论文部分内容阅读
With the widespread application of quality by design(QBD)and process analytical technology(PAT)many of t he l eading p harmaceutical an d b iopharmaceutical manufacturers u se multivariate methods f or m odeling t heir processes.In p articular multivariate s tatistical p rocess co ntrol(MSPC)monitoring has become a widely accep ted practice for real-time detection and diagnosis of process deviations.An active area of study more recently has been the u se o f o ptimization a nd ad vanced co ntrol.I n t his t alk multivariate monitoring case s tudies ar e p rovided demonstrating the application of MSPC on a solid dosage,wet granulation process and a cell culture biological reactor.In t he wet granulation s tudy multivariate monitoring is used as a basis for real-time r elease.For t he cel l culture example multivariate models are built to monitor cell productivity.It is shown how multivariate models can be used for prediction of future process performance and integrated into optimization routines to create a model predictive controller.The prediction and optimization method is applied to the cell culture process for forecasting of important parameters,such as cell productivity,and used for adjustment of nutrient addition to maintain the process on a desired trajectory and increase final yield.