News

Workshop makes numerical optimisation more accessible for water companies

Optimising distribution networks with Gondwana

More and more attention is being directed at the optimisation of drinking water distribution systems. But the complexity of these systems means that optimal solutions are hard to achieve using traditional (frequently handwork) techniques. Numerical optimisation techniques (with software) offer help, but a lack of suitable tools makes it hard to implement these techniques in practice. That is why KWR developed Gondwana: an optimisation software platform that specifically targets drinking water distribution systems. Gondwana thus unlocks powerful and frequently complex optimisation techniques for use by KWR’s hydraulic experts, and consequently also for the entire drinking water sector. At a workshop on 6 June 2018, staff members from water companies and a consultancy were given an overview of the developments in Gondwana during the last few years and of the results obtained to-date. Afterwards, they themselves set to work with Gondwana in designing District Metered Areas (DMAs) and in optimising pipe diameters. This was much appreciated, because there is a demand for more opportunities for the application of Gondwana.

Perspective on optimally designed distribution networks

The drinking water distribution network that we have in the ground today is the result of an evolution over more than 150 years in the materials used, and in the design philosophy and objectives. The design often incorporates experience-based knowledge and extra safety margins. Now that many drinking water companies are facing a mounting replacement challenge, it makes sense to make well-founded choices concerning the replacement network structure. It is important in this process to take a more holistic approach to the design process; that is, an approach that takes account of aspects such as supply assurance, water quality, energy and, of course, costs. Research shows that experts produce better network designs when they apply numerical optimisation techniques alongside the classical approaches.

Gondwana optimisation platform: status

KWR set up the Gondwana optimisation platform in order to make numerical optimisation techniques, developed in academic research, more accessible for water practice. Between 2015 and 2017, KWR researchers and others applied Gondwana to design District Metered Areas (DMAs): delineated areas that are purposefully designed with specific objectives in mind, and for which the volumes of water inflow and outflow are measured. Gondwana is also used in the design of target structures and of the transition to them. Various interesting results have already been achieved with Gondwana:

  • The division of the supply areas into DMAs offers insight into the trade-off between the number of flow meters required and the sensitivity of the DMAs for the detection of anomalies. Based on this information, drinking water companies can take well-founded decisions about the DMA division that best meets their requirements and expectations.
  • In the design of target structures, it is apparent that the currently existing distribution networks can be significantly reduced in size. This means that cost-savings can be made, while the performance, in terms of supply assurance and pressure, is improved.
  • The analysis of the desired transition from the current infrastructure to the optimal target structures clarifies the relationships that exist between the number of network locations subjected to annual work and the decrease in the number of failures or the improvement of the network’s hydraulic performance. This information enables the drinking water companies to make better-founded choices.

The experience has also taught us that the application of Gondwana in practice is very demanding with regard to data availability and good communication between researchers and drinking water companies: these are important prerequisites to producing good and implementable results in an iterative process.

KWR is developing Gondwana further and taking into consideration aspects like water-demand uncertainty and cyber-attack resilience.

Workshop for drinking water practice

On 6 June KWR organised a workshop in which the above-mentioned experience with Gondwana was explained to staff members from a Belgian and from several Dutch drinking water companies, as well as from a consultancy.

Following the presentations on developments in Gondwana, results obtained, lessons learned and ongoing research, a wide range of questions were aired: from the importance of having the appropriate data quality, to the size of the distribution network models that can be processed, and to Gondwana’s flexibility for incorporating new and company-specific conditions and objectives. The drinking water companies see many new application opportunities for Gondwana. They would like to use the platform in several important areas including the optimisation of pressure measurement points, new distribution network lay-outs and energy optimisation.

KWR researchers Peter, Ina and Mario help workshop participants in setting up optimisation problems in Gondwana.

Gondwana in the sandbox

The workshop participants then received a mini-course in Gondwana, and themselves set to work on two optimisation problems: determining the optimal pipe diameters for New York’s drinking water transport network, and dividing the Velsen supply area into DMAs, in such a way that a particular size of leakage would be measurable. The biggest challenges this involved were translating practical issues into formal optimisation problems and selecting the most appropriate functions in Gondwana. All the participants succeeded in generating results. By playing with Gondwana themselves, they developed a better feel for what optimisation problems involve and how Gondwana can be used to run calculations on them. Following their short afternoon with Gondwana, a number of the participants wanted more: they went home full of enthusiasm with the desire to test, explore and apply Gondwana further. KWR will work on making this possible.

 

share