Our Statistical Consultants have Decades of Experience.
Statistical Manufacturing Solutions is based in Columbia, South Carolina providing Statistical Consultants that help companies use statistics and models to identify problems in their processes, and improve them over time. I am a Certified Quality Engineer, Six Sigma Black Belt, with a Masters of Science in Statistics from DePaul. With almost three decades in the industry spent with companies like Cabot Microelectronics, Cook Medical and others, I have learned a thing or two about improving processes and increasing margins for businesses.
Statistics for Quality and Process Improvements
Most of my career has been in statistical quality control and the most rewarding experiences have been quality and process improvements. These projects have met with varying degrees of success from spectacular to merely OK. The most spectacular success was in metal stamping where we measured quality in Mean Time to Fail, i.e. how long could the machine run before producing defects. The MTTF was increased from 30,000 parts to 480,000 parts.
Meeting Compliance with BS5750 / ISO 9000
I started working in quality in 1982 at a small plant in Israel that manufactured industrial safety boots. The company made a decision to enter the lucrative market for injection molded mining boots for the UK mining industry. Insurance and government regulations required a British Kitemark showing compliance to the requirements of BS5750, a precursor to the ISO 9000 series of standards. This initiated my experience with quality system requirements. I moved back to the United States in 1985 and starting in the early 90s worked with companies seeking certification to ISO 9000. These experiences provided me with a good sense of what works and what doesn’t. Why some companies find compliance improves their quality system and why for others audits are a constants source of headaches.
David Patin – Owner/Lead Consultant
Although having officially retired in 2017 I have kept busy travelling, reading and keeping in touch with colleagues while at the same time developing a statistical consulting service, “Statistical Manufacturing Solutions.”
Statistical Manufacturing Solutions is a consulting service located in Columbia, SC and is based on helping companies use statistics to identify problems in their processes, and improve them over time. I am a Six Sigma Black Belt. A Certified Quality Engineer, Reliability Engineer and have a Master’s of Science degree in Statistics from DePaul University. With almost three decades in the industry spent with companies like Cabot Microelectronics, Cook Medical and others, I have learned a thing or two about improving processes and productivity.
I am interested in contract work of 3 to 6 months duration with the possibility of extension. From my 35 years experience I bring to the table a broad range of knowledge and experience in a variety of industries, and am able to translate the information from the numbers to the real world and relevant examples so that those with less of a grasp in statistics can understand what it means to the business.
Begin Consultation
Start moving in the right direction.
Design of Experiments
A very common technique in process improvement are Design of Experiments. A good DOE project will start out with a screening design to determine which factors from all potential factors do influence the process. One or more DOEs then measure the magnitude of these affects and potential interactions. A third confirming DOE is used to determine process setting. During my years employment I have overseen scores of DOEs. I have seen what causes failures, i.e. not achieving usable results, and what’s required to achieve success.
Documented Statistical Methods
Statistical methods should be documented. Sample sizes are determined through a balance of risk and severity. Degrees of confidence and reliability should be established in high level documents. How to analyze data and what statistical tests are appropriate to the type of data should be documented to ensure consistency across projects. At several jobs it was a main function of my job to write procedures for statistical methods.
Understanding and Using Acceptance Sampling Plans
There are a host of other common problems that companies experience. Many companies have problems understanding and using Acceptance Sampling plans either in their receiving inspection or when used for in-process inspection. A goal should always be to have measurable quality characteristics for which SPC, Statistical Process Control, have vastly superior advantages. But in the absence of measurable characteristics utilizing the correct sampling plan can both prevent defects from being shipped to your customers but also signify what are the biggest or most frequent quality defects.
Sample Size
Determining the proper sample size for testing to correctly document levels of confidence and reliability and analyzing the data from tests especially when the data is not normal or contains anomalous or contradictory values are techniques is have used almost weekly for those 35 years I’ve been in manufacturing.
Improving Quality with Better Measurements Systems
Prior to my experience with medical devices I had worked a number of years in metal stamping. First, what I thought at the time was a high volume stamping operation, 14 million a week, stamping mounting cups for the aerosol industry. It was here that I learned the importance of measuring the actual quality characteristic of interest, in that case, how many parts could a machine produce before requiring maintenance to correct a defect in production. At another metal stamping company, White Cap, that made closures for the food and beverage industry (lids to everyone else) I experience slightly higher volume at 14 million per day. To improve quality better measurements systems were required and to that purpose I designed and co-designed several machines for the automatic measurement of quality characteristics. I also wrote computer code that would analysis the resulting data and enter it into a database.