An Introduction to Manufacturing Intelligence Software (MI)

MI can realize synergies between the top-down approach of ERP and the bottom-up approach of MES.

Manufacturing Intelligence (MI), also known as Enterprise Manufacturing Intelligence (EMI), software delivers real-time information about manufacturing processes to help businesses optimize the performance of these processes as well as manufacturing yields. Manufacturing analytics within MI are statistical and rule-based. MI software gathers and analyzes production data, provides role-based visualization, and helps manufacturers reduce waste. The software also enables the improvement of manufacturing processes, identification of best practices, and the ability to respond to exceptions and events. Additionally, it aids in the prediction of potential problems before they impact critical factors such as cost, quality, or yield. MI software is frequently grouped with performance, plant, and factory management software.¹

Manufacturing Intelligence draws from the known concept of Business Intelligence (BI). While BI focuses on acquiring data and keeping it available for management, MI connects the various data sources, including proprietary data providers and BI systems, to pull the requisite information from them. This allows MI to realize synergies between the top-down approach of Enterprise Resource Planning (ERP) and the bottom-up approach of Manufacturing Execution Systems (MES). It also aggregates business management, production engineering, and technical product data to derive important performance indicators such as cycle times, order levels, and utilization rates.²

The term “Enterprise Manufacturing Intelligence” was first applied to the Lighthammer “Illuminator” product in 2001. (Lighthammer was a pioneer in Manufacturing Intelligence and integration; the popular SAP MII is a direct outgrowth of SAP’s acquisition of Lighthammer in 2005.)

AMR Research, now part of Gartner Group, identified five core functions that every MI application should possess:

• Aggregation: Provides availability to data from multiple sources, most often databases.
• Contextualization: Supplies a structure or model for the data that will help users find what they need (e.g., a folder tree utilizing a hierarchy such as the ISA-95 standard).
• Analysis : Enables the user to analyze data across sources (and manufacturing sites), providing a foundation for genuine ad hoc reporting.
• Visualization: Provides tools to create visual summaries of data (e.g., dashboards) to alert decision-makers and focus their attention on the most important information at the moment.
• Propagation: Automates the transfer of data from the plant floor to enterprise-level systems such as ERP.

Typically, MI uses embedded applications to perform device-level diagnostics for production equipment and manufacturing systems. As a rule, these are statistically based and used to identify significant exceptions and events. MI software also provides trend and threshold analysis, key performance indicator (KPI) alerts, downtime analysis, process verification, and process optimization. In some manufacturing environments, MI is used to develop predictive systems and knowledge bases. The software’s operational metrics can be used to drive down waste, determine causes of manufacturing inefficiencies, and combine operations data with ERP applications. Other uses of MI include process monitoring and improvement, regulatory compliance, and supply chain certification.⁴

Most MI software is Web-based or designed to run on personal computers and workstations. Embedded applications may use serial communications protocols such as RS232, RS422, and RS485. Other choices include Ethernet, DeviceNet, and controller area network bus (CANbus). Limits can be entered manually or controlled automatically, though manual entry is diminishing rapidly. MI visualization tools, principally dashboards, provide users with graphs, charts, and histograms for statistical analysis and predictive modeling.⁵

The foundation of any manufacturing organization is the plant floor. As the place where goods are produced, the plant floor contains a plethora of hardware, software, and processes that have particular connectivity. This underlying connectivity creates relationships that must be understood to turn data into information that can be used by enterprise systems. This is the core domain of MI.⁶

Simply put, the main objective of MI is to transform large amounts of disparate manufacturing data into usable, understandable, actionable knowledge that is subsequently used to achieve favorable business results.