The Second Digitalization: Ontology and Logs (Towards the Semantic Industry)
Industry produces millions of data points every day. Machine tools, sensors, PLCs, and control systems record their every event in files called logs . The problem? This data often remains trapped in machine language: it describes what happened, but rarely explains why.
The nature of the problem: The trap of proprietary languages
Until now, analyzing logs required an intimate knowledge of the machines, their protocols, and their proprietary formats. Since each system has its own structure and vocabulary, an isolated piece of data becomes very difficult to interpret when comparing it to others.
This is where ontology becomes a game-changer. By linking technical events to objects, processes, and the real constraints of the business, it transforms simple computer traces into genuine industrial knowledge. It doesn't replace existing systems, but introduces a layer of meaning that offers them a second, cross-cutting, and intelligible interpretation. An alarm is no longer a simple error code: it becomes the symptom of a deviation, a quality defect, or a tension in the production process.
The qualitative leap of AI: From anomaly to hypothesis
Combined with ontology, artificial intelligence takes a decisive step forward. Instead of only searching for statistical anomalies, it can now:
Bringing isolated events together: A measurement discrepancy, a temperature variation, and a maintenance intervention suddenly take on a common meaning.
Identifying invisible relationships: Revealing hidden connections between different silos in the factory.
To generate hypotheses: To explain the root causes of a problem rather than simply observing its effects.
The Bosch example : Unifying industrial reality
At Bosch, one of the major challenges lies in integrating data from highly heterogeneous sources (sensors, machines, production systems, management software). To address this, the group developed an industrial ontology called Bosch I4.0 Core Ontology , based on the concepts of the international standard IEC 62264.
This common model allows for the uniform representation of equipment, materials, and manufacturing processes. When a factory needs to analyze the causes of a quality decline on an electronic component line, only this ontological approach allows for cross-systems querying to solve the equation.
The future: The rise of the semantic industry
The industry has gone through two major digital revolutions:
Numerical control (CNC): It has made it possible to program industrial movements with extraordinary precision and repeatability.
Industry 4.0: It has connected machines and systems to make the factory communicative, measurable and predictable.
But a fundamental question remains: is an industry capable of producing millions of data points necessarily capable of understanding them?
The next industrial revolution will consist of unlocking the knowledge locked within machine data. Thanks to ontology and AI, logs cease to be the silent archives of production and become a source of active intelligence, capable of anticipating failures and driving innovation.
Navigating this semantic industry effectively and making your machine data speak becomes possible thanks to La Fonderie , which we developed as part of the NPR 575 project, and which performs this ontological diagnosis in just a few minutes. Your organization's silent archives finally become intelligible.
Don't let the future dictate your life. Shape it. Subscribe to our newsletter so you don't miss any of our analyses on the second wave of digitalization!
