Artificial Intelligence With all the hype surrounding Artificial Intelligence (AI) over the last couple of years, you would be forgiven for thinking it is something new. But AI has, in fact, been around for significantly longer. AI has many strands. The AI fuelling the current hype is called Generative AI, or GenAI. GenAI is the area where artificial intelligence creates new content – like text, images, music, or code – by learning patterns from existing data, sometimes referred to as large language models (LLM). LLMs power many of the tools you may be hearing of today, such as ChatGPT, Claude, Copilot and Gemini. These tools are fed unfathomable amounts of data to learn from to then be able to answer pretty much any question you can throw at them – not always completely accurately though. Alongside these very broad tools are more specialised generative tools that are trained on very specific sets of data to handle focused tasks. This may include customer service chatbots focused on the questions customers may ask or tools set up to enable users to query large, complex sets of data, such as technical manuals they are specifically trained on. These tools can be configured to avoid hallucinations – where if an AI tool can’t find an answer it makes one up as a best guess – overcoming one of the significant shortcomings of the more generalist tools. The history of AI As you might imagine, there are differing views on the history and origins of AI. In the 1950s, Alan Turing proposed a method for testing for machine intelligence referred to as the Turing Test, following his description of a Universal Machine “capable of computing any computable sequence” in the mid-1930s. But this was all just theoretical. AI, in the terms we understand today, really began to emerge in the 1990s as machine learning (ML) started to develop and become more accessible. Then, in the 2010s, deep learning powered another significant step forward, paving the way for voice assistants like Siri and Alexa and autonomous vehicles. The explosion in AI adoption we are experiencing today is supported by further advances in computing power, the development of Multimodal AI – models that can understand text, images, audio, and video all at once – the explosion in data generated by devices, apps and sensors, and further deep learning breakthroughs, among other things. In industrial settings, AI has also been in use for many years. Rule-based systems used in the 70s and 80s enabled systems to use ‘if-then’ logic commands to maintain operations by triggering actions based on inputs. In the 80s and 90s, industrial robots, while not ‘learning’ yet, were able to perform repetitive tasks based on set instructions, and computer vision was being used for relatively basic quality control. From the early 2000s, the rise of machine learning models powered areas such as predictive maintenance, as AI was able to spot patterns and predict failures before they happened, while a more accurate ability to forecast demand and optimise inventory supported supply chain optimisation. Then, as Industry 4.0 and the concept of smart factories took hold in the 2010s, thanks to the Industrial Internet of Things, cloud computing, and AI converging, huge advances were made. Real-time monitoring made instant decision-making possible, computer vision and robotics took on new powers thanks to the ability to react to new parameters, and digital twins – AI-driven simulations of real factory environments used to test equipment and processes virtually – became more powerful and widespread. Today, GenAI is helping to design optimised components and systems, and autonomous mobile robots are learning their environments and optimising materials handling and intralogistics. AI in fluid power So, what does AI mean for the fluid power sector now and in the years to come? I spoke to representatives from two BFPA members, Nils Beckmann, Director of Engineering Intelligent Automation at Emerson and Steve Sands, Technical Consultant at Festo, to find out. A significant benefit AI-powered systems can offer is energy efficiency and reliability improvements. Nils Beckmann explained: “AI-driven condition monitoring and prediction helps facilities optimise their energy use, improve sustainability and meet net-zero targets. Moreover, by reducing The role of Artificial Intelligence in the fluid power sector 42 www.bfpa.co.uk With Artificial Intelligence playing an ever-increasing role all our lives, Chris Callander looks at the history of its use in industrial settings and finds out from two BFPA members – Emerson and Festo – how it is already benefiting the fluid power sector, and what the future may hold.
RkJQdWJsaXNoZXIy MjQ0NzM=