..

An alternate approach that can be taken with the modeling process is to compare a system's existing performance with historical performance records. The vibration data shown earlier in Figure 6.4 is an example of this type situation. One or more baseline runs are made to establish an expected value for the system response at specific points in the manufacturing process. Then, this data is used to provide an on-line real time evaluation of the process performance. An advantage of this technique is that it is insensitive to the normal shifts in the information signal, as occurred in Figure 6.4, because only deviations from the baseline are considered important. Another technique for making the historical comparison is to utilize control charts, which were discussed in Chapter 4, to compare specific parameters at particular points in the manufacturing cycle. This method shares an advantage with the empirical model process in that an in-depth understanding of the process mechanics is not required. Instead, it is only necessary that the desired process variables be monitored during a period of stable operation to form the baseline characteristics of the system. A shortcoming of this approach is that an error condition that exists from the beginning of operations will not be detected; instead, it will be assumed to be a normal characteristic of the process.