Improving Railway Efficiency with IIoT System Analytics

---

East Japan Railway Company

---

Railways are an essential means of transportation in Japan. The East Japan Railway Company approached us about employing IIoT System Analytics to help improve train efficiency and safety for their more than six billion yearly passengers.

Objective

Improve train efficiency and safety
In recent years, the East Japan Railway Company, or JR East, has faced the challenges of an aging infrastructure, a reduction in the number of new train maintenance specialists due to the decreasing population, and rising costs alongside shrinking budgets. JR East was in search of an innovative solution to address these problems and ensure their trains continue to run reliably each and every day for their passengers.

Why us?

IIoT System Analytics solution
Condition-based maintenance (CBM) is a concept in which maintenance of machinery is performed when the need arises, as opposed to traditional practices of time-based maintenance (TBM) and reactive “fail and fix” or planned maintenance, which can be costly, prone to human error, and in worst case lead to downtime or accidents. JR East turned to us to deploy our MOXI™ technology suite, which employs model-based algorithms that enable 90% or higher accuracy and low false alarm rates, and need only minimal data sets in some cases. Together with their Open Innovation model and breadth of interdisciplinary expertise in system sciences and materials, we were the perfect choice for JR East.

“Now the knowledge normally obtained and accumulated among train maintenance workers through their work can be revealed with IoT technology.” – Atsushi Yokoyama, Head of Technology Planning, JR East

Solution

Customized fault detection and diagnosis software
With partner Nomura Research Institute (NRI), we interviewed JR East engineers, R&D teams, and maintenance technicians to understand the problem and discover how the MOXI technology could be used. After gaining an understanding of the available data to train the machine-learning algorithm, as well as the need for data-driven versus first-principle or hybrid models, we suggested an analytics approach to JR East. We then created dashboard mock-ups, gathered feedback from JR East end users, and began rapid iterations of algorithm and software development. The end solution? We developed customized fault detection and diagnosis pilot software, leveraging advanced machine-learning and model-based system analyses approaches, to test an initial pilot for JR East focusing on train door and railway track maintenance. Dashboards were also developed for JR East engineers to visualize and better understand the obtained data. We and NRI are now working with JR East Technical Center teams to test and implement these solutions on some of the most heavily used train lines in Japan.

Results

High true positive rates, low false positive rates
Initial tests for both the railway fault detection and train door detection indicated very high true positive rates and very low false positive rates. Further tests are currently being conducted to validate findings and prepare for field implementation. Once installed, the CBM dashboards with our solution would enable JR East to detect and repair upcoming train failures before they happen – all with remarkable confidence. This would allow JR East to improve efficiency, cost-effectiveness and most importantly, keep trains running safely, optimally, and on time. The MOXI IIoT System Analytics technology suite has great potential in transportation and many other industries like energy storage and smart manufacturing and is one step toward our broader quest to enable self-aware, self-adaptive systems.