I E 584
Time Series Control and Process Adjustment (3) Design of Time Series-based process controllers for Quality Engineering. Study of the effect of autocorrelation on control chart performance.
I E 584 Time Series Conrol & Process Adjustment (3)
With modern sensor technology, quality control data frequently exhibits dynamics due to the short time between observations. Quality specifications keep "shrinking", and process drift is less tolerated than before. Under these circumstances, Statistical Process Control (SPC) techniques cannot be applied, and the emphasis in quality control moves from monitoring a process to actively adjusting it. Time Series techniques are ideal tools for developing such process adjustment strategies. This course covers topics of recent interest both in academia and in industry, including: integration of feedback adjustment techniques with traditional SPC methods; the "run-to-run" control problem as it occurs in discrete-part manufacturing (e.g., semiconductors); and optimal design of proportional-integral and EWMA controllers. In addition, a detailed treatment of statistical identification and estimation of ARIMA and discrete-time transfer function processes is presented. The effect of data autocorrelation on the performance of SPC control charts is discussed, and process adjustment strategies are presented as an alternative. For this reason, ABIMA modeling is discussed in detail as a means to represent data autocorrelation. Use of the MATLAB and SAS software packages are encouraged. A book on the course subject matter is under preparation and has been accepted by John Wiley & Sons who will publish it in its Probability & Statistics Series. Given the heterogeneity of the students taking the course, the prerequisites are rather modest, and the course is almost self-contained. The prerequisite is I E 423, or a similar introductory course in statistical process control.
General Education: None
Bachelor of Arts: None
Effective: Summer 2013
Note : Class size, frequency of offering, and evaluation methods will vary by location and instructor. For these details check the specific course syllabus.