Detailed traceability without the overhead

Shaun McGinnis says the best way to visualize the large-scale data storage systems made by Hitachi Computer Products America is to think of "a refrigerator packed with boards and drives." That means the equipment is powerful, but at the same time, it's challenging to manufacture not only because of the assembly involved, but the testing and traceability requirements too.

McGinnis is production engineering manager for this Hitachi division, based in Norman, Okla. He says an ERP system from PeopleSoft drives production planning and material management at the plant, but achieving a single automated system for traceability and integrated quality management concerns drove Hitachi to implement a manufacturing execution system (MES).

Before the MES, says McGinnis, Hitachi had in place some point solutions for work-in-process (WIP) tracking and quality management, but lacked a single execution system to track the progress of work from its beginnings, and on to shipment, and record the configuration and genealogy of each product shipped. A survey of relevant vendors led Hitachi to Visiprise and its Visiprise Manufacturing suite.

"We use the MES for quality management and to support our Six Sigma efforts—and it's also good at traceability," says McGinnis. "If we ever have a field problem with a product, we can quickly search through that product's genealogy to pinpoint information and solve the problem," he says. "It's much faster than working with paper-based records and data from point solutions."

McGinnis says the ability of the MES to define and track the parent-child relationships between assemblies and subassemblies supports efficient data collection on systems that might contain 400 subassemblies when fully configured. "We can just scan the parent, and collect the data we need with one scan versus 400 scans," he says.

The implementation did hold some complexities, says McGinnis. For one, Hitachi emphasized operator retraining in new procedures such as tying serial numbers to specific shop orders. The deployment also was phased in, beginning in August 2002. The first phase covered the creation of shop orders and basic execution functions such as WIP tracking, as well as the collection of core inspection and test data.

Later stages collected further test data, automated integration of data from test equipment to MES, and instituted detailed genealogy tracking. Functionally, the deployment was broken down into five subprojects: genealogy, configuration, manufacture scan/WIP tracking, process control, and revision control. Says McGinnis, "We now have one production management system that forces detailed data collection against a serial number, but that doesn't require additional employees or negatively impact our lead time to customers."

The more granular data held in the MES, along with its reports, also helps Hitachi find areas of waste, says Joe Soliz, head of test engineering at the Norman plant. The MES deployment was preceded by a successful effort to reduce lead times, says Soliz, but "with the MES, we found areas of waste we didn't realize were occurring. Now we have the data resolution and the visibility into cycle times that makes small deviations apparent."