Local and district heating networks make an important contribution to a cost-effective and increasingly renewable heat supply. However, this requires the best possible operation, from the heat generators and distribution to the substations in the buildings. And since heating networks are large and complex systems, deviations from optimal operating conditions and faults can occur during operation. Early detection of such problems is crucial to avoid failures, maximize efficiency and reduce unnecessary costs.

Many network operators already have a good basis for this employing regular maintenance, operating data analyses and the specialist knowledge of their operating staff to find anomalies and take effective countermeasures. However, some faults develop gradually or remain undetected because not all system components are covered by measurement technology or the capacity to analyze all available data is not always given. This is precisely where digital fault detection comes into play: By using a digital twin, all available operating data can be continuously monitored, anomalies and deviations can be detected at an early stage and targeted measures can be taken before major damage occurs or inefficient operation causes unnecessarily high costs.

In this blog post, we use a specific example to show how our heatbeat Digital Twin detected a defective control valve in a heating network with around 100 consumers, what actions were taken as a result and how the network operation ultimately benefited from the intelligent analysis.

How does our digital twin help with fault detection in heating networks?

Our heatbeat Digital Twin consists of 4 complementary modules:

• Base:The digital image of your heating network and/or your municipal heating plan with all data and analyses in one view
• Design:The interactive design of heating networks for optimal planning from the initial idea to a clear target image for the implementation
• Insights:The analysis of measurement data in real time
• Controls:Active control intervention for optimum operation of heating networks

For this case study, the focus is on the Insights module. This enables us to comprehensively analyze measurement data, which is recorded and evaluated in real time via a live connection. Together with our Base module, the measurement data can also be placed in the spatial context of the heating network.

Depending on requirements and conditions in the heating network, we can record both measurement data from the energy center and data from remotely readable substations, whereby both areas can also be connected and evaluated individually. For our current example of fault detection in the substations, we record the power consumed, the volume flows and the flow and return temperatures at the substations in the affected network. We record and save the data in our digital twin via an interface to the heating network system and can thus derive helpful analyses and error messages. This allows anomalies, errors and deviations from the target state of the network to be detected at an early stage and the operation of the entire system to be continuously optimized.

Detection of a defective control valve in a transfer station

In this case, we were able to inform the operator of a heating network about a specific error in a substation shortly after setting up the live data connection. However, since the latest enhancements to the live data table in the web interface of our digital twin, it is also very easy to identify similar faults in network operation yourself with very little effort.

For this purpose, the overview table offers the option of sorting all consumers according to the volume demand. The volume demand is calculated from the total volume that has flowed through the substation during a defined period and the amount of heat transferred. Therefore, this value, given in m³/MWh, is a good measure of how much water must be moved in the heating network to transport one MWh of heat. A low value is an indicator of an efficient substation with a high spread between flow and return temperatures, while a high value is a very good indicator of a possible malfunction of substations.

The following figure shows an overview of the measurement data for the building in question at the time of the first occurrence of a fault. It can be clearly seen how the volume flow through the substation rises steadily without any reduction in output and then remains almost constant at a maximum value. As a result, a bypass is created which unnecessarily allows the high temperatures of the flow to pass through to the return without any heat being drawn off. While the heat consumption of the building remains largely unchanged compared to other days, this error pattern suggests a problem with the control valve in the transfer station. Based on this initial situation, we drew the network operator's attention to the malfunction via the heatbeat Digital Twin.

Replacing the valve and operation confirmation

Thanks to this warning from our Insights module , the heating network operator was able to initiate an on-site inspection. This check confirmed the result of the fault analysis and identified a defect in the control valve. A repair could then be carried out. After successfully replacing the valve, trouble-free operation was confirmed using the live data. The end of the constantly high volume flow is clearly visible in the illustration, directly followed by a better spread between the flow and return temperatures when heat is removed.

Conclusion: Error detection in the digital twin for ongoing operational optimization

In this case, the fault occurred in a substation during a period of low outside temperatures during winter. While the increased volume flow of a substation in summer would also have been more clearly recognizable in the measurement data of the feed-in, the defect could have been overlooked without a warning from the digital twin during operation. And in addition to the fault detection, the measurement data also shows that the average return temperature in the overall network was slightly higher in the period between the first occurrence of the fault and its rectification.

On the one hand, this allowed the fault in the building to be detected before an undersupply or other problems would have been noticeable on the customer’s side. And secondly, by rectifying the fault quickly, unnecessary additional operating costs could be avoided. This example shows how our heatbeat Digital Twin can contribute to the efficient and economical operation of local and district heating networks.

We will be happy to provide you with more information about the heatbeat Digital Twin Insights software module in a personal meeting and shortly on heatbeat.de.

The results of this case study were compiled as part of the research project “BeStWärmKI - Optimized operational control of heating networks using AI-based processes”. We would like to thank the Bavarian State Ministry of Economic Affairs, Regional Development and Energy for funding this project.