Objective

Assess process stability and capability to establish a baseline understanding of quality performance, before developing a process model to identify and control critical parameters for quality improvement.

Background

BLX Chemicals is a leading producer of high-quality resin that serves as essential building blocks across numerous industries. From advanced manufacturing and coatings to adhesives, composites, and innovative material solutions, resins help customers create products that are stronger, more durable, and more sustainable.

A critical quality attribute (CQA) of a resin is purity, which describes how free it is from contaminants or any unreacted components. High purity ensures consistent performance, stability, and reliability in applications such as coatings, adhesives, and polymer production. Even small impurities can influence color, strength, or reactivity, making careful purification and quality control essential.

In recent months, the resin manufacturing process operated at the best achievable stability, which was still not optimal. As a result, purity did not meet its specification for too many batches, which led to a poor baseline capability of the process. The chemical engineering team wants to address this issue.

To ensure acceptable product quality, an additional purification step is required. This initial purification step is applied before the final purification takes place, which leads to an increased processing time. It is further required to measure purity on each product. Since sufficient data for online monitoring of purity are not available, an offline measurement of purity became necessary, adding even more time to the overall process.

The resin production runs 22.5 hours per day over two shifts and uses two reactors.

The Task

• Explore the relationships between the critical process parameters and purity.
• Use control charts to analyze stability of the process.
• Check normality assumptions and compare process capability for both reactors.
• Build and apply a process model for optimization:
• Build a response surface model using the given experimental data.
• Control the ranges of process parameters to improve the in spec portion of produced resins.
• Simulate new data under the optimized conditions to estimate the expected capability.
• Compare the baseline capability with the improved capability by simulating new data under optimized conditions.