Statistical Process Control
Statistical Process Control is an industry-standard methodology for measuring and controlling quality during the manufacturing process. Quality data in the form of Product or Process measurements are obtained in real-time during manufacturing. This data is then plotted on a graph with pre-determined control limits.
Data Variance Reduction
This particular alignment process was dramatically improved by introducing a Nikon CMM, a visual measurement system into a quality checkpoint station. The improvement was largely due to the replacement of attribute data with variable data. The previous quality checkpoint was a pass/fail test to determine disposition. The new process makes dispositions based on physical measurements and pre-determined specifications and control limits.
My contribution to this project is setting up the needed software infrastructure to host and use incoming data streams. The Nikon CMM could output measurable data, but the data can’t go anywhere without my data transfer systems. They provide live data visualizations and realtime feedback to the technicians on how to adjust machines settings back to nominal.
I created the software applications and systems that pull and push data from the Nikon system into an internal database. The first of many software projects started with this GUI, graphical user interface. This project needed to quickly import txt files, provide technician feedback and store data for longterm historical use.
Identify Important Factors
We are also famiiliar with DOEs (Design of Experiments) to optimize a manufacturing process. Most of our DOE studies end up being three factors: time, temperature and pressure. These are some of the most common variables in a lot of manufacturing processes. However, the amount of factors can range to what ever is thought to be a significant variable to a particular output.
Explorirtory DOEs can be performed to learn the effects of factors and their combinations. Effects are sorted from largest to smallest. Any factors or combination of factors that have anova effects greater than the vertical red line in a DOE are considered to be statistically significant. DOEs are often conducted in manufacturing engineering to optimize a process in either production throughput or product quality.
