As the imperative for decarbonisation and network growth looms, district heating companies must evolve their operations to reduce heat losses and implement sustainable sources. Since their efficiency and output are sensitive to different operating temperatures, such sources (like heat pumps) complicate operations. This article explores the shift companies should make in their daily temperature operations. It shows the limitations of conventional heating curves and highlights the need for more advanced temperature controls to stay economically and sustainably efficient. 

The district heating sector is transforming. Companies face decarbonisation and decentralisation of heat sources, while their networks keep growing. On top of that, the integration of cooling and the electrification of heat production introduces complexities and exposure to pricing volatility in electricity markets, making operational decisions more intricate. In the upcoming years, the daily operation of the heating system will become more challenging. 

Since sources such as heat pumps – which operate more efficiently at low supply temperatures – will be the baseload for most future-proof networks, there is a clear call for network operators to start lowering temperatures. This concerns lowering both forward and return temperatures structurally over the entire year and dynamically within a single day. However, this should be done without compromising the reliability of the heat supply to end customers. 

How can companies achieve that? How can they start lowering temperatures if, for decades, they have been using the same control strategy to operate the heating network combined with operator experience? Decarbonising your system is one challenge, but changing operational habits is yet another.  

It's no longer a question of if but when companies will start using new temperature control methods to make a heating system resilient and keep heat affordable. 

In the following sections, we will explain why operating a district heating network with a conventional heating curve (also known as a control curve or control table) is not sufficient anymore when moving towards a more renewable, sector-coupled energy system. 

Even if you don't entirely rely on heat pumps or other renewable sources yet, moving towards a more resilient temperature control method is the first low-hanging fruit in transforming your system. You'll also be making it future-proof against all the changes in the upcoming years while already reaping the benefits of lower temperatures, lower heat losses and increased security of supply. 

The conventional control curve and its limitations 

Many European heating companies have been using control curves or tables to determine their supply temperature day-by-day. This is the simplest way of determining temperature setpoints, based on a static calculation and often dependent on one single variable: outdoor temperature.  

This static approach is often too simplistic and comes with two risks: 

You can see an example of a typical control curve below.  

Figure 1. A typical control curve used by a heating company setting supply temperatures based on outdoor temperatures. 

As you can see in the figure above: 

This heating control curve has a few downsides: 

Static control approach 

First, this control approach is relatively static and often not up to date. For instance, it has no connection with changing system conditions: if a certain network area reduces the total demand due to the introduction of decentralised storage or increases it due to network expansion.  

Such changes (which happen quite often) would require the operator to continuously update and validate the control curve to ensure that it's still in line with customer requirements: is the used temperature still sufficient to meet the demand at the end-of-net? Also, if the heating system has many loops, it may be challenging to determine where the critical delivery points exactly are. 

Lack of insight into the network hour-by-hour 

Secondly, during daily operations, a control curve lacks a feedback loop to what is happening in the network hour-by-hour. There is no validation based on real user demand, network dynamics, or whether the temperatures used are sufficient.  

This can cause unexpected increases in flow and pressure drops, resulting in higher pump costs than expected (or even malfunctioning or tripping pumps).  

Also, without validation, an operator might see that they're actually running the system closer to hydraulic limits than their pre-defined heating curve assumed (that is why heating curves always need correction from operator experience). 

Simplified demand assumptions 

The main problem with the control curve is that it simplifies the assumption of the demand in your network based only on the outdoor temperature.   

However, more factors impact the actual network demand. Every day and every hour, the demand across the network changes based on the time of year, week, and weather conditions (as you can see in the data from all the deployed smart meters in the network).  

For instance, using a control curve, you would use the same supply temperature with 5°C ambient temperature on Christmas Day, a regular Tuesday in January or an unexpected cold Friday in April. Demand is significantly different during these days!  

For instance, on Christmas Day, all the office and government buildings are closed, and households use more heat. During an unexpectedly cold day in April, customers might use their heating differently than on an expected cold winter day in January.   

Why control curve can no longer support district heating 

The heating curve is not perfect. You can see it when comparing the day-to-day usage of a heating curve with the theoretical control curve. There are clear deviations due to multiple reasons:  

This indicates that the temperature that is needed is more dynamic and difficult to determine from hour-to-hour and day-to-day than what the control curve suggests. 

Figure 2: Full year of measurements of actual operations of a control curve used by a heating company 

As the heating system is in continuous transformation, the limitations of a control curve are even more pressing due to three reasons: 

1. Hydraulics determine your capacity 

Network capacity depends on the amount of flow that your pumps can handle. If you want to add new users to your network, it's key to understand how heat flows are pulled throughout the entire network and where (potential) bottlenecks are. However, the control curve doesn't consider actual network hydraulics and resulting dynamics, such as heat propagation.   

This is also something that the control curve doesn't consider during day-to-day operations throughout the year.   

In summer, the flow in the network is usually very low due to a lack of demand. This means that it takes some additional time (and heat losses) until the heat produced at the source arrives at the end-user.   

That's why you'd likely need to increase your supply temperature occasionally during low flows.   

In winter, though, the network's flows are high, so temperatures can be lowered by pumping a bit more (if no hydraulic bottlenecks arise). 

2. New decentralised sources are temperature-sensitive  

As companies decarbonise district heating networks further, a higher penetration of more sustainable heat production sources – often electrified such as heat pumps – is underway. Such sources are more efficient when operating at relatively low supply temperatures.  

When these sources are commissioned to new locations across the network, different supply temperatures at each heat supply location might be preferable based on the network area it supplies.  

However, since a network is often a large interconnected system of pipes, it's difficult to determine which part of your assets supplies which area. You can't simply copy-paste your control curve to these new assets at different locations.  

3. Your heating will become more dynamic 

The bottom line of controlling temperatures is that it should never compromise heat supply reliability. The contractual arrival temperature at the end customers' needs to always be guaranteed. 

However, when customers start to actively control their demand based on price incentives, the entire network demand, heat propagation and temperatures across the grid will become more unpredictable. A heating control curve is too static to handle this dynamic behaviour, and you can't adapt it when demand starts fluctuating based on prices.  

Replace legacy control curves with real-time, data-driven, dynamic temperature control 

As a future-proof alternative to static temperature operations, a fully dynamic temperature control has been introduced by many heating companies across Europe.  

Figure 3: Full-year comparison of control curve and dynamic temperature control, including the arrival temperature constraint at the users 

The graph above depicts a comparison of a dynamic temperature control algorithm with a conventional control curve.  

The red cloud represents data collected during a year of control curve setpoints, while the black cloud shows a year of setpoints based on dynamic temperature control.  

Every hour, dynamic temperature control calculates the actual and expected demand in the network for the next few hours, the resulting heat flows in each pipe, and the time the heat needs to travel.  

This means that supply temperature setpoints can be different every hour and every day – even though outdoor temperature is similar!  

Instead of using a static control curve 'line' to operate temperatures, dynamic temperature control is 'free' to calculate the best temperature every hour, and that's why the data points are a bit more spread out.  

This immediately considers hydraulic bottlenecks so, sometimes, a slightly higher temperature is suggested to avoid too high flows and pumping. 

A proven solution 

Results across different networks show that dynamic temperature control lowers supply temperatures by 5 to 15°C compared to conventional control curves on average over the year.  

This is because a conventional control curve needs built-in safety margins to deal with varying demand at different ambient temperatures. These relatively large safety margins are required with conventional control curves to ensure security of supply.  

However, they are often set too broad 'just to be sure'. You can't blame the operators if they lack the right solutions to give them more insight and confidence in what is really happening in their network based on actual and calculated data, can you? 

How does dynamic temperature control work? 

What we refer to as dynamic temperature control has a continuous feedback loop based on everything happening within the network. The core of this solution is a real-time, live thermohydraulic model trained on historical and live data.  

All data is combined from sources to substations to all smart heat meters. The live model continuously calculates and presents all temperatures, flows, and pressures throughout the entire network, even where they are not measured. The solution can, therefore, continuously calculate the point in the network where temperatures are lowest and if all hydraulic limits are met throughout the network.  

Another key feature is the granular demand forecasts across the network based on historical and actual heating demand and live weather variables (wind, rain, outdoor temperature and more) so that they're always accurate and up to date.  

Based on the live thermohydraulic model and the granular demand forecasts across the network, dynamic temperature control can very accurately forecast the ideal temperature setpoints hours and days ahead. This means it can ensure sufficient heat supply across the network without creating any hydraulic bottlenecks and taking heat propagation into account. That way, it's immediately clear when a heating source 'oversupplies' and when lower supply temperatures are beneficial.  

Moreover, the adoption of dynamic temperature control solutions has proven to increase the reliability of the heat supply significantly since it reduces the risk of undersupplying. 

A pathway to lowering your supply temperatures  

Below, you see an example of what dynamic temperature control means over a full year in comparison to a conventional control curve for a heating company serving a few thousand buildings.  

An immediate observation is that dynamic temperature control almost always suggests lower temperatures during winter, spring and autumn. During these months, there is high circulation in the network and, hence, a room to optimize temperatures to match the heat demand. In summer, when circulation is lower, it takes more time to transport heat and higher temperatures are suggested to avoid undersupply. 

Figure 4: Full year of dynamic temperature control measurement compared to a conventional control curve for an existing heating network. 

Dynamic temperature control ensures an operator can run the system at lower temperatures, always considering hydraulic behaviour and without risking security of supply.  

Dynamic doesn't simply mean 'lower the control curve', but rather 'use the temperature that fits the situation best'. For heating companies that are already running close to their hydraulic limits, dynamic temperature control might even come up with some higher temperatures now and then to free up some capacity if that is needed.  

Or even more advanced, 'frontloading' the network with higher temperatures, using the storage capacity of pipes, to reduce peak supply temperatures later.  

For instance, heating companies who make use of dynamic temperature control that Gradyent offers with its Digital Twin, experience 5-10°C lower forward temperatures on average throughout the year, which leads to 10-15% reduction of heat losses on average.  

At the same time, security of supply for all of customers increases to almost 100%. This results in 1-2% lower fuel costs and CO2 emissions and a smoother integration of new renewable sources. 

Future-proof your operations 

Not having to rely on legacy control tables reduces heat losses and emissions immediately, making network operations futureproof.  

The advanced control of dynamic temperature ensures complexity can be managed better and that decentralised renewable sources and storages can be integrated. Insights into network dynamics, temperatures, flows as well as demand and supply matching provide a solid basis for all upcoming transformations. 

Starting with these insights today immediately delivers benefits from reduced temperatures, and improves the business case of all renewables that will be integrated in the upcoming years. 

Are you operating a heating network and want to learn more about what the benefits could be for you? Check how Eneco or Mantova implemented dynamic temperature control, reduced temperatures with ~10° C on average, saving on heat losses and CO2. 

Would you like to see how dynamic temperature control could impact your heating network? Book a demo with one of our specialists.