Big data analytics in complex production environments

Complex manufacturing processes with particularly high quality requirements can benefit enormously from a detailed knowledge of all existing data and from the resulting insights into influencing factors and variables. Analyses of large data volumes can serve to convert information into predictive models, which then allow the configuration of processes within a range of optimum parameters. These ranges are customarily defined to lie safely within the specification limits while, at the same time, allowing high yields.

Biopharmaceutical production processes provide a particularly suitable test environment for the technological potentials of big data analytics. One example is the technique of producing antibodies. Antibodies are first generated in botanical plants before they are converted into a biopharmaceutical product through a cleaning process. Since the growth of plants is determined by many factors including light and temperature and since the growth rate in turn determines key pharmaceutical parameters such as stability and effectiveness, a profound understandíng of the underlying processes is vital. Such an understanding can be gained from data-based models that allow a batch-specific adjustment of the conditions of the plant cultivation process.

The highly flexible and adjustable analytic tools and models that are developed by the ICNAP can be used in a wide range of fields where production conditions must be frequently adjusted, for example across the entire process industry including pharmaceuticals, cosmetics and agriculture. Intelligent analytical methods can also be highly useful in classical discrete manufacturing environments where large and highly heterogeneous data volumes must be processed. Under such circumstances, predictive models can serve to establish and to eliminate potential error sources at an early stage.

• Transformation of Big Data into Smart Data through contextualization
• Accelerated product release through comprehensive, integrated process simulations
• Strict compliance with ideal process requirements through automatic establishment of key reference variables and continuous target-actual comparisons even in highly complex production environments
• Faster and more precise adjustments to changing manufacturing conditions
• Improved utilization of production line capacities even for small batch sizes

Upstream production and downstream processing: from the sowing of seeds to the biopharmaceutical product

The so-called upstream production process, i.e. the process chain “sowing of seeds – growth – harvest”, benefits from Big Data Analytics, which allow the continuous optimization of plant growth and expression levels of the desired substances as well as the increased generation of recombinant protein. The later so-called downstream processing – the process chain “extraction – cleaning – biopharmaceutical product” – not only uses the process parameters from the upstream production but also additional parameters (such as turbidity, pH and conductivity) to improve the process models with the objective of achieving higher yields, higher quality levels of the biopharmaceutical product and more reliable monitoring techniques.