Relationship between hydraulic grade and head

Question

I'm modeling a pressurized wastewater system in SewerGems v8i. I am treating the design-phase pump station that begins the pressure section simply as a Pressure Node. The Maximum Hydraulic Grade at this start node (132 feet above msl) remains constant when I change only the elevation of the node and re-compute (GVF complex). The high pipe elevation of the system is midway and 70 feet above msl. Can this unchanging Max Hydraulic Grade be understood in light of the the fact that the static head is less with a proposed higher start node elevation (i.e. 70-10 = 60) and more with a proposed lower start node elevation (70-5 = 65)? It in an oversimplification to say Head Required = Maximum Hydraulic Grade - System Start Elevation?

Scott Kampa · Answer

Hello Philip, 
 We may need some additional information on what you are modeling. First, pressure subnetworks typically start with a wet well. If you don't have one, one will likely be needed. One of the SewerGEMS sample files will include a setup that includes a wet well. 
 The maximum hydraulic grade is simply the highest hydraulic grade value reported during the calculation. There may be a lot of factors to take into consideration for how this may change, but it will be related to the pump head and the elevation of the wet well on the suction side of the pump. 
 I'm not exactly sure what the "head required" you mention is in reference to. Is it related to system head curves ? 
 Regards, 
 Scott

Yashodhan Joshi · Answer

Hello Philip, 
 What is upstream of this "pressure node"? How do you get an HGL over here? Is there a pump upstream? 
 The hydraulic grade at a point is the sum of pressure and datum head. The hydraulic grade won't change if you change the elevation in a pressurized system because your boundary (maybe an upstream pump) is not changing. If the pump upstream is supplying at a constant head the HGL at downstream points won't change much. 
 Philip Kunkle said: Can this unchanging Max Hydraulic Grade be understood in light of the the fact that the static head is less with a proposed higher start node elevation (i.e. 70-10 = 60) and more with a proposed lower start node elevation (70-5 = 65)? 
 The static head is usually the difference of elevation from your point of delivery (pump discharge, tank outlet) and the elevation of your discharge point. The statement you have proposed is obviously true but I am not getting the point of it. 
 Philip Kunkle said: It in an oversimplification to say Head Required = Maximum Hydraulic Grade - System Start Elevation? 
 Yes. Because head required depends on what head are you computing. Is it a system head as Scott suggested? Is it pump head that you are determining for an upstream pump? Also, head requirement does not depend only on static head. It depends also on frictional headloss and minor losses. See the below article for details on determining pump curve for a system; 
 Estimating a pump curve for a model 
 If you can elaborate more on what you are trying to model we can help you better.

Jesse Dringoli · Answer

Hi Philip, I am not quite clear on some of the things you mentioned. For example when you say that you have a "high pipe invert elevation" 70 feet above sea level, do you mean that you have a local high point that the pump needs to overcome? This should be automatically accounted for in SewerGEMS and the pump should add enough head to overcome it, or you can use an air valve with the GVF-Convex (SewerCAD) solver as mentioned here . 
 However, if you are saying at the high point is at 70 feet and the calculated HGL is 132 feet, then it suggests that the pumps are adding enough head to keep this high point pressurized, in which case changing the physical elevation of the pressure junction would indeed not impact the hydraulic grade, only the pressure (as long as the node elevation is kept below the HGL), as the pump operating point will be based on the system head characteristics based on boundary conditions (upstream wetwell elevation and downstream gravity discharge point). See more in the system head curve articles previously provided. 
 
 If this does not help, please provide a copy of the model for review, and specific steps for us to follow with that model open, so we can see what you are looking at and better understand your concern. 
 As a side note, I noticed you mentioned that you are using SewerGEMS V8i. We recommend upgrading to the CONNECT Edition which also uses updated licensing, as Support for V8i will be discontinued (and we've made many enhancements that you can take advantage of in the CONNECT Edition.) See this and this .

Scott Kampa · Answer

Hello Philip, 
 The best way to see how the pump will operate would be to include a pump of a given design. Loading data alone will not include head in the system, just the flow. 
 You could try to use system head curves to help with this. 
 Regards, 
 Scott

Jesse Dringoli · Answer

Hi Phlip, 
 I have taken a look at the new model revision you have uploaded. I see that you are using the GVF-Convex (SewerCAD) solver with a steady state simulation and you have a pressure network starting at a wetwell at 1.0 ft and ending at an outfall at -3.39 ft. You also have several air valves between, with "treat as junction" set to "false". This means that the upstream boundary condition that the pump "sees" is 1.0 ft (wetwell HGL) and the downstream boundary is either the outfall at -3.39 ft, or one of the air valve elevations, of negative pressure occurs. From the profile view we can see that if the pump only considered the downstream outfall as a boundary condition (to "lift" the water to), then most of the network would have a negative pressure. Because of the air valves, the pump knows to add more energy to "lift" the water over the high point, thereby the downstream boundary HGL is the air valve elevation ~62 ft. You can see some of the downstream portion of the network is operating part-full as explained in this article . 
 However, the flow generated in this model is not from the pump, as the pump's initial status is "off". It is from a sanitary load of 3.09 CFS which you have entered on junction "Woolwich Connection". I assume that you want this additional tributary inflow in the calculation of the pump operating point. In this case, a pump (in the location where you placed the pump element) would need to add enough head to "lift" the water from the upstream boundary hydraulic grade (the wetwell HGL of 1.0 ft), to the existing HGL already in the system, due to the tributary inflow. which is around 88.35 ft as measured at the discharge side of the pump. So, the zero flow point for the pump to turn on would be 88.35 - 1 = 87.35 ft. 
 Now, if you want to see what this "static head" would be if the wetwell elevation were lowered by ten feet to an elevation of -9 ft, then that would simply be 88.35 - -9 = 97.35. The downstream HGL would remain 88.35 regardless of the upstream wetwell elevation because that 88.35 ft is generated from the fixed inflow at junction "Woolwich Connection". So, the pump would indeed need to add 10 feet more of head at the static/zero flow point, and you would see that in the pump properties by subtracting the "Hydraulic Grade (upstream)" from the "Hydraulic grade (downstream)". 
 As the pump adds flow to the system, that would raise the downstream HGL because of the additional friction losses from the increased flow, and the pump would need to add additional "dynamic head", but you would not see that in the model with the pump off since the flow is zero (and therefore no impact on the downstream HGL that is already present). You would need to use the System Head Curve tool to see this overall curve that illustrates the change in required head as the flow increases, and the pump operating point where it intersects with the pump curve. This will also reflect the increase in static head from changes to things like the upstream wetwell elevation, as it looks at the *difference* in HGL between the upstream and downstream side. However in this model, the system head curve tool may not generate a valid curve because of challenges related to the air valves and extra inflow. (and I believe the reverse orientation of the pipes adjacent to some of your air valves will also impact this). 
 As a side note, I noticed the model you sent was last saved in an old version, 08.11.05.58. If this is indeed the version you're using, I highly recommend upgrading to the latest CONNECT Edition, per our Support Policy and this article .

Jesse Dringoli · Answer

Hi Philip, 
 Since it sounds like you are just trying to size the pump, it is OK that the pump element is initially off, as the system head curve is independent from the pump. It just needs to be in place with an initial pump curve estimate, so that you can right click on it and choose the System Head Curve option. 
 The System Head curve then enables you to visualize the TDH as it will plot the required head needed to overcome a range of flow. It also shows where the pump operating point would be so that you can select a pump curve that meets your needs. This article has information on the general pump selection process. 
 In a sewer context, the following two articles may also be helpful, as well as our wastewater book : 
 Generating a System Head Curve in SewerGEMS 
 Unexpected system head curve in SewerCAD or SewerGEMS 
 As noted in my previous reply, I had some trouble generating the system head curve in your model, due to the configuration of the air valves and the adjacent pipe orientation. If you are also seeing an unexpected system head curve in the older version that you're using, please let us know.

Jesse Dringoli · Answer

Philip Kunkle said: 1 - The system head curve is dependent on the design flow and the design head that must be entered in the pump definition field, correct? (the intersection point changes) Am I getting away from an objective calculated value for TDH by providing an "initial pump curve estimate"? 
 No the system head curve is independent from the pump - it plots the required total head to overcome a range of flows. See the articles previously provided for more background on how system head curves work and how they can be used to select a pump. 
 Philip Kunkle said: 2 - the flow (maximum, 6.54 cfs) is greater than the inflow (2.85 + 3.10 = 5.95 cfs) What does this mean? 
 Can you clarify what this refers to? Are you seeing a user notification? (if so, please provide the details and steps to reproduce) If you mean that you are seeing that the maximum value in the system head curve is greater than the tributary inflow, that is because they are two separate things. The flow on the system head curve is the flow forced through that location (the point where the pump node was placed) and the resulting head needed to overcome that. So, if you are seeing a value of 6.54 cfs on the system head curve, that means that if 6.54 cfs was flowing through the pump (and then adding to the other tributary flow downstream), the head on the system head curve corresponding to that point is the head that would be required for a pump to overcome the static head and dynamic head losses at that flow. If this does not help. please elaborate on what you are seeing. 
 Philip Kunkle said: 3 - I am concerned by your (twice) comments about problematic air valves, preventing computation. While trying different relative inflows from the two inflow points, I sometimes get an 'air valve disconnection' error that seems to flow specific, but not necessarily from super large flows. Should I NOT be calling Treat Air Valve as Junction = False? The original intent of having the air valves in the model was quantify the various high elevations throughout the system. There is a degree of pipe elevation detail, specifically the lowest points in the design, that are not in the SewerGEMS model because we believed it served no purpose. 
 Choosing "false" for "treat air valve as junction" causes the air valve to be treated as a Pressure Sustaining Valve (PSV) behind the scenes, forcing the upstream pressure to be at least zero. Setting this to false for all air valves thus adds extra complication to the model, especially at zero or near-zero flows. Per this article , it is recommended that you start by choosing "true" for all air valves except for the one at the high point that you believe is likely to be open, compute the model, check the profile and enable additional air valves as needed.

Jesse Dringoli · Answer

Philip Kunkle said: RE: #1, when I change the inflow at the downstream Woolwich Connection, which has an effect on the hydraulic grade line throughout the system, should doing so have an effect on the system head curve at the upstream pump? A large inflow at Woolwich doesn't change the flow at the upstream pump, however I would expect it (a large flow at Woolwich) to make it more difficult to move the same volume at the upstream pump. 
 Yes, if the downstream system characteristics are changing, that would impact the system head curve if those changes result in a different head on the downstream side. The system head curve is simply forcing a flow through the pump location and displaying the resulting head difference - it keeps the rest of the model the way it is, so that when that flow is forced through, those other conditions are also happening at the same time, such as your additional inflow. 
 Philip Kunkle said: RE: #2, I am referring to the Flow (maximum) in the Detailed Calculation Summary, downstream of the Woolwich Connection, being 6.54 cfs, which is greater than the sum of the 2.85 cfs inflow into the wet well upstream of the pump and the 3.10 inflow at the Woolwich connection. The comment/question is independent from discussions about the system head curve. 
 In the model you had sent, there is only the flow of 3.1 CFS from the "Woolwich connection" as the pump is off. The Pipe Report section of the Calculation Summary shows 3.1 CFS for the max flow for me. Plus, the wetwell inflow is independent of the pump flow, so even with the pump on, unless the pump flow is exactly equal to the wetwell inflow, the pump flow would be greater than 2.85 and thus the combined flow would be greater than the sum of the connection inflow and wetwell inflow. If this does not help, please provide the latest version of the model, with steps to reproduce the problem in question. 
 Philip Kunkle said: One way to get the pump to move water at the same rate as the water is entering the upstream wet well (2.85 cfs) is to try combinations of design flow and design head in Pump Definitions until the pipe immediately downstream of the pump is flowing with 2.85 cfs. Is there a better way to achieve this (ensure rate of water entering/exiting wet well are equal)? 
 The System Head Curve tool can be used to select a pump whose pump definition intersects the system head curve at the desired flow. Why do you want the pump flow to be equal to the wetwell inflow though? Typically you would run a steady state for peak flow design, and an EPS to account for the pump turning on and off as the wetwell fills and drains. Since the pump flow would be higher than the wetwell inflow, you would have the pump turn off when the wetwell is near the bottom, then turn back on when it gets near the top. You could certainly replace the pump and wetwell with an inflow of 2.85 CFS (similar to what you did with J-12 for the "Woolwich Connection" - you may need to choose "true" for "treat air valve as junction" for the nearby air valves) but I am not clear on why you need to do this and what your ultimately need the model to help you decide.

Jesse Dringoli · Answer

Hi Philip, 
 I looked at the "4C" model revision (using the current version of SewerGEMS) and the pump flow is 3.17 cfs. I see a downstream flow of 6.17 cfs, which accounts for the woolwich connection inflow of 3.10 cfs. 
 I'm not quite certain from your response but I think you are saying that some steps need to be taken in this model to adjust the air valve settings in order to reproduce the issue with the flow not adding up - is that right? The model you sent only has "Sta 205+02 Air Valve" with "treat as junction" set to "False", but that air valve is well below the HGL. If I set that one to True, and set "Sta 489+10 Air Valve" as "False", the profile looks good and the pump flow is then 2.85 cfs, so I assume this is what you had done. In this case, the downstream flow changes to 5.95 cfs, which is the sum of the 2.85 pump flow and the 3.10 connection. Are you seeing something else, with that same air valve not treated as a junction? (and all others as junctions) If so, then it could be related to your older version. I examined the model with the latest version (10.03.01.08) whereas your model file suggests that you are using version 08.11.05.58. Upgrading is strongly recommended either way. (see this and this )

Jesse Dringoli · Answer

Philip Kunkle said: How do we solve for TDH, at the pump, given the 2.85 cfs + 3.10 cfs = 5.95 total cfs flows??? 
 This article has some guidance but basically you would find the difference in elevation between the upstream wetwell and downstream boundary condition (which you might consider to be the high point), then subtract that "static head" from the pump head (since the pump head is the total head that the pump is adding to overcome both the static and dynamic head).

Jesse Dringoli · Answer

Philip Kunkle said: In the most recent model I uploaded yesterday, WHAT IS THE TDH, at the pump, given the 2.85 cfs + 3.10 cfs = 5.95 total cfs flows in that specific set of files I provided? 
 With the air valve "Sta 489+10 Air Valve" activated and the pump flow of 2.85 CFS, the total head that the pump needs to add as 166.08 ft as seen in the "pump head" property. This 166.08 feet is made up of a static lift of 55.00 ft (elevation at air valve) minus 1.00 ft (upstream wetwell hydraulic grade) = 54.00 ft and the remainder is the dynamic lift (headlosses) of 166.08 - 54.00 = 112.08 ft. 
 If you wanted to simply assume this 2.85 cfs pump flow along with the 3.1 cfs connection flow, you could also see the same number by modeling the 2.85 cfs as an inflow on a junction instead of the wetwell+pump, then look at the calculated hydraulic grade at the inflow junction placed on the pump discharge side location. (and compare to an assumed upstream hydraulic grade). 
 Lastly, you could also see this by using any pump definition, generate a system head curve and look at the 2.85 cfs flow point on the system head curve (Data tab) to see the corresponding head. 
 Philip Kunkle said: Also, trying to get at the heart of the matter (from our POV), the sum of the calculated headloss values for pipe sections that are under pressure with 5.95 total CFS = ~112 feet. We have 71 feet of static head (between pump (-1) and pipe high point (70)). We are expecting a head at the pump of 112+71 = 183 feet. However both attempted model methods of introducing upstream flow (pressure junction or wet well/pump) show 167 feet of head. 
 See further above for my calculation. For your static head calculation, you assumed an upstream boundary HGL of -1 ft, but the wetwell HGL is -7 (base elevation) plus 8 (initial level) = 1.0 ft. You had assumed 70 ft as the downstream boundary HGL, but that is the elevation of an air valve further upstream. If you only enable that air valve, then part of the downstream system will have a negative pressure, and the pump will "see" the downstream outfall as the boundary since the air valve at 70 ft would have a positive pressure. As seen in profile view, it appears that the air valve further downstream (Sta 489+10 Air Valve) at an elevation of 55 ft is the one that would be open with part-full downstream flow and therefore acting as your downstream boundary condition (the pump is "lifting" the water up to that point). Compare profile view to visualize this.

Philip Kunkle said: Why is the system head curve at the pump independent of the downstream inflow at the Woolwich Connection? 
 The system head curve would be dependent on the system conditions including inflows like the "woolwich connection". If you remove that inflow for example and re-run the system head curve, you will see it change. The flow from this inflow causes additional headlosses, impacting the pump discharge head, thus impacting the system head curve (independent from the pump curve).

Jesse Dringoli · Answer

Philip Kunkle said: 1 - I see the pump discharge head value (in Pump Properties) change when I edit the downstream inflow value at the Woolwich Connection and re-compute, but I the system head curve does not change at all. We are in total agreement that it *should*, I'm just not seeing it. Could you attempt on the most recently uploaded file (for instance, double the Woolwich inflow from 3.1 cfs to 6.2 cfs) and post images in your reply. 
 Using the latest version of SewerGEMS (10.03.01.08), model file name "2010-10-27 Force Main Modeling Run #4C" uploaded October 27th at 2:18 PM Eastern: 
 Original, as-is: 
 
 Woolwich connection inflow doubled to 6.2 cfs: 
 
 This looks reasonable to me; more head is required (blue system head curve line) for all flows due to the additional headloss from the increased downstream inflow. 
 If you are seeing something different, I would recommend upgrading to the latest version in case you have encountered a problem that has been fixed in later versions. As mentioned earlier, you appear to be using an old version which will soon be unsupported. 
 Philip Kunkle said: 2 - I experimented with different pump design flows until I found a design flow (2.32 cfs) that allowed the fixed wet well inflow (2.85 cfs) to match the "externally imposed" flow into the system (2.85 cfs). Is there any computational downside or later consequence for having arrived at the 2.32 cfs design flow by trial-and-error? As I have said throughout this process, I'm trying to minimize data inputs to only what is necessary to arrive at an accurate pump discharge head. 
 As discussed earlier, if you only need to know the discharge head to overcome an assumed inflow (2.85 cfs + 3.1 cfs), the pump node is not technically necessary. I had outlined a few options earlier. 
 If you have used a pump in the model and used trial-and-error to determine the pump curve necessary to achieve desired pump performance, that is probably OK too (you would need to use your engineering judgement here), as long as a manufacturer can actually provide pump with the same performance. Typically you would use the system head curve to overlay a range of available pumps from a catalog and select based on which one intersects at the desired operating point. 
 Additionally (and it is possible I still be misunderstanding what you are trying to do), typically the pump flow will be much greater than the wetwell inflow, because controls will turn the pump on and off as the wetwell fills and drains.