A new revolution that has the potential to fundamentally alter industrial processes, including manufacturing, engineering, materials, supply chains, and lifecycle management, has been sparked by opportunities and challenges in the petrochemical industry and the emergence of highly disruptive technologies. The recently developed smart factory has become a significant player in the petrochemical industry and has embraced a disruptive manufacturing technique. The smart factory, which differs from the traditional petrochemical sector production processes, has to evaluate and position its future research agenda—including its definition, intentions, structure, and technology. Prioritizing a system’s thinking and its issue-solving is necessary for the smart factory.
A study suggests a lifecycle blueprint and a consensus-based operating and technological roadmap based on an understanding of the driving forces for the development of smart factories. The petrochemical industry is now facing new difficulties brought on by more stringent environmental protection and energy-conservation regulations, pressure to reduce product costs, a variety of raw material sources, specialized customer needs, combination and supply-chain optimization, and other factors. The Internet of Things, cloud computing, robotics, and big data technologies are all part of a new wave of scientific and technological advancement. Additionally, the petrochemical industry's production mode has undergone revolutionary changes due to the integration of information technology (IT) with operations and manufacturing technologies.
This article analyzes the latest changes faced by the smart petrochemical factory from the perspectives of the value chain, production and service mode, management and control chain, as well as energy constraints. The discussion on the definition and connotation of the smart petrochemical factory by comparing them with mainstream research proposes the technology system and roadmap of the smart petrochemical factory and summarizes key hotspot technologies to be researched.
Inference of Smart Factory in the Petrochemical Industry
Academic and industry communities define the "smart factory" from various angles. Industry 4.0 primarily defines the smart factory from the viewpoint of intelligent production focused on CPS (Cyber Physical Systems), while SMLC (Smart Manufacturing Leadership Coalition) mostly discusses smart factories from the views of knowledge and modelling. Smart factories are great when seen from the standpoint of integrating industrialization and IT applications, focusing on oil refining and chemical businesses. A smart factory in the petrochemical industry can be defined from two perspectives:
1. A smart factory in the petrochemical sector is focused on the entire industrial production chain of petrochemicals. To achieve the horizontal, vertical, and end-to-end integration of factories, it tightly combines next-generation IT with people, process/equipment operating technologies, and existing petrochemical production processes. Additionally, it improves the level of factory operation and management in a more delicate and dynamic way and facilitates innovation in manufacturing and business models by optimizing the four key capabilities of overall location awareness, forecast and early warning, collaborative optimization, and scientific decisions.
2. In the petrochemical sector, a smart factory creates a new generation of the petrochemical production environment that is characterized by ubiquitous sensing and intelligent services that link ubiquitous sensors, intelligent hardware, control systems, computing facilities, and information terminals into an intelligent network via CPS. This is done to comprehend how connected businesses, people, things, and services are as well as to produce, integrate, and use as much information resources, knowledge, and expertise as possible.
To put it simply, a smart factory in the petrochemical sector is a brand-new, cutting-edge facility that focuses on achieving operational excellence for the factory while also being environmentally friendly, highly efficient, safe, and sustainable.
Opportunities and Challenges in Petrochemical Industries
Operational Agility
Operational agility, or the ability to adapt quickly to changing conditions brought on by changes in feedstock, market demand, and pricing, is one of the essential characteristics of smart process manufacturing. Such perturbations will undoubtedly have a significant impact on plant performance. Therefore, switching operational methods, such as rebuilding the process flowsheet and changes in temperature, flow rate, and pressure, is crucial. The first issue here is how the lowest-level control systems can swiftly arrive at the correct set points.
Though existing information is unlikely to cover all conceivable operation instances, specialists' heuristic knowledge and operational expertise are the best options from an industrial standpoint. The accuracy of the knowledge-driven method is the second difficulty. Model-based real-time optimization is the primary strategy from an academic standpoint. The first concern is where to find trustworthy first-principal models for chemical-processing units, particularly complicated reactors; the second is how to solve the real-time optimization model effectively; and the third is if the industry would accept the optimization findings.
Adjustable Data-Driven Model-Building
The difficulties in solving an optimization model subject to a challenging first-principal model were discussed in the previous subsection. One potential option is using a data-driven surrogate model in place of the original first-principal model. Many statistical techniques have been used to create surrogate models for defect detection, process control, and optimization, including neural networks, support vector machines, and mathematical programming. When creating a surrogate model, it is important to consider its intended usage and implementation, such as whether it will be utilized for flowsheet optimization or trend prediction.
Adopting a neural-network-driven surrogate model for flowsheet optimization may result in significant computational challenges and the inability to produce high-accuracy results. External ductility is crucial, regardless of the technique used to create the surrogate model. The first question, in this case, is how many datasets are necessary to create a surrogate model with high accuracy; the second question is how to create an adjustable data-driven model. Industrial data sets should be used to validate accuracy.
Planning and Scheduling for an Entire Oil Refinery or Petrochemical Plant
Mathematical modelling and global optimization offer significant prospects for cost reduction, profit margin expansion, energy efficiency, and demand satisfaction in planning and scheduling various operations in an oil refinery or petrochemical plant. It has only recently become possible to fully plan and schedule an entire oil refinery or petrochemical plant—a crucial component of smart manufacturing. Crude oil blending and processing, processing unit activities, and product mixing and distribution make up the majority of operations.
Conclusions
The technology vision and blueprint for a "smart factory" in the petrochemical industry is built around petrochemical plant computing and extends traditional chemical engineering simulation. This allows for the control and management of an actual factory through the concurrent operation and collaboration of an actual factory and a virtual factory. A technology framework for a smart factory in the petrochemical sector believes that a holistic understanding of the physical environment is essential to the construction of a smart factory, and that smart business applications with optimization and decision-making capabilities can foster cross-sector cooperation and global petrochemical production optimization.
Additionally, a greater emphasis is placed on human factors to create the next generation of production and operation centers that are focused on people. This will allow the right people to access the necessary information and resources at the appropriate times and locations to innovate the business model. Finally, investigations primarily in critical technologies are required for priority development while considering the patterns in industrial development. Before a successful smart factory can be developed, a lot of effort must be done driven by application needs and related technologies because the research on this topic is still in its early stages and the relevant theoretical study has only recently begun.
