In the manufacturing industry, for example, the pharmaceutical industry, it is a goal to obtain consistently predefined end product qualities. A properly designed process monitoring and analysis system is needed to assure the predefined end product qualities. The design of process monitoring and analysis systems involve the identification of critical quality product parameters, selection of the appropriate monitoring tools and the determination of efficient control strategies to ensure that the selected critical quality parameters can be controlled and/or monitored. The framework for design of the monitoring/control system consists of three main parts: a knowledge base of methods and tools for monitoring/analysis, a model library (contains models for unit processes, sensors and controllers) to supplement the gaps in the knowledge base and an algorithm for design and analysis of the monitoring/control system. A set of design decisions need to be made in an integrated manner taking into account the interaction between product quality specifications, process operational constraints, cost of the monitoring and analysis system and the time needed for analysis.

The starting point of the design procedure for the monitoring/control system is the product quality specifications and process specifications (provided by designer of the system), followed by four analysis steps: process analysis, sensitivity analysis, interdependency analysis and performance analysis of monitoring/control tools. Process analysis provides an extended set of relevant process variables (identified through the use of a knowledge base) among which the critical quality parameters and the corresponding actuators have to be selected in subsequent analysis steps. It also provides a good understanding of the process-product rleationships. The critical quality parameters are identified through a sensitivity analysis (for this purpose the model library is used). The process variables, which violate the operational limits and have a major influence on the product quality are selected as the critical quality parameters. The pairing of critical quality parameters with the actuators is achieved through an interdependency analysis (model library is used). The process parameters, which are the most sensitive for the selected critical quality parameters are selected as the actuators. Appropriate tools for on-line measurement of the critical quality parameters are selected through performance analysis of the monitoring tools. In the process of selecting the monitoring tools, first the available monitoring tools together with specifications (e.g. accuracy, precision, operating range, time, cost etc.) for measurement of selected quality parameters are retrieved from a knowledge base and then their performances are compared to select (configure) the best monitoring/control system. In the final step, the most promising candidate systems are verified through simulation, ie., check for their monitoring/control performance.

The objective of this presentation is to demonstrate the application of the systematic framework and its design methodology for product quality monitoring and control. The identification of the critical quality parameters, pairing with actuators and selection of monitoring tools will be illustrated through two case studies, an insulin production process and a tablet manufacturing process.