Hydraulic Modelling Software

[Eliecer Brathwaite]

Hydraulicmodeling is a process in which a pipe network is modeled using physical attributes and equations. The network can be any type of network that transfers liquid or gas in pipes or open channels. In a hydraulic model, the medium is transferred via pressure difference or gravity in the network. In this article, mainly pressurized water networks are handled. Pressure difference transfers the medium according to the path of lowest resistance. Hydraulic models are used to study and analyze system behavior now and in the near future. In the below picture, a simple hydraulic model is depicted.

In a hydraulic model, all the components are converted into digital forms of points and lines. Typically pipes, pumps, and valves are line types of components that have a start point and an endpoint. Consumers, water towers, and water sources are depicted as point components. All the point and line components have x,y,z information so they can be placed over any common map type.

In hydraulic models, there is one component that you cannot find in real networks that is a node or junction component. It is an imaginary component that is in the end and beginning of every pipeline. Junctions are needed to solve the mathematical graph model of system flows. Graph models used in hydraulic models are deterministic, meaning that they yield the same results with the same input parameters. In the below pictures, there are a few results depicted: minimum pressures for junctions and maximum flows for pipes.

Building a hydraulic model can be time-consuming and tedious work if it is done for the first time. The first and foremost information to add to the model is network information. Typically the information resides in a network information system or GIS system. There you can usually find the following things:

A very important thing to understand about network data is, that all the pipes must have data model connections to each other. In other words, all the pipes must know the pipe before and pipe after. There cannot be pipes dangling without connection to nearby pipes. There are automatic methods to enhance the network data if you suspect it to be poor quality.

The next important system to dig information into the hydraulic model is the customer information system. This information can also be in the network information system or the GIS system. The following information is needed from customers to be input into the hydraulic model:

The driving force in hydraulic models, especially water and district energy models, are pumps, controls, tanks, and other operational units. They bring life into the model by moving the medium in the pipelines. The key software component in understanding how any specific network system works is the SCADA(Supervisory control and data acquisition). SCADA systems can be very complex depending on the system size. In a typical medium-sized (200 000 Inhabitat) network you would find dozens of pumping stations, a handful of water towers, and water sources. Digging into the SCADA system and its history database can be the most time-consuming phase of model building. From the SCADA system we need to find out the following data:

For the control parameters, we need to figure out whether the pumpings are pressure or flow-controlled or do we have some type of flow-compensated pressure control or vice-versa. The simplest control for a pumping station would be fixed pressure or fixed flow setting, but there will most certainly be more complex controls in place as well. If there are free-floating water towers in the system, the pumps have also shutdown limits due to water tower levels. Another important aspect that we can get out of the SCADA is flow and pressure history data. All the networks have a small amount of leakage and we can use the history data to calculate base leakage to areas when we know all the inputs, the water consumption, and all the outputs of the area. We can also use the history data to model the water consumption pattern for each area. Typical water consumption patterns look like the picture below:

This publication is licensed under the terms of the Open Government Licence v3.0 except where otherwise stated. To view this licence, visit nationalarchives.gov.uk/doc/open-government-licence/version/3 or write to the Information Policy Team, The National Archives, Kew, London TW9 4DU, or email: p...@nationalarchives.gov.uk.

This document is part of the flood modelling standards for river systems. There are 4 more documents that cover this topic. Read all the documents to make sure you have the information you need to start your modelling project.

After establishing suitable inflow boundary conditions, you need to apply them to hydraulic models. The hydraulic model represents pathways and receptors of a fluvial, tidal, open coast or pipe system.

This guide gives best practice approaches for aspects of hydraulic modelling that cannot be categorised as pathways or receptors. This document assumes some existing knowledge of hydrological and hydraulic modelling.

The Fluvial Design Guide (FDG) (2009) encourages you not to carry out detailed analysis unnecessarily. You should challenge the level of detail and the validity of results throughout the modelling process, so you know when further analysis is needed.

You can change your proposed approaches during the project. You should not consider this a failure of the method statement, as a partially developed model can identify flood characteristics that were not evident before modelling. You should agree changes to the approved approaches with technical reviewers and project managers before you submit the model for final review.

Three dimensional (3D) models are available but are more suited to localised hydraulic problems. They are not normally used for flood mapping because they need a lot of detail which causes excessive simulation times.

In steady state models, you would apply a single flow rate at each inflow boundary. Your calculations would be taken on this constant flow rate. These models do not represent volume and are not suitable in situations where you need to consider attenuation or storage.

You must be able to demonstrate that the software you choose is suitable for the intended use. Benchmarking tests are available to show software suitability. These tests are presented in the Defra and Environment Agency (EA) 2004 and 2010 article. If the tests are done independently, the EA will need to review the results before you use the software for a project.

Many software platforms are updated regularly. These updates can bring more functionality but may not be readily available for you to use. You should discuss software choice and version early in the modelling process and include them in your method statements.

A timestep is the time when a hydraulic model makes calculations during the simulation. Hydraulic models are computer representations of river systems. They simplify the system they represent in space and time. Models split the study area and run time into manageable portions. This is known as discretisation.

In an implicit solution, the model calculates results at the current and previous timestep. You do not need the courant number to be below 1 for an implicit model as the defined timestep is less important. But a model with a large timestep may not be stable and could miss calculations at a critical time during the model simulation. You must consider appropriate values to provide accurate results and maintain manageable simulation times and mathematical stability. The timestep will also be dependent on the scale of the model under consideration. The TUFLOW manual states the:

You should consider using double precision options. Double precision lets you increase the number of digits and bytes used to store numbers. For example, Flood Modeller stores numbers using 7 digits or 4 bytes. Using the double precision option increases this to 15 digits or 8 bytes.

If the model contains a lot of elevations, or has a water body with a large surface area, you can use more digits to record the integer-part of the number. This can lead to mass conservation issues or instabilities if simulated in single precision.

Where possible, there should be a flow or level gauge at the downstream extent to provide a suitable boundary condition. When there is not a gauge, you should place the boundary far enough away from the site of interest (read Modelling downstream boundaries.

You can use recorded rainfall to feed into direct rainfall models or estimate flows from ungauged lateral catchments (if no tributary gauge exists) using ReFH1/ReFH2 models. For open coast and estuaries, you can make checks against tide tables and the Coastal flood boundaries (CFB) dataset.

You should only use an in-channel event and a larger flood to allow calibration of both in-channel and floodplain model elements. Set out your choice of events in your hydraulic modelling method statement.

Before you start calibration, you should check if you need to make event specific changes. You may need to block structures, remove completed flood defence schemes or use historical topographic data. If you need to make changes to calibrate against historic flood events, you should use the present-day representation for your final design runs.

The verification event performing at the same level as the calibration events is a desirable outcome. You do not need to re-calibrate if there are differences in performance but you should investigate the likely causes.

For example, you could use a water level gauge on a study reach with no flow data. You can use the water level gauge to calculate median annual flood levels and check them against modelled 50% AEP events.

Testing usually sees a 20% change adjustment to inflows, hydraulic roughness, downstream boundary, structural and spill coefficients. When adjusting downstream boundaries, do not just adjust gradients associated with normal depth by 20%. You should extract a modelled rating at the boundary and use an adjusted version to test the effect of the uncertainty. You should document your changes in your hydraulic model report.

3a8082e126