Skip to main content

How your energy management benefits from artificial intelligence


Automated consumption monitoring and pattern recognition in IngSoft InterWatt

17.11.2022 6 Minutes

When talking about Artificial Intelligence (AI), many buzzwords come up: Machine Learning (ML), Deep Learning (DL), neural networks, Big Data and many more. Their meaning or the distinction between them is not always clear. Often, the terms are only used as attention-grabbing buzzwords. However, if you take advantage of the actual methodical approaches behind them, a wide range of opportunities open up – also in energy management. We will show you how you can use automated consumption monitoring and pattern recognition to set up your energy management intelligently with IngSoft InterWatt – without any prior knowledge of AI.

What does AI mean in energy management?

Basically, artificial intelligence refers to methods or techniques that transfer intelligence or intelligent behavior to computers or machines. In energy management, machine learning approaches are generally used – a subfield of AI. ML is composed of algorithms that determine regularities from available data in the training phase. These can then be applied to unknown contexts. A distinction is made between supervised and unsupervised learning.

While the learning process in supervised learning is based on predefined training data, unsupervised learning determines regularities without result templates – and is thus significantly more demanding.

Pattern recognition in IngSoft InterWatt belongs to unsupervised learning and is based on the application of k-means algorithms. The clear advantage: application flexibility. It does not require similar data sets with known results to determine structures for pattern recognition.

How can your energy management benefit from the use of Artificial Intelligence?

IngSoft InterWatt offers in-depth functions to automatically monitor energy consumption in your company using artificial intelligence and to identify anomalies in a self-learning manner. Deviations are automatically detected and classified.

If anomalies are detected in your company's energy consumption, they are transmitted by the system directly to the responsible persons so that they can take appropriate measures in a timely manner to secure or improve your company's energy efficiency. The resulting time and cost savings are immense – especially if the systems to be monitored are very large.

How can you use the automated consumption monitoring and pattern recognition in IngSoft InterWatt?

Pattern recognition can be applied to all properties, consumers, meters, key figures and energy factors that exhibit rhythms of some kind.

First of all, it does not matter whether you have a very large amount of energy data or only a small amount of data so far. If the inventory is small, the software will first use what is called "adaptive mode" to identify and expand new patterns. As soon as sufficient data is available, the pattern recognition switches to the classic mode.

For consumption monitoring, several patterns are always identified during the calibration phase, which subsequently support the testing of new values. For this purpose, the software automatically selects a suitable calibration period for you from the existing data and independently builds mathematical models for each recognised pattern.

In the event of a permanent change in consumption, e.g. due to a new consumer in the system or other major changes, a new calibration is automatically triggered. Thus, the changed patterns are identified and consequently the adjusted monitoring is continued easily and reliably.

It is also possible to pause the monitoring with a simple click, in case alarms are undesirable due to special events (e.g. company party or repair work).

You want to go deeper into the topic of pattern recognition? No problem! For interested people and tinkerers there is the possibility to let off steam and to take the automatic processes, like the determination of the calibration period, into their own hands.

What are the advantages of pattern recognition in IngSoft InterWatt?

  • Application flexibility: Since pattern recognition in IngSoft InterWatt is based on unsupervised learning, it does not require training datasets to identify structures for pattern recognition. Therefore, you can apply the method to your different consumption data without much effort. Our software reliably identifies completely new patterns of results as well as anomalies in your consumption.
  • Flexible evaluation options: You can decide for yourself to what extent you want to apply pattern recognition: e.g. for individual meters or per building or property. If this is too time-consuming for you, you can specify that pattern recognition be created and activated automatically. This also applies to elements that are created subsequently.
  • Simple operation without prior knowledge: Activating intelligent monitoring does not require any configuration or deep understanding of how it acts. You can simply get started.
  • Comprehensive display and documentation options: When monitoring is active, the evaluation can be displayed in both tabular and graphical form. The graph shows both consumption and all associated models at a glance. This makes it easier for you to identify and classify any anomalies. Classification" in this case means the evaluation of an anomaly: e.g. as a false alarm or as a permanent/temporary change in consumption.
  • Automatic consideration of influences: The dependency on various external factors is automatically detected by the system, including weekday types, vacations, public holidays or the outside temperature. For example, incorrect messages about a sudden sharp increase in heating energy consumption can be avoided if it is based on an abrupt drop or constantly low temperature.

What is the advantage compared to limit value monitoring?

Common monitoring mechanisms are very limited compared to intelligent consumption monitoring. With the usual limit value monitoring, limit values have to be determined individually and manually for the different consumptions or envelope curves with corresponding limit values as well as depending on different influencing factors have to be created. The effort for a large number of meters is high.

Moreover, this methodology is generally very limited in the detection of different types of errors. Temporal anomalies are not detected if the data are still within the limits. Isolated outlier values, however, are. This is because: limit value monitoring reacts to every limit violation. This means that you have to check even one-off or very short events for relevance. Not so with pattern recognition monitoring: It offers the advantage that it filters out such meaningless anomalies for you.

Take the first step towards intelligent energy monitoring now

The pattern recognition monitoring in IngSoft InterWatt is a highly functional composition of very complex algorithms. But don't worry – it is uncomplicated and easy to use and enables reliable evaluation of all relevant anomalies within your energy or substance system. In this way, it ensures the energy efficiency of your company and uncovers potential for additional savings.

Would you like to use our artificial intelligence in your energy, environmental or climate management system? Feel free to call us at +49 (911) 430879-100 or send us an email for more information:

Inquire now

Don't miss anything! Subscribe to our newsletter now:

By pressing the button "Subscribe to newsletter now" personal data will be collected. We will process these in accordance with our privacy policy.
InterWatt: News categories

IngSoft is certified according to ISO 27001

Careful handling of customer data and the security of information and systems used are top priorities for us. Therefore, we act according to the…

User reports

Success story: This is how the city of Kiel achieved 20 per cent energy savings

Markus Noldin and Jan Scheer, energy managers of the city of Kiel, told us in an in-depth interview how they achieved around 20 per cent energy…