Vertech’ has been specialising in a glass Manufacturing Execution System (MES) for over 25 years. Its CEO Ulas Topal* discusses the use of AI in the glass industry.

For several years, the glass MES market has been seeking for leverage artificial intelligence (AI) to improve product quality, to optimise production processes and to reduce costs. However, the use of AI is still limited. Nevertheless, the recent AI emergence has brought this topic to the forefront, prompting us to consider the risks and on the opportunities associated with AI adoption for glassmakers.

It is important to clarify that when speaking about AI in this article, we are referring to a solution that imitates the human cognitive abilities and even surpasses it substantially. We are not talking about conventional data analysis systems that have been used for a long time, such as statistical solutions used to determine machine maintenance needs.

It is crucial to acknowledge that the potential of AI in the glass industry goes beyond traditional data analysis systems, as it harnesses the power of machine learning and advanced algorithms to unlock unprecedented insights and capabilities for optimising glass manufacturing processes.

An on-going process

To be effectively used in the glass industry, AI must absorb a vast amount of data, and one of the main risks associated with AI in glass MES is related to the quality of data used for machine learning. As an MES software provider, Vertech’ is highly aware of this issue. If the data is biased or incomplete, it can lead to inaccurate predictions and to erroneous decisions.

To envision a beneficial use of AI, it will be essential to ensure the quality and the maturity of the data collected initially and to ensure that they are sufficient to feed the AI system. This is not a one-time effort, but an on-going process. Continuous monitoring and maintenance of data quality will be imperative to avoid potential pitfalls and maximise the effectiveness of AI-driven solutions in glass MES, allowing for real-time optimisation and informed decision-making throughout the production cycle.

The glass packaging manufacturing process is very complex, and currently an AI capable of producing the perfect ‘recipe’ to manufacture a glass bottle doesn’t exist.

There are too many variables to consider, such as the glass composition (the quality of the raw materials used), the temperature settings across multiple elements throughout the production chain and the quality induced by production machines.

Other subtle factors also impact production, such as air humidity, external temperature, or atmospheric pressure.

Before considering the involvement of AI in defining the glass production recipe, a more advanced digitisation of the production process is necessary to understand all the production factors.

The expertise and tacit knowledge possessed by skilled operators in the glass industry play a crucial role in ensuring the quality and precision of the manufacturing process. It’s challenging to fully replace their skills and intuition with AI technology alone as glass professions follow much fewer rigid rules than other disciplines. While it is possible to replace a doctor with AI quite easily, the role of AI in the glass industry’s Industry 4.0 is still uncertain.

A symbiotic relationship

Another potential risk is the dependence on AI. If operators rely too much on AI for decision-making, they may lose the ability to react quickly to changes and unforeseen circumstances. While AI can provide valuable insights and assistance in decision-making, it should not replace the knowledge and craftsmanship of experienced glass professionals.

Maintaining a symbiotic relationship between skilled operators and AI systems is crucial to leverage the benefits of both and ensure the preservation and the continuous development of glass craftsmanship. The use of inspection machines highlights the importance of human control coupled with machine controls. It is important to maintain a certain level of flexibility in the company’s processes and decisions and thus segment the use of AI.

Despite all these risks, AI can contribute to the identification of complex patterns in data production, predict equipment failures (predictive maintenance), and optimise real-time production processes through a high-performance MES.

However, it is crucial to not consider AI as a mere statistical database that triggers alerts, it should bring much more than that. By leveraging its cognitive abilities, AI must be able to detect inefficiencies and propose improvements at a speed and quality currently inaccessible to humans.

Another aspect of AI needs to be considered, it has the potential to revolutionise the glass industry by unlocking new levels of innovation and creativity. By analysing vast amounts of data and uncovering hidden insights, AI can inspire breakthrough ideas and enable glass companies to push the boundaries of design, sustainability, and product development.

In conclusion, while AI holds the potential to bring significant advantages to glass companies, its practical impact on improving glass production remains uncertain. Vertech’ acknowledges the evolving landscape of AI and looks forward to assessing how its expertise and the vast repository of glass data align with AI outcomes.

However, the realisation of a tangible AI impact in the glass industry appears to be a distant prospect at present. The technology must undergo further maturation to be effectively harnessed in conjunction with glass MES, in a synchronised and integrated manner.

*Vertech’, Chalon-sur-Saône, France,