Quality Solutions for Data Management, Process Monitoring, Process Optimization, Supplier Quality Tracking

Root Cause Analysis

Manufacturing processes can be complex. Often there are several process steps, assembly steps, and test steps. Various raw materials and subassemblies from an assortment of suppliers are brought together to deliver a finished product. Key questions are:  

• What are the most important variables that impact product quality? 
• Which ones should we measure and monitor?
• What are the defining characteristics of the manufacturing process that cause scrap and rework?

STATISTICA provides a suite of tools for root cause analysis, empowering engineers and other quality professionals with the capabilities for quickly identifying the few important factors from the many. 
Using root cause analysis analytics in STATISTICA, manufacturers have uncovered process understanding that was previously not possible. 

Opportunities for cost savings include:

• better investment by monitoring and control of the inputs and parameters that impact product quality along with less stringent oversight of inputs and parameters that do not impact product quality
• improved management of raw materials and suppliers by defining the monitoring of process inputs that are critical to finished product quality.

Process Optimization 

Once the critical few inputs impacting product quality are known, analytics in STATISTICA provide predictive models that characterize the processes. The predictive models make product quality predictions for various settings of the process, enabling ‘what if?’ scenarios to be performed by the engineering team to determine how the process behaves.

The next logical question is “What should the inputs be and how should our manufacturing processes be run to achieve the most cost effective and robust product quality outcomes?” STATISTICA provides structured tools and processes for engineers to perform process optimization. Engineers utilize their subject matter expertise to define which parameters are easy to change or expensive to change. Engineers are able to interact with and scrutinize these recommendations to assess predicted versus current process performance, including % beyond specifications and Cpk/Ppk, to estimate the cost savings and expected product quality improvements. Engineering user interfaces result in recommendations that are target optimal ranges for all of the critical variables, including raw materials or other inputs, that become the values for the limits for process monitoring.

Incorporate Multivariate Statistical Process Control  

With Multivariate SPC methods, you can find anomalies in high-dimensional data that otherwise are not visible.  When looking at each dimension by itself, many anomalies or manufacturing defects are hidden and missed. Incorporate Multivariate Statistical Process Control:

• in order to detect abnormalities (quality problems) across many parameters, when each individual measurement may be in control
• to identify outliers that would not be detected unless all variables are analyzed together
• to automatically generate multivariate alerts when a process is predicted to be “out of control” given the relationships across parameters