For more than 15 years KISTERS has been developing smart algorithms. From a combination of efforts in the water and energy industries, a comprehensive set of advanced mathematic functions has been matured and now serves as the backbone of KISTERS artificial intelligence (AI).
Today, freshwater supply is under stress of climate change, urbanisation and high water loss. New freshwater sources will not solve resource volatility. Rather, optimisation of current supply volumes, operational times and processes is the key for a better managed water supply. Artificial intelligence can assist the water industry, rapidly analysing data from operational processes and informing strategic planning.
With artificial intelligence and deep learning, KISTERS solutions are able to precisely predict the demand of water. In turn, accurate demand forecasts can significantly cut down supply costs as energy prices and pumping schedules are considered.
Cost savings can be reinvested in better customer service in terms of leak detection, infrastructure repair or extension, and water supply vulnerability reduction.
Comparing AI-driven demand forecasts with metered consumption can reveal significant deviations, especially at night, which are indicative of leakage.
Taken a step further, “digital twin” hydraulic models can identify the most vulnerable locations within metered areas before leaks occur, thereby saving maintenance teams precious time and prioritising repairs.
AI-driven planning requires the combination of projections, such as urbanisation and infrastructure growth to meet the demands of today and in particular the future. With the integration of climate change projections, the risk of future water availability can be minimised.
When set up properly, AI can support scenario calculation to better simulate extensions of the water supply network.
Water is susceptible to contamination by microbes and other pollutants. In particular, zones of zero flow can significantly compromise water quality. Similar to detecting leaks, AI-driven hydraulic models can identify stagnant and vulnerable spots in water supply infrastructure. KISTERS’ hydraulic optimisation tools can inform the operation of pumps and valves to keep water flowing even in challenging zones.
With AI-driven demand forecasts in combination with hydrological forecasts, KISTERS planning models can optimally balance the needs of flood control and opportunities to store surface water or recharge aquifers.
Limitations of AI
In the applications mentioned, artificial intelligence can only support water utilities successfully if a detailed understanding of the entire situation exists.
In the absence of that understanding, deep learning is extremely beneficial to generate intelligence of past behaviours of water supply systems. At this time, however, deep learning is not able to extend its intelligence and adapt to future conditions.
Since AI requires training, new conditions or the first instance of recurring events are unpredictable. Therefore, AI requires integration with high quality data sets to limit any misleading of deep learning processes. KISTERS offers water utilities a complete end-to-end solution that first upgrades the quality of its data and information before applying AI algorithms to enhance processes or assist in strategic planning.
The KISTERS team with experts from the water management and IT industries support the entire life cycle of your water solution right from the beginning with extensive consulting: We select the readymade standard modules to build and configure the exact solution that fits your business needs. Training offers and manuals tailored to your organization, round-the-clock internet support and a telephone hotline all serve to help you with any operational issues.