Representative End User Clients
Representative Automation Clients
Representative Software Clients
Clearly, to enable IIoT in the manufacturing environment and to facilitate the building of digital thread and digital twin, we need some systematic solution: a consistent framework to collect, manage and analyze data from the products, machines and processes and distribute the insights from the analytics instantly to the process applications so they can respond to events timely. The same framework also needs facilitate building and maintaining the digital twin throughout the manufacturing processes and even from the product usage and maintenance after they are deployed. Finally, it needs support a large scale computation to perform Big Data analytics, model building and simulation.
A comprehensive data and analytics stack is needed to meet the needs and satisfy the requirements. Figure 4 illustrates the essential layers in the stack and their relationship in a distributed computing scheme. The lower stacks are for fast and scalable streaming data collection, processing and analytics, the middle stacks for scalable and durable data storage and management, and upper stacks for large scale and intensive batch-oriented analytics, likely based on a Big Data framework.Evidently, distributed analytics is required: to perform analytics close to the data sources and decision points to meet the latency, reliability and security requirements – that means we need to be able to deploy streaming analytics at the edge at or near the plant level. We also need to deploy large scale batch analytics in the cloud to take the advantage of its abundant computing resources.
Now to put everything together, architecturally, in a digital manufacturing environment. At its bottom, we have the CPS plane consisting of manufacturing equipment, products and the physical facility and environment.
We then have our manufacturing planning, decision-making and execution plane – consisting of processes in the value chain and the product chain.
To apply IIoT and industrial analytics to the manufacturing processes, to enable intelligence for smart manufacturing, we need to insert an information plane to provide the functions for connectivity, data collection and analytics, in order to provide the necessary insights to the decision-making processes in the top plane.
To look at this architecture a different way, we can see that the conventional ISA 95 or Purdue model becomes a multi-facet pyramid, as shown in , with the value chain (ERP) represented in one facet, and the product chain (PLM) in another. At the middle, we are introducing an IIoT and industrial analytics in a new facet providing the insights to each of the other two facets to achieve their respective goals as outlined in the figure.
Moving up the pyramid layers, the industrial assets and processes are organized from small to large scale of CPS systems. These systems span from the edge to the cloud as just as analytics, other manufacturing functions, such as ERP, PLM and MES will be increasingly distributed from edge to cloud.
Furthermore, this three-plane architecture maps reasonably well to the functional domain of the Industrial Internet Reference Architecture published by the Industrial Internet Consortium, as illustrated in Figure 7
.
Finally, while perusing optimization to enable smart manufacturing, let us keep in mind the vision that we are seeking to achieve as outlined below:
- System performance and capability adaptive as demands and conditions change
- Supply are replenished in time and surplus production minimized
- Downtime is predicted and prevented
- Waste and defects are reduced or eliminated
- Worker safety is ensured and sustainability maintained
- Real time full visibility to operation status with actionable information – enabling a transition from operation management to mission control
To conclude, smart manufacturing is about optimizing our manufacturing processes to do more with less, better and fast! And the Industrial Internet of Things and industrial analytics will play major roles in enabling it.
Dr. Shi-Wan Lin is the CEO and a co-founder of Thingswise, LLC, a startup providing streaming industrial analytics solutions purposely built for IIoT systems and Smart Manufacturing – as a turnkey solution adapting and innovating key IT technologies to meet OT’s stringent requirements in reliability, performance and security, deployable from the edge to the cloud.
Dr. Lin co-chairs various technical groups for the Industrial Internet Consortium (IIC), the Architecture Joint Task Group between Plattform Industrie 4.0 and the Industrial Internet Consortium and the National Institute for Standards & Technology (NIST) Cyber-Physical Systems Public Working Group. Dr. Lin is a lead editor and contributor to the Industrial Internet Reference Architecture published by IIC.
Previously, he worked for Intel for 10 years last as a Principal Engineer/Technologist in the Strategy and Technology Office in its Internet of Things Group and before then Sarvega, Inc (a Web Service/SOA/Security startup), Lucent Technologies and Motorola. Dr. Shi-Wan Lin has 20+ years’ experience in system architecture, Big Data, analytics, enterprise software, Cloud Service, system security and trust, telecommunications and wireless data communications.
The opinions expressed in this series of blog reflect his personal view and observations and he is solely responsible for them.