Exponential innovation fuelling hope, disrupting the VUCA world with the onset of the arrival of industry 4.0 waves, building seamless bridges with entities: Digital world, the Virtual world with the Physical World. The digital technology has opened up plethora of opportunities created for full set of new age technologies stack of future e.g. Artificial Intelligence (AI), Advanced Analytics, Augmented Reality (AR), 3D Printing, Quantum Computing, Smart Automation, Robotics, Bioinformatics, Nanotechnology, Neurotechnology, 6G communication, Blockchain, Internet of Things (IoT), Autonomous System, and Cybersecurity are aimed at addressing human need, improving quality of life by way of speedy development in all spheres of life.

Chemical industry is very vast and all encompassing, comprising of commodity and specialty, basic chemicals, performance chemicals, synthetics, elastomers, API drugs, and pharmaceuticals will become the obvious beneficiary from these high-end technology development for serving human, the ecology, and the environment and in turn help in minimizing waste of scarce resource in every activity performed by human. ‘Process Digital Twin’s is uniquely positioned to utilize at its breadth and reap benefits through deployment of above technology stack depending on complexity of objectives for economic and social well beings with rapidity in future Digital twin entails three distinct parts - the physical products in real space, the virtual product in virtual space, and lastly the connection of data and information that ties the virtual & real products together. It was in the aftermath of NASA’s successful project launch in 2010, digital twin gained tremendous popularity in academia, government institutions and in industry settings across organizations as a ploy to advancing in technology for harnessing business performance enhancement, conceptualization in R&D prototypes in multiple business domain for knowing the unknowns.

On a simplistic view away from complexity humans face difficulty in visualizing in life situation what is happening inside from 2D drawing made from plain paper and pencil and the complexity increases with the multi[1]fold increase in process complexity that petrochemicals product inherits, required to follow a set of fundamental theoretical unit operations (Heat, mas transfer, fluid flow property changes, chemical reactions occurring leading to compositional changes, prediction of product properties occurring inside) for the entire conversion processes tied together, goes from one intermediate to another until the finished product is produced acceptable to customer for specific application. Hence new technology features like ‘Digital Twin’ makes people think better, relate their work better with better real-life understanding.

Design of ‘Process Digital Twin’ and its utility depends on the use case being deployed. Digital twin capability supports three of the most powerful knowledge-tools in human toolkit and these tools are basically conceptualization, comparisons, and collaboration. Taken together these attributes form the foundation of the next generation of problem solving and innovation. In Petrochemicals business, setting Physics and Chemistry led first principle centric models coupled with Artificial Intelligence (AI), Machine Learning (ML) capability i.e. sort of hybrid models will find better suited application in designing a set technology products e.g. Product Twins, Production Twins, Performance Twins, and

Asset twins and all put together gives the digital backbone and helps in separating the crowd from the technology oriented, knowledge driven organization of future and continue to thrive.

Use case deployment should be thought out well and then plan better in advance. Next step is to identify a specific area of business to start with, define the objectives and seek support and funding from stakeholder and top management. Look at technologies that will be necessary, it does not mean that all technologies that are available in the market apply nor are relevant to a particular segment of business that the organization is serving to its customers. Second, identify a team and specify a timeline of completion. 3rd start small and communicate successes, benefits and the value that project has delivered; share difficulties, group learning from the project and lastly be flexible to adopt continual improvement with time and refine it as it goes.

If we deep dive into the digital twin conceptual framework at the core of it is digital representation of a physical product in all its aspects including the functional, behavioural part of it. It represents the whole life cycle of a product starting from designing it mechanically, describing it through embedded software satisfying principles of flow mechanics, thermal, chemical and electrical aspects in it. Digital twin also describes how to manufacture a product, replicating every behaviour of the physical product in operation while in service and during its maintenance activity.

Digital twin represents a physical product in a digital world, allowing us to do experimentation; does the thing faster, for example it helps in timely simulation as opposed to building costly physical prototype; designed to do things more often until the optimal solution is reached. The virtual product so designed will use less and less of resources i.e. both money and human capital.

In essence its core capability is to do multiple simulations, build many different models instead of just one and then pick the one which meets the optimum, suits best for the application. Lifecycle of a physical product starts with designing a product or designing a plant in CAD models, then simulating those CAD models without implementing it real time and then engineered for coding it in detail with an aim to optimize from flow dynamics point of view, determining product attributes like property prediction etc. in essence turns out to be a game changer that provides broad set of models in design space for attaining higher order goals.

Next phase of lifecycle of Digital twin is manufacturing the product and naturally it becomes the longest phase of run where it utilizes the model capability to enhance operational efficiency, reduce emissions, minimize downtime of large machines (compressors, turbines, motors etc) that are integral to plant operations offers enormous visualization to people and processes at large. It is to be kept in mind for operations phase model’s order of complexity ought to reduce from the data processing perspectives, hardware needs, optimizing computer CPU capacity for longer term sustenance for in-line optimization in real-time. It is said that all models are bad but definitely some are good.

It is also to be mentioned here that digital twins are in operation, another very important aspect in product lifecycle is the maintenance activity. When predictability in maintenance comes into picture, it naturally gets linked to artificial intelligence (AI), essentially driven by data that goes into model development; it forms the very important foundation of the product life cycle. With faster adoption of artificial intelligence technology, widening of design space has become a reality; it offers multiple degrees of freedom in carrying out design simulations and helps in selecting the best simulated case for freezing the design, reducing cost and time for complex design practices.

Similarly, the mechanical aspect of de[1]sign, fluid flow dynamics, mass & energy balances based on first principle also becomes integral to it, built on embedded software comes into the model in play. As-built plant design should get accurately represented in the digital twin supported over a longer time horizon as part of product life cycle offering. On a comparative tone industry 4.0 is to gather cues from the environment and remediating it as information is received but on the contrary job of digital twin would be to predict it before an event occurs. In similar technical parlance artificial intelligence is also a technology putting together with digital twin generates data from the physical plant or product and fits into the real-life plant virtual models utilizes for decision making, helps in building predictive and preventive maintenance capabilities for plant equipment and machineries.

With digital twins in play, higher order and more granular attributes in product design can be attained. In the context of using artificial intelligence in the physical product allowed to roam around and if utilized in plant operation-maintenance activities, then it is necessary to deploy a digital twin to build standby capabilities to know what and how it is doing, what sort of emergent behaviour it does, meaning when it was designed we did not know it could do versus when connected to digital twin it will provide understanding what data it took to change such behaviour.

For autonomous entities deployment in industrial and commercial establishments, it is essential that we put digital twins together to gather all core aspects. Putting them together will help in navigating it scientifically.

Digital twin brings value in larger aspects of visualization, a big part of it today for example in 3D visualization gives enormous capabilities. Any intelligent digital twin aided with simulation of physical objects will build capability to augment human decision making and in no way guarantees replacement of humans.

Digital twin has 3 phases of deployment namely: prototype before the physical product is made, 2nd digital instance where an expensive objective is running or put into operation to gather all sorts of behaviour it does and then the 3rd ‘digital twin aggregator’ which takes all the steps that include sensorization of data, pull every information which human think he knows but did not.

Hence the whole idea of digital twin is going to evolve and create capabilities to develop prognostic and predict failures to fix it before they are occurring. It is also true with the phenomenal increase in computational capabilities at a lesser cost, digital twin is going to throw enormous perspectives to bring in futuristic view in hours, days and month ahead to human and help in alerting if same ways of doing the things continues as it is being done; digital twin predictive capability on failure will enhance multi-fold visibility and in a way it will be a real assistance for human going forward.

Digital twin for remote operation-absence of physical proximity can be leveraged simultaneously as an watch dog from any part of the world instantly with strict authorization, securitization etc as a result physical proximity can be avoided for example operations like a nuclear reactor, smelter in a steel melting shop where molten iron is poured in a reactor; close proximity to such operation is not at all desirable but humans are to be there and has to be involved, digital twins provides augmented intelligence. Make sure everybody understands remote operation what it can lead to, ought to be extremely cautious of it and its every outcome. Therefore, it should be tied to the larger cultural issue of the organization rather than a technology issue for handling such complex operation scenarios. There are plenty of opportunities and it becomes a double-edged sword if it is not designed thoughtfully, it may create havoc; a lot of process understanding to go in while building, articulating design aspects of virtual product digital twin. Data fidelity in terms of true representation of digital twin with the actual product or plant with time granularity of data is a key for instances where it ought to act in real instances need to act without physically being there e.g. exothermic reaction causing runaway in a reactor calling for the reactor to go to a fail-safe mode instantly to eliminate the cause of temperature run[1]away. Hence humans have to be in place in critical physical settings remotely to handle such scenarios in the aftermath of such an incident taking place. Intelligent digital twins will augment human decision making and not intended for replacement of humans in the entire loop.

Sustainability - Use of information in digital twins and predicting outcome will cost less and less as against wastage of a lot of money, reputation through prevention and elimination of occurrences; minimize wastages of resources, environment, etc. Digital twin harnesses larger insights of a physical thing, brings in fore what is going on inside, creates situational awareness for helping humans to take control of a situation, and plans for remediation better with proper planning. In essence it builds the foundation on which the data being collected now will help us do a much better job in terms of efficiency and effectiveness in product development and in the entire production process, enriching industrial artificial intelligence and machine learning capabilities.

It should not be allowed to let such complex systems to run on its own, human requires to take control of it. There appears to be no fixed blueprint readily available for which product or which part of the life cycle a company should start building a digital backbone in the process industry, a careful analysis is required in terms of dynamics of the market, differentiation a company wishes to create against its competitors.

Combination of feed forward loop where it is possible to automate some part of product lifecycle and then there is feedback loop from service to manufacturing in factory setting to redesign the product and so forth and its seamless integration with latest digital infrastructure starting with Edge computing,

Big data, and Cloud computing starting from planning phase to design to operation to maintenance will spur growth in technology, know-how’s aid in new age material development as a part of future business strategy.