I have a study area of 400 km2 where there are around 180 villages spread (See picture attached). I am trying to determine the best location of a number of Wastewater treatment plants (WWTP) in a way that the distance from each village to the WWTP would be the minimum possible, assuming that the pipes will be installed right under the main roads that connect the villages between them.
I need to make the network analysis many times with the variation of number and location of the WWTP to determine the cheapest way in terms of number of WWTP and their location. The costss in this case are directly derived from the lenght of the pipes - which I assume is the same lenght of the roads under which the pipes would be installed.
I am using the Location - Allocation option in ArcGIS Pro, but I can not figure out how to set the options in order for the analysys not to consider the impedance but only the distance of the roads (because in my case the pipes and transport of wastewater will not be affected by the velocity on the roads, so I do not need to save time in that kind of transportation).
Could you please guide me? Am I using an incorrect tool? or maybe do I need to build a model to do the analysis that I require?
The main goal of my research is to define what is the best option for the villages in the study area: a wastewater treatment plant for each of the 180 villages? or how many villages I should connect to one WWTP in a way that the pipes would be as shortest ad possible, therefore reducing the costs associated with the pipes?
Thank you very much!! I really need a guidance on this matter.
Att: Sara Ramos
Are you doing the workflow manually at this point? You could reduce the amount of work you need to do by automating the procedure using Python or Model Builder. I don't know that this would reduce the overall time, but at least you could press start and then take a lunch break. Does this seem like something you might want to do?
Hello dear Melinda,
All this time I have been doing the process manually, so far I have chosen service areas for the potential facilities and after that, within the service areas of each facility I have used the Location-Allocation tool to choose the final best location for a WWTP. But as long as I am trying different service areas and alternatives this is an iterative process that is taking a lot of time, so what you mentioned about creating a model with Model Builder sounds promising.
However, I have been watching tutorials about Model Builder but I do not know how to use it with these Network Analysis tools. Do you maybe know about a tutorial that I can go through to learn how to do this?
In summary, I would need the model to choose n numbers of cuttofs to create service areas, and for each of them to run the location allocation. It is abour creating different scenarios to at the end choose the best.
I hope I was clear enough. Thank you so much for your contributions,
This tutorial describes how to create a Model using Network Analyst tools:
The example uses the Route solver, but the basic procedure would be the same for Service Area or Location-Allocation.
Thank you for the link. It worked and I created a model for the location allocation analysis and service areas!.
Now I have the best location of the facilities, but it´s location is shown with straight lines that point the facility that was chosen. Is there a way to show the actual route of each of the demand points towards the chosen facility? (instead of straight lines).
In the attached image, I have the best location for different facilties, and I need to show the actual path (on the roads network) from the demand points allocated to each facility, but can not find the way to show it in this way. Can you please guide me?
Thank you very much once again.
The Location-Allocation solver cannot display the actual paths taken through the network from Facilities to Demand Points, even though it uses the network paths under the hood. It's just a limitation of the solver and is done this way to improve performance.
If you want to display the actual paths taken for visualization purposes, you will need to use the Demand Points and chosen facilities as input stops for a Route analysis (or maybe Closest Facility) and solve that analysis.
If the chosen facilities are reliably the closest to each demand point, or you only have one chosen facility, Closest Facility is the easiest thing to do. Otherwise, you can use Route, and you must load a copy of the facility for each demand point. You need to use the RouteName field to designate a unique identifier for each Demand Point/Facility pair to ensure that a separate route is generated between each demand point and its associated facility.
I have defined now the wastewater treatment plants (WWTP) required as well as which demand points are allocated to each of them (please see the image attached as an example). As long as I am using this tool to define a sewerage network, I can see the distance of each pipeline from each demand point to the WWTP in the atribute table of the lines of the Location Allocation or the Closest Facility layer. However I found the following issues or doubts:
1.Once the pipelines (which I assume are the roads in my ArcGIS project) of two or more demand points converge, the volume of water will increase, meaning that the diameter of the pipeline might increase as well, and I need to know the exact lenght of every section from the demand point to each of the convergences until reaching the WWTP. Is there a way to find these lenghts of the sections, additional to the total distance from the demand point to the WWTP?
2. Right now, in the atribute table of the lines I get the total lenght from each demand point to the WWTP. However, if I sum them up I will be counting more than once some of the lenghts because the routes from different demand points have common paths. Do you know how can I avoid these issue? This is required because at the end I need to make a cost analysis and this is based strongly on the length of the pipelines.
3. And the last question, just to confirm if I understand properly. The location allocation solver, type of problem Minimize Weighted Impedance (P-Median), chooses the best location considering both the length of the road as well as the weigh of each demand point? I ask because I need the WWTP to be located in the closest place to all the demand points but also considering the population of each town, because the higher the population, the bigger he amount of wastewater produced and therefore, I would not like to transport a huge amoun of water but to locate the WWTP as close as possible to big towns.
Thank you so much and I hope I explained myself properly.
For Question 1, I think you can probably use the Copy Traversed Source Features tool. For a Route or Closest Facility layer, this tool will output a copy of the network edges, junctions, and turns that were traversed by the route. You can look for duplicate SourceOID field values to identify streets that are used by more than one route.
For Question 2: I think Copy Traversed Source Features can help you with that, too. You will just have to eliminate duplicated segments and then count the length of what's left.
For Question 3: I think that's correct, but I am going to ask a colleague to look at this post to confirm.
Thank you for your answer. I will try what you suggested.
In addition to this, while doing my analysis, I realised that a critical variable for me is the slope of the terrain, as ideally the sewerage network should be sloped so that the water flows by gravity. I should avoid as much as possible connecting points of demand with facilities that are against the slope.
Therefore, I have the following question: is it possible to include the slope within the costs of travel mode so that the impedance is based on both length and slope? I tried to include a new cost called "slope" with units as "other"; but I am not clear how to configure this cost to calculate the slope between a demand point and the facility and assign the slope and consider it in the impedance.
I have the DEM of the area and therefore I have the elevation, which can be assigned to the different demand points and facilities to consider the slope between them.
Thank you very much again for your suggestions.
Yes, this is feasible but challenging. I do not have any magical suggestions on how to do it successfully. You need to somehow calculate the slope for each street feature in your network's source feature class based on the DEM. This can be tricky because you could potentially have a street segment that goes down and back up a huge hill, and the average slope is 0.
Anyway, assuming you have calculated the slope in a way you're happy with, you can use the slope field in your source feature class to calculate a cost attribute using the slope value. Or perhaps you could make a descriptor for slope and calculate the cost as a function of the descriptor and some constant or parameter.
Sorry, this is a very vague answer. Your problem has become quite complex, and it's reaching the limit of what I'm able to advise upon here on Geonet.
I totally understand, and you have helped a lot! Is there an option to contact someone that could support further in this topic? or another platform that I could use? I know it is complex and therefore I am trying to find help because by myself can not manage yet.
Thank you very much!