23 May 2017: Edgar Dietrich

Statistics: You Ask. We Answer.
Q-DAS software for the statistical analysis of data is a quality assurance standard in industrial production. Using Q-DAS software, various industries tab the potential of correct and reliable statistical evaluations. Q-DAS experts answer important questions about statistics in the following article.

Why do we need statistics in industrial production at all?

Companies produce many parts in relatively short periods of time, especially in mass production. Typical examples are the production of 800 engines a day or several thousands of cutting teeth for chainsaws a shift. It requires too much effort and it is too expensive and even unnecessary to check all these parts based on a 100% inspection. This is the reason why we monitor a manufacturing process based on samples. This approach is referred to as SPC or statistical process control.
 

How does statistical process control work?

The first step is to establish machine performance or manufacturing process capability based on statistical procedures. ISO standards, association guidelines such as AIAG and VDA and general reference manuals specify the respective approach.

After establishing process capability, you take representative samples of parts from the process at regular intervals. A sample normally consists of three to four parts. You measure these parts and plot the results on a previously calculated quality control chart. As long as the specified criteria are not violated, the process is assumed to be stable and the production just continues. In case of limit violations, the operator needs to be informed. He tries to find out why the process violated control limits and takes respective corrective action. This ensures that the on-going production process does not change significantly, and the process is considered to be capable or suitable to produce the single parts.
 

“Do not trust any statistics you did not fake yourself” is what Winston Churchill is supposed to have said. What is your opinion?