9621 Agnes Crossing, Lake Suzanneview, New Mexico Island 84604-9295.
One already has digital twins in production at various locations, the other plans to do so by 2030. Other pharmaceutical companies are still exploring what is possible. Cognizant
One already has digital twins in production at various locations, the other plans to do so by 2030. Other pharmaceutical companies are still exploring what is possible. Cognizant and CESPE bring together leading experts to discuss what is possible and how to scale knowledge. Everyone can learn something there.
Copper mortars, century-old books and glasses fill the decorative wooden pharmacy cabinets in the large meeting room of the Faculty of Pharmaceutical Sciences at Ghent University. Cognizant and CESPE, it Competence center for sustainable pharmaceutical technology and manufacturingThey couldn’t have chosen a more suitable location for their exclusive discussion Digital twins. The contrast is almost tangible: old pharmaceutical techniques form the background for lectures and discussions about the latest technologies.
Cognizant, together with CESPE, wants to inspire the pharmaceutical industry and show what is possible with the right technology. The event’s focus is on life sciences, but the challenges and lessons highlighted today could resonate across industries. Ultimately, digital twins as a technology are one of the top ten trends in manufacturing across all industries. Who doesn’t want more agility, flexibility and robustness in their organization?
A digital twin is a virtual replica of a component, system or process created from data. The purpose of Gemini is to display, predict and understand real-world performance. That’s a nice, but also broad definition.
Specifically, a digital twin is a model based on mechanistic information and data that represents a process in the physical world. For this discussion, think of a digital version of, for example, a chemical factory, a bioreactor, or a tablet production line.
Mechanistic equations describe physical, biological and chemical processes explicitly and as precisely as possible. They take physical laws into account, but are complex and never truly complete because they are based on inherent assumptions. All of reality does not (yet) fit into a useful equation.
The data comes from all kinds of sensors and can be used to train models through machine learning. Such models give very good results, but their creation requires a lot of time and effort. Furthermore, they are completely independent of the laws of physics.
“Practically speaking, simplified hybrid models are the solution,” says Prof. Jan Verwaeren from the Department of Data Analysis and Mathematical Modeling at Ghent University. “They use machine learning and data, but still take relevant physical rules into account.”
Prof. Matthieu Duvinage, Principal Data Scientist for AI and Digital Twins at GSK, gives the example of a fermenter, where a digital twin can initially help clarify what is happening inside and, in a next phase, can also predict what will happen becomes . The twins therefore make it possible to adjust parameters and choose the best approach.
Prof. Ashish Kumar, Head of the Pharmaceutical Engineering Laboratory in Ghent, further explains the importance of digital twins in optimizing complex production processes. “Several production concepts are often possible,” says the professor. “Which method ensures the best delivery of a drug to the body? This requires a lot of experiments, but when developing a new drug, the active pharmaceutical ingredient is in very short supply.”
Digital models can help in all phases, from design and development to manufacturing and control.”
Prof. Ashish Kumar, Head of the Pharmaceutical Engineering Laboratory Ghent
Even in more obvious situations, a digital image of reality is worth its weight in gold. Prof. Kumar: “You may have made a good tablet, but marketing suddenly wants a different shape and that creates cracks. Such a question sounds simple, but it is not, especially not when you take the time pressure into account.” Prof. Kumar wants to make it clear with his examples: “Digital models can help in all phases, from design and development to production and control.”
During the roundtable discussion, it became clear that not every organization knows how to start digitalization with digital twins. Large overall projects are underway here and there, but those present also speak of ad hoc initiatives on a project basis. An approach that delivers quick MVPs (minimal viable products) is the key, and Dr. Elisa Canzani, Data Science Lead at Cognizant, the solution.
“If you really want scalable technology, you need high-quality and automatic data flows,” she says. Cognizant and GSK have developed a solution for this with TwinOps, which Dr. Canzani wrote a paper. TwinOps is a platform consisting of building blocks for data ingestion, validation, transformation and storage. “These building blocks can be reused after validation.”
With the TwinOps platform there is a sustainable and scalable connection between the data on the one hand and the digital twins on the other. “Thanks to automated and repeatable workflows, the effort required to maintain compliance is up to eighty percent lower,” says Canzani. The result is impressive: At GSK, the TwinOps platform ensures that creating an MVP takes barely three weeks (minimal viable product) come.
A digital model without data flow is not a true digital twin.
Dr. Elisa Canzani, Data Science Lead Cognizant
TwinOps is not only important for scalability, speed and compliance, but also simply for getting the most out of digital twins. DR. Canzani “A digital model without data flow is not a true digital twin. Companies often have data, but it is stored in a colorful Excel.” When data flows from sensors directly to the models, the digital version of an asset is actually a digital twin.
This traffic can flow in two directions. The ultimate goal for most participants is to reconnect models to production environments. Data from bioreactors then feed the models, which then provide insights based on simulations. These are initially reported back to a human operator, but in the long term such an intelligent system can adjust parameters itself and optimize production.
This is where heritage collides with digitalization. Complex factories are controlled by software that is not always open. Several pharmaceutical companies are struggling with this closed approach, where control software represents a bottleneck to digitalization and automation. Retrieving data from the systems is sometimes more difficult than expected. CESPE would like to make a contribution here, among other things by providing good alternatives.
Not only is heritage a potential stumbling block, people also bring challenges. Even clear mechanistic models that do not suffer from the black box aspect of an AI model still seem like a black box to employees who have no experience with them. That creates mistrust. The timely involvement of the right stakeholders in a digital twin project is therefore a must.
A flexible data platform is important here too, as it provides the necessary freedom when developing a project. In general, everyone present strives for a balance between human knowledge on the one hand and the value from data and models on the other. The combination of IT, data, industry knowledge and academic expertise is not a given, but very important. “That’s exactly why Cognizant and CESPE are collaborating on this story,” said Sean Heshmat, data and AI lead at Cognizant.
We also see different views on the value of data and models among those present. Drug development is seen as a priority and digitalization is seen as little more than a means to an end. Not much thought is given to the intellectual property of data and models.
We hear from others a more nuanced approach. Some participants see value in the digital twin models and associated intellectual property.
As part of the round table, lectures and discussions often focus on niche applications of digitalization and digital twins in the pharmaceutical sector. That makes sense: Cognizant and CESPE are consciously putting in a lot of effort. However, a more distant look shows that the industry is confronted with the same problems and questions as many other industries, particularly in the chemical and manufacturing sectors.
Ultimately, everyone is looking for ways to digitally simulate complex processes in order to ultimately gain deeper insights and greater efficiency. An efficient and scalable approach that takes compliance conditions into account is essential. The TwinOps platform is the star of the day: the data layer is the sought-after basis that enables rapid experimentation.
The puzzle pieces are there. Now the challenge is to get started. There may only be a few in the industry First movers But there is no shortage of fast followers. “To realize the full potential of digital twins and accelerate innovation, it is crucial for companies in the chemical, pharmaceutical and biopharmaceutical sectors to collaborate and launch projects with multiple partners,” concludes Dr. Christoph Portier, Managing Director of CESPE. “By sharing resources, knowledge and expertise, we can jointly explore the possibilities of digital twin technology and create a foundation for sustainable growth with the associated competitive advantages.
Source: IT Daily
As an experienced journalist and author, Mary has been reporting on the latest news and trends for over 5 years. With a passion for uncovering the stories behind the headlines, Mary has earned a reputation as a trusted voice in the world of journalism. Her writing style is insightful, engaging and thought-provoking, as she takes a deep dive into the most pressing issues of our time.
