Page 18 - Summer 2025
P. 18

Industry News


                                                                                                                                 prediction and anomaly detection in industrial
                                                                                                                                 equipment, residential facilities and home appliances.
                                                                                                                                 ROHM has also released an AI simulation tool (Soloist-
                                                                                                                                 AI™ Sim) on its website, allowing users to evaluate
                                                                                                                                 the effectiveness of learning and inference prior to
                                                                                                                                 deployment of the AI MCU. The data so generated        World First for Largest Coral Reef
                                                                                                                                 may then be used as training data for the AI MCU,      Restoration Programme
                                                                                                                                 supporting pre-implementation validation and
                                                                                                                                                                                        The King Abdullah University of Science and
                                                                                                                                 improving inference accuracy.
                                                                                                                                                                                        Technology (KAUST) in partnership with digiLab is
                                                                                                                                                                                        developing an AI-embedded digital twin platform
                                                                                                                                 ABB launches Gen-AI Vault                              to support its KAUST Coral Restoration Initiative, the
                                                                                                                                                                                        world’s largest coral restoration project.
                                                                                                                                 ABB has launched a generative AI-driven tool to
          DEVELOPMENTS IN AI                                                                                                     prevent knowledge from being lost in industrial        The AI solutions are designed to tackle critical
                                                                                                                                 environments.                                          aspects of coral resilience, such as thermal tolerance
          APPLICATIONS                                                                                                           The Industrial Knowledge Vault can store and           and growth, reducing the time required for coral
                                                                                                                                 safeguard critical expertise, while transforming that   assessments and increasing the accuracy of data
                                                                                                                                                                                        that can be retrieved from environmental stressors
                                                                                                                                 knowledge into step-by-step workflows, allowing
                                                                                                                                 workers to access and apply best practices. It learns   such as temperature and light. They will help scientists
                                                                                                                                                                                        to closely simulate and predict the behaviour of the
                                                                                                                                 continuously from real-world applications, refining
                                                                                                                                 workflows and best practices in order to drive         real-world reefscape in real-time, identifying factors
                                                                                                                                                                                        such as algae growth and signs of disease.
                                                                                                                                 continuous improvement. By integrating with existing
                                                                                                                                 industrial systems it enhances collaboration between
                                                                                                                                                                                        The project focuses notably on the restoration and
          Isambard-AI goes live in Bristol                      supercomputer. It was constructed using low-carbon               field workers and control rooms providing real-time    enhancement of coral reefs through a large-scale
                                                                                                                                 visibility into job status and operational procedures,
                                                                modular parts and aims to incorporate waste
          The UK’s most powerful supercomputer, Isambard-                                                                        which can reduce bottlenecks and improve overall       coral nursery system to support coral propagation
                                                                recycling to support local heating needs. It ranks
          AI, is now operational at the University of Bristol.                                                                   process functioning.                                   and natural reproduction, with KAUST planning to
                                                                fourth globally on the Green500 list of the world’s most
                                                                                                                                                                                        plant two million corals around the reefscape by 2030.
          It will be used in conjunction with Dawn, a similar but   energy efficient supercomputers.                             As experienced employees retire, the loss of their     At peak capacity, the sea-based nurseries should
          smaller supercomputer in Cambridge, to create a
                                                                Reference: Cappella, N., ‘UK’s most advanced Supercomputer       decades of knowledge threatens operational             produce 100,000 corals a year, with the primary land-
          national AI backbone for scientific research.         Isambard-AI goes live in Bristol’, Data Centre World, 25th. July  continuity, efficiency and innovation. Traditional    based nursery producing about 400,000.
                                                                                                                                 documentation is typically fragmented, requiring
          Applications include:
                                                                                                                                 significant manual effort to compile, interpret and    An innovative sensor placement system will augment
            •  Analysing MRI scans to detect early signs of     Breakthrough AI-equipped MCUs                                    apply. The Industrial Knowledge Vault addresses this   data collection, with machine learning being used to
              Cancer                                            The first MCU capable of predicting equipment                    by capturing and structuring expertise in a secure     assess optimal sensor placement.
            •  Modelling protein structures to discover drugs   anomalies via on-device learning and inference                   central repository that can be accessed using natural
              faster                                            without a network has been launched by ROHM.                     language conversations.                                Professor Dave Suggett, Director of the KAUST
                                                                                                                                                                                        Restoration Initiative, states:
            •  Interpreting video data from wearable devices to   The device enables fault prediction and degradation            Structural procedures can be created in seconds,
              improve support for dementia                      forecasting using sensing data of various kinds, such            eliminating the need to parse documents manually.        “ The KAUST Coral Restoration Initiative represents
                                                                as that from industrial equipment like motors.                                                                            a significant step towards more effective
            •  Monitoring livestock to detect early warnings of                                                                  The Industrial Knowledge Vault uses ABB’s Genix          restoration of reefs globally. By partnering with
              disease                                           Equipment manufacturers are increasingly seeking                 CoPilot in conjunction with Microsoft’s Azure OpenAI     digiLab to develop an AI-enabled digital twin
                                                                solutions that allow real-time monitoring of                     to analyse and synthesise information into actionable
          It will also support the UK’s AI Safety Institute, helping   operational status while avoiding the drawbacks           procedures.                                              of the reefscape, we are pioneering innovative
          to assess and mitigate risks associated with so-                                                                                                                                approaches to upscale coral management
                                                                of network latency and security risks. Standard AI
          called frontier AI technologies such as biothreats and   processing models, however, typically depend on               ABB’s Global Head of Digital for Process Industries,     efforts, including restoration decision-making. This
          cybercrime.                                                                                                            Sanjit Shewale, states:                                  collaboration allows us to leverage cutting-edge
                                                                network connectivity and high-performance CPUs,                                                                           AI technology to monitor, predict and optimise
          Housed at the National Composites Centre at the       which can be costly and difficult to install.                      “Industries everywhere are facing a critical           our restoration strategies in real-time, potentially
                                                                                                                                   knowledge gap as a generation of experienced
          Bristol and Bath Science Park, Isambard-AI was        In order to help address this, ROHM has developed                                                                         revolutionising how we approach coral reef
          built using 5,400+ NVIDIA GH200 Grace Hopper          groundbreaking AI MCUs that allow stand-alone AI                   workers retire and take their invaluable expertise     conservation on a global scale”.
          Superchips and is the world’s eleventh most powerful   learning and inference directly on the device. These              with them. Our Industrial Knowledge Vault tackles
                                                                                                                                   this challenge head on by providing a secure and
                                                                deploy a simple three-layer neural network algorithm               accessible platform to enable businesses to do       Co-founder of digiLab, Anhad Sandhu, adds:
                                                                implementing ROHM’s proprietary on-device AI                       more with digital. By capturing, preserving and        “ digiLab’s partnership with KAUST on the world’s
                                                                solution known as Soloist-AI™, enabling the MCUs to                sharing this vital knowledge, businesses can ensure    largest coral restoration project marks a significant
                                                                perform learning and inference independently without               operational continuity and empower workforces          milestone in the application of AI technology to
                                                                the need for cloud or network connectivity.                        with in-the-moment information to perform their        solve the world’s challenges. digiLab is supporting
                                                                Incorporating a 32-bit Arm® Cortex®-MO+ core, CAN FD               jobs smarter, faster and more sustainably”.            revolutionary methods for coral reef monitoring
                                                                                                                                                                                          and restoration by developing these AI models for
                                                                controller, 3-phase motor control PWM and dual A/D
                                                                                                                                 Reference: ‘Gen-AI Vault will preserve Knowledge from Industry
                                                                converters, and achieving a low power consumption                                                                         the reefscape’s digital twin”.
                                                                                                                                 Experts’, Drives and Controls, May 2025
                                                                of 40mW, they are said to be ideally suited for fault
                                                                                                                                                                                        Reference: Wood, A., Electronic Specifier, 2nd. July

             18                                                                                                                                                                                                                   19
                                                                            Become a member: https://membermojo.co.uk/ie         Become a member: https://membermojo.co.uk/ie
                                                                    Become a subscriber: https://membermojo.co.uk/subscriber     Become a subscriber: https://membermojo.co.uk/subscriber
                                                                        Become a sponsor: https://membermojo.co.uk/sponsor       Become a sponsor: https://membermojo.co.uk/sponsor
   13   14   15   16   17   18   19   20   21   22   23