Extreme negative pressures and hydraulic grades

Question

Hello, 
 I have imported a model into WaterGEMS and entered as much info as I have from the client. When I run the model, I am getting some very extreme negative pressures and hydraulic grades. 
 The areas of the model in question are at high points in the system. Where it is the worst (south end of the model), the hydraulic grades are in the -20,000's ft and pressures in the -10,000's psi. At the other problem area (north end of the model), the hydraulic grades are positive but 100-200 ft below elevation, and pressures are between 0 and -100 psi. Both of these areas of the model are each directly downstream of pump stations (i.e. two separate pump stations, one serving the north end of the system, and the other serving the south end of the system) which when I run the model, I get the following user notifications for both pumps: 
 
 Message ID 40003: "Pump exceeds the maximum operating point specified for the pump curve." 
 Message ID 40008: "Pump cannot deliver flow or head." 
 
 I have looked through many other discussion pages and Bentley help pages but nothing I try is remedying this issue. I would like to know what could be causing this, and what do I need to change in my model to fix the issue? 
 
 I have uploaded the model to the Bentley Sharefile system. Additionally, the screenshots below are of my model. The first screenshot is elevation contours, ranging from about 200' (red) to 600' (green). As you can see, the north and south ends of the system where I am having the issues are the high points of the system.

The second screenshot is of hydraulic grade contours, ranging from about -27,000' (red) to +500' (green). The south end of the model has a very extreme negative hydraulic grade (stemming from the south pump station), while the north end is also slightly in the negatives.

Any help and advice is much appreciated!

Jesse Dringoli · Answer

Hi Dominick, 
 I have taken a look at the new model revision and found additional problems contributing to the negative pressures. I have attempted to order these by importance: 
 
 There are very large demands entered on a number of hydrants in the model. They appear to be fire demands, but are entered as base demands and the model will attempt to satisfy them during the steady state. This results in large flows, causing large headlosses, leading to negative pressures. To see these elements, go to Components > Demand Center > Demand Control Center > Hydrants tab. Or, you can run a query on hydrants with demand > 0. As seen in the Calculation Summary, this results in a total system demand of over 47,000 gpm. Please review these hydrant demands and delete if needed. Note that even though the hydrant "status" field is "closed", these demands are still applied as the hydrant status only applies to the hydrant emitter field as explained in this article . You may also want to check the junctions as all of them have a demand of only 1 gpm. I have added a note about this article on the topic of negative pressure troubleshooting. 
 Some elements are still not connected. For example if you zoom into "South Tank" and drag/drop the adjacent junction, you will see the tank is not connected. You may want to review the source data for this model and consider cleaning it up and re-importing, or run additional network navigator queries such as "nodes in close proximity". 
 There are still numerous overlapping pipes. Run the Network Navigator query "Duplicate Pipes" under "Network Review" to identify and highlight them (my count was 72). 
 There are some pipe diameters that seem very small. For example pipes "wMain-543" and "wMain-282" are only two inches.

Jesse Dringoli · Answer

Hello Dominick, 
 At this point if you are unsure about the demands, it will be a calibration exercise. If you do not have demand information it may be very difficult for you to get the model results to match the field measurements. Generally speaking if you have some spatial information about your demands, you can use LoadBuilder to distribute them. Read more about LoadBuilder here: Troubleshooting and Understanding LoadBuilder 
 If you only have a lump sum demand amount and want to distribute it to nodes equally, you can divide it by the total number of junctions and use the Demand Control Center to initialize demands for all of those elements, as seen here: Distributing a lump sum demand or load over a large area 
 However, if the assumed demands and distribution in the model do not match how they are in the real system, it may be difficult to get other results like pressure to match field measurements. On top of that, there are a large number of other factors that can cause the model results to be different, such as pipe roughness and pump and valve status. WaterGEMS has a tool called Darwin Calibrator to help with this, but we recommend reviewing the material in the following article first: Water Model Calibration Tips 
 I did take a brief look at the area of the model you just uploaded and I have one observation about the pump station - the downstream area that has low pressure is due to a sudden increase in elevation. The elements around the pump station are around 460 ft whereas the downstream negative pressures are at junctions that are at an elevation of around 500 ft. The pump is not turning on because both the upstream and immediate downstream boundary are both around the same elevation. The upstream tank is at 468 ft and there is a hydropneumatic tank immediately downstream at an elevation of 462 ft. Since the downstream boundary HGL is at a lower elevation, the pump cannot turn on and the hydropneumatic tank HGL of 462 is not enough for the elements at 500 ft. If you intended for this hydropneumatic tank to "float" on the system HGL, select "true" for the "treat as junction" property. You will still also need to fix the other tank, because there is a path for water to loop back to it from the hydropneumatic tank. You may have intended to close pipe "P-101(1)" to prevent this. If you close that pipe and treat the hydropneumatic tank as a junction, negative pressures no longer occur. There is still the aforementioned problem of the many other factors that can cause the pressure to not match the field measurements.