Road Map to the Statistical Process Control
Since 1920s, when Walter Shewhart first introduced the foundations for control charts, that several developments regarding new implementations of the statistical process control (SPC) have been presented in order to suit different situations that can be f
- PDF / 207,125 Bytes
- 9 Pages / 439.37 x 666.142 pts Page_size
- 47 Downloads / 212 Views
Road Map to the Statistical Process Control José Gomes Requeijo, Rogério Puga-Leal and Ana Sofia Matos
Abstract Since 1920s, when Walter Shewhart first introduced the foundations for control charts, that several developments regarding new implementations of the statistical process control (SPC) have been presented in order to suit different situations that can be found in several processes. It is noted, among others, the Short Run SPC, data non-normality, the presence of auto-correlation in process data, detection of small and moderate shifts in process parameters and the simultaneous control of various quality characteristics. This great diversity of situations is crucial for academic researchers and quality managers in making decisions regarding the choice of the best technique to implement the statistical processes control. For answering this diversity of situations in production systems, this paper presents a road map that allows the decision maker choosing the best technique for implementation. Various techniques are shown, such as the traditional Shewhart control charts, cumulative sums (CUSUM) charts, exponential weighted moving average (EWMA) charts, dimensionless Z/W and Q charts, residuals/forecast errors charts to processes with a significant autocorrelation and multivariate control charts. Keywords Statistical process control (SPC) Auto-correlation
· Process capability · Short runs ·
82.1 Introduction The implementation of traditional SPC is basically performed using the control charts developed by Shewhart [16]. Other important references are Woodall [17] and Montgomery [9]. The Shewhart charts have shown to be less sensitive in detecting J. G. Requeijo (B) · R. Puga-Leal · A. S. Matos UNIDEMI, Departamento de Engenharia Mecénica e Industrial, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal e-mail: [email protected]
J. Xu et al. (eds.), Proceedings of the Eighth International Conference on Management Science and Engineering Management, Advances in Intelligent Systems and Computing 281, DOI: 10.1007/978-3-642-55122-2_82, © Springer-Verlag Berlin Heidelberg 2014
949
950
J. G. Requeijo et al.
assignable causes of variation, particularly when the process shifts are small or moderate. For solving this situation, Page [11] developed the cumulative sums (CUSUM charts) and Roberts [15] the exponentially weighted moving average (EWMA charts). The developments of Gan [6] for the CUSUM and the developments of Crowder [5] for the EWMA must be considered. There are many developments of these charts, like the studies of Maravelakis [8]. Nowadays, production systems are characterized by a great diversity of products and low volume production. The classical SPC is difficult to implement, since Shewhart charts are designed for high production volume and few products to control. The most effective way to overcome this limitation is to apply dimensionless charts. Pereira and Requeijo [12] suggest the use of Z and W charts when it is possible to estimate the process mean
Data Loading...
