energy costs and preventing unplanned shutdowns, these solutions deliver a rapid return on investment. “Compressed air production accounts for 20–30% of total energy costs in industrial facilities, yet nearly one-third of compressed air is wasted due to leaks. By identifying leaks early, facilities can significantly reduce energy use, lower carbon emissions, and improve overall equipment efficiency (OEE). Preventing leaks also minimises unplanned downtime and reduces costs. “Machine learning enhances pneumatic system applications by learning ideal system behaviour and forecasting anomalies with remarkable accuracy,” Nils continued. “When minor deviations occur, AI can pinpoint leaks before they escalate, allowing personnel to address them proactively. At Emerson, we have developed a machine-learning model that predicts pneumatic system behaviour. This model forecasts significant anomalies and leaks, even detecting creeping leakage – a subtle but impactful issue that can reduce OEE.” Steve Sands expanded on this area, explaining how expert teams can help users fine-tune their AI systems: “Once electronics are integrated into what were traditionally 'dumb' components, onboard sensing and communication is enabled. Today, through systems such as IO-Link or the Festo Automation Protocol (AP-bus), we can monitor our pneumatic components and use this information to save energy or optimise our predictive maintenance regimes. Combining structured, machine-readable data with deep learning AI algorithms is developing very fast. “Festo has installed hundreds of successful AI projects monitoring complete machines where a specialist team supports our customers in collecting, standardising, and aggregating their machine data. The team also advises clients on the selection and tuning of their AI algorithms to create the actionable insights they are targeting, whether it be quality improvement, energy efficiency or timely maintenance interventions.” AI’s ability to enhance the skills of onsite staff is also a way to reduce the impacts of the current skills shortage, as Nils pointed out: “As industrial systems grow more complex, the expertise required to configure and maintain them is becoming increasingly specialised. With experienced engineers in short supply, GenAI will serve as an invaluable assistant, simplifying tasks and providing on-demand guidance. Whether through AI-driven chatbots that provide instant access to device manuals or AIpowered tools that assist in programming programmable logic controllers (PLCs), these technologies will make engineers’ jobs easier. AI assistants will work alongside engineers of all experience levels, helping to select optimal configurations, diagnose issues and suggest or even implement actions to enhance performance.” Another example of how AI can support engineers with traditionally complex tasks is Festo’s Motion Insights which the company describes as a smart function block which can be incorporated into existing control environments by in-house PLC programmers rather than data scientists, as Steve explained: “With Motion Insights for pneumatic cylinders all the user has to do is download and install the ‘containerised’ module into their existing program environment (e.g. Siemens, Beckhoff, Rockwell etc.), follow the guided script to map the relevant pneumatic solenoid valve outputs and cylinder end-of-stroke feedback switches and then allow the AI software to monitor and learn the standard operating cycle. Once it has self-taught, the module sets targeted alarms when it sees an anomaly trend developing based on the control output signals and the response times, giving plenty of time for planned maintenance interventions.” Systems optimisation is another traditionally complex area in many facilities, with large numbers of parameters involved, all influenced by the operating environment. As Nils added, this is also an area where AI can offer significant advantages: “Traditionally, devices are configured based on past practices or conservative estimates rather than data-driven insights. However, as manufacturers strive for greater efficiency, ML will analyse application-specific requirements to determine optimal device configuration. This ensures not only proper operation but also enhances performance – whether that be by reducing energy consumption, increasing throughput or balancing both without compromising quality. By leveraging pre-trained models, AI will be able to intelligently match parameters to individual use cases.” It is clear that AI is already having a significant impact on the fluid power sector, bringing positive enhancements and solving long-standing challenges, and with the fast pace of development in AI technology this is only going to increase. The BFPA fully recognises this and is exploring avenues to develop further content to support its members’ understanding and use of AI, and the Association is also in the early stages of setting up a team to look specifically at this topic to ensure members are able to benefit from the opportunities this technology can clearly offer. www.bfpa.co.uk 43
RkJQdWJsaXNoZXIy MjQ0NzM=