A negative pressure means that the calculated hydraulic grade is below the physical elevation of the element. You should closely examine your node elevations and boundary conditions (reservoir/tank hydraulic grades) to make sure they are correct.
It is common for this to occur on the suction side of the pump, due to the elevation that you entered for the nearby reservoir. If the pump is off, then it is likely that the reservoir elevation is set to be lower than the pump elevation. Creating a profile of the area in question (physical elevations and hydraulic grades) should give you a good visual idea of what is going on.
Note that negative pressures will not prevent the model from computing - the messages that you get about these are just informational. WaterCAD/GEMS still assumes that the pipes are pressurized and no vapor pockets will form. It basically just computes the hydraulic grade based on the system conditions and other data input, and then reports the pressure as the difference in head between that and the physical elevation. If the negative pressures occur at high points in the system, you probably would have an air valve at that point so you can fill the pipeline. However, once the pipeline is running, only those air valves that would have a pressure below 0 need to be modeled as active air valves (Treat as Junction? = False). The others can be set to 'True'.
Start by tracing the negative pressure upstream to the supply of the flow. If the supply is from a pump examine the pump's output flow and pump head to see where it's running during the time of the negative pressure. Next check the pumps definition (pump curve) and see where that point is located on the curve. It could be the pump is too small to deliver the required amount of flow or head the system demands. If the supply is from a reservoir or tank make sure the hydraulic grade (elevation) is high enough to supply the downstream demands and overcome the head losses in the pipes.
After you've found the supply of the flow, and if the above information didn't help, follow the flow downstream and check each node until you locate where the negative pressure first starts. Once the first negative pressure is found determining the cause should be a matter of looking at the data entry and results in the element just upstream of it.
A negative pressure occurs when the hydraulic grade is below the physical elevation of a node. If WaterCAD/GEMS says the pressures will be negative, then in all likelyhood you will have problems. Assuming all data input has been checked, there are usually two general causes of negative pressure:
1. Trying to serve a customer at too high of an elevation. This will show up as low/negative pressure at any demand. You need to increase pump head or adjust pressure zone boundaries.
2. Some restriction in the system. This will show up as good pressure during low demand and poor pressure at high demand. You need to look at the model results and see if the pipes are too small causing excess head loss or the pumps are inadequate such that they are running far off to the right of the curve. You need to upsize the pipes or pumps accordingly.
You should also check your demands for errors. Since your demands are likely based on historic averages, a significant decrease in pressure may skew the results, since the demands would likely be decreased in that condition. You may consider using pressure dependent demands or flow emitters. You may also consider conducting a transient analysis using Bentley HAMMER, if the problem occurs at a transmission main.
Another example of demand-related problems causing negative pressure are hydrant / fire flows. Check your hydrants (or all tabs of the Demand Control Center and Unit Demand Control Center) to make sure that you have not accidentally included fire demands when you did not intend. Note that even if the hydrant's "status" field is set to closed, any demands entered in the demand collection or unit demand collection will still be applied, as the open/closed hydrant status only applies to the hydrant emitter coefficient field. See more here: How to use the "Hydrant Status" property in WaterGEMS and WaterCAD
This user notification is occurs if there is a negative pressure somewhere in the model that is also below the vapor pressure limit (-14 psi by default). In some cases small negative pressures may be acceptable, but pressures below the vapor pressure limit are physically impossible, warranting a separate notification. This single notification helps to reduce the number of negative pressure user notifications that would otherwise display for each node where it occurred. If you'd still like to see those user notifications you can set up custom alerts described in the wiki link below. Note that negative pressure can occur not just at junctions but also at valves, pumps and other elements.
By double clicking on the user notification the properties window for the calculation options will open and you can adjust the minimum possible pressure (default value -14) if desired. Changing this will change the pressure at which the user notification will be generated.
1. By creating an annotation based on pressures for your node elements.
2. By creating color coding based on pressures for your node elements.
3. Using a network navigator (View > Network Navigator) query for negative pressures (see the screen shot below)
4. Creating a custom alert for negative pressures.
5. By using the flextable to view the pressures sorted descending to ascending.
By default, pumps only consider the boundary conditions (reservoirs and tank elevations) in your system. So, the pump will add enough head to lift the water to the downstream known hydraulic grade. It does not consider junction elevations in between. If you are using WaterCAD or WaterGEMS V8i, you should add an Air Valve element at the high point to properly model this situation. By placing an air valve at the high point, the pump sees the air valve elevation as its downstream boundary condition for instances in which pressure would have otherwise been negative at the high point. For any air valve that is expected to be open in this way, ensure that you select "false" for the "Treat air valve as junction?" attribute. For more on this, see this technote.
Another alternative to resolving negative pressures to take into consideration is installing a PSV at the downstream end of the system with a hydraulic grade equal to the highest point in the system.
If you get a negative pressure at a high point that means the system is a siphon at that location. It is best to put an active air valve at that point (Treat as Junction? = False).
If the pressure drops below the vapor pressure, the siphon will not work. You definitely need to install an active air valve there (Treat as Junction? = False). This will also help provide protection against a vapor pocket collapse that would cause a transient to occur.
This commonly occurs in cases where the user models the pump station (or connection to an existing system) using a reservoir and pump, with short pipe between them. Typically, the same elevation is used for both the pump and reservoir nodes. Since the reservoir elevation defines the boundary hydraulic grade and since there will always be some amount of headloss through a pipe, this means that the hydraulic grade at the pump node location will be slightly below the physical elevation. The suction pressure of the pump is derived from the difference between the hydraulic grade and the physical elevation, so that is why the pressure ends up being negative.
You can simply ignore this informational message, but if you'd like to remove it, the solution is to simply raise the elevation of the reservoir by a small amount. Make sure the pipe has a very large diameter and smooth roughness coefficient, too (to minimize headlosses.)
If the upstream hydraulic grade or reservoir elevation is correct, then this negative pressure message is accurate. To better understand what is happening, create a profile of the hydraulic grade and physical elevation for this segment of piping - you will see that the hydraulic grade is below the pump elevation. If this is an existing system, you may need to check your NPSH to ensure that cavitation will not occur at the pump. You can also investigate ways to increase the hydraulic grade upstream of the pump. Also consider:
1) Margin of error - the model results may not exactly match real field conditions. You may have made an assumption for certain model parameters that may not necessarily reflect the real system, and if they are skewed a certain way, it could result in even lower pressures than predicted in the hydraulic model - unless you have calibrated the model.
2) Demand conditions - what demand conditions were you looking at? (average day demand, peak hour, etc) Higher demand conditions would result in lower pump suction pressure.
3) You should also consider transient conditions. For example, upon pump startup, if the pump starts too quickly, the resulting transient may cause a "downsurge" wave on the suction side, lowering the pressure further. If it drops below the vapor pressure limit of water, a vapor pocket can form and cause subsequent damaging effects when it collapses.