Manufacturing plants aim not only to ensure safe operation, but also to maximize revenues and profits throughout the lifecycle. Besides these considerations, increasing competition and complexity of manufacturing operations is pushing companies to seek access to data mining capabilities so that they can extract hidden information and discover underlying patterns in large data sets. Data mining capabilities will empower companies to properly contextualize the information and carryout further analysis. That way, they can gain the necessary insights, which will help them initiate the required actions and test their appropriateness.
Through data mining, manufacturing companies will be able to garner the necessary insights that will help them comprehensively manage business, production, and process operations. The data sets may come directly from the plant’s physical assets and processes or extracted from operational technology systems, such as programmable logic controllers, distributed control systems, safety instrumented systems, etc., and information technology solutions, such as enterprise resource planning, supply chain management, etc.
digital twin’s Value Proposition
digital twin leverages a combination of industrial internet technologies, such as industrial internet of things (IIoT), artificial intelligence (AI), data analytics, edge and cloud computing, and more. A powerful tool that has the potential to address the requirements of manufacturing plants, a digital twin is a computational model of a real physical asset in a plant that can be kept updated across its entire lifecycle.
Manufacturing companies can leverage digital twin to study the physical asset’s behavior, predict their performance, and optimize production processes. digital twin can be applied to gain insight about the asset health and improve their availability, enhance business, production, and operational performance, reduce costs, minimize environmental impact, enhance agility and competitiveness, prioritize plant safety, and more. With its help, the design and engineering concepts can be validated and deviations and anomalies from the design criteria can be identified.
Typical Application Objectives of digital twins
digital twin may be built to serve numerous objectives. For example, energy intensive companies may use digital twin for energy conservation and for matching energy demand with availability; process plants, with its help, can also determine the optimum operating conditions, which do not transgress the economic and technical limits. digital twin can also be built for forecasting process anomalies, equipment failures, and exploring the preventive measures options. digital twin is also a powerful tool to carry out root cause and what-if analysis.
Let us take as a case-study a thermal power plant that is connected to a region’s common grid. The power station may typically have multiple boiler-turbine units of different generating capacities. When the demand for power increases, it will become necessary for the power station to decide about distributing the load optimally, but not necessarily equally. A digital twin can help in optimally loading the individual boiler-turbine units depending on their load-efficiency profiles.
In a thermal power plant, boiler feed water pumps have a critical role to play. Failure of a unit’s feed water pump can lead to its shutdown. Such instances can be avoided with the help of a properly designed digital twin of the boiler feed water pump, as it can predict the impending failures. Boiler tubes failures are quite common in a power plants and at times can result in the plant shutdown. Root cause analysis can help identify the causes so that corrective actions can be initiated.
Some original equipment manufacturers (OEM) are also building digital twins of their products, such as compressors and pumps, to develop new revenue streams from the services related to products they sell. They aim to create a new business model, which involves offering product performance along with related maintenance services. In this business model, described as product-as-a-service or pay-per-use, the product is supplied as a cyber-physical system embedded with sensors and the digital twin so that the OEM can track the performance of the equipment remotely on a real-time basis and guarantee performance.
digital twin Implementation Strategy
Each company that finds economic justification to harness the power of digital twin has to define the goals proposed to be achieved based on its own economic evaluation, as well as draw up the development and maintenance plans for each digital twin, and implement them on the basis of its priorities. Some of the digital twins may also be integrated to maximize their value proposition.
The computational model can be either first-principles based or data-driven:
- The first-principles based model, also called white-box model, is developed by using the laws of physical sciences that govern the physical asset that it replicates.
- The data-driven model, sometimes referred to as black-box model, is built by using the data of the replicated asset obtained with the help of data analytics and machine learning. The model can also be updated from time to time so that it remains current.
The ambit of the model may cover a single asset, such as the feed water pump of a thermal power plant or a single process unit such as a steam generation section of a thermal power plant. It can also be plant-wide, such as an integrated model of the boiler, turbine, and generator.
A report that sources LNS Research says that over 1,000 digital twin models may have to be built to represent the working of a typical refinery. While companies that want to leverage the power of digital twins to push their performance boundaries farther must be prepared to put in significant efforts to build and maintain them, their efforts will be immensely rewarded.