Software Sensors for the Monitoring of Bioprocesses

Abstract

This paper presents various software based approaches suitable for the design of knowledge-based monitoring of biotechnological production processes. These processes require special treatment with respect to the complexity in biochemical reactions which make the design and construction of reasonably complex and practically usable mathematical models rather difficult. Additional complexity arises from the lack of industrially viable sensors for on-line measurement of key process variables. Software sensors which often use tools from the field of artificial intelligence represent one of the suitable approaches for the overcoming of the above mentioned limitations for its ability to utilize effectively both quantitative and qualitative knowledge about the monitored bioprocess. This approach is shown in practice using two different case studies of knowledge-based software sensors.