Digitalisation is progressing in all areas of business and society. Track construction and maintenance is no exception. New possibilities of connectivity are constantly being developed, and intelligent systems in machines generate a large amount of data. This means that the partnership between infrastructure operators, construction companies, suppliers and service providers takes on a new dimension. Availability, transparency and traceability are key elements. Planning preventive measures brings with it many benefits.
Predictive maintenance has great potential for the maintenance of track construction machines. Machine systems gather valuable data used as a basis for evaluation and recommended actions. The service partner, who deals with the implementation of the recommended actions, plays a central role here.
“Predictive Maintenance” is the process of developing an individual assessment of the current machine condition based on collected data and work out concrete plans of action for condition-based maintenance.
Upon request by the operating company, relevant machine systems, e.g. PlasserDatamatic 2.0, can perform live monitoring of key data, such as working parameters, GPS position, direction of travel, engine data, operating fluid filling levels and hydraulic pressures. In doing so, the current status of whole machine fleets can be examined from one’s office. The gathered data can also be stored and offers a valuable basis for further analysis.
Dealing with this type of data is fairly new in the track construction industry. One company that is active in this area is P&T Connected, based in Hagenberg in Austria. The company deals primarily with the collection, processing and evaluation of data with the aim of implementing predictive maintenance for track construction machines.
For track construction and maintenance machines the regular, interval-based inspection on site or in a depot forms an integral part of testing and maintenance. Up to 250 sensors are integrated in these machines, providing a large amount of information to the responsible parties. This sensor data has been and is analysed for control purposes at fixed intervals and allows conditions and processes to be tracked at all times.
Furthermore, the following additional data can provide a wide variety of information about the condition of a machine:
- spare parts stocking,
- maintenance contracts,
- error messages,
- log files as well as
- weather and geo data.
Linking historical and current data will provide an improved analysis. The combination of various data sources allows for a completely new level of analysis. Its particular aim is to provide system information to be used for recommended actions.
An example of the many new possibilities is the replacement of wear parts at the most economically efficient point in time. This is based on load condition data provided by sensors. They indicate how much longer a wear part is likely to work perfectly. Thus, it is possible to diagnose early when certain parts will need to be replaced. Furthermore, downtimes and machine failures can be reduced to a minimum – a benefit for both machine operating companies and the infrastructure.
In principle, this method can be applied to the maintenance of whole machines. This will then pave the way for moving from the condition-based maintenance at regular intervals to predictive maintenance.
The latter offers a range of benefits compared to previous approaches (see illustration). Frequently occurring issues that do not have a major impact on the machine functionality can usually be solved by replacing spare and wear parts. Where issues that do have a major impact on the machine functionality occur frequently, the manufacturer will consider design changes. Predictive maintenance makes it possible to significantly reduce the frequency of issues and challenges that may have a major impact and thus improve the machine availability considerably.
By linking and analysing data from machines and systems, such as PlasserDatamatic 2.0, it is possible to draw better conclusions on the condition of a machine and make predictions about its further development. This enables the planning of maintenance intervals based on the actual machine condition, i.e. an individual maintenance schedule can be created for each machine and adapted, if required.
The benefit for the operating company is obvious: on the one hand it is possible to diagnose early when certain components will need to be replaced. The components can then be ordered in good time. On the other hand, the analysis may show that a wear part will function longer than usual and can therefore be replaced later.
Thus, predictive maintenance offers the following benefits compared to previous approaches:
- reduction of life cycle costs
- cost optimisation in maintenance strategy
- data-based recommendations for maintenance measures.
Fleet management gains a detailed insight into the condition of machines and can improve its repair and spare parts management. This results in a longer machine service life and has the potential to optimise maintenance costs.
One of the primary aims of this approach is to identify patterns in existing data. Everything revolves around analysis, forecast and data-driven modelling. After all, the aim is to provide tools for better and especially earlier decision making – and to make the railway system even more competitive.
Many companies have already implemented condition-based maintenance. With the use of sensors, the current condition of and in parts is identified. Maintenance is carried out as soon as it is required, or one or more indicators show a deterioration or exceedance of limit values.
Predictive maintenance goes one step further. Rather than just analysing the current situation, it predicts possible future events. It is possible to make forecasts using various approaches, such as mathematical modelling in combination with machine learning and neural networks.
Authorised representative, P&T Connected