We have the practical railways for commuting; do we have a predictive measure to ensure the safety rail running?

Derail a train is a disastrous event. There was a derailed oil tanker in Wales on 26 August 2020, causing flames and fuel spillage of around 330,000 litres. Ten of the wagons came off the tracks. Ensuring the safety of passengers, staff, and the environment in the surrounding areas, effective precaution measures are invented to assist the predictive maintenance.

Photos above: Detecting normal wheel and Noisy wheel

The use of electrical train service assures us to have an efficient way to commute around the city, even across countries. When we treat the train as if our bodies, wheels are just like our limbs and joints. They deteriorate over time with the frequent motion. Wheel defects, or wheel flats, on railway wheels would generate abnormal rolling noise causing potential noise nuisance in such a densely developed metropolitan city. In this case, ANewR has acted to help our client’s painstaking task. 

One of the world’s leading railway operators (The Operator) has about 5.6 million passengers traveling each day. Safety, reliability, customer service, and cost-efficiency are their objectives to maintain the industry reputation. Apart from the noise issue, one of their concerns is the derailed trains induced by the wheels’ imperfection in rolling stock. Derailed trains would lead to a painful death and multiple collateral damages. To prevent the derailment incident from happening, The Operator had spent over three million Hong Kong Dollars (over three hundred thousand British Pounds) for five-year manual inspection work, before applying ANewR’s acoustic system and Kvikoo advanced monitoring platform.

The past measurement method takes not only money but also human resources and time. The Operator conducted manual noise measurement at a specified location four times a year to inspect whether any abnormal noise occurred. The officer had to deploy to the site to measure the noise level during the standard operating time and analyse the noise measurement result with the video clips to identify any suspected defected wheel. The whole manual process is labour-intensive.

Kvikoo’s most significant advantage is to solve the manual scheduled inspection and maintenance problems by identifying potential flaw wheels without interrupting regular operation. It is a combination of acoustics and vision analytics technologies. The working concept is simple and straightforward: recording the noise with the acoustic camera and a CCTV at a specific spot, detecting abnormal stock wheels and tracks by identifying the high-frequency noise whenever the train passes the sensing system.

The sensing system comprises various hardware and software, including an acoustic camera, a high-speed camera, and an automated analytical intelligence system. The acoustic camera with one hundred and twenty-eight microphones and one camera will make an acoustic photo record indicating the noise source. The high-speed camera is to photoshoot the compartment number. When the acoustic camera identifies the noise source in the first step, the task of looking for the trouble compartment becomes effortless to follow. The cameras function well to capture the photos when the train speed at a 65 – 85 km/hr—the whole system works to start with the motion-detection camera to trigger an acoustic camera and high-speed camera. At the cameras’ back-end, we have installed a trackside server for storage and a data center for data analysis of compartment number and noise level. The transmission is secured through the 4G network.

At the other end, ANewR has set up the Kvikoo’s customised web-based monitoring platform for The Operator to perform various functions, such as managing devices and users, schedule measurement, customize train tables, and set email notification and trigger levels. The user-mode allows users to view measurement results and online video streaming. The monitoring platform will highlight noisy train data in red for easy tracking. Once the system discovers any abnormalities, the department in charge will receive an alert. The safety and environmental departments of The Operator is empowered to act quickly with follow-up action and repair to minimize accidents. Such preventive measures conform to safety compliances. No interruption will occur to the rails’ regular operation for the machinery is monitored and analysed in real-time. The promise of customer service with safety and no delay guarantees a positive corporate image. The designated user can retrieve data through filtering functions by tailored classifications of line, location, train type, noisy wheel, compartment ID, and measurement period. The records are traceable; the history of documents is saved, and reports are downloadable. The export function is great for the user to generate reports conveniently and look back through continuous monitoring records.

The outcome of adopting Kvikoo preventive maintenance system is remarkable. The automatic system reaches 95% accuracy; it also raises the inspection efficiency by at least 20%. In terms of cost reduction, Kvikoo system has saved a quarter of the original cost. Since the system is the customised automated system that does not require technicians to operate, The Operator has minimised workforce assigned to the scheduled investigation and maintenance works. Kvikoo contributes to a further reduction of inspection costs.

The Operator’s case demonstrates ANewR’s Kvikoo intelligent preventive maintenance system’s merits in guaranteeing the project owner with the passenger safety enhancement, the feasible maintenance schedule of rolling stock, and the maintenance checking accuracy. Ultimately, ANewR aids in detecting the noise source from the wheels and the rolling stock and other parts of the train. The customisable predictive maintenance system is non-intrusive – the measurement and monitoring works do not interfere with the regular train service and operation. It is trouble-free to determine whether the issue to fix originates from the problematic wheels, or compartment’s ventilation, or other mechanical pieces by catching the unusual sound while monitoring it with the acoustic and monitoring application. Last but not least, ANewR’s acoustic system and the Kvikoo monitoring system are a great deal in terms of cost-saving and maintenance.

Video – Automating metro wheel inspection while maintaining normal train operations by IoT / big data system: https://www.youtube.com/watch?v=nz8gG2X7370

By: ANewR Consulting Limited, a digital environmental consultant headquartered in Hong Kong since 2008. Our expertise has grown into the context of air and water qualities, noise, green building, waste management, and remediation. With extensive know-how in environmental planning and assessment, feasibility study and policy review, ecological design, monitoring, and audit (EM&A), ANewR has matured to be a leading management consultancy. Standing in the digital transformation reign, ANewR has participated in various environmental digital projects – interactive 3D visualisation, immersive automation virtual environment, Virtual reality, automation system, and monitoring platforms.
(Website: https://anewr.com, LinkedIn: ANewR Consulting Group, Twitter: ANewR – Everyday Newer,
YouTube: https://www.youtube.com/channel/UCnpvmxnR9hbNxytSfBdfV8Q/videos)