Wednesday, July 15, 2009

Statistics in the Manufacturing World






Before, when production quantities only amounts to tens and hundreds and most of product manufacturing is done either manually or in a small shop using few machines, quality control is quite easy. All you need is to check all your finished products or parts if it meets all the required dimensions and its tolerances. But as manufacturing enters “mass production”, quality check was elevated to a superlative degree.




Before, 99% quality signifies a high quality production process. So if the total quantity produced is 100, only one is defective. Fairly acceptable, isn’t it? Today, however, 99% is just far from acceptable. Now that production quantities amounts to millions, 99% quality translates to 10,000 defective products in a 1 Million produced!


Also, the complexity of the manufacturing process has increased dramatically. Before, creating one finished product can only take up to 10, 20 or even 50 steps and the number of components only amounts to a maximum of a hundred. Now, there are products that take up to 100 or more steps and components up to 1000 in number! Imagine the outcome if each step or each part is 99% in quality!


(99%) 10=90% (100,000 defectives)


(99%) 100=37% (630,000 defectives)






This concern has been one of the greatest challenges facing the manufacturing industry today. Indeed, 99% is just not enough…even far from what is acceptable.


A selection of Normal Distribution Probability...Image via Wikipedia

Improvement is a continuous process the manufacturing world is taking. Many techniques on quality have been developed. Most of these quality concepts involve one branch of mathematics we sometimes take for granted-Statistics.


Yes, you cannot check all the 1 Million products, do you? The quality inspection group depends on samples or small portions to predict the quality of the entire amount. Also, various statistical approaches and techniques have been developed to aid in quality improvement. Among these are Six Sigma, Design of Experiments (DOE) and Taguchi Methods.

This picture was reworked by the Bilderwerksta...Image via Wikipedia

Yes, never before had statistics played a vital role like these in the world of manufacturing. In the future articles of this blog, we will discuss some of these quality improvement techniques.

Reblog this post [with Zemanta]



No comments:

Post a Comment