By Nikolaus Fecht
Customers who purchase wind power or industrial gears do not consider predictive maintenance in an isolated manner. "They often see predictive maintenance in the context of digitalization, the Internet of Things or Industrie 4.0," says Wilhelm Rehm, member of the Board at ZF Friedrichshafen AG as well as Chair of the VDMA Power Transmission Engineering Association.
Basis for autonomous driving systems
In April 2017 at the Hannover Messe, ZF Friedrichshafen showed how they enable industrial applications to see, think and act. Among other things, the company presented a control system for artificial intelligence, which it developed together with a leading American developer of image processors. The system allows vehicles to "understand" their environment by using deep learning technology to process and interpret sensor and camera data.
According to ZF Friedrichshafen, this control system provides an optimal basis for the automated, autonomous driving systems of tomorrow, both for cars and in an industrial setting.
More efficiency at lower costs
In all predictive maintenance applications, sensors monitor important values such as the oil level, temperature or acoustics of the gears. "From the data generated there, recommendations for action can be derived for users and the optimum time for maintenance can be determined," explains Rehm, adding: "In connection with our drive technology, these functions help to increase efficiency and reliability while lowering maintenance costs."
Roller bearings collect information
"For predictive maintenance that extends far into the future, the data obtained with sensor technology is a prerequisite," explains Dr. Stefan Spindler, Member of the Board of Schaeffler AG, Industrial Division, Herzogenaurach. "Particularly exciting are our customer-specific configurable, integrated sensor bearings, which can simultaneously record several measuring units." According to Spindler, rolling bearings hold the most potential for gaining this information. They are used where mechanical forces are applied.
Predictive maintenance offers big advantages to customers, such as the avoidance of sudden and costly failures. In addition, they can plan maintenance intervals better, because predictive maintenance allows the actual remaining time of individual machine components to be calculated in advance. This allows maintenance to be optimally adapted to the operations: Manufacturers carry out maintenance on production plants or drive systems when customer demand is low.
Domain know-how is customer-friendly
At the Hannover Messe, Schaeffler presented a smart Industrie 4.0 system as part of the VDMA special presentation "Predictive Maintenance 4.0". It allows the user to leverage the domain know-how of Schaeffler as digital cloud services in the cloud.
Many customers of Bosch Rexroth AG, Lohr am Main, are already taking advantage of the possibilities of predictive maintenance to reduce the downtime of their machines and systems. "With a new service portfolio, we are offering a cloud-based big-data solution for predictive maintenance of large-scale plants," says Dr. Steffen Haack, member of the Board at the company and responsible for the Industrial Applications business unit, continuing: "For this purpose, the sensors integrated in hydraulic units continuously measure temperatures, vibrations, pressures and oil quality, among other things."
A data acquisition unit in the units continuously sends the sensor data via an encrypted connection to a Bosch Rexroth subsidiary server. "The system uses machine learning methods in the cloud to detect any critical errors or significant changes in the operating state early on. It is proven to reduce the risk of a plant standstill," says Haack, in praise of the system.
How positive the influence of predictive maintenance can be is shown by an application in the rubber industry: Hydraulic motors from Bosch Rexroth move the rolling mills in which rubber mixtures are created. In its new system, the plant also uses an online diagnostic system from Bosch Rexroth. It is designed to enable the user to perform maintenance work through the interaction of sensors, cloud-based applications, and machine-learning methods before a standstill occurs. Sensors first record detailed data in the new system, which is then analyzed by a secure Bosch Rexroth server in compliance with the company's strict data protection guidelines.
Early but not unnecessary warning
A machine learning algorithm determines the normal condition of the rolling mill after a longer training phase. If only a single measurement value exceeds the tolerance range for a short time, this does not necessarily lead to a warning since wear and tear can rarely be detected with one single signal. If, however, the health index deteriorates because the data from several sensors changes, the system warns of a problem - even if the individual changes are within defined boundaries.
Cleverly replacing parts
"Customers have realized that they can save money with predictive maintenance," notes Dieter Michalkowski, Global Account Manager and Industrie 4.0 expert at Aventics GmbH, Laatzen. This is ensured by increased plant availability as well as clever replacement: The user only replaces parts if they absolutely have to be replaced, instead of replacing all parts preemptively in a fixed cycle. Industrie 4.0 makes it possible to generate the necessary information for predictive maintenance. For this purpose, Aventics monitors the sensors present in the plant parallel to the control system and evaluates the information in an intelligent monitor and intelligent valve systems. This information is made available to the user through standardized data protocols.
Actively identifying changes
"Customers expect systems to actively identify and react to changes," says Roman Cecil Krähling, Head of Condition Monitoring, Fluid Management & Electronics at Argo-Hytos GmbH in Kraichtal-Menzingen, speaking from experience. "In the past, predictive maintenance usually did this by providing only a simple warning of a future threshold violation. Nowadays, customers demand that the optimization of the machine operation is derived from the detected change in condition," describes Krähling. This could be done, for example, by changing the load of a system or by planning a service with an automatic spare parts order placement.
Predictive maintenance rests on two pillars
Here, predictive maintenance comes into play which in principle rests on two pillars: the reliable recording and provision of status information and the optimized operation and maintenance based on this information. "The sensors and intelligent algorithms developed by us enable our customers to reliably monitor the state variables, such as oil aging and component wear and tear online," says Krähling, concluding: "Changes can be identified in time and appropriate measures taken before damages or a standstill occur." Argo-Hytos' sensor technology makes it possible to develop tailor-made predictive maintenance solutions that have already proved their worth in a variety of applications. The company presented new developments and systems for condition monitoring at the Hannover Messe.
Customers demand greater energy efficiency
"Driven by European and international legislation, customers are calling for further improvements in energy consumption and noise emission," says Dr. Klaus Roosen, Systems Engineering Manager at Parker Hannifin GmbH in Kaarst. In addition, there are often restrictions with regards to space for installation and maintenance cycles, whereby oil status monitoring is of particular importance.
However, according to Roosen, the customers attach particular importance to the fact that the machines and systems have a long service life without downtime. The service life of the machines is determined by the optimum maintenance as well as preventive and predictive maintenance. Parker Hannifin therefore recommends regular maintenance and inspection as well as the requirements-oriented replacement of filter elements for example, and condition monitoring of the systems.
The company pursues a holistic approach based on the principle of "Total System Health Management." It not only considers individual components, but also complete systems. This includes diagnosis and prevention as well as the maintenance and rapid repair of complex systems, such as hydraulic systems and fluid power drives. The combination of diagnostic tools and processing systems allows a realistic view into the system. Ideally, continuous monitoring would make corrective measures unnecessary or calculable in order to avoid unpredictable failures.
Access in real-time
Parker Hannifin presented a modular hydraulic unit with integrated cooling and electrical control at the special presentation "Predictive Maintenance" at the Hannover Messe, which shows active status and cycle monitoring in real time. The company also demonstrated remote monitoring which entails monitoring and operation from a distance.
Identifying changes early on
"From a maintenance perspective, condition monitoring sensors and systems form the basis to recognize changes in system behavior at an early stage and to derive suitable recommendations for action from them," explains product manager Christian Meindl from Hydac International GmbH in Sulzbach. For this purpose, the individual measuring units of sensors are correlated, then either evaluated centrally or in submodules and potentially automatically transmitted to the maintenance planning and control systems. Depending on the degree, this can lead to automated order placement in the company's merchandise management system. The basic prerequisite, however, is a closer networking of all plant and system components as well as a standardization of interfaces and processes.
At the Hannover Messe, as part of the special presentation, Hydac International presented an exhibit that demonstrates the possibilities of networking and system integration. It is a hydraulic adjuster for rotor blades (pitch control) of wind power systems consisting of two hydraulic cylinders with an integrated distance measuring system and a control block with control valve and piston accumulators for emergency functions.
Networking with data logger
In addition to the sensor-monitored system components, the central hydraulic unit comes with fluid condition monitoring. Apart from solid particles, it also determines liquid impurities such as water as well as fluid aging states throughout the entire hydraulic system. Both the condition monitoring components and component-specific sensor technology can be networked individually or in groups with superordinate system and process control systems and integrated into the maintenance planning. "This is primarily made possible by modern data loggers and network technologies, which, thanks to flexible interface communication, are ideally networked with the customer's systems," says Meindl.
More effective maintenance
Customers of J. Schmalz GmbH from Glatten are also focusing on making maintenance as a whole more effective, targeted and thus more cost-effective, under the motto "Service on Demand". Walter Dunkmann, head of the vacuum automation business field at the company, says: "The use of external smart services through predictive maintenance data is also becoming increasingly interesting for customers."
J. Schmalz therefore tackles this on various levels. The company offers smart field devices, for example, which collect, interpret and provide energy related and performance-relevant data in the network. In addition, the data can be compared with further data from the process - this is also something which is being demanded repeatedly by customers. They place emphasis on integrated solutions with predictive maintenance as part of the automation and industrial IT environment.
The seamless communication of the data from the field to the cloud is also important and made possible with intelligent components. For example, J. Schmalz has a vacuum and pressure switch that also uses "Near Field Communication" (NFC) as a communication channel. This means that all the important data is directly available to the system's maintenance person on his or her smartphone. At the Hannover Messe, as part of the special presentation and among other things, J. Schmalz presented intelligent vacuum components with comprehensive functions for energy and process control. "We show how data can be communicated in real-time and at the same time create usable value-added data, such as for displaying in the cloud. Gripping systems are therefore becoming an important component for Industrie 4.0," summarizes Dunkmann.
VDMA Fluid Power Association | VDMAimpulse 02-2017: "Predictive maintenance is becoming ever more important" | VDMAimpulse 02-2017: "Machine Learning - A journey to the year 2030" | Argo-Hytos | Aventics | Bosch Rexroth | Hydac International | Parker Hannifin | Schaeffler Technologies | J. Schmalz | ZF Friedrichshafen