Insights & curiosities

TELEMATICS FOR RAIL TELEDIAGNOSIS: THE EVOLUTION OF MAINTENANCE THROUGH DATA ANALYSIS

The world of rail, and more generally that of transport, is a sector in constant evolution and growth, where themes and technologies are being introduced which are increasingly innovative and revolutionary. It is obvious that transport operators are not always ready to acquire these technologies and seize the opportunities associated with them; in Italy, this happens due to problems related to backwardness of infrastructures, Little propensity to change and a form of lies that is most often based on outdated dogmas. In this scenario, the maintenance of transport means is crucial to ensure safety and reliability for the travellers who use them daily. It is easy to understand how, thanks to proper maintenance of the vehicles, the risk of in-service failures, line stops, passenger dissatisfaction and overcrowding due to a low percentage of vehicle availability is considerably reduced.

In this scenario, the maintenance of transport means is crucial to ensure safety and reliability for the travellers who use them daily. It is easy to understand how, thanks to proper maintenance of the vehicles, the risk of in-service failures, line stops, passenger dissatisfaction and overcrowding due to a low percentage of vehicle availability is considerably reduced.

Types of maintenance in the railway world

Focusing on the railway environment, different types of maintenance are distinguished according to their nature; leaving out the interventions of an extraordinary nature, we can summarize the maintenance interventions in four distinct categories:

Preventive or cyclical maintenance: carried out on the basis of a time-based or mileage-based schedule, or at set intervals of time, in accordance with the vehicle manufacturer’s manuals and maintenance schedules.
Corrective or failure maintenance: carried out following the detection or appearance of a fault, aimed at eliminating the resulting failure and restoring the correct functioning and condition of the vehicle.
Condition maintenance (on condition)based on monitoring of performance and data from a particular system concerned, is performed when it is necessary. It allows you to intervene if a particular condition occurs on a component or system analyzed, and possibly anticipate the appearance of a fault.
Predictive maintenance: allows intervention to be carried out even before a failure occurs or symptoms develop. This is done by analysing the data, mainly diagnostic, from the medium, combining mathematical models and graphical simulations of the trend of these data in order to determine the cause/effect relationships that cause a failure to occur. This allows you to predict how long it will take for a given failure to occur, so that you can prevent it and intervene before it occurs.

From the points listed above, it is clear that today predictive maintenance is the most important thing to try to minimize the occurrence of failures that can cause more or less serious problems during the operation of a rolling stock. This type of maintenance is increasingly linked to the concepts of machine learning and artificial intelligence, which allow you to obtain amazing processing and results with significantly reduced times compared to human analysis of thousands of data.

Telediagnostics - what it is and how it applies

In this context, what is called Telediagnosticaor to analyse real-time diagnostic data from the vehicles, in order to know the actual state of health of them even before they return to the workshop for repair. Through this concept, it is possible to obtain essential information to optimize the downtime of the vehicle during a maintenance period, thus increasing the levels of availability to the service. This information, and therefore the data associated with it, can be obtained, visualized and analysed through modern tele-diagnostic systems which may be more or less complex as required, but that in any case are based on a dedicated Software structure with supervision and control of highly qualified personnel physically located in operating centers often called "Control Rooms"Control Roomor even remotely located.

The on-board equipment stores its own diagnostic data (e.g. a fault or an alarm) as well as process data (e.g. temperature, pressure, speed), thus creating a database andBig DataThe on-board network provides ground transmission to processing centres and/or suitable data loggers installed on the train for local storage. The ground processing centre is able, by means of appropriate algorithms, to analyse the data and detect in a predictive manner any anomalies which could lead to failures or malfunctions during operation, and alert in real time the authorized personnel who can thus organize the maintenance operations, not only according to what is already programmed in the field of preventive maintenance, but also on the basis of the actual train conditions detected in operation.

Before we go into the details of Telediagnostics and its possible applications, it is necessary to understand the evolution of maintenance concepts that can be obtained through it. The result will be a shift from maintenance with slow and unreliable repair times to advanced maintenance with optimisation of associated time and costs.

The current scenario - traditional maintenance

Traditional maintenance is certainly sequential. As you can see from Figure 1, each step associated with it requires the completion or verification of the previous one to which it is linked. So, the maintenance staff, only after having found each single point can complete the repair of the vehicle, with poor results of optimization of time.

Figure 1: Traditional maintenance

It is clear that such a maintenance typology can nowadays be modernized and improved in many aspects, by placing a series of activities in parallel, and perhaps even while the rolling stock has not yet been returned to maintenance, The entire flow of interventions is optimised.

A look into the future - Innovative maintenance

Modern and advanced maintenance allows to overcome the obstacles present in traditional one. By carrying out the unloading and data analysis activities in parallel with the definition of the interventions, it is possible to save time from the downtime of the vehicle, since these activities can be carried out while the vehicle is still in operation. These activities are now carried out on the agenda by the so-called Control Rooms, or sections or divisions dedicated to such operations formed by highly qualified personnel with extensive experience in the sector. The real-time data sent by the vehicles are analysed by the operating staff, and supplied to advanced artificial intelligence systems to verify failures on the vehicle and possible predictions of future failures. In this way, when the vehicle is returned to storage, the interventions to be carried out are already known, as they have been defined even before its return, thus obtaining the first time savings. Control Roomor sections or divisions dedicated to such operations formed by highly qualified personnel with extensive experience in the sector. The real-time data sent by the vehicles are analysed by the operating staff, and supplied to advanced artificial intelligence systems to verify failures on the vehicle and possible predictions of future failures. In this way, when the vehicle is returned to storage, the interventions to be carried out are already known, as they have been defined even before its return, thus obtaining the first time savings.

Figure 2: Innovative maintenance

As can be seen from Figure 2, parallel activities allow a dual output to be achieved simultaneously:

Monitor the actual occurrence of a fault on board the train, thus generating corrective maintenance. This allows the operational team to be prepared for the intervention even before the vehicle is returned to storage, thus eliminating the phase of an eventual inspection of the fault and thus saving time and money.
Analyze the trend of data sent from assets to monitor their "health" status, and predict any failures that may occur in the near future. In this case, the failure has not yet actually occurred; for example, if we are referring to a component, it may not have failed or broken down, but worn out or present breaking principles, which can be defined by the name of "symptom".

The benefits of telediagnostics

It is clear that to obtain these benefits, assets must be able to transmit information and data to the control rooms. Modern trains are already being built with the most advanced sensors and equipment to send data from the various subsystems on board. The situation is different on the most dated means, which are however in great majority on the Italian territory to travel thousands of kilometers every day on all railway lines. In most cases, these facilities do not have a large number of sensors capable of storing and transmitting data, or are not even present at all. It is therefore necessary to carry out operations of revamping of the vehicles or installation of sensors at specific points of interest on board, and to transfer the information to the CCU (Central Control Unit) of the vehicle logic; Control RoomAt this point, the information stored in the vehicle logic, in a database internal to the vehicle, must be sent to the ground via a device such as a GSM-based transmission modemR, which allows the data to be downloaded on board and sent to the operating room.

Although it may seem expensive to carry out such operations, they can be economically and thanks to the constant monitoring and analysis of data, it is possible to reduce the number of vehicle failures that generate a downtime in line of the vehicle, resulting in a reduction in the maintenance times of the vehicles.
As can be seen in Figure 3, despite an initial increase in costs due to the equipment of train-board monitoring systems and Control Room, Control RoomThis investment will soon be repaid to save on maintenance costs and lost earnings due to unexpected vehicle downtime.

Figure 3: Comparison of investment and savings in ongoing maintenance

In summary, therefore, the use of telediagnostics is capable of:

Limiting the occurrence of in-service failures:data provided in real time by the on-board system, thanks to processing of information acquired from dedicated sensors and correlation logics between available diagnostic signals, They allow the potential state of degradation and the residual useful life of the monitored components to be obtained and predict the possible occurrence of failures before their operational occurrence. An additional advantage of this methodology is that the repair or replacement of the component can only take place when it is really necessary and not at pre-defined time or distance intervals, resulting in economic benefit.
Optimize preventive, corrective and predictive maintenance processes by reducing the Mean Time To Repair (MTTR):operating staff are able to continuously monitor the status of the fleet in real time. Maintenance planning is optimized and spare parts procurement can be done according to operational priorities and constraints. The availability of the fleet increases, the running costs decrease and finally the quality of the service offered is improved.

On condition and predictive maintenance

We have seen how on condition maintenance, often called CBM (Condition Based Maintenance), and predictive maintenance allow to improve the results achieved through the concepts of classic maintenance, Obtaining numerous economic performance benefits and savings.

To better clarify a possible use case of applications in real scenarios in the context of railway maintenance on condition and/or predictive we can consider a typical example:

Modern passenger entry and exit doors are usually formed by a pneumatic unit and an electronic control unit necessary to send the opening and closing input. The manufacturer generally defines the time at which the door opens or closes completely, and you can also define the current absorption that generates this event under optimal conditions during tests in production. If such data are available for each opening or closing during the operation of the medium, it can be analyzed whether the time needed to open or close increases and the current absorption increases; monitoring the evolution of this information allows to predict if conditions are about to be reached where a failure or rupture could occur which would cause the door to become blocked in the closed position, or even worse in the open position, forcing the immediate stopping of the vehicle, since obviously a passenger train cannot travel with an open door locked.

It can be clearly seen how this context can be analysed and applied to CBM and predictive maintenance, In the first case, the occurrence of the conditions described is exploited to verify the fault, while in the second case this information is obtained by a massive analysis of data transmitted by the system concerned and processed by artificial intelligence algorithms.

The role of telediagnostics in an application context

In Italy, several railway operators have a Control Room control room capable of analysing data from the rolling stock. In addition to the large national operators, regional operators are now also working to form their own Telediagnostic divisions. In addition, the main railway manufacturers also provide their own technicians specialized in Control Rooms, Control Roomas part of the provision of Service in the contract for purchase of rolling stock by a customer.

Axcent has acquired extensive experience in this context, developing skills on new logics of telediagnostic over the years thanks to its "Data Analyst" and "System Engineer", that is analysts and engineers with a global knowledge of the railway vehicle, Capable of creating the diagnostic rules necessary for optimization of the maintenance process.

Thanks to this, Axcent System Engineering has conceived and produced its own idea of Control Room Control Room, developing the remote diagnostic monitoring software system, called MiKO. Through a simple and intuitive web interface, the operator can monitor the received data completely remotely. The train diagnostic files are processed by the system, which allows the user to perform different operations, including analyzing data in real time, creating diagnostic rules through a Drag&Drop tool shown in Figure 4, monitoring strategic KPIs, Geolocate assets in real time using GPS data, create geofencing areas to signal the entry of trains into strategic areas, view the status of the fleet, open maintenance alerts to the ERP system or directly to the maintenance depot.