Five use cases for Machine Learning in Industry 4.0

from | 30 May 2019 | Basics

Machine Learning in Industry 4.0 is one of the key drivers and an enormous opportunity for economic development. In this article, we therefore look at five concrete use cases for machine learning.

Machines learn to think: the use of robots, sensor technology, Big Data and artificial intelligence are making machines in industrial production smarter today in Industry 4.0 than ever before. In particular, the data science method Machine Learning is becoming increasingly important due to the enormous advances in data processing and computing speed in recent years.

Machine learning is even becoming the norm in certain areas of industry. Innovation driver. On this occasion, we present here five concrete and illustrative Use cases for Machine Learning in Industry 4.0.

Reading tip: In these articles we go into Machine Learning methods.

1. smart manufacturing: better understanding and controlling the production process

With Data Science-methods such as machine learning, individual Production processes view and transform in a new way. For this purpose, data is collected and evaluated as part of the production process. This allows individual processes to be better understood and subsequently optimised.

For example, one of our customers wanted to improve the painting process for car parts. The problem was that incorrect painting made a lot of manual reworking necessary. Our solution was to digitally record the painting process. On the basis of Training data on paint thickness, PH values and drying times, the process as a whole could be evaluated and optimised with regard to specific targets such as gap dimensions. Machine learning algorithms bring two major advantages to the production process:

  • Improving the quality of the products
  • Flexibilisation of the production process

Data evaluations can thus lead to processes being continuously adapted to the current production conditions. Smart Manufacturing is characterised accordingly by the fact that optimisations are carried out automatically and adjustments can be made at the level of individual components.

2 Predictive Maintenance: The intelligent, predictive & preventive maintenance.

Components such as sensors are not only becoming smaller and smaller, they can also be used more and more cost-efficiently. This makes the Monitoring of machines is becoming increasingly interesting. Many thousands of measuring points within a vehicle, a machine or an entire Machine parks can be monitored in this way.

Up to now, it has been the case in many cases that the measuring accuracy of individual sensors was very strongly influenced by other environmental influences. The more points of a machine with Sensors equipped, the more accurate the Measurements. The more Data are collected, the better. In particular, incorrect measurements from individual sensors can be identified and evaluated more clearly if two sensors measuring at similar points provide different values.

Sensors provide a view inside machines

Sensor data provides valuable information about the condition of machines. Over the course of time, an accurate picture of the "healthy" state of machines emerges. Machine learning algorithms can then be trained with the data sets from this healthy state. In the many petabytes of sensor data, the machine learning algorithms search for patterns that indicate malfunctions or the possible failure of parts. In this way Machines are repaired even before they are defective (Predictive Maintenance).

Reading tip: In this article we look at the Potential of Augmented Reality and Mixed Reality for Preventive Maintenance.

3. optimised energy management thanks to machine learning methods

Climate change and Energy transition are two of the greatest current challenges - not only for politics and society, but also for industry. Even today, the energy market is more complicated than it was just a few years ago. The current energy mix of conventional and renewable energy sources such as wind leads to fluctuations in the power grid. Electricity providers face penalties for both overproduction and underproduction, which must be avoided at all costs (cf. chart).

Data science methods such as machine learning make an increasingly complex energy market manageable. To ensure that demand can always be optimally satisfied, it is necessary to keep a close eye on both the framework conditions of energy production and the expected consumption. For this task, in which Knowledge from experience must be derived, machine learning offers itself as the ideal solution. Machine learning algorithms help to match demand and supply, or to Anomalies in power consumption to recognise. This brings three major advantages:

  • Based on historical energy consumption patterns, the projected demand can be derived
  • Intelligent control ensures a price-optimised strategy for power generators
  • Production control is possible in real time
Cost savings through lead forecasting

Cost savings through lead forecasting

4. "Test automation 2.0" brings the reversal of conditions in quality control

In the past, the quality of products was judged by End of the production process checked. Through the use of sensor technology and the continuous evaluation of data at component level, the quality of workpieces can be checked and ensured during ongoing operation.

A key to this new form of Quality control is the use of Machine Learning. Its great strengths can be found in this group of Algorithms especially in environments where not only individual predefined sources of error are to be investigated. Through their Ability to learn machine learning algorithms can detect previously unknown sources of error in the data. Especially in the Mechanical Engineering Machine Learning is therefore becoming increasingly relevant in Industry 4.0.

Autonomous vehicles in manufacturing and logistics are inconceivable without machine learning.

Without machine learning, developments in the field of the autonomous driving hardly conceivable. The capabilities of machine learning to learn rules independently and to assess new, unforeseen situations on the basis of these rules are what distinguish these algorithms.

Link tip: Learn more in this article about the connected car.

Road traffic is only one of many environments in which new situations constantly arise that have to be assessed on the basis of the trained rules. Through autonomous vehicles many systems within the industry are being completely restructured.


By loading the video you accept YouTube's privacy policy.
Learn more

load Video

Autonomous systems transform the industry

One of the most prominent examples of the transformation of the Manufacturing process by autonomous vehicles is production in the networked factory. The exact demand for material and placement can be perfectly coordinated and partially automated. Also the entire area of Logistics can be taken to a whole new level through machine learning and controlled more efficiently than ever before.

In this article you will learn more about the differences between Brownfield and greenfield plants.

One of the most important trends within machine learning that contributes to autonomy is Deep Learning. The level of intelligence that can be achieved with Deep Learning is necessary for autonomous driving vehicles to correctly recognise and interpret their environment.

[bctt tweet="Autonomous driving is one of the key use cases of #MachineLearning in #Industry4.0″ username=""]

It's time for Machine Learning in Industry 4.0

Whether large corporation or medium-sized company: Machine Learning in Industry 4.0 is one of the most important trends in the coming years. The prerequisites for the success of data science methods have been created in recent years. Favourable data processing and large data volumes form the optimal framework for the application of machine learning in Industry 4.0. In this way, much previously unused information can become part of the value chain and be used for the digital transformation of companies.

On the one hand, it is important to find concrete application possibilities as early as possible. Since machine learning algorithms are learning algorithms and not out-of-the-box solutions, it takes time for the solutions to develop their full effect.


Michaela Tiedemann

Michaela Tiedemann has been part of the Alexander Thamm GmbH team since the early start-up days. She has actively shaped the development from a fast-moving, spontaneous start-up to a successful company. With the founding of her own family, a whole new chapter began for Michaela Tiedemann at the same time. Hanging up her job, however, was out of the question for the new mother. Instead, she developed a strategy to reconcile her job as Chief Marketing Officer with her role as a mother.

0 Kommentare