OPC UA Information Modelling: The Glue That Binds Industry 4.0

2021 Update

OPC UA. It's a standard that I feel is not fully understood or explored, especially when it comes to its information modelling capabilities. There are many online articles that talk about OPC UA but most of the time, it seems to be a comparison with other protocols such as MQTT. An argument which I believe is misplaced.

I mean, there are many other industrial communication protocols that are being touted as Industry 4.0 enablers. Can't we just pick one and go? Why all the pomp and circumstance with OPC UA?

Here's the thing, most of the existing industrial communication protocols cannot fulfil the extreme requirements for the perpetuation of Industry 4.0. But they do have to exist as a transportation mechanism for OPC UA modelled information that is rich in semantic content, a key enabling attribute for virtualisation and Industry 4.0.

In this article, we are going to break down OPC UA's information representation methodology that makes it an indispensable part of Industry 4.0.

OPC UA Address Space Model

By now, I'm pretty sure we all agree that a seamless exchange of information between physical entities at the plant floor and ICT systems is essential to enable Cyber-Physical Production Sytems in manufacturing. Fundamentally, the goal is to create an environment that makes information represented by process management systems at the plant level indistinguishable from that produced and consumed by business management systems. And to also be able to virtualise real-world objects such as pumps, boilers, robot arms and the underlying real-time processes in order to interact with them in their wholeness in cyber space.

But as it turns out, the communication frameworks that are predominantly used in manufacturing plants fall far too short from fulfilling this and many other requirement of Industry 4.0. Here's why. Common industrial communication protocols such as Modbus typically use Request/Response frame pairs, meaning that you can only get one piece of information per transaction e.g the steam temperature. Data about that piece of information(metadata) is missing, for example, what are the units of measurements?

As if that's not enough, there is no depiction of the relationships between relevant objects, for instance it would be useful to the client requesting the information to know that the pipe carrying the steam has a flow transmitter that loops back to a flow controller. No doubt about it, it would be a mammoth task to try to represent a complex object like an entire plant using this methodology, you could try to request the piece of data and rebuild the bigger picture on the other end but that would be a very expensive process resource-wise and there would be no data integrity, or robustnesss at all. And bear in mind that for it to be a truly universal standard the clients need not have prior knowledge of the server's data structure.

Simply stated, it's entirely a different set of concepts or standards that are required for this task. What is required and OPC UA brings to the table instead is an information architecture that is able to map the information content of a complex object and its underlying processes for modelling of its digital equivalent. A digital twin. To overcome the limits outlined earlier, OPC UA introduces the browsing mechanism and the reference concept to make information internally consistent and available to clients as a large whole.

To make a point, I will use an example that you can easily relate to, as the Address Space Model concept can be a bit too abstract to grasp. When you visit a website for the first the only information that you need to know about it is its web address, you don't know how much content is there , how it is structured or how to find it even. But when you type in the address you are presented with a menu with links to various pages and sections. you actually get to discover the structure of the website step by step using links and consuming content of interest as and when you find it.That is exactly how the OPC UA Address Space Concept works (Well, maybe not exactly but I'm sure you get the point).???

OPC UA uses an Object model i.e each variable, method, and event in the underlying real-time process is represented as an object. Interconnected by references and collectively exposed as information to clients. Information that describes the state and behaviour of the process. The set of objects that the OPCUA server makes available to clients is referred to as its Address Space. Even better, this means that integrity is not compromised as data information about a complex object can be processed and transferred as one item, represented as uniquely as possible and all the while allowing it to be mutually understandable by independent parties. The OPC foundation provides rules governing the creation and modification processes of the Address Space.

Think about it, this is the true definition of interoperability. The kind that Industry 4.0 can rely on to enable a plug-and-produce functionality where devices can be plugged onto a network and be able to navigate through the network and obtain all the formation they need about information sources unknown to them before without configuration. And as they say, the OPC UA Address Space is capable of displaying movies, and not just static pictures. But wait there's more, let's take a look at the OPC UA information model.

OPC UA Information Model??

If you are familiar with computer programming you've most likely used or at least come across Object Oriented Programming. A concept that uses four pillars namely, inheritance, polymorphism, data abstraction and encapsulation to build computer programs. Th good news is that OPC UA Information model employs the same concept. To achieve that, OPC UA provides a way to define and expose object types (classes with member variables and member methods) and object instances. Right at the start, the OPC foundation defines standard types from which you can inherit to create your custom types. So for example if you wanted to expose information about a compressor you could inherit OPC UA's BaseObjectType to define a new type called CompressorObjectType with member variables such as Motor Temperature, Vibration, Motor Current and Member Methods such as Start Compressor and Stop Compressor.

What does this mean? This means that engineering effort is significantly reduced as it allows development against a type definition providing dedicated functionality, for example graphically displaying the compressor as expected by the client. This also means that reusability is greatly improved as type definitons by one OEM manufacturer can be inherited by another with appropriate modifications. Which brings us to the next point. Companion specifications.

There is currently widespread collaboration across industry whereby organisations are using OPC UA provided base information models to define standards in their own context. The collaborations span from industrial communication protocol modelling where protocols such as Profibus, BACnet etc are being mapped to OPC UA, to vertical specific collaborations such as in Oil and Gas, Energy and Factory Automation. This approach enables common processing of information by variety of systems in various industry verticals and brings about a universal and uprecedented standard on which Industry 4.0 can thrive.

In conclusion, OPC UA Information Modelling is such a broad and compex subject. So much so that this article does not even begin to scratch the surface. All the same it is a concept that I believe is truly fundamental in transforming how information is interchanged in industry, so I thought to share.

Do you agree that OPC UA Information Modelling is an indispensable part of Industry 4.0? Please share why (or why not!) in the comments!