Experiments

More efficient drug production using cloud-based CFD simulation of bioreactor

Motivation

Many people need drugs and therefore an efficient production is important for an affordable health system. An efficient production of biopharmaceuticals is important intermediate manufacturing step for high quality drugs. They are isolated from natural sources and are widely used for the abatement of illness, such as cancer and leukaemia. Aerated stirred reactors, the most common type of both small- and large-scale bioreactors, are used for performing microbial fermentation or mammalian cell culture unit operations for the production of the biological therapeutics, such as vaccines, hormones, proteins and antibodies. Usually, basic design criteria have been adapted in such a way as to meet the requirements of biopharmaceuticals. In particular, the shear sensitivity requires consideration in impeller design, aspect ratio and aeration. Sufficient oxygen transfer and carbon dioxide removal are very important criteria in selecting a bioreactor system. Optimization of these criteria can dramatically improve biopharmaceutical quality and therewith the drugs, which are finally delivered as end product to patients.

Computational Fluid Dynamics (CFD - is a branch of fluid mechanics that uses numerical analysis and algorithms to solve and analyse problems that involve fluid flows) is a simulation approach that can be successfully used for the characterization of bioreactors by evaluating above mentioned process parameters. Useful process information can be obtained using CFD without the need for building-up a prototype already in an early development stage of the devices, which can save lots of costs. Furthermore, successful prediction of the optimal process conditions can clearly improve biopharmaceutical quality and finally the drug quality.

Goals of the experiment

The main challenge in the calculation is the treatment of multiphase systems and long process time of several hours, which leads to long calculations, which are not suitable for industrial applications. In order to overcome these limitations and to apply CFD simulations in the development process a highly optimized work flow and huge computational resources are required. SMEs usually are not able to develop such workflows and to use such simulations without HPC infrastructure and a huge number of licenses. In this experiment, a workflow for multiphase flow and mass transfer simulation for a process engineering application is developed by using HPC/Cloud infrastructure, with special emphasis on pharmaceutical industry needs for bioreactor simulations.

According to Rader and Langer, the worldwide markets (expenditures) for bioprocessing facilities and equipment are estimated at 10.5 billion dollars (or about 7.2 percent of total biopharmaceutical sales). It is also reported that the overall bioprocessing market and its major niches will continue to grow at a steady 15-18 percent. Nevertheless, manufacturer of bioreactors are usually equipment manufacturers, which are usually SMEs. Therefore, the Cloud based simulation technology is ad-dressing SMEs and not big pharma companies.

Economic and technical impacts

Using Cloud based simulation technology it is possible to carry out analysis of bioreactors with clearly reduced costs due to reduced calculation time and full cost control. Through parallel calculation of a huge number of variants in a very short calculation time it enables the opportunity to get new customers. SES-Tec is expecting about five new customers, only due to the decrease of simulation time in the next three years. SES-Tec also expects to increase the turnover, which is obtained from bioreactor simulation from 15 percent to 30 percent. Furthermore, SES-Tec is planning to hire a new employee, which deals exclusively with cloud computation. The AVL-List as independent software vendor in the experiment benefits by new customers or more sold licences. Moreover, a new Cloud based business model was created as well.

The computational time for design of experiments (DoE) analysis was decreased from five weeks to one week.  Due to huge computational resources in the cloud, all 25 simulations variants can be run in parallel and not one after the other. This number of variants are typical for DoE analysis, but are not limited anymore thanks to Cloud based technology. Furthermore, the number of simulation variants is no longer related to the hardware in-house resources and therefore no investments are needed. Finally, each bioreactor manufacturer can benefit from proven and validated simulation technology and workflows for this kind of application, which prefers to use AVL-software and HPC/Cloud Infrastructure.

 

Partners:

SES-Tec OG  – SME, Austria

AVL List GmbH  – industrial company, Austria