Merger of Substring and LeanBI

We are delighted to announce an important milestone for our company. From 01.01.2024, LeanBI and Substring will go hand in hand into the future as Substring AG. The name LeanBI will continue as a brand.


This merger enables us to bundle our competencies and offer you a comprehensive range of services in the field of data, AI & digitalization. Substring brings impressive expertise in the areas of Business Intelligence, Data Warehouse and Lake House.

Blog: New optimization potentials thanks to AI and physical models

New optimization potentials thanks to AI and physical models

Physical models and Artificial Intelligence can work together to simplify and accelerate the development of complex products and processes. In this blog, we continue to show possible applications of the two technologies.



The development of products can be tedious, time-consuming, and costly. Artificial intelligence and Machine Learning can accelerate the development process. With the help of simulations, efficient and optimal solutions can be predicted more quickly. The technologies can find different applications. Physical models can feed the AI, the AI can provide data for physical models, or both run in parallel – hybrid models.


In this second blog post, we show you three more application scenarios for the use of artificial intelligence and machine learning:


Damage detection of structures is carried out using the finite element method (FEM). With the help of a structural analysis, the stress distribution is identified, for example, within tower and bridge constructions. Monitoring can be carried out directly on the construction or with the help of imaging techniques such as drone cameras. Both methods are supplemented with artificial intelligence. This makes it easier to identify risk levels of damage and to derive necessary repair or renovation measures.


In the field of indoor presence control, the use of sensors can be used to record CO2 profiles while maintaining data protection. Since rooms have different ventilation characteristics, these must be taken into account in the evaluation of results. AI analytics can calibrate physical models to the specific conditions of the respective interiors and thus achieve meaningful results.


AI models can also support the training of robotic arms. With the help of Reinforcement Learning, gripping strategies are learned so that robotic arms can grasp objects of various shapes in an undefined position. Costly real test runs with a robotic arm can be substituted by simulating the robot movements through AI models.


The combination of artificial intelligence and physical models can facilitate data collection in many application scenarios and at the same time generate new areas of application.


Have we piqued your interest?


We would be happy to advise you on this topic and show you how you can use artificial intelligence and physical models in your area of application.

Blog: Acceleration of development through AI and physical models

Acceleration of development through AI and physical models

The development of products and processes can be cost- and time-intensive. Artificial intelligence and physical models can contribute to simplify and accelerate development. In this blog, we will show you two application scenarios.



Aircraft wings should not only provide the aircraft with sufficient lift and stability, but at the same time be designed to minimize air resistance. The development resembles an aerodynamic marvel that can only be achieved by using vast amounts of data. Artificial intelligence and physical models can make the development process faster and more efficient. With the help of physical simulations, the optimal shape of aircraft wings can be predicted.


In this blog, we will show you in which two scenarios the combination of artificial intelligence and physical models is used and how this opens new optimization potential:


For production planning, the interaction of AI and physical models can be the decisive added value for companies. The optimal utilization of production lines and their units is a complex process that can be simplified by artificial intelligence. Production planning is simulated and set up on a daily basis based on different situations. In this way, a high number of variants, such as in a paint shop in automotive production, can also be covered and planned accordingly thanks to AI.


In the field of predictive maintenance, critical components can be monitored with the help of artificial intelligence and Machine Learning. Sensors detect the smallest damage to the components and can filter them out at an early stage. Thanks to AI and ML, noise due to environmental influences can be excluded if they are fed with physical algorithms.


AI/ML models open completely new practical options for application that can be used in a wide variety of processes.


In the second part of our article “New optimization potentials thanks to AI and physical models”, we will explain which other application scenarios are possible with AI/ML and physical models.

Blog: Data analytics solutions to improve the OEE key figure

Data analytics solutions to improve the OEE key figure.


Overall Equipment Effectiveness (OEE) is the most important indicator for industrial companies. Many companies rely on data analytics solutions to optimize the overall equipment effectiveness. In this blog, we show how an improvement of the OEE key figure is possible with a digital Retrofit.



Optimizing effectiveness is one of the highest priorities for many companies. Overall Equipment Effectiveness (OEE) measures whether machines and systems are running effectively. Composed of the maximum availability, quality, and output, this determines the overall effectiveness of plants and machines.


With the help of data analytics solutions, the generated data from machines and systems can be recorded and evaluated by machine learning algorithms. This gives companies an overview of the status of plants and machines and enables them to predict malfunctions or plan maintenance work.


In the second part, you will get to know other data analytics solutions that contribute to improving the OEE key figure. Thanks to a digital retrofit of machines and systems, these solutions can also be used in legacy systems:


Predictive quality applications use the collected data to predict malfunctions in machines and systems that could have a negative impact on product quality. By intervening promptly and changing process parameters or machine settings, industrial companies can avoid the occurrence of scrap and reduce added costs.


Predictive performance applications draw on the detected anomalies of predictive maintenance and predictive quality. The data is used to continuously optimize and train the machine learning algorithms. This enables continuous improvement of performance.


Legacy systems pose a particular challenge in the realization and implementation of such data analytics solutions. With the help of a digital retrofit, machines as old as 20 or even 30 years, can be upgraded with sensors and data interfaces. In addition to conventional vibration, temperature and current measurement technology, optical and acoustic sensors can also be installed. Industrial companies can rely on proven technical frameworks to implement data analytics solutions such as predictive maintenance, predictive quality, and predictive performance to improve overall equipment effectiveness.


Have we aroused your interest?


We would be happy to advise you and show you various ways in which you can improve your overall equipment effectiveness with a retrofit of your machines.


Blog: Digital Retrofit and Overall Equipment Effectiveness

Digital Retrofit and Overall Equipment Effectiveness

To optimize overall equipment effectiveness, companies are turning to data analytics solutions. For this, they do not necessarily need modern machines with built-in sensors. In this blog, we show you how a digital retrofit of your machines can contribute to optimizing overall equipment effectiveness.



Industrial companies rely heavily on the Overall Equipment Effectiveness (OEE). Based on the metrics availability, quality, and output, this most important key figure measures the overall effectiveness of plants and machines. At the same time, conclusions can be drawn about the performance of the production processes. Optimizing this key figure is inevitably one of the main goals of companies.


Data analytics solutions offer excellent opportunities for this. The data from machines and systems can be recorded and evaluated with machine learning algorithms. In this way, predictions can be made that not only enable timely intervention in the event of faults, but at the same time contribute to the continuous improvement of OEE.


In this blog, we first focus on predictive maintenance applications. We will show you how all OEE key figures can be covered with the help of these applications:


Predictive maintenance applications can predict the wear of machine and plant parts from data generated by these machines and plants. This allows companies to detect the wear and tear of parts before planned maintenance work and take maintenance measures in suitable time. In this way, unplanned downtimes of machines and entire production lines can be prevented and fines due to missing goods are avoided.


One challenge in the realization of these applications are legacy systems. Compared to modern machines and systems, they do not have the necessary sensors. With a digital retrofit, these production systems can be retrofitted with sensors and the corresponding interfaces for data acquisition.


In the second part of our article “Data Analytics Solutions for Improving the OEE Key Figure”, you can read which other data analytics solutions can contribute to the optimization of the OEE key figures and with which sensors legacy systems can be equipped.

Blog: Predictive Quality: Edge or Cloud Solution?

Predictive Quality: Edge or Cloud Solution?


Predictive quality applications are becoming increasingly important in production.  The cloud offers a wide range of new possibilities, but when it comes to speed, it makes sense to rely on edge computing. In this blog, we show you how to use edge systems despite their lower computing power.



With predictive quality applications, companies can continuously optimize the quality of their products and processes. This also makes it possible to identify quality problems. The prerequisite for this is a machine learning process that can quickly predict problems such as jams on a machine or overheating of an electric motor. The analytics calculations of the algorithms required for this can be carried out both via edge computing and via a cloud platform. In the first part “Predictive Quality and Edge Computing“, we have shown you reasons why it is not advisable to use a cloud platform for time-critical use cases.


In this part, we explain how you can use edge systems despite lower computing power by optimizing machine learning models:

  • Pruning: By simplifying, shortening, and optimizing the decision trees, the complexity of the input parameters can be reduced, thus requiring less computing power.
  • Frameworks: The range of frameworks is wide and offers different advantages and disadvantages. With a framework such as Google’s Learn2Compress, the layers of deep learning models can be reduced in their breadth and depth, thus saving further computing power.
  • Programming language: Machine learning models are usually written in the Python programming language. However, Python requires high processor and computing power. Compilers can be used to translate the code into a language that requires less processing power, such as C#.

With these tips, machine learning models can be optimized, and edge systems can be used. However, the efficiency of the cloud cannot be made possible by an edge solution. This is a major disadvantage, because the cloud often requires less computing capacity due to load balancing and is therefore more cost-effective than the edge solution.


Whether you choose a cloud or edge solution should depend on each specific case. A combination of the two solutions is also conceivable and possible. For example, the data can be prepared and pre-filtered at the edge and further processing can be done in the cloud.


Do you have any questions about the different options? Please feel free to contact us.

Blog: Predictive Quality and Edge Computing

Predictive Quality and Edge Computing


Everyone is talking about the cloud, but especially when things need to move fast, it is good to rely on edge computing for predictive quality applications. In this blog, we will show you the advantages of edge computing over the cloud.



Predictive quality applications give industrial companies the opportunity to optimize both the quality of products and the process flows.  At the same time, quality problems that arise at short term can also be identified. For example, an imminent blockage of a  valve or a suddenly occurring unusual behavior of an electric motor of a machine can be detected. Immediate intervention, for example by changing the process parameters or machine settings, can therefore often prevent major damage, provided that the used machine learning method can predict these problems as quickly as possible.


We show you why, for time-critical use cases, outsourcing the analytical computations of algorithms to a cloud platform can be problematic:


  • Data transmission to the cloud platform: The sensor measurement data and process parameters recorded by the machines and systems must be transmitted to the cloud platform for the calculations. This results in time delays. In extreme cases, this delay can be too big and results in the inability to react and intervene in time.
  • Data exchange with the cloud platform: In contrast to edge computing, the exchange with a cloud requires a stable Internet connection. Companies may find it difficult to ensure this, as production and manufacturing halls are often located in remote locations. As a result, companies are constantly exposed to the risk of a connection failure, which in turn inhibits timely intervention.

With analytics calculation directly on the machines and systems, companies can avoid these problems. The hardware for an edge solution is increasingly available as there are more and more sensors with their own processing devices and industrial computers today are more often equipped with high-performance processing units.


The choice between an edge or cloud solution must be made individually, there is no general recommendation. Companies should consider important factors such as response time, network availability and costs when deciding. We are happy to support you in this.


Find out how edge solutions can be used despite their lower computing power and what major disadvantages still exist compared to the cloud, in our second blog „Predictive Quality: Edge- or Cloud-Solution“.

Publication: Computer vision in industrial companies

LeanBI in the media on the topic:


Computer vision in industrial companies


The quality of computer vision systems strongly depends on the recording technology used. Each industrial company has its own requirements and conditions to consider.


With our checklist you will find the right solution.
 Computer Vision in industrial companies



Blog: Digital retrofit for legacy systems – paving the way for Industry 4.0


Digital retrofit for legacy systems – paving the way for Industry 4.0

Machinery and plant parks are indispensable for many companies, but many of them are no longer state of the art. These so-called legacy systems have a shortcoming, they lack the connectivity to integrate them into digital workflows. In our blog, we show you the positive effects of a digital retrofit on three application scenarios.


Leidensdruck Retrofit I


As digitization progresses, many companies in the industrial environment are facing a major challenge: How can legacy systems be integrated? They lack two essential factors: the data interface and the sensors for collecting the data. A digital retrofit enables upgrading the legacy systems with the necessary sensor and control technology. This not only increases performance, but also improves the quality rate.


The application scenarios for a digital retrofit are broadly diversified. We have already given you a first insight in our blog «Integration of legacy systems through digital retrofit». Using three further examples, we will show you the practical use of the retrofit and its positive effects:

  1. Metal production: Detecting quality fluctuations in the manufacture of metal products such as tubes, components or body parts at an early stage is crucial. Thanks to a digital retrofit with which old systems are equipped with the appropriate sensors, fluctuations in production can be minimized. Production control can be designed to be reactive; adapting to fluctuations measured by the sensors. As a result, continuous traceability can also be ensured throughout the entire production process. The combination of decentralized data acquisition and centralized data analysis can also contribute to scrap reduction and more cost-effective production.
  2. Surface treatment: By combining digital retrofitting with predictive maintenance, wear times for machining tools can be better predicted. The use of additional sensors paired with the analysis optimizes the processes in several ways: on the one hand, highly stressed components can be used longer in surface treatment, on the other hand, this reduces material consumption and downtime. Productivity can thus be increased.
  3. Electronics production: With the upgrade of legacy systems through a digital retrofit, highly complex, AI-supported test methods can be used for the early detection of component failures in the production process. Not only does this increase productivity rates, but it also simultaneously helps to understand and learn to fix emerging anomalies.

The sustainable modernization technology retrofit is an important enabler for Industry 4.0. Without them, valuable and in some cases indispensable legacy systems and production facilities would be lost, which would be accompanied by an immense loss of value.


Have we aroused your interest? Then contact us and we will be happy to answer your questions.

Blog: Integration of legacy systems through digital retrofit


Integration of legacy systems through digital retrofit

To be able to digitize machinery and plant parks without carrying over any burden? Which company does not hope for this? Often it is not a matter of starting from scratch, but a question about the conversion and integration of various legacy systems into digital workflows. In this blog, we show you the positive effects of a digital retrofit using three exemplary application scenarios.


Leidensdruck Retrofit I


Digitalization is in full swing – one of the biggest stumbling blocks in the industrial environment is the integration of legacy systems. Two decisive prerequisites to digital operations are missing, thus hindering optimization of planning and changeover times, increase of performance and availability, improvement of the quality rate and implementation of predictive maintenance: The sensor technology for collecting data as well as an interface for forwarding it is missing. With a digital retrofit, legacy systems can be equipped with the necessary sensor and control technology.


In the first part, we show you three scenarios for practical use and the potential positive effects of a digital retrofit:

  1. Intralogistics: With the booming online trade, the shipping volume has also risen rapidly, accompanied by the high stress on logistics facilities, which cannot keep up with this growth due to their (partially) analog structures. In contrast to a retrofit of existing systems, the construction or acquisition of additional buildings and/or systems cannot be achieved at the same time. This means that a digital retrofit is not only the faster, but often also the more cost-effective and resource-saving option. As an icing on the cake, the service life of the existing systems is also increased.
  2. Assembly lines: Many product suppliers such as automotive suppliers or bicycle manufacturers are faced with major challenges: How can delivery times be met despite high demand and at the same time growing number of product variants? A failure in the production chain leads to complete plants coming to a standstill, delivery dates not being met and thus high penalty payments. Through predictive maintenance, possible failures can be detected at an early stage. Thanks to the retrofit of legacy systems, making this possible in the first place, delivery reliability can be ensured, production output increased, and customer satisfaction enhanced.
  3. Packaging industry: The fact that many machines are connected in series distinguishes the packaging industry, but also carries the risk that a complete line will come to a standstill in the event of a failure. Here, too, unplanned downtime and planned maintenance work can be reduced with the help of retrofitting and predictive maintenance. For example, legacy systems can use sensors to automatically check the packaging quality of different products and thus reduce the reject rate.

With a digital retrofit, not only the important prerequisites for the integration of legacy systems into digital workflows can be fulfilled, but also the way for Industry 4.0 can be prepared We’d be happy to support you.


Would you like to know more about the integration of legacy systems through a digital retrofit? In our second blog post «Digital retrofit for legacy systems – paving the way for Industry 4.0», we outlined further application examples.