (e.g., with one or three pipe connection), InfoSurge will create the above error message. The easy remedy for this case is to create a new artificial junction near the original place and then install the protection device (see the following figure)



  • 13. What is the error message: "Surge node connections exceed limit (10)?"

The error means that there is a node in the model that in connected with 10 or more pipes. A node in Surge can not be connected with more than 10 pipes. A dummy node (see below) has to be created to split the pipes.



  • 14. What is the pressure-sensitive demand in Run Manager?  

InfoSurge provides two different demand approaches: demand-driven (pressure-insensitive) and pressure sensitive demands. Demand-driven analysis assumes that the demands are independent of pressures and can be met under all operating conditions. However, under transient conditions the resulting positive or negative pressure surges can drastically alter the local pressures and affect the demand magnitude that can be extracted. Pressure-sensitive demand representation is to assess the impact of pressure changes and produce more accurate transient results.

InfoSurge has a feature of pressure sensitive demand based on the orifice relation (Q = Cv• [dP]0.5). This feature can be applied by clicking Pressure Sensitive Demand in the Run Manager shown below. Otherwise, demands remain constant during the transient. System demand are usually best modeled using this type of demands so the demands will response to pressure changes in the pipe system.


  • 15. How do you simulate the leakage?

Two approaches are available. First, side discharge orifice can simulate a leakage in a local location. Second, using pressure-sensitive demand with leakage factor or constant can estimate the global effect of the leakage. See more the application of leakage factor or constant below.


  • 16. What is the Leakage Constant, Leakage Factor and Intrusion Integral? How to use the intrusion modeling?

The Leakage Constant is defined using the relation Q(leak) = Lc*(pressure difference)^0.5. This is same concept as the orifice equation. For example, if Lc = 0.1, a leak of 0.1 gpm with a pressure difference of 1 psi occurs.

The Leakage Factor is defined as the percent of leakage in the distribution system. For example, a leakage factor of 0.1 means that the leakage is 10% of the demand. From this factor, a leakage constant is calculated for each demand junction, which will discharge 10% of the flow at the initial pressure. For example, at a junction where the demand is 10 gpm at 64 psi (pressure difference) assume that 1 gpm (10% of the demand) is leaking through an orifice, the leakage constant is computed as Lc = 1/(64)^0.5 = 0.125.

The intrusion (volume) is estimated based on either the leakage constant or the leakage factor assigned when the pressure gradient is negative (Pexit > Pline), where Pexit is the exit pressure and Pline is the pressure inside the pipeline system. The Intrusion Integral is defined as the negative area under the time plot of the curve of dP(Pexit – Pline)^0.5. For example, if a node experiences the negative pressure of 4 psi for 6 second (0.1 minute), the intrusion integral is 0.2 [= 0.1*(4)^0.5]. Therefore, the estimated intrusion volume is calculated by multiplying the intrusion integral by a leakage constant. For example, if the node has Lc = 0.1, the intrusion volume through the node is 0.02 gal (=0.1*0.2). Please refer the example of Intrusion modeling.


  • 17. Why intrusion is important in surge analysis?  

One of the challenging problems in distribution system water quality management is that contaminants can intrude into pipes through leaks from reduced or negative pressure transients. In reality, since all pipeline systems leak and hydraulic transients occur more or less continuously in distribution systems, it is not surprising that low pressure transient events (e.g., pump trip and valve opening) introduce a considerable risk of drawing untreated and possibly hazardous water into a pipeline system. In fact, soil and water samples were recently collected adjacent to drinking water pipelines and then tested for occurrence of total and fecal coliforms, clostridium perfringens, bacillus subtilis, coliphage, and enteric viruses (*Karim et al., 2003). This study found that indicator microorganisms and enteric viruses were detected in more than 50% of the samples examined. These and other results suggest that during negative or low pressure events, microorganisms can directly enter the distribution system through pipeline leaks.

*Karim, M.R. et al., 2003. Potential for Pathogen Intrusion during Pressure Transient. Jour. AWWA, 95:5:134.


  • 18. What is the initial air volume in the air valve?

The initial air volume is the air volume that has accumulated at the location of the valve at the starting point of the simulation. The air in the pipe system tends to be accumulated at the high locations (e.g., highly elevated dead end and air valve) and the initial air volume in InfoSurge is purposed to consider the effect of the air pocket. Normally, it is assumed, unless a field test is executed to estimate the air volume, that no air pocket is initially present (assign 0 in the initial air volume).


  • 19. I am going to run a surge analysis with a huge water distribution network. To save the computation time, I am planning to use skeletonization techniques to reduce its size. What is the effect of the skeletonization on surge analysis?

The traditional rules of steady-state model skeletonization (Reduction, Pipes in Series and Parallel, and Trimming) ignore the complex interaction of transient pressure waves in the different pipe properties and characteristics of a water distribution system. The hydraulic equivalency theory used in model skeletonization and derived from steady state network equilibrium, is not applicable to surge analysis. At pipe junctions and dead ends, wave reflections and transmissions occur, which often magnify or attenuate the surge waves. Conducting a surge analysis with a skeletonized model may not be conservative and may not be suitable for estimating transient pressure extremes in a distribution network system. 

Two approaches are suggested to minimize the impact of the skeletonization on the surge analysis.  First, the interior nodes of all series pipes with same diameter can be eliminated. Please note that it doesn’t make any effect on the surge analysis as well as the steady state analysis. Second, a small dead end pipe (e.g., less than 10 m) can be trimmed. Trimming a dead end causes a difference on surge analysis; however, its effect of using a small pipe would be minor.

  • 20. What data is reported for a surge protection device?

The Surge Protection Device report includes the following:


Inlet Pressure: pressure at node

Outlet Pressure: pressure at surge protection device

Side 1 Flow: flow from upstream

Side 2 Flow: flow to downstream

External Flow: flow to surge protection device

External Volume: air volume of surge protection device (e.g., surge tank and air valve)



  • 21. How does InfoSurge handle non-TCV valves like pressure reducing valve or flow control valve?

InfoSurge calculates the local loss coefficient, k, based on the initial steady state condition and the valve is modeled as a local constant-loss component during transient analysis. This is based on the assumption that any valves are not sensitive enough to act on the change of head and flow in the short period of transient analysis. If a valve responds to the changes of head of flow in a sensitive way during the transient analysis, the user can model PRV and FCV as a Regulator Active Valve, and model other valve types as a TCV Active Valve with a corresponding valve operation change curve. 


  • 22. How does InfoSurge model vapor cavitation?

InfoSurge uses Discrete Vapor Cavity Model (DVCM) for simulating vapor cavitation. The DVCM is the most popular model used in transient analysis due to its simplicity and ease of implementation with reasonably reproducing many features of the physical characteristics of water column separation.


  • 23. I am getting very severe pressure spikes in my model. What is happening?

Very rapid high and low pressure swings in the simulation are usually an indication that cavitation is taking place during the simulation. If the pressure in the system drops to the vapor pressure of the liquid, the liquid changes state from the liquid phase to the gas phase. When the pressure in the system increases above the vapor pressure, the vapor returns to the liquid phase and resulting vapor cavity collapses. The liquid column moves very rapidly to close the cavity. When the vapor cavity collapses it causes a severe jump in pressure and the pressure spikes upward. If conditions in the morel are just right, the positive and negative pressure spikes can continue for a long period of time.



  • 24. Where do I look for cavitation problems?

Cavitation occurs at areas in the system experiencing very low pressure. Look for minimum junction pressures that are equal to the liquid vapor pressure. The liquid Cavitation Head (liquid vapor head) is defined in the Run Manager - Surge Tab. Cavitation head for water is -33.21 feet (-10.30 meters)


Check valves that are usually associated with pumps can also contribute to cavitation problems by rapidly opening and closing and contributing to the pressure spikes. If type of action occurs during a pump shut down at the check valve, the pump is operating in an abnormal fashion (possible flow reversal). In this case, a pump file is required and a pump trip should be modeled.

When this occurs in a model, the magnitude of the pressure changes are very sensitive to system data and a minor change can have major affect on pressures. It is important to realize that the response is volatile and unstable.

  • 25. What can I do to determine if cavitation is causing a problem in my model?

You can eliminate the effects of the cavitation to see if it makes a difference in the model solution. Go to the Run Manager - Surge Tab and change the Cavitation Head to a negative value that is well below the actual vapor head of the liquid.


Run the simulation again with the new Cavitation Head value. If there is a significant change in pressure values and fluctuations, the model is probably experiencing a cavitation problem.

In the case of cavitation caused by check valves, you can either remove the check valve from the model or set them to non-reopening so that they will not continue to open and close.