Detect defects before they occur

Unforeseen breakdowns are the worst-case scenario for fleet managers, machine operators and service staff and are almost always associated with high costs - for example, with agricultural machines that are in use day and night during harvest. Here, an unplanned shutdown can wipe out an entire season's work. The fact that a combine harvester or forage harvester breaks down is something that should not happen in the first place. Prevention is the motto in modern agriculture. Early detection of faults, risk minimisation of unplanned downtimes and proactive maintenance are topics that will consequently be high on the agenda of SYSTEMS & COMPONENTS, the B2B marketplace for agricultural machinery and the entire off-highway sector. Additional input is provided by the trade fair's diverse technical programme, which introduces visitors to current and future trends in maintenance - for example, in the expert lectures and discussion rounds in the "Future Lounge". The buzzwords "condition monitoring" and "predictive maintenance" are being used more and more frequently: The goal of condition monitoring and predictive maintenance is to proactively maintain machines and minimise downtimes.

Machines always in view

Predictive maintenance differs from traditional reactive maintenance approaches. The process uses real-time data to derive maintenance information. This is where advancing digitalisation is leaving its mark. It allows new inspection possibilities to continuously record the actual condition of mobile work machines. In addition to powerful hardware that can collect data at high frequency, this requires intelligent diagnostic software and IoT gateways. Connected via the vehicle's internal interfaces such as CAN bus or Ethernet, they collect the data and transmit it to the cloud via WiFi, Bluetooth or mobile communications.

Here the focus is particularly on condition monitoring of the drive train. Because wherever there is movement, sooner or later there will be signs of wear and tear. Modern condition monitoring systems that integrate intelligent plug & play sensors directly into the critical components of mobile machines such as excavators, wheel loaders or bulldozers, for example into the drive shafts, will be presented. As the sensors are exposed to extremely harsh conditions in the off-highway environment, they must be very robust. In addition to operating hours, temperatures, speed and torque, vibrations and oscillations are also measured. If, for example, the amplitude increases in a specific frequency range, this is an indication of an impending defect in the cardan shaft, roller bearing or gearbox if the sensors are correctly parameterised.

Intervene before everything grinds to a halt

Condition monitoring offers a number of advantages. The collected values form the basis for further Big Data analyses that filter out conspicuous deviations from the large amount of data. They enable permanent self-monitoring of the important subsystems in the engine, the coolant and lubricant circuit, exhaust gas aftertreatment or air circulation. If values deviating from the norm are detected, the system makes recommendations for action to the driver via the on-board computer or intervenes directly. Impending failures in rotating components can be prevented, for example by the lubrication system increasing the amount of lubricant when required. In the future, this will enable necessary maintenance and unscheduled service to be detected in good time, thereby preventing downtimes.

In addition, the data from a vibration analysis provides indications of the time frame within which the machine operator should replace the affected components. For this purpose, the operational strength designed during construction is compared with the loading registered by sensors and the wear is calculated from this. Such intelligent predictive algorithms are a central component of predictive maintenance. They make it possible to carry out the necessary maintenance "just in time".

Machine breakdowns as a virtual phenomenon

The topic of predictive maintenance is closely linked to the digital twin, as the data for the prediction can also be obtained through simulation. The digital twin is the virtual model of a machine or one of its components, which can process the same sensor-based real-time data as the physical original. In the event of a defect, the service technicians can use the digital twin to find the cause of the malfunction. The advantage: As early as the design stage, the simulation can be used to run through different degrees of utilisation or to generate sensor data for different fault conditions that train the predictive maintenance algorithm. The digital twin makes it possible to map operating processes and predicts how the condition of a component will develop over the entire life cycle as the operating time progresses. Machine learning algorithms ensure that the digital twin learns from every event enabling it to make increasingly accurate predictions over time. As a result, machine failures in the field or on the construction site become a purely virtual event.

At SYSTEMS & COMPONENTS, IT-supported enhancement of human perception - i.e. augmented reality - will also find its way into service and maintenance. This displays the interactive content with comprehensible instructions for action via a smartphone, tablet or data glasses. Image and video-supported diagnosis is intended to support the driver in the future. To do this, he/she can contact the service desk and is guided from there to potential sources of error via audio, visual and video information. In addition, the expert at the service desk can use augmented reality tools to support a service technician directly on site. The aim is to significantly improve the quality of service so that the technicians already have the right spare parts with them on the first visit.

On the road to Smart Maintenance

Scientists at the Fraunhofer Institute for Cognitive Systems (IKS) in Munich want to apply these smart maintenance application scenarios to an entire fleet of highly complex harvesting machines together with the agricultural machinery manufacturer Holmer and the telecommunications manufacturer Huawei. The sensor information is sent to a central maintenance service in the cloud via a mobile network. "In many fields, however, it is not always possible to realise a stable and powerful radio connection," knows Michael Stiller, who coordinates the INVIA project as the responsible Business Development Manager at the Fraunhofer IKS. In order to reliably and seamlessly monitor the work of the harvesters, even when a constant real-time connection to the internet is not guaranteed, part of the intelligence required for this was therefore relocated directly to the communication gateway on the machine. "Depending on the application, this makes it possible to process the data from the harvester directly on site and to carry out important partial analyses without transferring them to the cloud," said Stiller.

INVIA relies on a combination of cloud computing, mobile edge computing (MEC) and fog computing. A special Edge Embedded Control Unit (ECU) forms the core of the fog cloud in the event of a poor or temporarily absent connection to the mobile phone base station. In this way, basic functions of the assistance system can be provided even without a connection to the service centre. In addition, the project will investigate the advantages of integrating an MEC component instead of or in addition to central cloud services. The background: MEC components offer very low latency and high computing capacity due to their proximity to the machine, making the provision of augmented reality or video services possible. Furthermore, INVIA is to expand the service offer with online on-the-job training. Stiller said: "This allows the company headquarters to support the driver with a trainer in situations that are still unfamiliar."

SYSTEMS & COMPONENTS shows where the journey is heading

Maintenance in the off-highway sectors is changing. Engineers, designers and original equipment manufacturers will find everything they need to know about condition monitoring and predictive maintenance, as well as assistance systems for the maintenance and repair of mobile work machines, on the trade fair grounds in Hanover. New tools such as smart sensors, digital twins and AR glasses complement the classic spanner. Interrupting a construction project or the harvest because maintenance is due or a spare part is missing for an urgent repair - in future, this should increasingly be a thing of the past.