Manufacturing and construction industries can more efficiently address maintenance problems with remote monitoring and instant alerting, according to Roman Pavlyuk, vice president of digital strategy at Intellias.
In early 2018, an oilfield services company approached us with a problem to solve, Pavlyuk says. Their equipment was installed at a depth of 1.5-2 kilometres underwater, and the regular maintenance procedure involved raising, disassembling the equipment at least every six months. The process resulted in extremely high costs, amounting to millions of dollars. Therefore, the need was to find an approach that would require equipment to be raised only when necessary.
In response, we proposed an AI-based digital twin for predictive maintenance and analytics, which would enable continuous monitoring of the equipment and provide timely signals if any anomalies were detected. The company decided to proceed with developing a PoC, which later resulted in very cost-effective preliminary estimates.
At that time, in 2018, AI technology was just entering an active era, and building a digital twin was not an option for every business due to the high costs of sensors. Even for such expensive industry as energy. However, fast forward to 2023, the rapid development of microelectronics has significantly reduced costs, making digital twin technology much more accessible. As pioneers in the field, Intellias, knowing how beneficial this technology might be for businesses, have been working to refine and enhance predictive maintenance capabilities to meet the evolving needs of the industry.
The rise of the internet of things
Maintaining machinery in the chemical and manufacturing industries is costly. Global Fortune 500 manufacturing and industrial companies claim to lose over 3 million hours a year due to unexpected downtime, resulting in an $864 billion loss — the equivalent of 8% of their combined annual revenue. Moreover, according to Gartner, the average cost of machine downtime is $5,600 per minute.
However, with the rise of the Industrial Internet of Things (IoT), artificial Intelligence and machine learning, and data engineering, companies have shifted towards tech-powered solutions to save money. Smart sensors, predictive analytics, remote asset management, equipment tracking and many more all have roles to play in proactively managing maintenance and reducing downtime.
Introducing proactive prevention
PreFix, Intellias’ latest predictive maintenance predict-n-fix concept, is designed to help companies maximise uptime and limit potential complications in production. It’s an industrial equipment monitoring solution for detecting anomalies that can identify heat exchange system breakdowns, liquid leaks in chemical storage tanks, pump system failures, motor vibrations, and more.
Before implementing a large-scale IoT implementation, companies have the option to evaluate the financial benefits and rapidly validate initial ideas within 2-3 weeks with the PreFix proof of concept (PoC).
Using an edge computing system with pressure sensors and machine learning algorithms for data acquisition and aggregation on edge gateways, PreFix monitors and ensures the health of industrial equipment. Working with high-frequency data (around 10,000 records per second), it can detect various anomalies to provide early warnings and prevent potential failures of critical equipment.
Integrating DAQ cards and time-series streams for accurate detection and identification of unexpected issues enables PreFix to provide real-time visibility into asset conditions while creating a single point for high-resolution data analysis and instant leakage or breakdown alerts.
Real-world application
A world-leading science and research centre responsible for a range of novel technology platforms for companies and governments worldwide needed a production-ready scalable IoT predictive maintenance solution to prevent malfunctions and unplanned downtime.
The intelligent PreFix predictive maintenance solution-concept for remote facility monitoring and management in the manufacturing industry eradicates equipment failures and unplanned downtime by offering predictive analytics.
A network of sensors connected to business-critical assets fed a set of interactive dashboards, where users can continuously monitor equipment performance and health in real time. At the same time, the Intellias engineering team created an agent and server of the edge computing system to capture and read analogue sensor signals on edge gateways, storing plant condition data at up to five million entries per second.
Furthermore, this proactive approach enabled manufacturers to save substantial costs associated with assets and resources. The system achieved this by promoting efficient asset utilisation and maintenance practices, resulting in improved operational efficiency and cost savings.
The real secret sauce at the system’s heart is the predictive algorithm. This physics-based forecasting model processes the sensor data to calculate the ‘remaining useful life’ (RUL) of each asset. It then generates alerts and warnings when it anticipates hidden defects and unexpected events.
Tangible benefits
The new proactive maintenance IoT solution was developed in just twelve months and now tracks 36 assets rather than the initial four. It is bringing real value to the research centre by identifying anomalies and alerting the diagnostic team. In addition, Intellias is now adding a portal to give enterprise-level manufacturers a macro-view of all their installations across facilities and geographies.
This is just one example of how predictive technologies and IoT transform how we manage our maintenance more efficiently. As technology continues to develop, we expect many more examples to become evident. If you work in the chemical, manufacturing, energy and utility, and construction industries, now is the time to get on board.
