HOT|COOL NO. 1/2017 - "System Integration"


consumption in comparison to the real energy consumption. As can be seen, the forecaster has a mean absolute percentage error of about 8% during the tested period. This is considered as good for these model-free algorithms.

First results of the project In the first year of the project, the focus was on the development of a so-called forecaster. This is one of the important modules in the controller algorithms, and tries to forecast the energy consumption in the network for the next 24 hours. This forecast is very important to optimize the behaviour of the network. A number of different machine learning techniques were implemented and tested in the Rottne network. As an input, the forecaster gets the day in the week, hour in the day, consumption of the network in the previous days and a forecast of the outside temperature. Output is the forecast of the consumption in the network. The figure above (figure 6) shows some first results, indicating the forecasted energy Figure 6: First results of the forecaster module in the Rottne network. Red is the real network heat consumption in kW, blue is the forecasted consumption, green is the error of the forecast.

Figure 5: Installation of a cluster unit in Mijnwater, The Netherlands

Conclusion In STORM, a generically applicable controller is developed for the operation of DHC networks. It focusses on both new, 4th generation networks, but also on existing, mostly 3rd generation systems. By maximising the use of sustainable energy sources, STORM is a stepping stone in making DHC networks more energy efficient and competitive. Intelligent control of DHC networks is indispensable for the transition of our energy systems towards zero carbon solutions. Johan Desmedt Project manager DHC networks For further information please contact:

Smart heat metering is the key to unlocking the full value of data The more data you have, the more value you can create

The wealth of data generated by smart meters has a number of uses that add value to energy metering – and as demands for higher energy efficiency and the integration of renewables increase, so does the value of frequent meter data. From revenue protection and improved customer service to network optimisation and shaping peak demands, smart metering provides the necessary transparency for heating utilities to make informed decisions in their day-to-day operations today and in the future. Learn more at

J O U R N A L N 0 . 1 / 2 0 1 7

Made with FlippingBook Ebook Creator