During the 1970s, the process industry was at the forefront of digitalisation. At the time, the vision of a single integrated system running on a database became what would later be known as ERP (enterprise resource planning). The earliest large-scale deployments of ERP systems originated in the chemicals sector and were quickly adopted in industries such as oil and gas and mining. ERP has continued to evolve until it now reached a point where it is impossible to run any modern business at scale without it. At the core of ERP is a single integrated database. Initially focused on financial management and MRP (materials resource planning), ERP has evolved to embrace the whole value chain, from vendor management, manufacturing, supply chain/logistics, customer relationship management and more.
However, anyone involved in an ERP implementation will realise that it is not a simple exercise. In practice, ERP only partly fulfils the promise of a “single, integrated enterprise system”. Despite the systems having wide-ranging functionality, silos within the business are never completely eliminated. The lack of integration standards also meant that a single model of the total business value chain is only partly effective. It was also realised that in a tightly integrated system, data errors quickly propagate throughout and can lead to incorrect information and poor decision making. Many businesses, despite having implemented state of the art ERP systems still operate as disconnected silos with a significant percentage of processes at the interfaces being manual. After some quick wins, the cost/benefit of more integration becomes more difficult to justify against the increased complexity and rigidity of a more tightly coupled system.
Notwithstanding the many failures in ERP, digital business systems still continue to evolve and get more complex. Industry 4.0 has raised the bar on connecting data from multiple sources across entire supply chain networks. Notwithstanding the effectiveness of their ERP system, manufacturers need to take these new technologies on-board just to remain competitive.
The digital twin and the digital thread
Two important Industry 4.0 concepts are starting to demand more attention. These are the digital twin and the digital thread.
A digital twin is a representation of the physical entities, processes and people in the business. This includes plant and business data from ERP systems. The digital twin model can be fed with near real-time data from sensors in the plant or through the Internet. In theory, this allows the wide-ranging optimising of operations, as well as lifecycle tracking of products as they are designed, made and consumed by customers.
The digital thread is the framework that supports the flow of data that makes up the digital twin. The digital thread is fundamental to the digital twin concept. The digital thread will support data flow between disparate systems and across silos within the business.
Without this common thread, many of the concepts essential to Industry 4.0 will not be possible.
Not only should the digital thread be able to support near-real-time data such as data from the IIoT it should also be able to support the original design information. This includes, for example, process and instrumentation diagrams (P&ID), CAD models, equipment datasheets, maintenance instructions, and so on. The digital thread should also accommodate historical information from the construction, commissioning and operations.
A digital thread is, in short, the seamless digital connection that weaves its way through the entire operation.
The complexity of implementation
However, while the concept of a digital thread may seem captivating, it is by no means easy to implement. As an example of the complexity, consider the process industry. The product (for example a speciality chemical) is conceived in the early engineering stages of a project.
Various stakeholders are involved, including owners, business development, engineering and procurement (EPC) contractors, commissioning teams, operators and so on. Most of the equipment design takes place during the project execution phase, with the deliverables of this phase being final piping and instrumentation diagrams (P&ID), equipment datasheets, mechanical design, drawings and instrumentation drawings. In a typical EPC environment, disparate computer systems are often used by the various disciplines to produce the design. The actual equipment is then manufactured by third parties, who in turn use another set of digital technologies. During construction of the plant there might be further changes, where again, another set of systems might be in use. And so it continues, throughout the commissioning of the plant and the subsequent operations.
Considering the challenges associated with ERP projects, can we not then expect the same challenges when implementing the digital thread? Is the digital thread concept simply too complex to be of any use?
I would argue that it is very early days still. However, several of the major control and instrumentation vendors are coming up with viable platforms for the digital twin/thread and I believe it important to take note of these offerings.
In the process industry, there are several initiatives underway to define common standards for data exchange. One such initiative is DEXPI (which stands for Data EXchange in the Process Industry). DEXPI is a working party of the ProcessNet initiative in Germany. The objective of DEXPI is to address the interoperability challenges between computer-aided engineering and other systems in the process industry. One goal of DEXPI is seamless interoperability of design, operations and maintenance information between EPCs, owner-operators and vendors. DEXPI is currently working on a data exchange model based on the ISO 15926 standard.
So, what is DEXPI actually? It is an information model that represents the objects that appear in a P&ID; diagram, such as vessels, heat exchangers, pumps, instruments, motors etc. The model is represented as entity classes, similar to UML in programming. The result is a standardised representation of all the physical entities in a manufacturing operation; in other words, it can serve as the foundation for the digital thread.
It takes time for big shifts to take place, and traditional process manufacturing can sometimes be slow off the mark. But with the support of major owners, vendors and EPC contractors, the investment in a universal model for the industry will make the goal of implementing a digital thread framework that much easier. I would recommend keeping an eye on these developments if you are in any way involved in manufacturing IT.
This article was first published on SA Instrumentation and Control.
