The system is designed to be “fit-and-forget” to sense, analyse and transmit data, removing the need for trackside activities to replace batteries and improve safety, plus environmental credentials. This system offers sufficient quantity and granularity of sensor capability for data to meet requirements in delivering rail modernisation, enabling affordability of ownership for greater sensor density in building a trackside “intelligent Network Sensing Services” (INSS), and enabling deployment of wide-area rail track monitoring for predictive and preventative maintenance. The system can be used to improve the decision-making process for discovery, prediction and advice in maintenance, consequently, reduce maintenance cost, track downtime, site visit and unplanned maintenance, and eventually achieve better, faster and reliable journeys for passengers.
Rolling stock and railway tracks play a vital role in moving people and goods around. Thus, ensuring a smooth and undisrupted rail journey for people or freight movement is crucial in maintaining productivity and economic activity. This means rolling stock and rail tracks have to be well maintained. Trains, which are mobile, can be taken to a maintenance depot, but the whole country’s rail network must be inspected and maintained in situ. Currently, There is insufficient sensor capability deployed as either fixed sensors on the rail network or mobile train and vehicle mounted sensors, such as the dedicated measurement trains that run “in-traffic” on a multi-weekly or multi-monthly scheduled track inspection and measurement. The data and information obtained are insufficient to enable a rail modernisation that can detect trends and deviations from acceptable operating parameters for predictive and preventative maintenance. Consequently, service providers consider installing a large number of distributed sensors on the widely distributed area rail network critical to effectively monitor track condition and report status for such a purpose. However, the need to install new equipment such as transformers to power 3.3V low voltage sensors from the existing high voltage 25kV rail electrification system is not economically justifiable. Routing cables alongside the rail track for sensors is a costly exercise. Wireless sensors are a possible alternative option, but regularly replacing or charging depleted batteries is a cumbersome and expensive maintenance project due to the vast number of sensors that would be required over an extensive rail network. For safety reasons, the rail track cannot be used during battery replacement by maintenance teams and would be very disruptive to the railway system. Therefore, harvesting energy from a rail track’s local environment to power wireless sensors is a desirable solution.
What capability can the energy harvesting powered wireless sensor system being developed by the Group deliver?
A novel, deployable electromagnetic flux energy harvester has been developed and the whole energy harvesting powered wireless sensor system is at the developing stage. The principle of the idea is to harvest ambient electromagnetic flux energy generated by the return current when electric traction trains operate on the track. Such an idea has come out of Network Rail and a system harvesting such an energy source to power wireless sensors is being developed through a collaboration where the know-how research and development at the Energy harvesting Research Group at the University of Exeter and Network rail system knowledge have been paired together, which enables the research and development to produce a fit-for-purpose technology demonstrator. The system comprises three core technologies: (1) an electromagnetic flux energy harvester outputting sub-watt to watt levels of power; (2) an adaptable power management circuit embedded with an intelligent wake-up circuit that detects approaching trains and an energy-aware function to manage unpredictable and variable energy harvesting conditions; (3) a low-power and long-range communication wireless sensor node embedded with a high-performance computing capability for local data processing. The electromagnetic energy harvester, power management circuit and wireless sensor are all contained in an IP68 waterproof rated enclosure and installed under the rails, as shown on the figure. This ensures that the system can survive even extreme weather events such as a storm or flood. Installation is quick and straightforward, using a clamp to attach to the underside of the rail. Current passes through the track when there is a train on it. The current passing time can vary from one to a few minutes, depending on the train speed and the distance between autotransformer feeder stations.
The system harvests electromagnetic flux energy and converts it into usable electrical power and voltage supply for the wireless sensor nodes. The sensor nodes monitor track condition and associated health- and safety-related parameters such as vibration, temperature and loading conditions, perform local data processing, and wirelessly transmit the measured and/or processed data to a gateway. The sensors in the developed system are temperature, humidity and power-hungry accelerometer sensors; however other types of sensor can be added or substituted such as acoustic emission, LIDAR, radars, and image sensors to identify hazards on and around the rail track such as cars, boulders, human and animals depending on specific requirements. A high-performance computing microcontroller (MCU) and Sub-1GHz wireless MCU processor have been designed into the system for local data-computing and long-range communication capabilities. The sensors are activated by an intelligent wake-up circuit when a train approaches and a periodically scheduled timer for wake-up and measurement during times when the rail network is not in use. A significant event such as the onset of a crack in the rail can be identified when there is a train passing by due to a change in the weakened rail structure’s response towards the high loading exerted by the train. The sensor wakes up in time to capture data as the train passes the section of the track where the wireless sensors are installed. The system’s train detector follows an intelligent wake-up protocol, responding only to a gradually increasing signal as the train gets nearer to the sensor to avoid false triggering.
How many AA batteries of equivalent energy can the system generate over a year?
The system has been tested under emulated currents of 100-500A through the rails in the Energy Harvesting Research Group’s lab at the University of Exeter and harvested 0.2-4.2W of power. AA batteries typically have a capacity of 2500mAh. The question of how many AA batteries of energy the system can produce during one year of operation depends on many factors, including (1) the supply current in the contact wire, (2) how much supply current is returned to the rails, (3) how far from the energy harvester the autotransformer feed stations are located, (4) what speed the train is travelling, (5) how frequently trains pass over the section of track, (6) how many days the rail track is operated per year, and (7) the power management circuit energy conversion efficiency.
Looking at real-world situations, the autotransformer feeder stations are located every 5-8km; the supply current is 100-500A, and the return current is 53%-59% of the supply current; standard and high speed trains typically travel at 100-300km per hour in the world; the power management energy conversion efficiency is 60-75%. It is assumed that there are 6 trains per hour, the rails are operated for 15 hours per day, and there are 360 days per year for train operation for electrified routes. If we consider the worst case return current of 53%, the shortest autotransformer distance of 5km, and the worst power management efficiency of 60%, and 100km per hour of trains travelling, the energy harvested and available in energy storage for powering wireless sensors is about 10-162 J per train passing the section of rail with the system. So, the system will need to harvest either 14 days, 7 days, 3 days, 1.5 days or 1 days to obtain energy equivalent to one fully charged standard AA battery, at supply currents in the contact wire of 100A, 200A, 300A, 400A and 500A, respectively. Therefore, based on the calculations, the system can produce energy equivalent to at least 25, 51, 129, 233, 388 standard AA batteries, respectively, during a year of operation.
Can the system supply sufficient usable power to routinely monitor trackside sensors?
The answer is yes. Furthermore, the Group has studied the power consumption of a wireless sensor node they have developed for wide area rail track monitoring to confirm using the system, there is no need of batteries. For example, a node using a wireless protocol of Sub-1GHz and with a power-hungry accelerometer consumes around 14J for 53 seconds, sampling 30 seconds with 2 kHz sample frequency, performing a local data analysis, including a signal FFT operation, and using 17 dBm to transmit one set of 2048 sampled data and FFT results to 1Km away from a gateway. Look at energy harvested, about 10-162 J for per train passing the section of rail with the system at different return current, stated above, the system can supply sufficient usable power to routinely monitor trackside sensors without needs of batteries when the current in the contact wire is larger than 200A. In reality, the current in the contact wire is normally larger than 300A and so the system does not need batteries and mains to power most trackside sensors.
What key benefits can the system bring to predictive and preventative maintenance?
Permanently installed under the rail, the system will allow highly frequent and regular accurate measurements and monitoring. Therefore, a large amount of real-world data can be obtained to do trend analyses for smart modernisation of rail maintenance and to create simulations that predict how the rail track may perform and the optimum schedules for maintenance. Unlike using a mobile measurement train or vehicle, it has no operational interruption on the rail network. We expect that such a system will form a key part of an ambition to build a trackside “Intelligent Network Sensing Services” (INSS), and move the railway toward remote inspection and monitoring. Thus, enabling condition-based maintenance to replace the current practice of periodic on-site inspection and maintenance. We also expect that the system will allow the transition to predict and prevent maintenance regimes driving efficiency and minimising the downtime for maintenance and repairs, optimising the railway’s availability for customers. Additionally, the system can reduce the carbon footprint generated by using mobile measurement trains or vehicles. This is because there will be a need for much-reduced dedicated measurement runs and measurement trains when using the energy harvesting powered system. Prof. Zhu said: “our system supplies clean, sustainable, climate neutral energy, enabling wireless sensor nodes to operate for years without the need to charge or replace batteries, eradicating concerns related to energy deterioration affecting sensor node operation. Once installed, it is a truly fit-and-forget solution for gathering data. We are very excited by this technology; our system can completely replace batteries in wireless sensor networks, eradicating concerns about batteries going flat.
What is the current technology readiness level?
This system has yet to make it into everyday rail sector service but the Group’s technology has now reached TRL 4 -7 proof-of-concept in the laboratory environment with the consideration of real-world specifications and requirements.
What is the next for system development?
Prof. Zhu’s Group has filed a patent and plans to bring the energy harvesting technology to market via a spin-out company. So the next logical step is to develop a commercially deployable technology demonstrator to a pre-production built standard capable of undergoing assessment for regulatory approval. One of Prof. Zhu’s Group aims is to provide a self-powered rail modernisation capability for track condition monitoring. Meanwhile, Prof. Zhu’s Group will explore how to further improve the system and application of the system for large-scale monitoring in the rail sector. If you are interested in learning more about the technology or the work of the Energy Harvesting Research Group, please contact Prof. Meiling Zhu (m.zhu@exeter.ac.uk).
Acknowledgement
Prof. Zhu’s Group would like to acknowledge the Engineering and Physical Science Research Council (EPSRC) support, the UK, through the £1.38m standard grant ‘Zero power, large area rail track monitoring’ (EP/S024840/1).
Get to grips with the future factory: sign up now for our Advanced Manufacturing briefings (19-23 July), part of the Engineering Futures series.
Content published by Professional Engineering does not necessarily represent the views of the Institution of Mechanical Engineers.