Model Build the Smart Way

- By:

Courtesy of Courtesy of Innovyze

Wallingford Software’s InfoNet is used widely now as a data repository for collection systems and water supply systems, with a number of applications for managing, accessing, and displaying that data. However, InfoNet has other uses, including preparing input for model building, specifically the building of collection system models. Richard Allitt Associates (RAA) has developed a specific process for input from InfoNet to InfoWorks CS models, and it is now central to their model building.

Step 1 – Preparing asset data for the model

The first step in the model build is to obtain the sewer records, which may be on paper, electronic, or of various other sorts and sources, to provide the asset data that is the basis for the model. This asset data consists of physical items such as manholes, sewers, combined sewer overflows, pumping stations and storage tanks. 

Asset data usually requires checking and cleaning, including connectivity checks, and checking that sewers flow downhill and are connected to the right manholes. System type should also be checked – it is important that foul is going to foul, not to surface water somewhere else, unless it is an overflow. Another common error is duplicate data entries - perhaps two sewers on top of each other - one with 95% of its data complete, and the other 5% complete. The modeler must establish which record is correct, which data is correct, and whether data in both records is correct but different. Cleaning data involves putting it together correctly and then deleting the incorrect duplicate sewer or manhole.

The next stage with asset data and data clean-up is identifying missing data, and filling in the gaps using inference, interpolation, estimation (much of which InfoNet can undertake) and engineering judgment. InfoNet has many features, some fixed and some user-defined rules, to assist extensively in identifying and correcting suspect data. However, key manholes may need to be surveyed, if that is possible, rather than using the less accurate inferencing.

Step 2 - Surface data for the model

Once the asset data in InfoNet is complete, but not necessarily cleaned, the initial version of the underground model can be built as long as all or most of the nodes to be modeled are included. The next stage of the model building process is to define and input the surface , or above ground data. This typically includes defining the contributing areas, adding populations, trade flows, impermeable areas, run off surfaces and land uses. These are all aspects of data that show how flow reaches the sewers, through run-off over the ground. The population data enables calculation of dry weather flows, and the contributing areas show how each area is connected to each node and where run-off comes into the system. Whilst the surface data is being added to the model in InfoWorks CS the data clean up in InfoNet can continue.

The use of InfoNet to shorten project durations

In the past, the process of model build at RAA was a single stream process. First sewer records were imported into the modeling software- usually InfoWorks CS- followed by defining the surface data. The resulting model would then be verified with a flow survey, resulting in a working, calibrated, verified model.

The problem with such a process was that only one person can work on the data at any one time, because of the check out and check in process on an individual model. That meant the elapsed time of this phase of a project could not be reduced by increased staffing using solely InfoWorks CS. However, by introducing InfoNet to the project this constraint is relaxed.

This can be exemplified using a hypothetical project, with various tasks such as data clean-up, flow survey, verification and reporting. Using a modeling package only, the elapsed time of this sample project is 26 weeks to completion, with resourcing small but constant over the timescale.

Single stream Project Plan with elapsed time of 26 weeks

With the introduction of InfoNet, it is possible to use a parallel stream process on this model build. Its use brings a few extra very small tasks into the project plan, but there is a big change in the critical path and the elapsed time. This project now has a timescale of 19 weeks, by using more resources through the parallel working phases. The savings on elapsed time are not consistent across all projects, but the ability to work in parallel always gives real reductions in elapsed times.

Parallel stream Project Plan with elapsed time of 19 weeks

It is important to note here though that the two projects involve the same number of hours of work. There is no saving on work, or on resource costs. However, it is a more effective way of working, allowing RAA to better manage and balance resources.

Comparing the critical paths of the two project approaches, different work segments appear on the critical paths. With the parallel stream process the flow survey becomes a critical path work element, (this example has a five week flow survey). However, the single stream process has ‘define above ground data’ on the critical path, whereas it is not on the critical path in the parallel approach. This is the key reason for RAA’s use of InfoNet in model building. This task can be started earlier, allowing more efficient management of the project.

The InfoNet/InfoWorks link

The asset data was taken into InfoNet from sewer records or a GIS package. The stage 1 transfer into InfoWorks involves just the nodes. This allows simultaneous working of as many people with InfoWorks and InfoNet as there are licenses for working in parallel, with data clean-up of the other asset data continuing in InfoNet. At the same time, another team member works on the above ground data.

As the above ground contributing area data is defined, it is used to populate the model. Because the nodes are already defined, the contributing areas can be linked to nodes. As more cleaned asset data, such as conduits, becomes available in InfoNet, it is moved across to updated InfoWorks, providing an increasingly populated model over time. There is a pattern of periodic transfer from InfoNet to InfoWorks – every week, or more or less frequently as the project demands it.

InfoNet as an aid in pruning, merging, and roughness factors

The next stage of pruning and merging is very important. This depends on the client’s requirements. Some clients require all pipe models, needing no pruning and merging. Other clients just want a Type II DAP model (a UK specification from WaPUG), which needs some pruning and merging to speed up computational analysis, or even to bring it under the license limit of the software.

In addressing the issue of pruning, InfoNet can indicate which assets are critical. It is possible to bring into InfoNet details of flooding and other incidents, and thus to identify that a specific manhole, or perhaps sewer, might be really critical because it has a certain number of reported incidents associated with it. That can be flagged in InfoNet as critical and an essential component in the model. This is denoted in a user field in InfoNet that can be brought into InfoWorks and noted during pruning and merging.

Another benefit of using InfoNet is its data on asset condition. If an InfoNet database includes all CCTV data, it will indicate where infiltration occurs in the catchment, where fractures are, and where there is debris or roots. This is helpful information for the verification process. Knowing the location and volume of debris, and the sediment levels, roughness factors- and therefore the model itself- will be more accurate.


In summary, model building at RAA is now a parallel process. This allows much better and more efficient working, with better resourcing and resource balancing, targeting when to put more staff onto the project to give the maximum impact. By using this parallel process, office based tasks have generally been removed from the critical path, and fieldwork - flow surveys and manhole surveys – are becoming critical. This is not always going to be the case, but is an observed trend that will have a beneficial effect on the modeling industry.

This paper was presented by Martin Allitt of Richard Allitt Associates Ltd at Wallingford Software’s 2006 International User Conference

Customer comments

No comments were found for Model Build the Smart Way. Be the first to comment!