There has been a growing interest in Smart Manufacturing techniques including the use of artificial intelligence (AI) to automate routine decisions, enhance process performance and coordinate supply chain activities. This has resulted in a strong demand for quality data. Manufacturers understand that the insights generated by these new analytical methods are only as good as the data they use.
When we check the weather or traffic conditions on our phone apps, we expect to see real-time information collected automatically through sensors. We don't expect someone manually calculating in a spreadsheet and providing insights based on data from days ago. However, when it comes to factory information, the latter scenario still seems to be prevalent.
Fortunately, the technology to collect and organize real-time production data for Smart Manufacturing techniques is available today. This includes data collected directly from machines and processes, as well as data entered by operators during their tasks, product inspections and material handling. It encompasses structured data from sensors and measurements, as well as unstructured data like logs of customer complaints.
Organizing and contextualizing this data is essential for effective mining and analysis, surpassing its initial use for process control or regulatory compliance. Moreover, its crucial to ensure that the raw data is stored with additional contextual data and organized using standardized operational level information models for advanced analysis.
To illustrate this, let's examine how different departments track data related to a mechanical press in a production line. Process engineering monitors the live data stream from the press using a programmable logic controller (PLC). They collect various data points such as temperature, pressure, vibration and forming time. All the machine and sensor data are stored in a time series historian database.
Operations management tracks the start and completion times of production runs, material consumption and product quantities. They utilize a manufacturing execution system (MES) with a relational database for production job tracking.
In this typical scenario, each department tracks the data necessary for their specific record-keeping and reporting needs. The system owners of each department have a comprehensive understanding of their data users, how they use the data and their preferred ways of accessing it. However, the goal of Smart Manufacturing is to organize and make integrated production data available for optimizing the overall production process.
To achieve this, a Smart Manufacturing information platform can use a data mesh or data fabric architecture. This provides shared data infrastructure, governance, integration and access to data consumers across the enterprise. At the same time, it maintains decentralized data responsibility for the domain owners who manage the data and systems for various departments.
Establishing this kind of information platform does require some assembly, but there are IIoT edge gateways available for connectivity, as well as cloud data storage for scalability.
This information platform should also incorporate a metadata catalog for data discoverability and access through standardized information models. This helps achieve greater interoperability between data producers and data-consuming applications, including AI-based monitoring processes.
In the case of the mechanical press process example (see graphic below), the information model would encompass the historian data, along with the relevant data from the production jobs in the MES. Digital dashboards can monitor performance and autonomously schedule tasks based on real-time data. Meanwhile, Smart Manufacturing apps and AI technologies are employed to analyze the integrated data, identify patterns and determine the optimal production conditions.
With the technological advancements available today, manufacturers can bid farewell to manual spreadsheets and embrace real-time data and AI-driven methods. This shift allows for the optimization of internal processes and grants the ability to participate in highly integrated digital supplier networks.
For more information on Smart Manufacturing techniques, visit CESMII.org and MESA.org.
Conrad Leiva is vice president of ecosystem and workforce development at CESMII – the Smart Manufacturing Institute. Conrad has more than 30 years of experience in manufacturing as an industrial and systems engineer and more than 10 years working on Smart Manufacturing systems and practices with industry leaders at MESA International and CESMII.
