IoT-enabled Platform for Rail Assets Monitoring and Predictive Maintenance (i-RAMP)
In 2016, there were more than 233,000 cancelled rail journeys in the UK out of a total of 7.2 million planned journeys according to the Office of Road and Rail (ORR, 2016). The majority of these disruptions were caused by asset failures and unscheduled maintenance on the rail network.
This project will develop an IoT-enabled Platform for Rail Assets Monitoring and Predictive Maintenance (i-RAMP). The i-RAMP system will employ techniques in Artiﬁcial Intelligence (AI), IoT and Augmented Reality (AR) to enable predictive and preventive maintenance. The project's outputs will signiﬁcantly reduce time to ﬁx failures and provide better value to the end users. In addition, the i-RAMP system will enable a deeper understanding of customers' behaviours and to predict how services can be dynamically scaled to meet the demand of the passengers.
- Mobibbiz Ltd.
The UK rail network experienced over 233,000 cancelled rail journeys in 2016 (Ofﬁce-of-Road-and-Rail, ORR, 2016), which were mostly caused by asset failures and unscheduled maintenance. As a result, over £28 million was claimed for service disruptions in 2016. Delays due to asset failures are increasing and maintenance expenditure is not decreasing as expected (Network Rail, 2016). Data from rail network assets and users has not been leveraged for predictive and preventive maintenance (Network Rail, 2016).
Currently, there are no holistic solutions that leverage asset data, and novel Artiﬁcial Intelligence (AI) and Augmented Reality (AR) technologies for predictive and preventive maintenance, e.g. to identify anomalies promptly that could remain undetected until regular inspections are undertaken and to reduce maintenance costs by supporting maintenance staff.
This project will address this business need by developing an IoT-enabled Platform for Rail Assets Monitoring and Predictive Maintenance (i-RAMP). The consortium will leverage their existing connections (Network Rail, Cross Rail, London Underground, HS2) to boost this market opportunity and increase its market share by ~4%. This project will provide an operational edge to the consortium by enabling ~50% time reduction in asset tracking, which represents ~30% cost reduction in maintenance tasks. It will extend the consortium's portfolio through a Spin-out company set-up to provide high-value services.
Our approach is to integrate IoT sensors into rail station assets for monitoring and predictive maintenance. These IoT sensors will be connected to an AI-based simulation platform to virtually simulate and evaluate rail assets. The platform will use novel predictive models (i.e. deep-learning) to provide alerts of potential asset failures and suggest optimal maintenance-plan in real-time. The AI simulation platform will also incorporate pedestrian-ﬂow-models to provide insight into customers' behaviour and to predict how and when station services could be improved. The solution will also support operatives with in-situ asset information and maintenance procedure using AR.
- AI-based Simulation IoT Platform (ASIP) where all the asset information and sensor data are overlaid on the digitised 3D-models of rail stations. The ASIP will enable virtual evaluation of the entire railway station in real-time to support decision-making. The ASIP will leverage historical data and AI techniques to develop predictive maintenance models that will alert of potential failures and suggest optimal maintenance plans. Additionally, pedestrian ﬂow models will be incorporated into the ASIP to provide a deeper understanding of customers' behaviours and predict how services can be dynamically scaled to meet demand.
- AR-Toolkit consists of a mobile device and a head-mounted display equipped with an AR version of the ASIP. The AR-Toolkit will provide the same functionalities as the ASIP desktop version. This will allow stakeholders, technicians, etc. to make evaluations in-situ with all the asset information in real-time.