Some of the common misconceptions surrounding Network Modelling are:
- You need boffins to set up and run the models
- Once completed the models soon become obsolete
- It is difficult to translate model results into capital expenditure planning
- Models cannot be used effectively as an operational tool
- The models are expensive to produce and rarely justify the investment
- Probably the main reason for these misconceptions is that in the past models have often been poorly built using software of limited capability, and the models have rarely been used in anger. This shouldn’t be the case any more, as sophisticated software is now available, enabling models of high quality to be constructed within sensible budgets and timescales. The latest generation of models addresses and refutes these criticisms with the following functionality
- ease of use (intuitive, Windows type functionality) means users can concentrate on engineering issues, rather than relying on computer science boffins
- multi-user functionality makes them a team tool, removing the bottleneck of single user ownership
- links to other technical systems (such as GIS) increases productivity of modelling (and of other systems)
- links to office systems (such as spreadsheets) increases productivity, and allows the modelling results far broader reach and implementation
- automation of modelling processes – demand allocation, settings of network nodes elevation and pipe diameters,
- tracing connectivity, data cleansing – greatly reduces the cost of modelling
- automatic housekeeping – such as audit controls – increases security and productivity, and extends the life of models from one off exercises to maintainable, ongoing tools for a range of engineering planning and operations tasks.
This should enable Network Models to become the Operational and Management tools they have often promised to be but until now have failed to deliver.
Justifying the Investment in Modelling:
The cost of developing a well-calibrated fully dynamic model is often only a fraction of the cost of any single scheme it is used to design. If the model is used on a regular basis the initial cost of development soon becomes insignificant. We have to consider the potential overspend if errors are made in the design as a result of not fully analysing the problem using a Network Model. The cost of re-design, asset replacement, customer compensation and adverse publicity are all factors that should be taken into account.
Models can also be used to optimise the use of pumping plant, and can achieve considerable saving in operational costs. These savings in power and sometimes in reduced maintenance costs alone often far outweighs the investment required in the model construction.
Can models link to other systems?
Network Models can be linked to other data and business systems in use by the Water Provider. These include Geographical Information Systems (GIS), Asset Databases, SCADA and Telemetry systems to form an Integrated Network Management System.
The water industry has invested heavily in developing these systems, but historically there has been limited integration and therefore few examples of a successfully implemented integrated network modelling and management system. Linking Models to GIS.
Water Providers have invested significant time and money in developing comprehensive GIS systems, yet modellers have made little use of this valuable data resource and technology until now. Modellers have often seen GIS data as full of errors and not organised in a suitable way for model building purposes. However the latest software includes a series of data extraction and cleansing techniques that are enabling models to be built more accurately and efficiently. Using the latest tools the modeller can build an all mains model in less time than it used to take to manually build a strategic model in the past. Often the process of building a Network Model from GIS is an invaluable way to clean up the base data, to the benefit of all users of the GIS.
GIS brings with it a wealth of data that can be used by the modeller, including digital elevation data, pipe material and age information, individual customer point information, and Spatial data such as burst records, customer complaints, low pressure records, and water quality failure records.
Yet with all this wealth of data GIS remains essentially a repository of static system information, whereas modelling is fully dynamic, looking at past, present and future system performance. The only way to fully realise the potential of GIS is to integrate it with the dynamic medium that is Network Modelling.
Linking Models to Telemetry and SCADA systems
Linking Network Models to Telemetry systems provides an ongoing check that our theoretical model of the system is performing satisfactorily. This gives us confidence to use the model to make operational decisions and plan for future system development. The Model can also be used for contingency planning, reflecting or indeed guiding changes made in the real system to react to emergency situations.
Information from SCADA systems, Telemetry and temporary or permanent logging can be used essentially in three ways:
- To enable the latest system information to be available to aid model calibration
- Controlling model operation based on data directly obtained from the real system
- Controlling the real system based on the model results – truly integrating Network Modelling and SCADA systems in an operational environment.
- How recent improvements in the modelling process have greatly reduced the cost of modelling
The latest software developments have been designed to greatly streamline the modelling process, automating where possible routine and repetitive tasks, and providing flexible links to the source data. Some examples of the tasks that can now be automated include:
- Demand Allocation
- Setting elevation of nodes and customer points
- Setting internal diameter and initial roughness values from look-up tables
- Network connectivity tracing and data cleansing
The latest software has therefore developed considerably beyond just simulation. The automation of previously time-consuming tasks has led to significant savings in the model building process. Figures in the order of a 40% saving in the total cost of model building and calibration using the most modern software compared to traditional model building techniques are now being achieved. This together with the improved model accuracy and complexity has enabled Network Modelling to answer the criticism that models are expensive and time consuming to construct, and of little practical use in planning and operational decision making due to over-simplification because of limited software capability.
A Case Study:
Since January ’96 East of Scotland Water (ESW) have been using network modelling to drive their infrastructure improvements as part of their Integrated Network Management Strategy (INMS).
A fundamental part of their strategy is to be able to substantiate a value for money service to its customers. A major part of charges to customers is repayment of capital investment costs. It is essential, therefore that an audit trail is provided to substantiate infrastructure improvements and associated capital expenditure.
An infrastructure improvement plan was created, based entirely on improvements to customer levels of service and improved security of supply. A 10 year demand growth plan was simulated with a capital works plan for the period, derived. A considerable saving has been made, with substantial capital being reallocated to improving security of supply for customers. The steps of the audit trail can be summarised as:
- Set levels of service criteria (LOS)
- Measure service to customers accurately
- Record non-compliance of service to customers
- Produce a calibrated model to simulate the hydraulic behaviour of the supply and distribution system.
- Identify options for resolving non- compliances both now and into the future.
- Select best option to satisfy LOS criteria at minimum cost.
- Implement infrastructure improvements
- Monitor the ongoing performance to the customer
ESW recognises the audit trail as the key to providing the engineering substantiation required to request funds to carry out new works. Previously capital investment was planned considering the ‘views’ of the operators and the local requirements and in many cases was based on ‘myth and legend’ about the operation of the supply and distribution system. The fully dynamic network models form an integral part of the management plan, and are key factors in the decision making process, driving down costs and improving service to the customer.
In order to improve levels of service and reduce costs we must fully understand our networks. We cannot afford to plan new infrastructure or make operational changes without the benefit of a fully dynamic and calibrated network model.
It is no longer acceptable to produce skeletonised, static or un-calibrated models. Modern Network Modelling software is now capable of simulating all-mains models thus negating the need to skeletonise models. The latest software has enabled Water Providers to reduce the cost of modelling substantially making it both affordable and justifiable to build realistic, operationally complex all mains models.
Integrating Network Models with GIS, Asset Management and Telemetry / SCADA systems can increase the value our models widening their use beyond the traditional planning environment enabling them to be used operationally. Modelling can form the missing link, enabling the other data systems to be used to their full potential.
Comprehensive model management is the key to extending the investment in your network model. The model needs to be kept “alive” if it is to provide the maximum return on the initial investment.
A word of warning though; when considering Network Modelling, look at the whole cost of the project not just the cost of the software. Cheap and cheerful is usually not the best investment long term.
The lesson I have learnt over the years is that “Not undertaking network modelling is often far more expensive than undertaking network modelling!”