Artificial Intelligence in the manufacturing line for the pharmaceutical industry

How can pharmaceutical manufacturers benefit from the latest machinery control technologies? What if it were possible to develop a Machine Learning model that was capable of predicting critical moments in the production chain or risk of collapse and anticipating the request for spare parts? Or what if you had access to an industrial controller with embedded Artificial Intelligence, which uses pattern recognition to predict an unwarranted deviation in your fill line that you otherwise could never have detected?

These examples that may sound utopian can be a reality thanks to the installation of Artificial Intelligence applications on the production floor. The discipline investigating Artificial Intelligence is already several years old but, thanks to its exponential technological development and growing interest in data, it is becoming one of the most promising technologies in Industry 4.0.

Taking this context into account, the debate on the incidence of Artificial Intelligence in the pharmaceutical industry has also come to the fore. AI is already being used in many pharmaceutical companies, helping in the analysis of big data that allows them to improve their R&D processes and plant management. In addition, AI also offers significant advantages in machine-level packaging applications, allowing manufacturers to redesign their processes to be more flexible and to adapt to any possible change.

Transforming production by implementing Artificial Intelligence can help industry manufacturers extend the life of their equipment, improve OEE, and detect unexpected events to prevent failure.

The importance of operational excellence

Operational excellence is an essential requirement to maximize capital expenditures (Capex). At the same time, pharmaceutical manufacturers face the challenge of leaving behind traditional processes and moving to more homogeneous production cycles, directly linked to production, and at the same time more flexible in order to adapt to ever-changing demand.

In addition, the new pharmaceutical serialization regulations impose greater demands on manufacturers to guarantee the correct labeling of products, more precisely and with all the data necessary for compliance. The customization of the final stage of pharmaceutical manufacturing includes, for example, the inclusion of variable data on labels or the use of previously printed packaging. This allows the manufacturer to customize product labeling for certain markets, customers, or products without having to store individual packaging materials for each possible variant. In this way, the flexible production of smaller batches imposes new requirements for automatic and rapid changes.

Pharmaceutical companies can achieve operational excellence by implementing Artificial Intelligence on the ‘Edge’ or the same manufacturing line. However, they also require flexible and autonomous production support, as well as IIoT automation solutions ranging from data collection to smart production for seamless integration of the IT and OT worlds. In the case of the advances necessary for Industry 4.0, such as predictive maintenance and network generation, efficient production, the use of adaptive algorithms offers enormous potential. Many pharmaceutical companies are realizing that AI offers an opportunity to increase not only the Global Efficacy of Predictive Equipment (OEE) and, therefore,

According to an Aberdeen study (*), although the leading companies in the industry have reached OEE values ​​of 89%, the still use of traditional systems has led to figures of 74%. But what if we go further and add Artificial Intelligence solutions in automation? By improving quality and using predictive maintenance to reduce machine downtime, very significant efficiency gains can be achieved. In any case, the really important thing to know about the OEE figure is to obtain information on the entire process and to be able to alleviate the weak points.

‘Edge and Cloud’ in the pharmaceutical industry

What do we mean by AI in the ‘Edge’? We can define Machine Learning with Small Data as the backbone of Artificial Intelligence. At this level, lines and devices are monitored with real-time sensors, and data is collected and processed at high speed for rapid anomaly detection. To turn information into action, manufacturers need efficient control and monitoring for a more natural and proactive relationship between operator and machine, implementing a technology that brings together all areas of automation, including logic, motion, vision, security, and visualization… This helps companies to increase productivity in a very flexible way.

Big Data processing in the cloud can be described as the ‘brain’ part of AI. This requires open and secure standards, such as the MQTT protocol and the unified architecture OPC communications standard for the safe and easy conversion of machine and system data into high-value information.

Many of the AI ​​solutions advertised on the market, which are often cloud-based, have significant requirements in terms of infrastructure and IT. These solutions also work with an overwhelming amount of data that is laborious and time-consuming to prepare and process. The issue of value added is still a bit tricky for vendors, who cannot determine whether or not the AI ​​investment will generate a rate of return. The fact that system designs for the production industry are generally complex and unique is another factor to consider.

Taking these conditions into account, how are we going to design and integrate an AI that generates tangible added value in the production process? While the cloud is best suited to handle big data and perform massive analytics over the long term, AI on the ‘edge’ is crucial for real-time applications. This approach offers more flexibility and faster response times, so production environments can get better use of data analytics rather than relying on cloud computing. Rather than laboriously analyzing a large volume of data for patterns of behavior, in addition to the processes that are running, it is important to approach things from the other perspective. Technology is necessary when the required AI algorithms are integrated into the machine’s control system,

What does AI in the ‘Edge’ offer to pharmaceutical companies?

With AI at the ‘Edge’, pharmaceutical manufacturers can better control the complexity and safety of processes. Although the amount of data is still enormous, companies need fewer resources in terms of hardware, communication infrastructure, and enterprise-level processing capabilities.

Technology that provides seamless integration into the cloud is required, directly from the machine controller with secure and integrated IIoT protocols. With AI on the Edge, organizations can expand their processing and analysis capabilities with machine-level AI algorithms, benefiting from faster and more efficient information for the decision-making process. In addition, users can analyze and intervene in the process in real-time, exactly where the action occurs, autonomously improving performance with AI and Machine Learning.

By leveraging Artificial Intelligence at the ‘Edge’, pharmaceutical companies get most of their cloud computing with pre-processed and aggregated data at the machine level, reducing the IT infrastructure required for optimal data flow.

How does AI work on the ‘Edge’?

In a packaging machine, a machine controller offers synchronized control of all of the machine’s devices and advanced functionality such as motion, robotics, and database connectivity. An AI-equipped machine controller further fuses machine control functions with real-time AI processing. An AI controller features adaptive intelligence, bringing you closer to action and helping you learn to distinguish normal from abnormal patterns at the machine level.

This type of solution is mainly used in the packaging and production process at the points where the customer is experiencing the greatest efficiency problems (bottlenecks). The processes acquire intelligence based on previous findings and improvements that have been made and, subsequently, lead to a holistic optimization of the entire manufacturing process.

A good example is a bottling application: the bottles are transported through a conveyor belt and filled. The AI ​​controller learns what a normal situation looks like when no failure occurs. In the event of a disturbance, such as a lock causing selective friction, the AI ​​controller will detect these anomalies. The filling process will stop briefly and resume when the process stabilizes. This prevents, for example, liquid spills from occurring on the packaging line, which in turn helps reduce machine downtime and waste.