This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Load Builder for Demand Allocation question(s)

All,

I am tasked with taking a water model, performing hydrant flow tests, and calibrating said model.  I am looking for help in how to define the demands for the various buildings within.  Without further ado...

1)  It seems to me that the best option is through assigning residential demand to the nearest pipe.  I created Thiessen Polygons and, uponvisual inspection, these are simply too inaccurate in the distal regions of the network.  I then transitioned to using the load builder to assign homes to the nearest pipe.  For model pipe data, I am prompted to enter a method for load assignment.  The options are: "equal distribution, distance weighted, closest node, farthest node".  Our goal for a worst case scenario is to assign all usage to the dwonstream node, but this doesn't appear to be an option.  Which of the available four options would be best?

Coupled with the above question, I am also prompted to choose a polyline distribution for the billing meter layer, with the following options: "equal distribution, proportional distribution".  I'm really not sure what the choice even means.  What is a polyline distribution?\

2)  My last question is a little more broad-viewed.  In all, we have geocoded data for 5 different customer types: residential, commercial, industrial, government, and hospital.  My thoughts are to assign:

residential usage by an average household demand

commercial usage by square footage

industrial the same as commercial

government will be case-by-case

hospital by number of beds

Does this seem like a logical, efficient way to accurately assign the various demands?  Are there sample projects/designs available as examples? 

Thanks for any help you may provide,

Marshall

  • Hello Marshall,

    1) There is information in the Help documentation and at the following link on the different LoadBuilder methods:

    communities.bentley.com/.../10862.how-do-each-of-the-loadbuilder-methods-work

    Included are the definitions for the items in the load assignment field. As for the best method, it will depend greatly on the placement of the meters. One thing that you can do is to add the shapefile with the loading data as a background file. This will give you some idea on the placement of the meters in relation to the pipes in the model. Then you can choose the option that makes the most sense. I will note that the "downstream" node is a bit relative in a pressure pipe, since if the hydraulic grade in the system is higher at the stop node of a pipe, you can have flow moving toward the start node.

    For the polyline distrution, information on this field is also in the Help documentation. When a polyline meter layer is selected, this field will be activated. When multiple pipes are associated with (overlapped by) a polyline meter, the option chosen in this field determines the method that will be used to divide the polyline meter load among them. For Equal Distribution, this will distribute the load equally among the pipes associated with (overlapping) the meter. For Proportional Distribution, this will divide the load proportionally according to the ratio of the length of pipe that is associated with (overlapping) the meter to the total length of the meter.

    2) The method used to distribute the demand data is generally up to your engineering judgment. That said, there may be other users on the forum that can relate through own experience on the methodology they used in their modeling cases.

    Unfortunately, there are no Quick Start Lessons or sample files related to LoadBuilder at this time. But the Bentley colleagues in the forum can assist with issues or questions you come across in the use of the tool.

    Regards,
    Scott

    Answer Verified By: marshall grossell 

  • Thanks for the help Scott. I'm still unclear as to the definition of "polyline meter".
    Also, I was wondering if you have knowledge of any research that has been done on the different Load Assignment options when using the "Nearest Pipe" allocation method. Specifically, anything that discusses the implications of : Equal Distribution vs. Distance Weighted vs. Closest Node, etc.
    Thanks,Marshall
  • Hello Marshall,

    What part of the polyline meter definition is unclear? The only time polyline meter allocation is used is when this field is activated. If it is, the data in the layer is allocated based on the two available methods: Equal Distribution or Proportional Distribution. The definitions for these two methods were provided in my earlier post.

    With the nearest pipe method, there is no documented research, but the developers go through rigorous testing during the development process so you can be assured that it is working properly. If you have any questions or concerns on a specific model, we can look into an issue.

    Definitions for the difference "Nearest Pipe" methods can be found in the Help documentation. "Equal Distribution" assigns an equal portion of the total load assigned to a pipe to each of the pipe's end nodes. "Distance Weighted" assigns a portion of the total load assigned to a pipe based on the distance between the meter(s) and the nodes at the pipe ends. The closer a meter is to the node at the end of the pipe, the more load will be assigned to it. "Closest Node" assigns the entire total load assigned to the pipe end node that is closest to the meter.

    Regards,
    Scott
  • Scott,
    There is no definition given for 'polyline meter' in the manual, that is where my lack of clarity lies. My using this software coincides with the first time I've ever seen the term.
    Regarding the Load Assignment options, I've seen the help file, and I certainly assume that each algorithm is tested thoroughly by the developers. All four are approximations and, as such, all four introduce errors. My question was in regards to advice for when to use one method versus the other. Is one better for sparse vs dense neighborhoods, etc. I'm assuming from your answer that there is no information relative to their actual selection and usage.
    At any rate, thanks again for you time,Marshall
  • Hi Marshall - when we say 'polyline meter' we mean meters that are repesented in your GIS data as polylines (i.e. a line of one or more segments - en.wiktionary.org/.../polyline ).

    Normally a meter is represented as a point in GIS data, so it's pretty easy to figure out from that what the nearest pipe is. Things are a bit more complicated if the meters are represented as polylines, so we need some additional information. Hope that helps clear things up.

    Regarding the various LoadBuilder methods, I haven't done a detailed study of it but I would tend towards the 'Distance Weighted' method in most cases. It doesn't matter so much if all the demand are small (e.g. a typical house) and the network has a lot of loops, but if you have one or two large demands and long branched pipe networks, I prefer to assign some of that demand to each end of the pipe.

    One thing you may also wish to consider is adding 'Customer Meter' elements in the model to represent each connection in your system (or a subset - e.g. the very large users). Then you can explicitly select the junction that each meter is assigned to.

    Regards,

    Mal

          

    Answer Verified By: marshall grossell