Minimize Variation through Statistical Process Control

ARTICLE

Written by Chris Draska


The perfect manufacturing process may seem impossible to achieve but it is still a worthwhile goal for manufacturers. A step in striving towards this goal is through the use of Statistical Process Control (SPC) systems. Using an SPC system, users can find and eliminate the factors holding back their plants. SPC is a functionality that allows users to acquire and analyze statistical data using powerful statistical tools. With the help of an SPC system, a plant can move closer to achieving the goal of perfection.

Scrap, waste, and rework are some areas that influence the overall efficiency of a plant. You can expect to see a dip in overall plant performance if any one of those areas is deficient. In order to fully understand why any one of those areas is subpar, you must look at the underlying factors involved in the production process affecting them. An SPC system is designed to do just that, give users the necessary visibility to correct issues before they become a problem.

There are two overall indicators that should be monitored during a production process, process data and product data. Process data are the characteristics that are part of the production process. This data can be a characteristic such as the speed of a machine, temperature of an oven or cycle time of an injection mold. The other type of data that should be monitored by an SPC system is product data. Product data are characteristics concerned with the quality of the product produced. Such characteristics can be color, dimensions or weight of the product. Together, these two data types give quality managers the needed visibility into their process to optimize production runs. Such optimization is only possible by comparing process settings with the quality of the product being produced across multiple production runs.

Another key benefit of an SPC system is the ability to stop a problem before it starts. Good SPC systems allow users to define control rules (custom or standard) and control limits; both of which help keep a process from producing defective parts. Control rules look for shifts or trends in characteristics plotted on charts, such as an x-bar chart. The goal of monitoring these rules is to forewarn operators of potential problems. Violations of these rules act as pre-cursers to potential problems. Control limits are limits set on charts such as x-bars and are the “voice of the process”. Control limits are calculated using process data and help monitor the process to ensure machines are running acceptably. A violation of a control limit means a process is out of control. In either case, the system should be able to notify the correct personnel to adjust the process before too many defective parts are produced.

SPC alone cannot assure the highest quality of finished product for your customer (that depends greatly on the quality of the input materials), but it can assure the consistency of the finished product by eliminating common causes of variation and detecting special causes of variation. Some customers might have strict quality requirements and request documentation to prove production quality. An SPC system can deliver reporting tools such as Histograms and Pareto charts to satisfy such requests. These may also be used to satisfy federal manufacturing regulations (e.g. food and drug administration) to prove product quality and reduce hazardous risk.

Because of the high demand for automation to cut long-term costs versus the high overhead costs to install such automation, SPC data collection can be achieved using a wide variety of methods. An SPC system should offer manual input (e.g. reading from a handheld measuring device), semiautomatic input, and fully automatic input. With Manual input, an operator enters a value from a non-electronic measuring device or a count of visual defects on a semi-finished or finished good. Semiautomatic input takes advantage of some automation by collecting data directly from the manufacturing equipment while requiring confirmation from an operator. Fully automatic input requires no operator input and seamlessly collects data behind the scenes. Processes can use any mixture of these data collection methods.

Through our work with customers in both discrete (auto parts) and continuous (food & beverage) industries, Engineering has achieved a wide range of expertise in selection and implementation and is ready to help your company with the right SPC solution.

Interested in learning more? Contact us at info@indx.com.


Contact Us