Detecting leakages should be fast and precise. There is no use in knowing there is a leak if you still need to send out dozens of people to pinpoint the exact location within your DMA/network. The state of the art is creating small enough networks, but this is not cost efficient. We offer detection with a precision of a few hundred meters, no matter the size of your network.
Our society is evolving to become climate neutral and zero waste. A lot of invesments go into optimising the local production and consumption of resources with the aid of solar panels and batteries. At Agrippa we help you to optimise the usage of your water infrastructure and (collective) water gains as well. Connecting water basins across (and beyond) the industrial terrain offers the chance to combine the need for flood buffers and water storage.
A solution that is more expensive than the problem is no solution. We do not require fancy and expensive sensors or invasive infrastructural works. We only need the data you already have; pressure, consumption and a hydraulic model. The model even may be uncallibrated, that's how smart our Artificial Intelligence is. Let us start from the problem, not the solution.
Since we build Software as a Service, our solution is extremely scalable. Whether you have one DMA/network or thousand, the initial investment is small. You can turn monitoring for a particular network on and off whenever you want.
Detecting leakages in the vast network of water pipelines is hard. Almost as hard as delivering clean and vital water to everyone everyday. You focus on delivering clean water, we on detecting leakages. Without worries and big investments. Let's just use the infrastructure you already have.
We know water companies already do a great job on monitoring their network. As a first step we set up an integration. This enables our Artificial Intelligence and your data to do the work for you. Not doing anything and finding leakages. Sounds almost too good to be true, doesn't it?
Our Artificial Intelligence will start to learn how you network behaves based on the consumption, pressure and GIS information of your system. After a couple of weeks, it will know how data flows through your network and be ready to detect leakages. We do not require callibrated models or extra sensors. Most water companies already do enough monitoring to feed our AI. Let's solve a problem, not sell a solution.
After the training phase, we will permenantly monitor your network. We do not use statistics or consumption monitoring during the night. We know how your network behaves and what leakages look like. As soon as we have found a leakage, you will receive a notification. You can then use our Software as a Service to check your network and see the location and size of the leakage. And when we say location, we mean street-level precision.
Connecting water basins across (and beyond) the industrial terrain offers the chance to combine the need for flood buffers and water storage.
Combining both the need for buffer capacity and water storage requires thorough knowledge and modelling of both the expected precipitation and consumption needs.
By linking rain water storage to industrial process with a purification unit, the result becomes even more promising yet challenging.
Simultaneously optimising buffer capacity and process demands requires knowledge of the system. Intelligent models are necessary to describe both buffer needs and consumption demands. Where physical models alone may fail to describe the system accurately, hybrid systems integrating Artificial Intelligence in those models can help fill the gap.
Model Predictive Control allows us at Agrippa to leverage the knowledge captured in the obtained models. MPC combines the ability to optimise the systems controls based on the state of the entire system while still coping with and correcting for unexpected events.
AI expert and entrepeneur