From a quick AI analysis to a complex hydrodynamic simulation: there are different ways to understand a water system. Together, we choose the model that best fits your issue and translate the results into reliable insights for better water management decisions.
A model is not an end in itself; it must provide insight and support decisions. That is why we start with your question and determine what speed, accuracy and level of detail are needed.
When every second counts during an imminent flood, a machine learning model can offer a quick solution. For a detailed study, a physical model may be more suitable. HydroLogic weighs speed, reliability, explainability and cost, and chooses the model that offers the most value.
We start by clearly defining the issue, the users and the intended application. We then determine which processes need to be described, which data are available and what level of detail is required. We also consider requirements for computation time, timeliness, uncertainty, maintenance and explainability.
We work independently of models and platforms. We can develop a new model from scratch, improve an existing model, or combine different models and data sources. Among others, we work with D-HYDRO, SOBEK, 3Di, Ribasim, HEC-RAS and MODFLOW–MetaSWAP, but also with statistical, GIS and machine learning models.
From local studies to global analyses. Our models provide reliable and explainable answers at the scale your question requires.
AI and machine learning offer new possibilities for processing large amounts of data, recognising patterns and producing forecasts faster. Especially when speed is essential or many scenarios need to be calculated, these techniques can add significant value.
We only use AI when it demonstrably fits the application better. In doing so, we assess the quality of the training data, performance under different conditions, uncertainties and explainability. Where appropriate, we combine AI with physical models and hydrological knowledge. This way, we harness the speed of data-driven techniques without losing sight of our understanding of the water system.
Would you like to develop a new model, improve an existing model application, or explore where AI adds value? We would be happy to help!Laura Janssen
