Sensor Data as a Basis for Industry 4.0
Precisely tailored sensor technology in production is a basic prerequisite for the implementation of Industry 4.0. It captures process data and the machine status and makes them available.
The costs for sensor technology and the variety of possible applications often make their direct economic benefit for the individual user not immediately apparent.
The VDMA has compiled tools for practical application in the Sensor Technology for Industry 4.0 Guide. The aim of the guide is to show users and manufacturers of sensor controls possible levers and ways to reduce expenses for sensor technology. Appropriately compiled questions and toolboxes help in implementing possible strategies. The guide was developed by the VDMA Forum Industry 4.0 in cooperation with the wbk Institute for Production Technology of the Karlsruhe Institute of Technology (KIT) and a VDMA industrial working group consisting of 13 leading sensor manufacturers and users.
Lubrication Requirements in Ball Screws
One of the main initiators of the guide is Prof. Jürgen Fleischer from the wbk. "Sensors are the interface between the digital and the real world and therefore one of the most important links in the implementation of industry 4.0. Without adequate sensors, all higher-level data interpretation systems are blind," says Fleischer.
He uses project examples from KIT to describe which data can be captured and usefully processed with the use of sensors: “In drive components of machine tools, data can be collected to monitor their condition and optimize operation. For ball screws, the axial force and the friction torque at the ball screw’s nut can be measured. The lubrication requirement can then be determined by comparing a model for the friction behavior in order to lubricate the component as required. This adaptive lubrication has significantly increased the service life of ball screws in tests conducted by the KIT"
By detecting structure-borne noise, different drive components can be monitored, e.g. ball screws. “These signals change over the lifetime of the component and thus allow conclusions to be drawn about the state of wear. The goal is predictive, condition-based maintenance."
At EMO the wbk will present the combination of a camera system with a machine learning algorithm, which makes it possible to monitor the wear of a ball screw.
Software Facilitates Analysis of Different Data
The implementation of algorithms for the analysis of sensor data and the determination of quality-relevant features suitable for automatic evaluation, however, often requires a lot of time. The Xeidana software, which was developed at the Fraunhofer Institute for Machine Tools and Forming Technology IWU in Chemnitz, provides a solution package that can cover tasks ranging from data acquisition to automated quality control.
Among other things, the scientists identified quality-determining features of components. The software is able to detect surface defects reliably and in real time on the basis of optical sensors, such as multi-camera systems. In a further step, the researchers plan to feed this data back to the production system. This allows countermeasures to be taken in good time if, for example, process parameters get out of hand.
Further examples for the real-time acquisition of sensor data at the IWU are forces that are recorded in the tools of forming machines, for example pressing, punching and cutting forces.
Correct Interpretation of Acquired Sensor Data
Whether the sensor data needs to be captured in real time depends on the specific application. "You have to answer the question of how much real-time you actually need. There is also the question of how to synchronize data. But it is also important to know which sampling rates are necessary for a sufficient process description," explains Dr. Jörg Stahlmann, Managing Director of Consenses GmbH in Roßdorf. The company supplies industrial measurement technology and digitalization solutions. Against this background, the use and development of suitable sensors and the interpretation of data is one of Consenses' core competencies. "We use 3D step models to understand our customers' designs. This is necessary in order to be able to correctly classify the sensor data, such as the expected force and temperature flows or kinematics. Understanding these correlations is important in order to be able to interpret sensor data at all," is how Stahlmann describes a Consenses approach.
Fleischer emphasizes that: “Thanks to simulations of components, assemblies and machines, we can broaden our understanding of mechanical effects in the production plant. We use this knowledge to install sensors in a targeted manner and to interpret the captured data efficiently".
Not Every Application Requires Real-Time Data
Referring to the subject of real-time, Stahlmann explains: "In any case, it is a misconception to assume that the best possible quality will be achieved when they are labeled as real-time. Real-time data is often provided by controls that originally collected them to control certain actions of machines." This objective sometimes does not coincide with requirements for actual sensor data. Therefore, it is important to understand which signal is generated in a concrete case, before far-reaching analyses or decisions are derived from this data.
Fleischer uses an example to explain when real-time data capturing is superfluous: "A quick reaction to captured data is not necessary for condition-oriented maintenance. The results of the data evaluation may also be provided several hours after the data has been captured. However, in many cases sensor signals change very dynamically, for example in the case of structure-borne sound signals. The capturing of such signals requires a high sampling rate and particularly fast, real-time data acquisition. In this case, however, data can be memorized in a buffer allowing for a whole set to be evaluated later. The evaluation can also be outsourced to a more powerful server.
Dr. Thomas Päßler, Group Manager Forming Machines at IWU, added: "Everything that cannot be justified for economic reasons does not require real time. For instance, trend analyses covering a extended period of time do not require real-time data. It is not necessary to keep all data, -only individual parameters are generated and archived. Production parameters such as output or energy data are real-time capable, but it is not essential to have them available in real time. Regarding energy data, for example, it is sometimes sufficient to take a value every 15 minutes. In addition, it makes little sense to collect data in real time that is relevant to the management, for example certain parameters on the economic efficiency of production, such as how many components of a grade were produced on one plant."
In its position paper Industrial Workplace 2025, the Scientific Society for Production Engineering (WGP) also dealt with the question of meaningful and proper automation: “The design of a profitable value-added process should exploit all technical possibilities. This means that it is not always necessary or useful to implement the highest possible degree of automation."
Real-Time Helps Prevent Damage
One thing's for sure: Real-time sensor data are necessary wherever the protection of machine, tool or workpiece or the stability of the process is at stake. Päßler explains: "Real time capturing becomes indispensable if it is the only way to prevent hardware or workpiece damage. This applies, for example, to cases such as tool breakage or excessive stress on assemblies such as bearings or frame components. In order to avoid waste right from the start, it also makes sense to use appropriate sensors to determine the properties of the starting material in real time".
Fleischer gives another example of real-time capturing: "By detecting deviations in the production process in real-time, damage can be avoided or limited. For example, errors during setup of machine tools or in the NC program can cause collisions. If these errors are detected quickly enough, the machine can be stopped, and the material damage can be reduced."
Connected Sensor Data Provide Useful Status Parameters
Scientists at the IWU monitor forces, paths and expansions on forming presses in real time. However, this data is not evaluated individually, as usually. Among other things, they are fed into Smart Stamp, a software-based analysis module. This is where the data is merged and analyzed. Does the press work within normal parameters? Or is the RAM on which the upper tool is mounted tilted unfavorably, which would result in the workpiece not being optimally formed or the tool wearing out more quickly? "While individual sensor data are often not meaningful, such questions can be answered by merging the data," says Päßler.
This article was first published by MM MaschinenMarkt.
Original by Annedore Bose-Mund / Translation by Alexander Stark
This article is protected by copyright. You want to use it for your own purpose? Contact us at support.vogel.de (ID: 46111280)