“Domain expertise is the secret to distinguish industrial AI from more general AI methods. Industrial artificial intelligence will guide innovation and efficiency improvements in capital-intensive industries in the next few years,” said Willie K Chan, Chief Technology Officer of AspenTech. Chan was one of the original members of the MIT ASPEN research project, which later became AspenTech in 1981 and now celebrates the 40th anniversary of innovation.
Combining expertise in this field, industrial AI applications have an inherent understanding of the context, internal operations, and interdependence of highly complex industrial processes and assets, and take into account design features, capacity limitations, and criticality for practical applications Safety and regulatory guidelines. World industrial operations.
More general artificial intelligence methods may produce specious associations between industrial processes and equipment, resulting in inaccurate insights. General AI models are trained on large amounts of factory data, which usually cannot cover all potential operations. This is because for safety or design reasons, the factory may work in a very narrow and limited range of conditions. Therefore, it is impossible to infer these general AI models to respond to market changes or business opportunities. This has further exacerbated the barriers to productization of artificial intelligence programs in the industrial sector.
In contrast, industrial artificial intelligence uses domain expertise specific to industrial processes and actual engineering, based on the first principles explaining the laws of physics and chemistry (eg, mass balance, energy balance) as a guardrail to mitigate risks and comply with all necessary conditions Safety, operating and environmental regulations. This helps to achieve a safe, sustainable and holistic decision-making process that will produce comprehensive results and trustworthy insights in the long run.
The digitization of industrial facilities is critical to achieving new levels of safety, sustainability, and profitability, and industrial artificial intelligence is a key driver of this transformation.
Industrial artificial intelligence in action
It is one thing to regard industrial artificial intelligence as a revolutionary paradigm; it is another thing to truly see its role in the real industrial environment. The following are some examples that show how capital-intensive industries can use industrial artificial intelligence to overcome digital barriers and improve the productivity, efficiency, and reliability of their operations.
Processing plants may deploy advanced industrial artificial intelligence Mixed model, Using deeper collaboration, machine learning, and first principles between domain experts and data scientists to build more comprehensive, accurate, and high-performance models. These hybrid models can be used to optimize the design, operation and maintenance of plant assets throughout the life cycle. Because of their reliable relevance over a longer period of time, they can also better represent the factory.
Chemical plants can use industrial artificial intelligence to obtain real-time insights from integrated industrial data from the edge to the cloud. Internet of Things (AIoT) To achieve agile decision-making throughout the organization. Using a richer and more dynamic workflow, the supply chain and operating technology are seamlessly linked to detect changes in market conditions and automatically adjust operating plans and timetables in response.
Refineries can use industrial AI to simultaneously evaluate thousands of oil production scenarios across a set of different data sources to quickly identify the best crude oil slate for processing. Combining rich AI capabilities, enterprise-wide insights, and integrated workflows to improve executive decision-making, this approach enables employees to allocate time and energy to more strategic, value-driven tasks.
Next-generation industrial facilities can use industrial artificial intelligence as a “virtual assistant” for factories to verify the quality and efficiency of production plans in real time. Cognitive guidance that supports AI will ultimately help reduce the reliance on individual domain experts to make complex decisions, but instead institutionalize historical decisions and best practices to eliminate barriers to expertise.
These use cases are by no means exhaustive, but just a few examples to illustrate how the universality, innovation, and wide applicability of industrial artificial intelligence can serve the industry and lay the foundation for the digital factory of the future.
The digital factory of the future
Industrial organizations need to accelerate digital transformation to remain relevant, competitive, and able to respond to market disruptors. Self-optimizing plants represent the ultimate vision of this journey.
Industrial AI embeds domain-specific expertise together with the latest AI and machine learning capabilities into AI applications suitable for the purpose. This can realize and accelerate the autonomous and semi-autonomous processes of running these operations-realizing the vision of a self-optimizing factory.
The self-optimizing factory is a set of adaptive, self-learning and self-sustaining industrial software technologies that work together to predict future conditions and take corresponding actions to adjust operations within the digital enterprise. The combination of real-time data access and embedded industrial AI applications enables self-optimizing factories to continuously improve themselves—using domain knowledge to optimize industrial processes, make easy-to-implement recommendations, and automate mission-critical workflows.
This will have many positive effects on the business, including:
To curb carbon emissions caused by process disruptions, unplanned shutdowns or start-ups, and help achieve corporate environmental, social, and governance goals. This reduces production waste and carbon footprint, and promotes a new era of sustainable industrial development.
Improve overall safety by significantly reducing hazardous site conditions and reassigning operations and production floor employees to safer roles.
Unlocking new production efficiencies by tapping new areas of profit margin optimization and production stability, even in the economic downturn, they can obtain greater profitability.
Self-optimizing factories are not only the ultimate goal of industrial artificial intelligence, but also the ultimate goal of the digital transformation journey of the industrial sector. By realizing the popularization of industrial intelligence applications, the digital factory of the future can improve safety, sustainability and profitability, and empower the next generation of digital labor-providing companies with a future-oriented guarantee under turbulent and complex market conditions. This is the real potential of industrial artificial intelligence.
To learn more about how industrial artificial intelligence can support the digital workforce of the future and lay the foundation for self-optimizing factories, please visit
www.aspentech.com/accelerate, with www.aspentech.com/aiot.
This article was written by AspenTech. It was not produced by the editors of MIT Technology Review.