Graice Suite - AI-based Software solutions for power distribution grids

The aim of this module is to create a consistent and load-flow capable CIM (Common Information Model) network model. For this purpose, various checks are performed and, if possible, corrections and additions are made. These include checking the logical relationships of the grid topology, checking the parameters of the grid elements and the convergence of the load flow calculation.

The management of measurement data involves combining measurement data from different data sources and preparing and evaluating the measurement data for further processing in grid planning and operation. If necessary and possible, the measurement data are corrected and supplemented. The evaluation of the measurement data takes into account the measurement errors of the measurement chain. The analyses of the measurement data include, among other things, the evaluation of the status messages of the measurement devices, the verification of the voltages with respect to the defined voltage bands, the verification of compliance with the limits with respect to nominal current and power, and much more.

For grid state monitoring with minimal measurement technology, the optimal placement of measurement technology in the grid is an important sub-step. gridhound combines engineering approaches with AI algorithms to generate the optimal placement of measurement technology, taking into account existing technology and its constraints, if any.

Grid monitoring is a solution for determining all electrical operating parameters (voltage, current, power) of the electrical grid resources (lines, busbars, transformers, consumers, generators) in a (sub)grid area of a distribution grid at the medium and low voltage levels. The gridhound solution uses AI algorithms for this purpose, which reduces the need for measurement technology to a minimum and still delivers at least comparable results to the classical methods (WLS state estimation).

gridhound has developed AI methods to predict power for individual RE plants (PV, wind) but also for entire local grid stations. The prediction method is trained with historical master data, power measurements and weather data and can then predict power values with current master and weather data. This even works across other grid areas. Predictions can be generated for both the short and long term and are significantly influenced by the quality of the weather forecast.

Based on real-time grid monitoring as well as power prediction for RE plants and local grid stations, gridhound has developed AI methods to predict load for entire grid areas. These facilitate redispatch measures as well as grid connection planning, e.g. for RE plants, charging stations and heat pumps.

With the energy management system developed by gridhound, flexible generators and consumers can be controlled both in a way that serves the grid and is optimized for self-consumption. Grid monitoring and grid and power forecasts from gridhound serve as the basis. The grid-serving use cases include, for example, the avoidance and elimination of bottlenecks and voltage deviations in the distribution grid. Self-demand optimization includes use cases such as peak reduction, load/generation shifting, self-consumption optimization, etc.

gridhound has developed a simulation platform that combines all other Graice modules and enables the running of different scenarios. This makes it possible to generate data required for planning and to implement measures necessary for the energy transition.