ABSTRACT
Statistical process control Statistical process control (SPC) means controlling a process statistically. It is the use of statistical techniques to analyze a process in order to monitor, control, and improve it. The objective is to have a stable, consistent process that produces the fewest defects possible. SPC originated from the efforts of the early quality control researchers. The techniques of SPC are based on basic statistical concepts normally used for statistical quality control. In a manufacturing environment, it is known that not all products are made exactly alike. There are always some inherent variations in units of the same product. The variation in the characteristics of a product provides the basis for using SPC for quality improvement; with the help of statistical approaches, individual items can be studied and general inferences can be drawn about the process or batches of products from the process. Since 100% inspection is dif‰cult or impractical in many processes, SPC provides a mechanism to generalize concerning process performance. SPC uses samples generated consecutively over time. The samples should be representative of the general process. SPC can be accomplished through the following steps:
• Control charts (X--chart, R-chart) • Process capability analysis (nested design, Cp, Cpk) • Process control (factorial design, response surface)
Control charts Two of the most commonly used control charts in industry are the X-bar charts and the range charts (R-charts). The type of chart to be used normally depends on the kind of data collected. Data collected can be of two types: variable data and attribute data. The success of quality improvement depends on two major factors:
1. The quality of data available 2. The effectiveness of the techniques used for analyzing the data
• Control charts for individual data elements (X) • Moving-range chart (MR-chart) • Average chart (X--chart) • Range chart (R-chart) • Median chart • Standard deviation chart (σ-chart) • Cumulative sum chart (CUSUM) • Exponentially weighted moving average (EWMA)
• Proportion or fraction defective chart (p-chart) (subgroup sample size can vary)
• Percent defective chart (100p-chart) (subgroup sample size can vary) • Number defective chart (np-chart) (subgroup sample size is constant) • Number defective (c-chart) (subgroup sample size = 1) • Defective per inspection unit (u-chart) (subgroup sample size can
vary)
The statistical theory useful to generate control limits is the same for all aforementioned charts with the exception of EWMA and CUSUM.
