August 8, 2025

12 MIN.

Data science for domestic hot water installations – how advanced analysis saves energy and resources

When considering building energy efficiency, we often think first of heating, insulation, or lighting, but domestic hot water (DHW) preparation is frequently overlooked, despite being a silent but significant contributor to energy consumption. This is especially true in public and commercial buildings such as hotels, sports facilities, or care homes.
In well-insulated new builds, where energy costs for heating are falling, DHW now accounts for an increasingly large proportion of the total energy consumption. In the planning phase, basic assumptions are often made about hot water requirements – user behavior, demand profiles, and draw-off quantities and times – that are not reflected in the real world. An approach based on modern data science and data analysis paves the way for more energy-efficient system sizing without compromising on performance.

DHW performance vs. sustainability – A trade-off

The requirements for DHW systems are high.

• They must work reliably and hygienically, even at peak times, such as in the morning when many people shower in a hotel, or at a youth tournament in a sports facility.
• At the same time, there is growing pressure to conserve energy and design more sustainable systems – for example, when a heat pump is used, as they are particularly efficient when the flow rate and temperature differences are low.

Between these two aims – reliability and sustainability – planners face a central challenge. How can realistic assumptions about the actual user behavior in the specific building be made when it varies greatly depending on the building type, size, fixtures, and other factors?

Standards & guidelines – not a perfect guide for planning

An important starting point in planning is the DIN EN 12831-3.
which provides methodical guidelines for the design of DHW heating systems. This standard recommends designing systems based on the maximum volume flow averaged over 60 seconds.

The problem is that,
in actual use, demand peaks often last considerably less than 60 seconds and are are therefore somewhat “absorbed” by the calculation. To compensate for this, specifications contain generous assumptions about how much hot water could be needed at any one time.

The result is
that many systems are significantly oversized. This leads to unnecessarily high energy consumption, high resource consumption, and, in the long term, increased costs due to the provision of too much hot water. Additionally, storing large quantities of water that are used infrequently can cause hygiene issues.

A new approach with precise data analysis: Using scenarios instead of maximum values

The good news is that
there is an alternative.
In the current research project, Technical requirements of instantaneous water heaters increasing energy efficiency and comfort of big, renewable heat supply systems by the Institute for Solar Energy Research (ISFH) in collaboration with data scientists from TRIOVEGA, real usage patterns and demand profiles are understood in much greater detail by using shorter measurement intervals – with 10 seconds identified as the ideal averaging interval.

The project revealed that most peak loads are extremely short, in the range of a few seconds.

• Which raises the question: Do each of these peaks really need to be covered by increased system sizing?
• Or is it enough to guarantee full performance in 99.9% of cases?
• What happens in the 0.1% of cases not covered?

By considering this, it becomes possible to design smaller heating systems that better align with practical requirements, without users noticing any perceivable difference due to the thermal inertia of the water system. Instead of relying on overly cautious standard load profiles, simulated usage scenarios based on measured data allow for system design that reflects actual demand, rather than assumptions.

It also becomes clear how much the demand profiles differ depending on the building type. While sports facilities often experience occasional peak loads, for instance during a weekend tournament, hotels tend to have more regular but less extreme peaks in mornings and evenings. Blanket planning is insufficient here.

Conclusion: Embrace customized solutions

What can we learn from this? The path to an energy-efficient DHW system is through data, not standardized blanket values. Anyone who understands and analyzes the actual tapping patterns of a building can work with leaner systems, reduce costs – and contribute to decarbonization.

Our data scientists from service.factoryINSIGHTS provide the decisive expertise: As in the TA-DTE-XL research project, we work closely with our customers to develop concrete recommendations for action, based on the individual data situation of their project

We combine our many years of experience in the industrial environment with modern data analysis methods – whether basic statistical evaluations or innovative machine learning and AI methods – to identify the optimal solution in every scenario.

In an individual consultation our experts will be happy to show you how modern data analysis and data science can be used to increase efficiency and optimize processes in your industrial company.

Autor: Dr. Matthias Zahn

Dr. Matthias Zahn is a Senior Data Scientist at Triovega GmbH. He is responsible for the statistical analysis of customers’ production data and the development of predictive models for use in industrial production — enabling customers to increase their production efficiency permanently by leveraging data-based process improvement.

Book your individual consultation here!