ML4Heat - Tools for optimized operation of existing district heating networks based on machine learning methods

Project goals

The global objective of the ML4Heat project is the development of methods and software tools for the optimization of the operation of existing district heating networks under energetic and economic aspects. For this purpose, sensor and operating data of the district heating transfer stations as well as the heat supply shall be collected to a large extent and evaluated by means of machine learning methods.

Project result

Tools are to be realized on three levels:

1. Optimized operation of the district heating stations through performance and condition monitoring: the aim is to achieve the most automatic possible performance monitoring and optimization of the control in the district heating stations.

2. String optimization: Methods are developed which, based on the measurement data of the district heating transfer stations, can quickly recognize or ideally already predict the utilization of sections (strings). For this purpose machine learning methods in combination with basic physical equations are used.

3. Network optimization: First of all, methods are developed to predict the energy demand for the entire district heating network more accurately than before. For this purpose, the findings on local capacity planning gained during the string optimization will be combined with external data (especially weather forecasts). Taking into account the use of renewable energy sources (e.g. solar energy), the demand for fossil fuels can be significantly reduced through more accurate real-time forecasts.