The US railroad industry has one of the most sprawling and immensely critical pieces of infrastructure in North America. When failures or other issues occur, maintenance is highly manual and primarily reactive. When a failure is identified, a maintenance crew must be physically deployed to correct the problem. These maintenance corrections take time, causing delays and potentially large penalties for the railroad operator due to unmet SLAs, loss of productivity or other contracted stipulations between the industry and its customers.
These penalties can range from thousands to millions of dollars for unexpected downtime scenarios. Consequently, both the operations and business sides of the industry suffer the high costs associated with reactive maintenance. As the railroad and shipping industries continue to see decreased margins and increased regulation, it is critical to reduce excess costs and increase efficiency to stay competitive in the marketplace.
From Reactive to Proactive
Due to regulations, safety requirements and lack of real-time data across its vast infrastructure, the US railroad industry has historically been forced to follow a reactive maintenance strategy. This approach has a number of significant challenges, primarily around the cost of mobilizing parts and labor, as well as the direct and indirect effects of equipment being down or unnecessarily unavailable. But, there is a transformation occurring within the US railroad industry; one that is moving from a predominantly reactive operating model to one that is forward-looking and proactive.
This transformation is being greatly accelerated by the introduction of new technologies. These range from low cost sensors and trackers that allow for better control and placement of field equipment, to self-operating drones that run up and down the railways scanning for anomalies and providing information to maintenance crews. These new technologies ensure that the US railroad infrastructure meets the mandates of federally required inspections. As these technologies continue to evolve and are dropping in cost, more opportunities are arising for railroads to cost-effectively transition from preventive maintenance to predictive maintenance.
A New Kind of Predictive Technology
RailPod, a Boston-based infrastructure inspection system provider, is looking to make global rail infrastructure safer and more cost effective to maintain, operate and monitor through the use of automated ‘inspection’ drones. RailPod drones increase the frequency and quality of day-to-day railroad track inspections through advanced sensing technologies and data as a service. These daily track inspections are performed in an automated fashion with a high degree of precision and accuracy – saving both time and money. By providing railroad operators with repeatable and quantifiable data in real time, railroads are able to make timely decisions that improve bottom line revenue and ensure the safety of its passengers and cargo.
After spending more than a year running demos for clients, using a web portal locally hosted in its office, RailPod realized it needed a cloud based IoT platform that could demonstrate how its technology could ingest, process and visualize the data from the drones in real-time. RailPod wanted a platform that required minimal maintenance and leveraged managed services, yet would be cost-effective. RailPod also wanted a solution that could scale with its customers and was architected to enable new data sources to be ingested as they were identified. Furthermore, RailPod needed to closely control the flow of data to its customers and ensure governance and entitlements were properly maintained.
RailPod engaged Cloud Technology Partners (CTP) early in its ideation process to provide expertise around cloud, big data and IoT. After helping RailPod create a list of requirements, CTP created a solution architecture, UI/UX prototype, sprint backlog and a resource plan for future execution. In less than two months, CTP built out a prototype solution with viability for scalable growth.
An Enterprise-Class IoT Solution
The goal of the first phase of the project was to build a prototype cloud-based solution that would enable RailPod to accomplish three key goals:
- Store and archive data captured by the drone.
- Visualize the measurement data in a web-based portal.
- Secure data so that users only see the data they are authorized to view.
The portal was built with login and logout functionality, a map of rail lines and charts to exhibit sensor data. CTP began with a detailed experience design plan and then proceeded to create the infrastructure, develop the application code, implement security, performance and durability measures, and provide RailPod recommendations on IoT and data integrity.
The IoT platform was developed using Python, SQLAlchemy, React, Leaflet, and PostGIS and is currently hosted on an EC2 instance. The following AWS services are currently being used to support the solution: VPC, EC2, S3, RDS, SQS, SNS, and Snowball.
“Cloud Technology Partners is helping us build an enterprise-class IoT solution on AWS that enables RailPod to be a global leader in infrastructure information production to ensure safer railroads across the global railroad market.”
– Brendan English CEO, RailPod
The next phases of the project will enable integration with the AWS IoT platform and leverage serverless computing. Once the RailPod team is able to integrate the IoT SDK into their drone, MQTT messaging will replace the need for parsing large binary files and AWS Lambda will eliminate the need for EC2 instances, reducing maintenance overhead and costs. Future phases will also include performance improvements, including the building of a tile server for rendering maps more efficiently.
Deploying on Amazon Web Services
RailPod selected AWS to host its solution because of the smooth integration between the AWS IoT platform and Railpod’s existing system. While the RailPod team was integrating the AWS IoT SDK into their drone, AWS S3, RDS, and EC2 provided an inexpensive way to deliver the initial prototype. AWS Snowball also provides an easy way to quickly transfer hundreds of GigaBytes of existing data into S3.
The Data Backbone
Due to the high variety of data measurements and their frequency, the data collected by the RailPod drone is stored in large data files that are transmitted to the cloud either directly or through a proxy device. As the drone performs an inspection of a rail line, new files are automatically uploaded to AWS and processed as per RailPod’s predetermined business rules. When the data has been successfully ingested and processed, or if it has failed, an email notification is then sent to the appropriate groups. In order to handle RailPod’s security requirements, CTP integrated CTP Central, a proprietary containerized multitenant and multi-user software, as the authentication mechanism for the platform.
An Industry-Changing Service
RailPod is enabling the railroad industry to take a huge leap forward in how it maintains and operates its vast infrastructure. By allowing railroads to continuously and autonomously inspect rail lines and predict maintenance needs far ahead of failures, RailPod is providing not only the US railroad infrastructure, but now the global railroad industry with a game changing tool that can significantly cut maintenance costs, while also increasing capacity and revenue.