الفهرس 
LITERATURE REVIEWOnly 14 pages are availabe for public view
 |  | from | 155 |  |  |
| | | from | 155 | | |
Abstract
Statistical Process Control (SPC) is an effective approach for monitoring and improving quality, which aims at quality improvement through reduction of variation. One of the primary techniques of SPC is the control charts, which represents a real tool in reducing variability. When the measurements of quality characteristics of a product consist of a number of interrelate variables, the process follows a multivariate probability distribution. In this situation, classical univariate control charts could not be used to achieve the required effective control on the process. Consequently, multivariate process control is applicable in many production processes.
In this thesis, the applications of univariate and multivariate control charts in the field of steel industry are studied. The complex nature for steelmaking processes makes classical SPC methodologies not optimal when used to monitor and control steam boiler generation that used to supply the required steam for vacuum degassing process, which is an important step in steel production. This process includes a large number of variables that need to be monitored and controlled, while classical SPC requires a control chart for each variable.
Traditionally, control charts are developed assuming that the process observations are independent and follow normal distribution. Unfortunately, this assumption is unrealistic in practice. Thus, the effect of one variable can be confounded with effects of other correlated variables. Such a situation can lead to false alarm signals using univariate control charts to monitor each variable. Univariate charts are also difficult to be managed and analyzed because of the large numbers of control charts of each variable that are typically chosen.
An alternative approach is used to construct a single multivariate control T2 chart that minimizes the occurrence of false process alarms as well as identifies real process changes that not detectable using univariate charts. It is necessary to simultaneously monitor and control these variables to achieve high vacuum degassing process performance to remove harmful gases such as Nitrogen (N2), Oxygen (O2) and Hydrogen (H2) from the molten steel before casting.
The raw data of seven quality variables, which were pre-determinated, are collected to construct control charts for detecting the mean shifts and the variability conditions for each variable. The studied variables are: boiler water level (Wl), boiler water input temperature from economizer (Tbwi), fume gases temperature from boiler (Tgb), boiler pressure (Pb), steam accumulator pressure (Pa), steam header pressure (Ph¬), and steam header temperature (Th).
The univariate control charts of Shewhart control charts are constructed for each variable as a phase I for estimating the process parameters to monitor the specified process for improvement. The monitoring process is conducted for the boiler process to improve it. The assignable causes are studied and the suitable corrective actions are conducted to improve the performance of the boiler process for steam generation
Furthermore, Exponentially Weighted Moving Average (EWMA), and Cumulative Sum (CUSUM) control charts are used for monitoring the boiler process. The number of out-of-control signals of each variable for each univariate control charts is identified for comparing their performances. As well as in multivariate control charts, the historical data set and Hotelling T2 control chart for phase I operations are constructed, including the historical data set and analyzing data problems to purge outliers. The monitoring process, as a phase II is also conducted.
Finally, the performance analysis for each univariate and multivariate control charts is studied using the Average Run Length (ARL). A comparison of the univariate out-of-control signals with the multivariate out-of-control signals is also investigated to illustrate the performance of univariate (Shewhart, EWMA, and CUSUM) and Hotelling’s T2 statistic control charts.
corrective actions were carried out. A significant improvement is attained in the performance of boiler steam generation that consequently improves the performance of Vacuum Degassing (VD) process and hence the required vacuum pressure could be reached which leads to decrease the gas content.
Moreover, it was concluded that, the EWMA and CUSUM control charts are, in general, much better than Shewhart control chart for small shifts. In addition, Shewhart control chart has the highest false alarm signals. The multivariate control charts have less false signals than that for univariate control charts; they are more effective in detecting mean shifts.