Predictive maintenance (PdM) is big business – and getting bigger. What’s propelling it upward at such a rapid rate? Several factors, it would seem, including: the rising adoption of emerging technologies for gaining valuable insights into industry’s maintenance challenges; the growing influence of machine learning and artificial intelligence.
Predictive maintenance is certainly witnessing several innovative trends – and predictive analytics is one reason why, allowing organisations to leverage advanced algorithms and Machine Learning (ML) techniques to analyse vast amounts of unstructured data that can identify patterns and trends indicative of their equipment’s current state of health. By harnessing such power, businesses can anticipate maintenance needs, optimise resource allocation and mitigate risks associated with unexpected failures.
PRODUCTION GAINS
“Manufacturers are adopting advances in sensing, AI and machine learning (ML) to boost profits and cut production times,” says Peer Schumacher, head of electronics manufacturing solutions, Panasonic Connect Europe. “AI and ML enable smarter production processes with predictive maintenance, improved quality control and automated decision-making.”
These advances are not just about keeping up with the competition, he adds – they’re about making smarter choices that keep machines running smoothly and prevent costly breakdowns. “At Panasonic Connect Europe, we’re using AI and ML in our production machines and management software solutions to help our customers transform their Surface Mount Technology (SMT) manufacturing. These technologies are helping us implement predictive maintenance, improve quality control and automate decision-making, all of which contribute to more efficient production.”
The company’s latest predictive maintenance tool, APC-5M, is like “having a crystal ball for your production line”, states Schumacher – with ‘APC’ standing for Advanced Process Control, while ‘5M’ considers the variances of huMan, machine, material, method and measurement.
Another exciting development, he says, is how AI helps track real-time movements of conveyor systems, heads and axes in the company’s SMT equipment. “Think of it as a machine with a sixth sense, which learns from past operations to better manage maintenance efforts and alert manufacturers to any abnormal conditions before they escalate.”
Integrating AI and ML into SMT manufacturing is a huge leap forward for the industry, he adds. “It not only speeds up time-to-market, but also cuts production costs and boosts product quality. As manufacturers continue to adopt these technologies, we can expect even more innovations that will push the boundaries of what’s possible in electronics manufacturing.”
AN EXTRA GEAR
According to Sensemore, which specialises in optimising machinery performance and reliability, the advent of Predictive Maintenance as a Service (PdMaaS) is adding an extra gear to maintenance practices in 2024. “Cloud-based solutions democratise access to predictive maintenance tools and expertise, allowing organisations of all sizes to harness the benefits of advanced analytics and predictive algorithms,” the company states. “By outsourcing predictive maintenance functions to specialised service providers, businesses can streamline operations, reduce costs and focus on core competencies.”
Another area gaining traction are immersive technologies, such as Extended Reality (XR), it adds. “By integrating virtual and augmented reality tools into maintenance workflows, organisations enhance visualisation and interaction with equipment. XR technologies enable immersive training, remote assistance and virtual simulations, empowering maintenance teams to identify issues, troubleshoot problems and perform inspections with unparalleled precision and efficiency.”
Focusing in on the demanding landscape of oil and gas operations, for example, Sensemore identifies several advantages that predictive maintenance delivers.
“Predictive maintenance empowers companies to anticipate potential equipment failures before they occur, enabling them to schedule maintenance activities during planned downtime. This proactive approach minimises the risk of unexpected breakdowns, reducing both downtime and the associated repair costs.” It also enhances asset reliability and availability by identifying and addressing potential issues before they escalate into major failures.
PUT TO THE TEST
Which technologies are empowering predictive maintenance? One predictive/preventive measure now to the fore is ultrasonic testing. This involves the use of high-frequency sound waves to detect potential and actual flaws and defects in equipment components, such as pipelines, valves and storage tanks. By capturing ultrasonic signals and analysing their amplitude and frequency, operators can identify leaks, corrosion and other integrity issues before they escalate or where they are most likely to become serious issues.
Among the tools that are tried and trusted is vibration analysis, assessing as it does the condition of rotating machinery, such as pumps, compressors and turbines. By measuring vibrations and analysing frequency spectra, operators can detect abnormalities indicative of impending equipment failure, such as imbalance, misalignment or bearing defects. Equally notable is thermal imaging technology, which enables the detection of abnormal temperature patterns in equipment, signalling potential issues, such as overheating, insulation degradation or electrical faults.
We have entered a new era where AI and ML algorithms, with their ever-growing complexity, are enabling more accurate failure predictions to be calculated. However, for all the hype around AI and ML, there are other major influencers at work where predictive maintenance is concerned. Take Edge Computing, for example. This is rapidly gaining prominence, allowing the capture, processing and analysis of data at the farthest reaches of an organisation’s network: the ‘edge’ itself.
“This allows organisations and industries to work with urgent data in real time, sometimes without even needing to communicate with a primary datacentre and often by sending only the most relevant data to the primary datacentre for faster processing,” says Microsoft. “This spares primary computing resources, like cloud networks, from being glutted with irrelevant data, which lowers the latency for the entire network.”
An oil drilling rig operating in the middle of the ocean is one example. “Sensors that track information like drill depth, surface pressure and fluid flow rate can help keep the machinery on a rig running smoothly, and help keep workers and the environment safe. To do this without slowing down the network unnecessarily, the sensors send only the data about critical maintenance needs, equipment malfunctions and worker safety details over the network, and this makes it possible to identify and react to issues in close to real time.”
Meanwhile, the integration of IoT devices and cloud-based platforms is expanding, facilitating remote monitoring and maintenance. These trends promise to add even greater impetus and direction to predictive maintenance, enhancing equipment reliability and operational efficiency to levels not previously attainable.