Statistical Process Control (SPC)
What is Statistical Process Control (SPC)?
Statistical Process Control (SPC) is a scientific method for quality and process control. It is an inexpensive tool to monitor and predict the performance of almost any process. It is most often used in manufacturing industries. SPC may be applied to almost any repeatable process including health care, finance, and back office operations.
Statistical Process Control (SPC) techniques were developed in 1924 by Walter A. Shewhart. Shewart was an engineer in Bell Laboratories. SPC came into wide-spread use in the United States during World War II. The techniques were largely forgotten in the US after the war until the 1980s. That was when American manufacturing leaders woke up to discover that Japanese manufactures were gaining market share due to superior quality and lower costs.
How does Statistical Process Control (SPC) work?
SPC works by periodic sampling of a production process, rather than 100% inspection. Sampling is much less expensive than 100% inspection. It is also more accurate because it is much less likely to produce fatigue in the inspector.
Engineers (or inspectors or machine operators) use computer software or pencil and paper to plot the sample data on a graph. This graph is called an SPC Control Chart. There are many types of control charts, depending on the type and source of data.
With a very small statistical sample, trained engineers (or computer software) can evaluate the statistical probability that the process is “stable” or “unstable.” Stable processes are left alone. Unstable processes are said to be “out of control.” Out of control processes are examined and stabilized.
What about meeting customer specifications?
One of the most common misconceptions is that a stable (in control) process is “in spec.” The truth is that a process can be stable and produce out-of-specification parts that don’t meet the customer’s needs.
The SPC Control Chart tells the user that the process is predictable. It doesn’t tell the user if it can meet the customer’s needs.
To know if a process is meeting the customer’s needs, the data sample is plotted on another graph – the Distribution Histogram.
The Distribution Histogram is sometimes called a Capability Histogram (or a Cpk Chart, or several other names) because it tells the user whether the process is capable of meeting the customer’s requirements.
Why use Statistical Process Control?
Statistical Process Control gives you immediate, objective knowledge about your processes. With this knowledge you can:
- Respond to problems early and faster
- Empower people to make better decisions
- Reduce scrap, waste and rework
- Reduce material costs
- Increase throughput
- Increase customer responsiveness
- Increase supply chain flexibility
- Shift from an opinion-driven culture to a data and fact-driven culture
For stories where people like you have realized these benefits, see these case studies:
- Multi-National Electronics Manufacturer Improves Quality Across Supply Chain – Disparate Systems Impact Manufacturing Quality Due to Lack of Visibility Across Silos and a Reliance on Manual Approaches
- Multi-plant Packaging Company Drives Revenue Improvement by Boosting Line Speeds 15-20% Over Name Plate Capacity – Frustration Exposes Opportunity to Harmonize Quality System
- PLZ Aeroscience Improves Production Control, Increases Sales – Improvement in control leads to sales growth, more consistent products, and line speeds increasing by up to 20%.
- Creating a Data-Driven Culture Results in 90% Reduction in Scrap – Multi-plant Manufacturer Reduces Escaped Defects from 8% to Less Than 0.5%
The sooner you know about problems, the sooner you can take corrective action. Sampling frequency matters. People who deploy manual systems may only have the time to collect the data once or twice a shift. If they detect a problem during the inspection, then they probably have to isolate and inspect all of the product produced since the last inspection. This can be very expensive and time-consuming, and result in high levels of scrap or rework.
More frequent inspections (hourly or twice-hourly) greatly reduce the time a process can run amok before a problem is addressed.
Automated, real-time systems such as GainSeeker Suite SPC Software reduce the effort and time required to inspect product, record data, and analyze the process for stability.
Why drill down?
Real-time product quality and process performance data is a valuable corporate asset. It becomes even more valuable when it is married or tagged with detailed information about the context of the data. This context (or tag or traceability) data makes it possible to drill down to root cause of problems.
This context information might include:
- Material lot
- Serial #
- And so forth
This contextual information can help you answer these kinds of questions:
- What happens to this output variable when I change material suppliers?
- Does this product perform better on machine A or machine b?
- Do I get the same product regardless of who is operating the machine?
- Do my customers get the same product regardless of which facility in my supply chain produces it?
Tagging detailed information with contextual information is best done through automated integration with systems of record, or collected through bar codes and other equipment directly from the machine operator or inspector.
Why GainSeeker Suite SPC Software for Statistical Process Control?
GainSeeker Suite is the real-time manufacturing intelligence platform with very robust tools for automating Statistical Process Control. Some of the key capabilities include:
- Connectors to any automated data source to streamline automation and data connectivity
- User-configurable scripting to empower your team to manage your deployment
- Robust charting and analytics tools
- Data wizards that quickly and automatically make the root cause of variation visible
- Role-based dashboards that deliver the right information to the right people in real-time