Creating benefits by digitalization for the bioprocess industry – The importance of Data Science as central enabler

Creating benefits by digitalization for the bioprocess industry – The importance of Data Science as central enabler

Projecting digitalization to the bioprocess industry

The bioprocess industry is currently being transformed and disrupted through the capabilities of digitalization and Internet of Things (IoT), as summarized under the buzzword Industry 4.0.  Digitalization in the bioprocess industry create new forms of innovation and result in new business models: The bioprocess industry is about to go through a revolution by taking full advantage of what innovative digital technologies have to offer. Industry 4.0 is present in all aspects of the bioprocess industry, as it will have impact throughout its value chains, from logistics, process- and materials design, planning, plant operations, plant safety, monitoring and maintenance of factory equipment to marketing/sales and supplier/customer integration. This has been nicely summarized in a recent DECHEMA white paper.

In more detail, digitalization for the bioprocess industry will mainly act on two dimensions: A) the process chain and B) the product life cycle.

The effect of  digitalization for the process chain

The enhancement of the process chain by digitalization for the bioprocess industry will act on the supply chain, the logistics and on predictive, rather than preventative, maintenance. There is a strong need to provide solutions to the following tasks/objectives:

  • Link manufacturing with peripheral activities
  • Develop horizontal analytical techniques (PAT) along the process chain
  • Identify process parameters across unit operations and along the process chain.

Identification and data collection of process parameters across all unit operations in product development and operation will enable:

  • continuous manufacturing,
  • process robust and
  • consistent quality.

Digitalization for the bioprocess industry: Effect on the product life cycle

Digitalization for the bioprocess industry covers the complete product life cycle and will lead to flexibilization of production, by analyzing sales as well as basing quick product change overs on platform knowledge. Both should act on marketing strategies. Hence, the tasks/ objectives are to

  • Integrate data from facilities, sites, suppliers and clients
  • Increase process & manufacturing transparency by Quality Metrics
  • Allow feedback loops inside of the life cycle for Continuous Improvement
  • Create multi product facilities
  • Flexible resource management
  • Establish transparent and flexible business process workflows

Software roll-out and implementation strategies have to be carefully considered.

What is the main strategy to achieve this?

No matter which dimension you are working on: The strategy to find the way to a sustainable solution for digitalization for the bioprocess industry is knowledge:

No doubt, enough data is already available. But there is the strong need of generic software tools to generate knowledge. Present tools range from data visualization to data-driven or mechanistic models and provide ontologies and taxonomies.

Models are currently getting increased intention in the bioprocess industry. This is using the next step in the chain, from data over information and knowledge: Intelligence and wisdom. In technological language, this is a multiparametric control strategy to the above mentioned objectives, applied in a real time context. The tools here, such as model based control, model predictive control (MPC), software sensors is not new, but so far, hardly used in the value added process industry such as the biopharmaceutical sector. Why not used so far?

Once knowledge is provided to the process in a real time context, we have to check if the model / knowledge is still valid via knowledge management tools. We need computational model life cycle management (CMLCM) strategies, which is also an integral part of the product life cycle management, enabling feedback loops and continuous improvement of the process chain and product life cycle. This was also strongly emphasized in the recent guidance document ICH Q12 of the regulatory agencies for bioproducts (www.ich.org).

The follow up of the above knowledge based strategy will lead to a intelligent manufacturing which is worth to be named an Industry 4.0 solution. But what is needed, what are the enablers to do so?

What are the key enablers to digitalization for the bioprocess industry?

The key enabler for digitalization for the bioprocess industry is the complete spectrum of data science! Data science strategies are mainly applied in two main sectors, A) in a data management system and B) in a real time environment.

A)    Data Management System

  • Connectivity: Connectivity of process sensors, but also of all other data sources from the process chain and the product life cycle, hence including sales, marketing, maintenance, logistics etc. The data types may vary from meta data, time value pairs to 2D and 3D data. The data needs to be established in a central data base as shown in the following figure.

  • Data Contextualization: The data sources may vary in data density and occur at different time points. Data need to be aligned and put into context to that the following steps in data analysis and knowledge gathering can be executed
  • Data Analysis: Before data can be analyzed prior knowledge may be entered and detection of outliers need to be executed, of course fulfilling the entire spectrum of data integrity. Consider best practices for bioprocess data analytics.
  • Model Development, Software Sensors, Knowledge Management and Model Maintenance: For a sustainable solution, such as for being independent of modelling experts, we need to have full transparency on the procedures: We need automated model building workflows and business process workflows for model maintenance.

Conclusively, there is the need that the models have predictive power as they are going to be used in the real time environment.

B)    Real Time Environment

  • This system needs to be much more than a Programmable logic controller (PLC) but it may need to be connected to it!
  • Real-time architectures provide correct process information to operators in real-time for decision making.
  • Using Model predictive controls based on multiparametric control strategies, we allow feedback control to detect deviations and automatically adjust operations, decision support and advanced operator support.
  • The system allows real time execution of a ‘Digital Twin’ (virtual process/plant models) to predict the impact of (design) decisions and to anticipate bottlenecks as well as allow efficient upfront training for new processes and advanced operator support, for example through augmented reality.

There is the strong need to establish Data Management System (DMS) covering the complete value cycle form process development to production in a real-time environment. Organizations need to implement knowledge management processes and manage its knowledge within a software efficiently to become a leader in Industry 4.0.