Achieving Useful Plant Information Made Easier

Yet another Holy Grail in any manufacturing business is having a true enterprise-wide information system for better operational control.

Yet another Holy Grail in any manufacturing business is having a true enterprise-wide information system for better operational control. Human/machine interface (HMI), supervisory control and data acquisition (SCADA), and manufacturing execution systems help, but problems still exist. First, making sense of the massive volumes of plant data is hardly “non trivial.” Second, linking conventional business systems, such as enterprise resource planning (ERP), to real-time plant floor data is a nightmare in terms of custom coding, multiple information and communications technologies, and inconsistent plant implementations.

Some approaches at solving these issues (and getting closer to the Grail) have emerged. One, VisualPlant from Executive Manufacturing Technologies, Inc. (EMT, London, Ontario, Canada), is an “off-the-shelf,” Internet-based application that builds a data warehouse of machine states, production data, downtime incidents, process variables, and worker productivity based on the raw data from the plant floor. It then applies logic and business rules to those data. Web browsers dynamically display the resulting information.

VisualPlant does not perform equipment control per se, so it’s more like HMI than like SCADA. But unlike HMI, points out John Dyck, EMT’s vice president of marketing and business development, VisualPlant generally connects to any number of plant devices, and it provides visualization, analysis, and reporting—all through a web browser. Also, VisualPlant provides the plant floor data to a standard, open database where data mining tools can uncover production and equipment trends. Moreover, VisualPlant does all this without the need to write code or edit a database. In fact, VisualPlant is “non-intrusive” in that it “maps directly to plant equipment—right to the source,” says Dyck, and after some further configuration, it operates transparently.

Dyck is also quick to point out that VisualPlant is not a portal, which typically just points to different data sources around the plant and displays data extracted from those sources in a pretty screen. Portals often bog down when they become awash with data—and no way to analyze those data.

VisualPlant meets the latest buzzword checklist requirements starting with it being built upon Microsoft .NET architecture. It can be integrated to third-party applications, including wireless paging systems, and it can be extended using snap-in capabilities to third-party ActiveX components.

The application consists primarily of four software modules. VPCollector collects data using OLE for Process Control (OPC) as the glue between VisualPlant and the sources of plant data, such as programmable logic controllers (PLC). Plant data collect as standard database objects in a central database, VPServer. VPWeb delivers both historical and real-time data from VPServer to web browsers. The fourth module, VPAdministrator, is used to configure and manage VisualPlant so that it matches your plant’s control and equipment hierarchy (using the names and descriptions used in your plant)—no programming knowledge or database experience are required.

Once VisualPlant is installed, you can monitor the plant from its production lines and work centers on down to individual equipment and controllers through an Explorer-like, two-pane window displayed in a web browser. A tree view in the left pane lets you quickly navigate to specific areas in the plant. A graphical display on the right shows details, including production data, key performance indicators, data summaries, device reference data, and machine faults and alarms (historical or real time). You can stretch measurement scales, pan through time frames, and apply statistical tools to the data, as well as mouse click your way through OLAP tools to further analyze the data. The software includes about a hundred canned reports that can be modified and saved for future use.

EMT claims many major automotive OEMs and automotive suppliers are using VisualPlant. It also claims that VisualPlant has an ROI of six months or less. VisualPlant costs $1,250 per equipment asset, or node. A node is a logical grouping of equipment and equipment controls. If, for example, you have five operations on an automotive transfer line, they can be considered one asset. (While each of these operations produces parts, if one fails, the entire transfer line might stop.) While assets can be just about anything a plant wants it to be—PLC, robot controller, strain gauge, even a standalone database—nine times out of ten, says Dyck, one piece of equipment equals one PLC, which equals one asset.

But get this. For that price, any number of people can look at the information from that asset; the price of VisualPlant is based on assets, not the number of people using the data.

If your Grail is plant information, this might fill your cup.