What is predictive maintenance? - Definition, application and examples

But what does the term actually mean? Predictive maintenance means checking devices as soon as the first irregularities occur, but before faults occur. Unlike fixed maintenance cycles, predictive maintenance is intelligently predicted. Indications of this intelligence can be, for example, an increased operating temperature or higher energy consumption. This is based on data analysis of machine data and predictionsthat are made by analysing the database and sensor values.

In the context of predictive maintenance, you often hear the term preventive or predictive maintenance. This maintenance is particularly relevant for manufacturing companies, as their added value depends on fault-free devices and machines. It's hard to imagine how much one minute of downtime costs one of the major German car manufacturers. This quickly answers the question of whether intelligence for preventive measures is worthwhile.

Predictive maintenance gives companies the opportunity to significantly increase their operating efficiency through the following benefits:

• No unexpected production downtime: This is probably the greatest added value that predictive maintenance brings, as this is where most of the money can be burnt. But it is not the only advantage, and that is what makes predictive maintenance so attractive.
• Efficient personnel planning: In times of a shortage of skilled labour, skilled workers who can professionally maintain a specific machine are becoming increasingly scarce and therefore more expensive. If, for example, a specific maintenance requirement arises due to an increased operating temperature, the specialists no longer travel to the site on suspicion, but only when something actually needs to be done.
• Lower material costs: The consumption of wear parts and consumables can also be made more sustainable. If a saw needs a new saw blade, the increased consumption of cooling water indicates this. Thanks to predictive maintenance, the saw blade can be replaced at exactly the right time and the cooling water is not wasted unnecessarily.

Some of the most common predictive maintenance application examples can be found in the automotive industry, aviation and in manufacturing companies. In the automotive industry, for example, predictive maintenance is used to continuously monitor the condition of machines, engines and other vehicle components and to recognise maintenance requirements in advance. The technology is also used in aviation, for example to monitor engines and other safety-critical components in order to proactively prevent failures.

So what do you need for predictive maintenance? First and foremost, there are the sensors that measure a value. These sensors are either already present on the device or can be retrofitted. In the second step, the measured values are transmitted to an online platform. Depending on the application, a cost-effective Wi-Fi connection, a reliable 4G/5G connection or a long-range LPWA connection is most suitable here. With such a IoT network, many parameters such as power supply, network coverage, security requirements, etc. must be taken into account.

Once the data has arrived in an online platform, the magic happens. For predictive maintenance, for example, threshold values can be defined which, when exceeded, trigger a maintenance appointment with specific tasks. This allows maintenance measures to be controlled as required, with the aim of minimising the time, resources and downtime required.

Ziehl-Abegg, a family-owned company specialising in ventilation and air conditioning technology, uses predictive maintenance to minimise the downtime of its fans and reduce unnecessary maintenance work. Together with Deutsche Telekom, Ziehl-Abegg has developed a cloud-based IoT platform. This platform enables fans and third-party devices to be networked and monitored in real time. Customers and manufacturers can thus manage all machine information and operating data efficiently and centrally. Predictive maintenance is a great added value, especially for critical applications such as the ventilation of CT devices.

Like Ziehl-Abegg, RUD Ketten Rieger & Dietz GmbH & Co. KG is also a manufacturing company. The leading manufacturer of high-performance chains and industrial conveyor technology relies on predictive maintenance through the Internet of Things (IoT). Telekom's IoT cloud enables RUD to monitor the conveyor chains in power plants and industrial facilities remotely. Sensors on the systems send data on load, wear and heat development to the cloud, where it is analysed and visualised.

Thanks to predictive maintenance, RUD customers can monitor the condition of the chains in real time and plan maintenance precisely. This is particularly helpful in the areas where RUD chains are used. Unacceptable heat, dust-covered rollers or unusual speeds can cause chains to stall in harsh industrial environments. Predictive maintenance minimises unplanned downtime and significantly reduces maintenance costs, thereby increasing the efficiency and reliability of the systems.

How cost-effective is the use of predictive maintenance as a modern maintenance strategy? The potential is great, but the implementation costs in particular continue to pose a key challenge. Installing the sensors and setting up the dashboard is time-consuming, but there are already initial optimisation approaches. The solution is called Dormant Connectivity and is a cost-effective way of networking devices. With Dormant Connectivity, it is not only worth monitoring the machines on large production lines in industrial companies, but also smaller devices such as cleaning robots.