Department Urban Water Management

New methods for longterm planning of drinking water supply and drainage systems

by Max Maurer

The level of investment needed for infrastructure renewal – of urban drainage, for example – is largely dependent on the age of the system in question or on its prospective structural condition. In addition, systems will also need to be adapted as a result of urban development and climate change.

Shortage of data

Up until now, infrastructure planning has been difficult due to a lack of good models for predicting the ageing and decay of pipes. New generic models have therefore been developed in SWIP that can deal explicitly with incomplete records. (Scheidegger et al., 2011; Scheidegger und Maurer, 2012). This is especially significant for infrastructure planning in smaller communities, where records on pipe damage and condition may have only been kept over the past few years, or not at all, and where historical records are discarded.

Deterioration modelling

Statistical methods are employed to forecast the current and future condition of pipes. The models take into account previous replacements and thus minimise the risk of systematic overestimation of lifespan. For calibration of the model using the Bayesian approach, either expert knowledge or knowledge gleaned from other networks is combined with the available but incomplete records.

Predictions

The models allow predictions to be made about the future condition and deterioration of pipes. This feeds, on the one hand, directly into the MCDA (Multi-Criteria Decision Analysis), whereby the condition of the pipes has a significant bearing on various important objectives. For example, the reliability of the system can decrease drastically if drinking water pipes are not regularly renewed, which in turn pushes the burden of cost on to the next generation. Similarly, if sewers are in poor condition, this can lead to the ground water being compromised with nutrients and micropollutants.

Climate change

The longevity of sewers also gives rise to another relevant question, which asks to what extent our sewer systems will need to be adapted to cope with future climatic conditions. One particular key factor here is how the extreme precipitation events that are responsible for the flooding of urban infrastructure will change. For this purpose, we used a stochastic rainfall model that was adapted to both historical data as well as ten different climate predictions. The results gave no clear indications. It is thus not possible to make any clear recommendations for climate adaptation measures. The results did show that even under current climatic conditions, extreme precipitation is subject to a high degree of variability. This is not yet taken into account in current planning measures. The significance of these factors is therefore being investigated in two selected sewage networks in the case study region of “Mönchaltdorfer Aa”. This forms a basis from which recommendations can be derived for real-life situations.

Case study

By combining the ageing predictions with the replacement strategy and using simulated precipitation data, the required level of investment can be determined. These calculations are being run in parallel for the different possible future scenarios in the case study region of “Mönchaltdorfer Aa“. In the process, new infrastructures for urban areas to be built in the future are being created and included in the calculations.

Important attributes of various different planning alternatives emerge from these calculations, and these attributes are significant in terms of the MCDA. Replacement strategies and other infrastructure options can thereby be identified that are particularly attractive to the stakeholders. The uncertainty of these results and their degree of robustness in terms of the future scenarios are also taken into account.

Outlook

Although the models designed in the project are generic and not inextricably bound to the case studies, further implementation steps are required before the planning tools are fully ready for practical application. The implementation of the models must make provision for various exceptional circumstances (consistency checks), and, where necessary, be adapted to incorporate these. The models have to be improved in terms of their usability.

Further information

• Sub-project 2: Dissertation by Christoph Egger: Determining the performance and condition of future wastewater systems.
• Sub-project 3: Dissertation by Lisa Scholten: Multi-criteria decision analysis for water supply infrastructure planning under uncertainty.