The Case for DIPS and Distributed, Intelligent Automation

Summary

The Industrial Internet (II) represents a new stage in the advancement of human society. Its technical architecture is best summarized as a Distributed Intelligent Production System (DIPS), which is a mesh of various II “nodes.” These nodes range from the traditional—various automation systems, devices or independent sensors and various enterprise operation systems—to new II nodes such as Industrial Internet of Things (IIoT) platforms, artificial intelligence (AI) platforms, edge computing platforms and more. DIPS transcends the practical limitations of staying on a single II platform and solves shortcomings caused by suppliers neglecting the importance of merging their various advanced products into one unified automation architecture. 

History 

The manufacturing industry is classified into two categories: discrete manufacturing and process manufacturing. These two categories were once unrelated to each other, but now they have become part of the whole in the era of intelligent manufacturing.  

In the era of Industry 3.0, the focus was on how to use raw materials to make products; in the era of Industry 4.0, further consideration should be given to auxiliary materials, by-products and pollution. The representative system of Industry 3.0 is the distributed control system (DCS). What is the representative system of Industry 4.0? Is it an II platform? We believe it should be the Distributed Intelligent Production System. 

Besides product manufacturing, DIPS addresses four common categories of concern, as summarized by authoritative international institutions: safety concerns, production efficiency, equipment maintenance and pollution emissions. 

According to the ISA95 standard, the various daily operations of manufacturing enterprises are divided into five layers: The first and second layers, L1 and L2, include various production equipment and control systems involved in automated production. The third layer, L3, is about the production operations, including manufacturing execution systems (MES) and production management. Layer L4 above it includes various business, management and decision-making applications. The top L5 layer is about presentation and decision making. A good enterprise not only needs to achieve optimal efficiency for every application in these layers, but also requires coordination and synchronization between applications, as well as interconnectivity of data. 

So, in the era of intelligent manufacturing, we need an architecture in which the system can control basic production automation while serving safety concerns, production efficiency, equipment maintenance and pollution emissions, as well as accomplish all the operations described by ISA95. That architecture is DIPS. Here’s why. 

DIPS architecture 

The industrial systems architecture in the era of intelligent manufacturing should be a distributed one. The core of the DIPS architecture is its nerve center, which manages all the II nodes and the entire distributed network. As mentioned, traditional II nodes include familiar automation systems: the DCS system, programmable logic control (PLC) system, safety system, as well as upper layer ISA95 systems including manufacturing execution systems (MES), product lifecycle management (PLM) systems, customer relationship management (CRM) systems and more. Simple sensors may also be II nodes.  

The new market growth point of industrial manufacturing lies outside of automation systems, however. It is difficult to fully integrate safety, environmental protection, energy conservation, process optimization and other advanced II services into traditional automation systems. It requires a special class of nodes to perform those duties, which we call II Platform nodes. These nodes complete tasks beyond traditional manufacturing tasks.

Here is more on three common II Platform nodes. 

Central nerve system. The DIPS nerve center is a special II Platform node. It manages coordination among different nodes and all the operational data. The organization and management of data is already very difficult, and with distributed management, it becomes even more difficult. An important component in the nerve center is the “data sharing configuration.” Any DIPS data can be accessed through the nerve center and data that is not physically stored in the nerve center is indexed by the “data sharing configuration” module.

AI nodes. We place the data, algorithms and computing power that represent “intelligence” in a separate AI node. Considering real-time factors, the online operation of DIPS should not be affected by offline AI computation.  

Stars. A “star” is a productized general-purpose IIoT platform, so Stars in the DIPS architecture are IIoT Platform nodes. Most of the popular II platforms in China are in fact Stars. A Star can also run in a industrial computer workstation.

Implementation 

Standardization and productization are the only ways to achieve DIPS. Standardization enables seamless integration of products from different suppliers.  

The IEC30165 standard, published in 2021, describes the IIoT as a real-time system. Multiple II nodes mean that the interconnection between them and the management by their nerve center are very important—and no less important than the II platform itself.  

DIPS hardware products include chips, smart devices, communication devices, control devices and servers. DIPS software products include embedded software in the hardware products and software in the servers. The software in the server is the II platform software. 

One example of a successful application of DIPS is a digital transformation project for a leading global phosphorus chemical industry enterprise. Several key modules of the project have been developed. The most important one is the establishment of the DIPS nerve center to support continuous DIPS integration.  

The following are two items related to the project’s central nerve system:  

1. The “Data Operations Platform” is the project’s DIPS nerve center. It achieves data interconnection and interoperability between production systems and ERP, EAM, OA, SMS platforms, the government emergency management platform and other systems.  

2. There is a preliminary attempt at intelligent production. Chemical enterprises rarely just produce a single product and the by-products of one product are often used to produce other products. We built production scheduling recommendation software and implemented it on the DIPS nerve center. 

DIPS can be implemented in different formats to support different applications:  

Enhanced version of the automation system. DIPS, if it contains only one automation system, could be considered as an enhanced version of that automation system. In that case, an II node with advanced service is considered part of the automation system. DIPS with only one DCS node can also be called enhanced DCS. DIPS with one PLC in it becomes an enhanced version of a PLC system. If computer numeric control (CNC) is inside, DIPS becomes an enhanced version of CNC.

Enterprise digital transformation. The overall digital transformation architecture of an enterprise or a conglomerate is a complex DIPS system.

Smart industrial park. It’s a standard practice in China nowadays to integrate relevant manufacturing companies inside one industrial park. The internal DIPS systems of various companies in the park are connected to form a park-level DIPS system. There needs to be a DIPS nerve center at the park level.

Smart industry supply chain. The internal DIPS systems of various enterprises on the industry supply chain, integrated together, form a supply chain DIPS system. There needs to be a DIPS nerve center at the supply chain level.

Smart city. The current smart city practice is still based on the Internet and needs to be improved to be based on IIoT. It can integrate urban infrastructure DIPS, urban building DIPS, urban enterprise DIPS, etc. to form a larger level of smart city DIPS. A new global nerve center is needed.

Summary 

DIPS transcends the practical limitations of staying on a single II platform, which is commonly practiced in China. Thus, it redefines the role of the II platform. DIPS surpasses the shortcomings of international suppliers in neglecting system integration to unify into one architecture their various advanced products that only serve segmented requirements.  

Whether from the data level or the computing level, Industrial Internet must be of a distributed DIPS format. AI analysis cannot interfere with the real-time operation of DIPS. Finally, standardization/productization is the only way forward.

_{This feature originally appeared in the April 2024 isue of InTech digital magazine.}