ABSTRACT
As described briefly in Section 1.3, statistical process control (SPC) of a produc-
tion process can be roughly divided into two phases. In the initial stage (i.e., phase
I), we usually do not know much about the performance of the production process
yet, and our major goal in this stage is to properly adjust the production process to
make it run stably. To this end, we usually let the production process produce a given
amount of products, and the values of the quality characteristic(s) of interest of the
manufactured products are then recorded and analyzed. From the statistical analysis
of the collected data, if we find that the production process does not seem to run
stably, then the root causes responsible for that unfavorable performance should be
figured out and the corresponding adjustment of the production process should be
made as well. After the adjustment, another set of data needs to be collected and
analyzed, and the production process should be adjusted again if necessary. This
analysis-and-adjustment process is then repeated several times until it is confirmed
by the data analysis that the performance of the production process is stable. Once
all special causes have been accounted for and the production process is in-control
(IC), we collect an IC dataset from the products manufactured under the stable op-
erating conditions, and this IC dataset is then used for estimating the IC distribution
of the quality characteristic(s) of interest. Based on the (estimated) IC distribution,
a phase II SPC control chart is designed, and it is used for online monitoring of the
production process. When it detects a significant shift in the distribution of the qual-
ity characteristic(s) from the IC distribution, a signal is delivered and the production
process is stopped immediately for root cause identification and removal. This online
monitoring stage of SPC is often called phase II SPC.