Using data and local weather forecasts instead of drawings of the supply network and simulations to control supply temperature can save the district heating sector money and reduce CO2.
In Denmark, 1.7 million households (about 64 per cent) are heated by district heating running through 60,000 kilometres of district heating networks. The journey from the district heating plant to the radiators typically takes several hours, which is why the heating requirement must be predictable.
Increasing heat production above requirement levels is undesirable, as it costs money and wastes energy—just as the temperature loss in the pipes is greater at higher temperatures. At the same time, the water must be sufficiently hot at the so-called critical points on the periphery of the supply network. Managing district heating production optimally is thus a science.
At DTU Compute, Professor Henrik Madsen and his colleagues are working on data-driven energy and temperature optimization. Several research projects show that digitalization improves the forecast for heat demand and even helps to achieve Denmark’s 2030 climate goals.
A study carried out by Damvad Analytics together with Denmark’s largest smart city project—CITIES (led by Henrik Madsen)—and the Green Energy Association think tank established by the Danish District Heating Association shows that the district heating sector can save between DKK 240 and 790 million by introducing data-driven temperature control of the flow temperature, as the temperature can be lowered three to ten degrees. Lower temperature also reduces CO2, as well as heat loss in the supply network.
“There are great potential benefits in moving from experience and simulation-based management from diagrams of the supply network to data dynamic optimization of district heating. Our projects show that when the flow temperature is based on several here-and-now data sources, including weather data from local measuring stations, we are able to optimize production and accelerate the green transition,” says Henrik Madsen.
Lower heat prices
The district heating company Svebølle Viskinge Fjernvarmeselskab supplies district heating to 535 households and is one of the utilities that has increased digitalization. Since October 2019, the district heating company in Northwest Zealand has been using processed data from the DTU spinout company ENFOR to optimize flow temperature control.
In the space of a few months, the flow temperature has been lowered by more than 20 degrees. Whereas the temperature used to be 80.9 degrees, it was first lowered to 68.1 degrees—and subsequently to 60 degrees—resulting in an estimated saving of at least 550 MWh and a reduction of DKK 110,000 in annual production costs. The company expects to reduce the heat loss in the supply network to below 30 per cent and the price of heat by 47 per cent between 2015 and 2025. Before Svebølle Viskinge Fjernvarmeselskab moved to data-driven operation, the company used so-called simulation-based operations that relied on knowledge of the district heating network, experience, and a little forecasting.
"At the time, we were working in the dark because we didn’t know what the temperature of the district heating network really was. We never imagined that digitalization could deliver such huge benefits."
Svend Müller, bestyrelsesformand for Svebølle Viskinge Fjernvarmeselskab
“At the time, we were working in the dark because we didn’t know what the temperature of the district heating network really was. We never imagined that digitalization could deliver such huge benefits,” says Svend Müller, the company’s chairman of the board.
The digital setup itself is a simple and inexpensive IoT solution for temperature measurement. IoT stands for ‘Internet of Things’ and means technology that connects things to the internet.
The district heating company’s solution, which was developed locally by Svend Müller and the energy company SEAS-NVE, consists of battery-powered temperature meters that are positioned inside two valve wells in the district heating network. The meters relay data at five-minute intervals, ensuring that there is constant temperature control within the supply network. At the same time, consumption data from all 535 remotely read household meters are displayed on a map—a dashboard developed with the engineering company ABB—in order to provide a full overview.
Improved heat forecast
In another project—IDASC—DTU Compute is collaborating with HOFOR, among others, to digitize the heat supply in Copenhagen’s Tingbjerg district. Here, 25 apartment blocks are connected to the same heat exchanger, meaning that the area in practice functions as an isolated district heating network. Using feedback from smart meters installed in the buildings that automatically transmit consumption data from the end users to the utility company, the project is working to find the optimal temperature and flow in the system in order to lower the flow temperature, reduce CO2, and save money.
“It’s a small demonstration project, but it will give other district heating companies and HOFOR an idea of how much can be gained from digitalization and better use of data,” says PhD student Hjörleifur G. Bergsteinsson.
The experience of the IDASC project, for example, shows that the accuracy of predicted heat demand can be improved by about 40 per cent by combining real-time data collected from a local weather station and several different weather forecasts.
The self-learning and data-driven methods are suitable for dividing cities into temperature zones, thus enabling further savings, says Henrik Madsen:
“For example, evaporation from large green areas helps to keep the temperature down—while outdoor areas covered in asphalt retain the heat that is transmitted to the buildings. So the temperature can easily be five degrees higher in different parts of cities because of the buildings, asphalt, machinery, and the people.”
CITIES has developed several state-of-the-art methods, and according to Henrik Madsen, following the publication of the results in various scientific journals and at conferences, the methods are now playing an important role in the development of new software for the energy system of the future.
“Most energy companies, large and small, will be able to benefit from our experiences and solutions.”
• Denmark’s largest smart city project CITIES—Centre for IT-Intelligent Energy Systems—researches efficient, integrated, and intelligent energy solutions that support the transition from fossil fuels to sustainable energy.
• CITIES’ methods rely on artificial intelligence (AI) to analyse large amounts of data to predict, control, and optimize the interaction between different energy forms.
• CITIES has 24 Danish and Swedish industrial partners and cooperates with 15 international knowledge institutions from the EU, Korea, and the USA.
• Innovation Fund Denmark supports CITIES.
• The CITIES project ends on 31.12.2020. Several smart energy projects are based on CITIES’ research, including Center Denmark, Fed (Flexible Energy Denmark), HEATman, IDASC and Smart Cities Accelerator, as well as EU projects SmartNet, Syn.ikia, eBalancePlus, TOP-UP, and FLEXCoop.
• CITIES will hold its final conference from 9-10 November 2020 at DTU.