The basic functions of statistical process control (SPC) include the use of statistical tools to characterize processes, determine their inherent variability, monitor whether or not the processes are in control, and ultimately optimize processes and their control. Successfully implementing SPC requires an understanding of several fundamental concepts:

• Variability is inherent in manufacturing. We must analyze and try to minimize this variability.
• Ceramic manufacturing is very complex, with many poorly understood process interactions. We must begin to better understand these interactions.
• Effective ceramic process control requires personnel versed in both SPC and ceramic technology, a weakness for many manufacturing organizations.
• Making measurements to characterize a process will add measurement variability on top of true process variability.
• It is essential that upper management be highly committed to SPC.

The Variable Process of Ceramic Manufacturing

Each process in our manufacturing sequence has inherent variability. No piece of equipment, and certainly no human worker, does exactly the same operation each time. We can certainly work to minimize variability, but we will never eliminate it entirely. Ceramic manufacturing is far more complex than assembling widgets. Processing, properties of final products, yields, and profits are all dependent on a large list of variables and their complex interactions, which are often poorly understood or characterized.

Given the number of processing steps and the inconceivable complexity of the interactions that can and do occur, it is a wonder we can make anything resembling consistent product. Statistical process control is a powerful tool that can help us do just that.

If You Can’t Measure It, You Can’t Control It

Before we can analyze and ultimately control process variability, we must make measurements of the properties that characterize the process. We may be interested in dimensions, mass, density, specific gravity, hardness, mechanical strength, electrical properties, magnetic properties, particles size, moisture, shrinkage, temperature, porosity, surface area, viscosity and others. The measurement process itself contributes some level of variability to the data we will use to characterize the process.

We have two separate concerns regarding the instruments or procedures we use—accuracy and precision. Accurate measurements give values that closely characterize the actual value. Precise measurements provide values with little scatter. Our measurements of the process are only as good as the accuracy and precision of the instruments and procedures we use. If we have a measurement that has good accuracy but poor precision (a lot of scatter), then some of the variability we perceive in the process will actually be from the measurement and not inherent in the process itself. Additionally, we will probably use inspection data to help us control the manufacturing process.

In God We Trust—All Others, Please Bring Data

For routine and effective process control, the data we collect must be available in a database or a spreadsheet so we can use computer-based software for analysis. Much data can now be collected and stored automatically. Kiln temperature profile data is a common example in the ceramic industry. We will need to consider the sheer volume of data we can collect, how useful it will be, and how it can be stored for use. Computer technology allows us to easily collect, log and analyze data, but we need to be judicious in doing so.

In future columns, we will discuss the issues of personnel training for SPC, as well as management commitment and philosophy. Next time, we will review the basic statistical concepts needed to characterize and help control ceramic manufacturing processes. And we should always keep in sight the seven ultimate goals of effective statistical process control:

• Generate greater profits
• Improve product quality
• Increase manufacturing capacity
• Enhance customer satisfaction
• Improve product delivery
• Cut manufacturing costs
• Build employee pride