|M.Sc Student||Kobo-Greenhut Ayala|
|Subject||Integration of Feedback Control and Statistic Process|
|Department||Department of Civil and Environmental Engineering||Supervisor||Professor Raphael Linker|
|Full Thesis text - in Hebrew|
SPC - Statistic Process Control was first developed by Dr. Walter Shewart in the 1920s. The purpose of SPC is to ensure, in a cost efficient manner, that the product shipped to customers meets its specifications. One way to implement process control is through Control Charts. In general, a control chart contains a center line that represents the mean value for the in control process and two control limits. These control limits are chosen so that almost all the data points will fall within the limits as long as the process remains in-control. The control chart used in this work is the X bar control chart, which is one of the most commonly used control chart in the industry.
In this work, the ability to monitor a dynamic process using the X bar control chart and its standard decision rules is examined, meaning that the maximum percentage of variable output values allowed outside the control boundaries not caused due to a malfunction. Deviations from within the control boundaries are referred to as "false alarm". In this work, the conditions for which it is possible to monitor a continuous process by using the X bar control chart are derived. More specifically, it is showed that for any given system bandwidth, it is possible to determine the maximal sampling rate such that the process is monitored properly. This assumption does not always hold when monitoring a closed loop controlled system.
Feedback control is used to modify the control input of a plant in such a way that the plant remains stable, and that the plant output remains as close as possible to its desired value. The SPC methodology is not always suitable for feedback-controlled plants. We discuss the detection of malfunctions and the conditions for detecting such malfunctions using X bar control chart in closed loop controlled systems. It is shown there is a need to add to the control chart information from the control signals. When the controller is compensating a malfunction, its action appears as a change in the controller output signal. Therefore, monitoring the controller output signal gives a good indication of malfunctions in the process and allows the identification of malfunctions. In this work, the possibility to monitor the controller output signal using an X bar control chart is examined and tested.