May Feature

Technology Helps Manufacturers Create a Manufacturing Compliance Platform - continued

A technological look at capturing manufacturing and production data The successful integration of multiple data and analytical sources relies on the architecture of manufacturing execution systems (MES) technology (see Figure 3). For example, control systems are integrated to allow field device communication between a relational database management system (RDBMS) server, a Microsoft Windows client such as Visual Basic, and shop floor data collection systems. In addition, a SCADA or human machine interface (HMI) allows the MES to be connected to the controls platform through application brokers or application program interfaces. Another control system integration strategy is to connect an MES—through object linking and embedding (OLE) for process control (OPC) interfaces—with back-end and front-end clients. Such a setup will permit all controls to be viewed on a single screen on the shop floor. On the back end, the OPC interface is configured to connect with OPC servers, and is triggered by an MES system to transmit data to the OPC server and vice versa. On the front end, the OPC client links to Web screens for real-time displays of data such as weights and temperatures. Control system integration also addresses recipe management. Recipe definition mastered by the MES is revision-controlled. Recipe management allows production recipes to be maintained along with a comprehensive revision history and associated details. Recipe synchronization to the PLC or digital control system (DCS) can be accomplished through SCADA/HMI or OPC technology. By using a Windows-based client to manage data analyses, manufacturers also can interactively define statistical process control (SPC) charts to perform in-depth quality analyses. Several statistical methods can serve as analytical tools. Descriptive statistics, including univariate statistics, frequency distributions, frequency tables, multiway univariate statistics, tabular reporting, and graphics capabilities can be used to describe the data in a given sample set. For example, descriptive statistics can help define the level of data variability. Inferential statistics with one- and two-sample inference, enumerative data, analysis of variance (ANOVA), correlation analysis, regression analysis, multivariate analysis, time series analysis, and reliability and survival analysis can be used to draw reasonable inferences about groups of data sets. Statistical analysis should be command-driven with a graphical user interface so manufacturers can develop statistical controls to help read any data source and access data from anywhere. With centralized access to multiple data sources such as databases, spreadsheets, uniform resource locators (URLs), and data files, manufacturers can develop complete and accurate metrics. Engineers and scientists are accustomed to exporting and importing data, but for regulatory compliance, data access must be direct, because compliance requires both validation and revision control. In the past, manufacturers had to download data using software. Today, technology allows data to be integrated directly into an MES using analytical tools such as Statit, Minitab, SAS, and Microsoft Excel. In addition, integral charting capabilities mean that operators can receive interactive feedback (e.g., through SPC charting). Web clients also are available and can provide customizable viewing so end-users can operate proactively, with a complete audit trail. The successful integration of existing software forms the backbone of a reliable data retrieval and analysis system. Proven technology such as an MES can tie together disparate data and enable more-efficient production. By viewing the current technology in new ways, manufacturers can implement innovative business processes while meeting regulatory compliance requirements. (continued)