The role of AI within industrial operations
Author : Greg Hookings, Stratus Technologies
02 April 2024
For maintenance and control engineers, AI is not a new concept. Many, especially those more advanced in their digital transformation journeys, have become very familiar with, for example, machine learning (ML) technology that has been helping with predictive and preventive maintenance strategies for some time by analysing large datasets from numerous sources to help make better-informed decisions.
Similarly, the use of large, contextualised datasets for a range of applications is helping digitally advanced organisations with energy reduction, supply chain optimisation, quality control, and any number of other optimisation efforts.
Management and interrogation of data for improved decision-making sits right at the heart of digital transformation, and it relies on server technology to deliver reliable data, often in real-time, that is not only available, but is also complete.
Data downtime can fundamentally undermine any digitalisation initiatives, and I’m proud that our technologies are relied upon by many of the most important applications in the world, where data-criticality can quite simply be a matter of life and death. From handling calls for emergency services to security at airports, to safety systems in the energy industry and process automation for utilities, Stratus Technologies has built its reputation delivering fault-tolerant servers, uptime, and data integrity to companies of all sizes and shapes.
As for the future of AI in industry (and society as a whole), we are approaching an exciting period of exponential growth for its application with the recent advent of early Large Language Model (LLM) AI, such as ChatGPT. This type of natural language-based capability to interrogate potentially limitless and varied public data sets will have huge implications across industry and society, most of which are yet to unfold, but are likely to enable humans in industry to access and handle data very differently.
Early industrial use cases benefitting from LLM (or ‘generative’) AI, to help maintenance engineers assess issues at remote sites, are already close. So, an operator with access to the control-layer software for an asset at present can already see if, for example, a bearing is running hot, and can prepare the maintenance team to replace it (preventive maintenance).
But with generative AI pulling extensive additional information from various public and private (secure, IP-based) resources, the engineer could effectively ‘talk’ to the system, asking for likely causes and likely implications, as well as requesting relevant pages of machinery manuals and historic data from the application in question so that they could prepare the team for a visit not simply to replace a defective part, but to understand associated issues and have a full statement of works (SOW) laid out.
If, for example, the asset was on an oil rig, it could help plan all of this with the required logistics of cargo loads, weather, staffing and any other variables in mind. The whole process of issue identification, resolution planning, and SOW, which would normally take several people hours, days or even weeks to coordinate, could, in theory, be completed by a single person in a chat window in minutes, reducing the downtime of the asset in question dramatically, as well as the overall cost of maintenance.
That’s just one example of course but gives a hint at what is possible. It also points to the importance of the servers that are handling the critical information from within the system.
Other examples are perhaps better exemplified by considering the impact of AI in areas such as healthcare, where an AI-empowered future will make personalised healthcare a reality. A future where a person’s DNA is vital to the exact formulation of the medication, which is manufactured specific to them and their needs.
We are a long way from realising this possibility, but with AI enabling the crunching of data at the scales required to deliver it, as well as helping to coordinate the production processes required to realise it, the capability is very close.
Naturally, there are many, many hurdles to overcome, including legislative, privacy and data-sharing regulations and standards, but the impact of AI in the years to come will be bigger than the impact of the World Wide Web, and it will require world-class data handling and server technology – which makes my industry an extremely exciting place to be!
Industrial enterprises looking to leverage AI and advance digital transformation must deploy the compute power necessary to unlock the full benefits of improved operations.
Contact Details and Archive...