Advanced analytics reduces process failures and improves yields and quality too
23 April 2013
The need to reduce costs and improve product safety and sustainability is resulting in a shift of tools, traditionally used by analytical groups, to the plant floor in a bid to minimise failures, reduce variability, increase yields and improve quality.
CAMO Software specialises in multivariate data analysis software and solutions, working with manufacturers in the pharmaceutical, medical devices, food and beverage, chemicals and animal feed industries.
Multivariate data analysis (MVA) is the investigation of many variables simultaneously, to understand the relationships that may exist between them. Multivariate data analysis methods have been around since the 1960’s, but until recently were primarily used in laboratories and technical groups, rarely being applied to production processes.
Today, this is changing as manufacturers looking for a competitive advantage by using the data collected during production operations to provide a greater insight to help improve product development and process performance.
Because manufacturing processes are typically highly multivariate in nature they require multiple measurements to fully understand them.
Yet most Statistical Process Control (SPC) systems rely on univariate methods, which only look at single variables, one at a time. Univariate statistics tend to fail when analysing complex systems because they cannot detect relationships between the variables, often the cause of process upsets.
Multivariate analysis tools allow manufacturers to better understand process behaviour and implement more robust control strategies. This enables them to run processes closer to limits, use lower cost components, reduce energy use, reduce cycle times and minimise waste and rework.
Brad Swarbrick, vice president of business development at CAMO Software, explains further: “We worked with a food manufacturer who had a quality problem which they could not identify the cause of using their SPC tools. After analysing the data with multivariate methods, they worked out what the issue was and adjusted their process accordingly, saving them around 1 million euros per year in scrap, rework and energy costs alone.
“Another client in the pharmaceutical sector applied multivariate analysis together with Near Infrared (NIR) spectroscopy to monitor their fluid bed dryers to achieve consistent product quality using timeline and process trends.”
CAMO’s solutions allow multivariate models to be developed by Technical Service groups or CAMO’s consultants, and then applied to real-time production processes. The solutions can be used standalone to analyse off-line data, connected to databases or scientific instruments such as spectrometers, or integrated with control systems for use by process operators.
Contact Details and Archive...
Most Viewed Articles...