Course on Common Statistical Analysis Methods for Optimization of Chemical Process Processes and Formulations

Thread Content

Here is a course designed for chemical engineers: When conducting data analysis for the development of materials/formulations as well as for the optimization of chemical processes, chemical engineers often face various challenges. Process optimization and quality stability: Chemical processes are complex systems that involve dozens or even hundreds of variables such as temperature, pressure, flow rate, and ratios. Complex relationships such as non-linearity, time variability, uncertainty, and coupling may exist among these variables, making it difficult to establish accurate data models for precise control and effective monitoring. Formulation development and screening are costly and time-consuming: in the process of developing chemical products, there are numerous combinations of formulation variables, making it difficult to quickly and accurately identify the optimal formulation or direction for optimization, which can result in extremely high costs associated with experiment design. There are a wide variety of data types to analyze: these may include high-dimensional data, image information, and molecular structure data, which require proper integration, extraction, dimensionality reduction, and application in the analysis process. On Thursday, February 27, 2025, from 14:00 to 14:50, JMP will host a free live webinar. Over 50 minutes, 4 different cases will be used to cover various data analysis methods that are common and frequently utilized by chemical engineers, helping them identify the optimal formulations more quickly and improve chemical processes and product manufacturing techniques. Registration link: https://www.jmp.com/zh_cn/events/live-webinars/non-series/2025-02-27.html
Course outline: Common statistical analysis methods used in the chemical industry and the challenges associated with their practical application.
Case ① Process chemistry | Dynamic process optimization based on Tennessee Eastman’s chemical process data. Statistical Process Control (SPC): Process screening, model-driven multivariate control charts for monitoring complex processes. Multivariate analysis: Using Partial Least Squares (PLS) for dynamic process monitoring and optimization.
Case ② Personal care/organic materials | Improving the performance of surfactants through multivariate analysis. Data extraction and screening: Utilizing JMP’s free material information database plugin to obtain material properties directly from the RDKit database for data analysis and material selection. Multivariate tools for correlation analysis and modeling: Using PCA to develop covariance experimental designs, as well as PLS models for formula selection and optimization.
Case ③ Petrochemicals | Analyzing fuel octane number using NIR spectroscopy to improve fuel quality control efficiency. Using the Function Data Analyzer to extract principal components of functions for analyzing spectral changes and peaks. Applying Function Design of Experiments (FDOE) to describe spectral information.
Case ④ Agrochemicals | Optimizing the stability of agrochemical liquid formulations using image recognition. Employing deep learning and image recognition for high-throughput formula screening. Use of JMP Pro 18’s deep learning and neural network plugins