Statistical Process Control (SPC)

The goal of the article
Understand the purpose of Statistical Process Control (SPC)
Understand Transform data and Properties and chart types used for each.
Understand basic statistical concepts.
Understand how to choose a control chart.
Understand how to build a control chart.
Understand process competency.


Objective: To review the requirements of a company's quality management system.
Guige: A company's quality management standards (whichever) (ISO9000, IATF 16949:2016…) what is required to do for the different processes? 

Volatility management

- Volatility Management is the application of statistical methods to track and control a process to ensure it operates at its full potential to create the right product.
- A statistically stable process will operate in a predictable way to produce as many suitable products as possible, and produce the least amount of waste.
- While Volatility Management is most frequently used in production, Volatility Management also applies well to any process with a measurable output.
- The main tool for management is control charts, with an emphasis on continual improvement and design testing.

Required according to IATF 16949: 2016

 Production process design output
The organization shall document the production process design output in a manner that allows verification against the input to the production process design. The production process design outputs will include but are not limited to the following information: ...
machinery and equipment for production and control, including studies of equipment and process capabilities "
 Verification of job set-ups
The organization should: use the statistical test, if appropriate "
Monitoring and measurement of manufacturing processes
The organization will conduct process studies for all new production processes (including assembly or sequencing) to verify process competence and provide additional input to regulation control. process, even for special features. "
Identification of statistical tools
The organization should determine that the appropriate use of the statistical tool is appropriate. The organization should confirm that the appropriate statistical tool has been taken into account in the high-level product quality planning process (or equivalent) and taken into account in design risk analysis (such as DFMEA) (in the case of process risk analysis (such as PFMEA), and control plan.
Application of statistical concepts
" Statistical concepts, such as volatility, control (stability), process capacity, and the consequences of over-regulation, will be understood and used by the staff involved in the collection and analysis. and manage statistical data "

Special characteristics

Special Characteristics (SC) - A group of product characteristics or manufacturing process parameters that can affect the safety or regulatory compliance, fitness, functionality, performance, requirement or the corresponding treatment or product.
Check with customers regarding groups and acceptance criteria.

Data types

- Variable data
+ Quantitative data, which is the data for which the measurement result is used for the analysis
+ Data can be measured using a measuring instrument.
- Attribute data
+ Qualitative data, which can be counted for recording and analysis in 2 or more groups.
+ Data is often collected in the form of an inappropriate unit.

For example
Attribute or Variable data?

- In an engine cylinder manufacturing process, each cylinder is measured and checked before shipment, including:
+Diameter of the cylinder
+ Weight of the cylinder
+ Are there scratches or impact?

- In a process of painting plastic parts, parts are measured and checked prior to shipment, including:
+ The number of paint defects of each part
+ Paint density
+ Coating thickness at a defined area of a part

- In a molding process, the casting parts are measured and checked prior to shipment, including:
+ The rate of defects per day
+ The weight of each part
+ Rest periods of each work shift.


Over time, the output may form a pattern and be called distribution
3 4

Standard "bell-shaped" distribution

For a stable process, the data collected from a process will form a "bell-shaped" distribution, often referred to as a normal distribution.
For a fully weighed process:
Weakness = Median = Average
The value from lowest to highest in the distribution is called dispersion.

Data distribution

Data distribution can vary in Location, Dispersion, and shape
- Location (typical or central value)
- Dispersion (about or "width" of minimum to maximum value)
- Shape (pattern of volatility - balance, deviation ...)
Or any combination of these distributions

Location, Dispersion and shape

  • Position: Located on the nominal value
  • Dispersion: Less than Tolerance
  • Shape: Standard
  • Position: Located on the nominal value
  • Dispersion: Greater than Tolerance
  • Shape: Standard
  • Position: Located on the nominal value
  • Dispersion: Less than Tolerance
  • Shape: Standard

How can volatility be measured?

The most common ways to measure volatility include:
Variable range ®.
The maximum measured value minus the minimum measured value
Range of variation = Xmax - Xmin
Standard deviation (s)

Average distance between each value and mean


Note that there are several ways to estimate standard deviation.

Common cause

The most common cause is many sources of volatility that always affect the process. Cause-induced volatility is common in a process that produces a repetitive and stable distribution over time.
This is called "under statistical control".
If there are only common causes of volatility and no change, the output of a process is predictable.
These are events that are likely to occur due to a potential error in the process and appear as random fluctuations. These events will affect all of the individual values of the process.
Management Action
Usually done to address common causes of volatility.
It is almost always necessary to take management action, for example, changes to a plant, equipment or process
Usually capable of correcting about 85% of process problems.

Special Cause

Often referred to as outliers, is any factor that causes variation that affects only some of the process outputs.
'If there is a Special Cause of the volatility, the process output will not stabilize over time.
'These are discontinuous, unpredictable or unstable events that affect individual process values.

Action in place

Usually done to address particular causes of upheaval.
Usually by those most attached to the process, for example by the operator.
Usually capable of correcting about 15% of process problems.

Select-control chart



Author: Phan Nguyen

Note: The above article reprinted at the website or other media sources not specify the source is copyright infringement

Older articles



Definition: Carburizing increases the surface hardness of low carbon steel or...

Methods for determining the specifications of gears

Methods for determining the specifications of gears

The Helix angle is what differs helical gears from spur gears and it is...

Sign up to download the profile

Complete your information in the form below

You did not use the site, Click here to remain logged. Timeout: 60 second