Calibration transfer in pharmaceutical process engineering

我的研究项目侧重于开发监测和优化工业过程的方法。具体来说，我的工作是过程分析技术（PAT），例如近红外探头（NIR），以提供产品的实时信息。这些探针对于适当的监测、理解和控制至关重要。为此，他们需要多元数学模型将原始数据转换为感兴趣的变量（例如，将近红外光谱转换为成分）。然而，PAT的实现受到大规模校准模型的时间和成本的限制。在小型系统上校准模型并将其应用于工业系统要简单得多，成本也低得多。然而，这种转移并不容易完成。即使采集条件（如照明强度）和材料特性（如样品密度）的微小差异也可能足以防止直接模型转移（CT：校准转移）。因此，多功能CT的发展对工业规模和建模至关重要。我建议开发更好的CT方法，从小规模应用到大规模应用。这对于提高生产过程的理解、安全性和效率至关重要。它对新生产方法的发展也至关重要，例如用连续生产方法取代间歇生产方法。我擅长过程工程和通过使用多元数据分析进行在线监测。我的短期目标是开发PAT来监测制药过程。我的项目包含两个相互关联的主题：-主题1集中于用于生产药片的压片机单元的校准转移。新颖之处在于利用粉末在压机中流动时的空间信息（如分离）来改善校准传递——主题2集中于制药产品中用于掩盖味道的流化床涂层装置的校准传递。新颖之处在于使用批处理过程的时间信息（其动态）来改善校准转移。虽然该项目是针对制药环境量身定制的，但它适用于任何制造部门。除此之外，改进的过程监控方法，特别是那些低成本的方法，容易影响任何工业生产。本项目旨在让学生掌握质量控制、过程监控和应用统计学的先进概念。这些关键概念对我们行业的福祉至关重要，下一代工程师必须接受培训才能掌握它们。

My research program focuses on the development of methods to monitor and optimize industrial processes. Specifically, I work on Process Analytical Technologies (PAT), such as near infrared probes (NIR), to provide real-time information on the products. These probes are critical to proper monitoring, understanding and control. To do so, they require multivariate mathematical models to convert the raw data into a variable of interest (e.g. a NIR spectrum into composition). However, PAT implementation is limited by the time and cost to calibrate the models at large scales. It is much simpler and cost efficient to calibrate models on small systems and apply them to industrial ones. However, this transfer is not easily done. Even minor differences in acquisition conditions (e.g. lighting intensity) and material properties (e.g. sample density) may be enough to prevent direct model transfer (CT: calibration transfer). The development of versatile CT is therefore critical to industrial scale up and modeling. I propose to develop better methods for CT from small scale applications to larger ones. This is critical to the improvement of manufacturing to improve process understanding, safety and efficiency. It is also critical to the development of new production methods, such as the replacing batch processes by continuous ones. I specialize in process engineering and on-line monitoring through the use of multivariate data analysis. My short-term objectives seek to develop PAT to monitor pharmaceutical processes. My project contains 2 inter-related themes: - Theme 1 centers on calibration transfer for tablet press units used to produce pharmaceutical tablets. The novelty lies in the use of the spatial information present in the flowing in the press powder (such as segregation) to improve calibration transfer - Theme 2 centers on calibration transfer for fluid bed coating units used for taste masking in the pharmaceutical products. The novelty lies in the use of the temporal information of the batch process (its dynamics) to improve calibration transfer While tailored to a pharmaceutical context, this project pertains to any manufacturing sector. Beyond this, improved process monitoring methods, especially those available at low cost, are liable to impact any industrial production. This project seeks to put students in a position to grasp advanced concepts of quality control, process monitoring and applied statistics. These key notions are critical to the well-being of our industries and the next generation of engineers must be trained to master them.