Can the behaviour of a water supply network be predicted during a blackout or panic-driven water consumption amid missile attacks? Can its most vulnerable assets be identified and effective response strategies developed in advance? For Ukrainian cities operating under wartime conditions, these questions have become particularly pressing.

Tetiana Starovoit, a postgraduate student at the Educational and Scientific Institute of Applied Systems Analysis of Igor Sikorsky Kyiv Polytechnic Institute, is researching these issues. She has been working on mathematical modelling of engineering networks at Kyivvodokanal PJSC for more than ten years, seven of which she has spent heading the Geographic Information Systems (GIS) Support Department. She spoke to Kyiv Polytechnic about her work.

– How did practical engineering challenges develop into a research idea?

– My research grew out of my work with GIS, digital models and the hydraulics of water supply networks. A pump shutdown, a change in valve position or a power outage can affect pressure and water distribution far beyond a single facility. This led me to the idea of analysing not individual failures, but the behaviour of the entire network, how disruptions propagate and what opportunities there are for recovery. After the full-scale invasion, this challenge took on particular significance.

– How has the war changed the demands placed on engineering research?

– Shelling, prolonged power outages and a shortage of backup capacity undermine the conventional assumptions about stable infrastructure operation. A power outage at a pumping station can cause pressure drops, flow redistribution and water shortages in different districts. This is how cascading failures occur – a sequence of interconnected disruptions. To study them, we need to consider network topology, physical processes, time-dependent dynamics and recovery as parts of a single system.

– You completed a training programme in Japan. What did you gain from the experience?

– In 2024, at the invitation of Kyivvodokanal, I took part in a water quality training programme in Kobe and exchanged views with Japanese specialists on the resilience of water supply systems. Japan has extensive experience in preparing infrastructure for earthquakes and tsunamis: backup systems, damage forecasting and setting recovery priorities. The threats facing Ukraine are different, but the objective is the same – to keep the system operational under emergency conditions. This experience helped me develop methods for modelling crisis scenarios.

– What is the essence of the intelligent decision-support system you have developed?

– A conventional hydraulic model shows how a network will operate under specified conditions. Our task is broader: to predict the consequences of disruptions, how they spread, and to find solutions that minimise losses. To do this, I combine systems analysis, hydraulic modelling, GIS, graph theory and artificial intelligence. Graph neural networks help analyse the connections between network components, while physics-informed neural networks (PINNs) make it possible to incorporate physical laws. Artificial intelligence does not replace the laws of hydraulics; it complements them.

One practical area of application is blackout scenarios. The model makes it possible to assess the consequences of pumping station shutdowns, available water and power reserves, alternative operating modes and opportunities for redistributing flows. If there is no electricity, the model cannot keep the pumps running, but it can help determine the limits of the network’s operability and prepare a well-founded response plan.

– Has the model undergone practical testing?

– Yes. It has been tested on part of an actual hydraulic zone in Kyiv. This made it possible to validate the algorithms using a real network, with its specific topology, operational constraints and interdependencies. Such testing is an important step towards broader practical application.

– How is international cooperation developing?

– I exchange experience with European water utilities, including HAMBURG WASSER and Berliner Wasserbetriebe, as well as with representatives of DHI, Newcastle University and IHE Delft Institute for Water Education. This helps us compare approaches to digitalisation, hydraulic modelling and infrastructure management. I also provide training for Ukrainian water utilities on resilience modelling and the application of systems analysis to practical challenges.

– What role does your studies at the Educational and Research Institute of Applied Systems Analysis at Igor Sikorsky Kyiv Polytechnic Institute play in your research?

– Systems analysis makes it possible to view a water supply network not as a collection of pipes and pumps, but as a complex interconnected system. My professional experience helps me formulate realistic challenges, while my studies at Kyiv Polytechnic provide the tools to address them using mathematical optimisation, graph theory, complex systems analysis and artificial intelligence methods. The main goal is to turn technical, spatial and operational data into well-founded recommendations for engineers.

– What is the value of such developments for Ukrainian water utilities?

– A digital map of a network alone is not enough. We need to know how it will respond to a failure, which facilities are most vulnerable and how limited resources can best be used. Intelligent systems help shift from responding to the consequences of disruptions to predictive management: testing scenarios, identifying critical facilities, and planning backup capacity and recovery. Naturally, the final decisions are made by specialists, taking into account the actual condition of the equipment and safety requirements.

– What are the prospects for applying this methodology in other sectors?

– Following the initial trials in Kyiv, the next step is to extend the methodology to the city’s entire water supply and wastewater systems. Many of the graph-based algorithms, optimisation methods and intelligent data analysis techniques can also be adapted for heating, gas and electricity networks. The long-term prospect is to jointly study the resilience of interconnected engineering systems.

– And finally, why did you choose systems analysis?

– Hydraulics describes the physical state of a network, but choosing the best crisis response strategy requires comparing numerous scenarios, risks and constraints. We have to consider safety, reliability, resource availability, the number of consumers supplied with water and the speed of recovery all at once. This is a multi-criteria problem that calls for systems analysis.

My research brings together the practical experience of Kyivvodokanal, the scientific methods of Igor Sikorsky Kyiv Polytechnic Institute and international knowledge exchange. Its goal is to make mathematical modelling not merely a tool for engineering calculations, but also an instrument for forecasting crises, assessing risks and supporting decisions that are vital to the functioning of cities.

– Thank you for the interview.

Volodymyr Shkolnyi

Дата події