Condition monitoring prevents unplanned downtimes in transport systems

Transport systems such as escalators, conveyor belts, monorail conveyors and storage and retrieval machines must not fail. Suitable condition monitoring could prevent unwanted downtimes at an early stage; conventional vibration analyses are not suitable for the very low-frequency vibrations of these systems, though. HARTING and Formsmedia GmbH have worked out a new concept for condition monitoring which has already proven itself in practice.

Unplanned downtimes are a recurring and expensive nuisance in transport systems. Monorail conveyor transport systems used in automobile manufacturing show this in an exemplarily way. The supporting structures transport heavy components or whole vehicle bodies. The whole load is borne on plastic coated wheels. At some point the coating will come off the wheels and the monorail conveyor will have wheel damage. Production is interrupted and repair of the monorail conveyor is time-consuming.

Ausfälle in Transport-Systemen kann zum Stillstand kompletter Betriebsabläufe führen. Ein Condition Monitoring System auf Basis des Edge Computers MICA® erkennt Verschleiß frühzeitig und verhindert so ungeplante Ausfälle.

Condition monitoring for a monorail conveyor transport system also available as a retrofit

Vibration analyses for fast running engines or transmissions have proven themselves as a reliable monitoring instrument for about 20 years. Slow moving transport systems such as monorail conveyors require a considerably more sensitive vibration detection in the milli-G range due to weak and very low-frequency vibrations. Formsmedia, a company for measurement technology, pulse and data analysis, has developed a monitoring solution with HARTING for this requirement:

  • Highly sensitive sensor boxes with MEMS acceleration sensors capture the movement of the transport system capacitively and detect the vibrations at the wheels, collect data on the motor current and on the temperature of the drives.
  • The sensor data is collected and transmitted via Modbus assigned to the Edge Computing System MICA from HARTING. The MICA is a networkable and secure mini-computer with a Linux based operating system and a virtualised application environment consisting of Linux containers.
  • For condition monitoring of the transport systems, the MICA uses an analysis software from Formsmedia to aggregate, store and visualise sensor data included locally onsite. MICA is suitable for assessing data quickly directly onsite so that no unnecessary data needs to be passed on.
  • The fast Fourier transform (FFT) is used to analyse the spectrum of the vibration and to evaluate the non-harmonious vibrations as well.
  • The condition monitoring system can also be added on for transport systems with slow moving components as a retrofit.
  • The data is transmitted to higher-level SCADA control systems or cloud based IoT platforms to carry out condition monitoring of several systems and provide extended functionalities, such as predictive maintenance.

Condition monitoring for difficult environment conditions

The condition monitoring of transport systems must be frequently set up under difficult environment conditions. There is either not enough space, the distances that need to be bridged are large or there is dust, heat and humidity. The sensor boxes and MICA can be used in a rough environment and are also protected against heavy EMC loads to degree of protection IP 65/67.

All local requirements are recorded in detail in a proof of concept to preconfigure the sensor box and MICA for installation. The attachment of the sensors is also checked for mechanical suitability so that weak vibrations are recorded reliably. A Modbus checkout program can be used to double check all the preset functions on-site during commissioning and ensure the correct measurement ranges. This means false alarms can be largely be ruled out.

Depending on the environment, the preprocessed data can be transmitted to higher-level IT systems wirelessly or wired in exchange formats, such as MQTT or OPC UA, using the gateway functionality of MICA. The wireless version of MICA is used for the monorail conveyor transport system to directly evaluate the sensor values in each mobile transport rack and be able to transmit them via WLAN.

An early warning message prevents plant downtimes

Threshold values are defined for evaluating the sensor data so that the production engineering department receives a message about the early need for maintenance. In practice, it has proven helpful to install a monitor on-site that is connected to MICA via Modbus RTU (Remote Terminal Unit). The current characteristic values and critical system states are displayed visually on the monitor in the form of a traffic light function. Selected receivers are also sent an alarm message if threshold values are exceeded. In addition, employees on-site have the possibility at the touch of a button to send events such as breaks, malfunctions, or missing material to MICA. This means the sensor values can be connected to real operational events.

The condition monitoring solution can be enlarged to include machine learning for predictive maintenance in order to predict when the next maintenance service is required. Cloud services are generally used for the necessary analysis of historical and current data. Machine-learning algorithms are developed and trained for this using the collected data about the vibrations of the wheels, the fluctuations in the motor current as well as the temperature development of the drives. Predictive maintenance is capable of recognising slight changes in the sensor data as anomalies and able to reach conclusions about the current conditions and forecasts about future conditions.

For topics such as predictive maintenance, it is recommended to engage in cooperation with specialised IT system houses who have experience in the area of data analysis and who develop algorithms, rules and dashboards individually to the user's requirements and are capable of integrating it into existing control or ERP systems. The was founded for this by HARTING with IT companies.

Quick amortisation of the condition monitoring solution

The cost of launching a condition monitoring system pays for itself relatively fast. Thanks to a standardised condition monitoring concept, the planning, system and installation costs are low and the benefit to the system operators is high. This means that wear of critical components can be recognised early and a failure of the transport system avoided. This increases system availability and the overall equipment effectiveness (OEE). In addition, fewer repairs and as-needed maintenance reduce the maintenance costs. This ultimately improves the service for the customers concerned/fields of use.

An additional effect of condition monitoring can be seen in a retrofit situation. Thanks to digital monitoring, old facilities can become state-of-the-art again. The maintenance cycles, which are otherwise more frequent due to the product lifecycle, can be reduced down to the actual requirement, thereby extending the usage time of the facility.

Durch die Schutzart IP 65/67 sind die Sensorboxen und die MICA auch im rauen Umfeld einsetzbar und außerdem gegen hohe EMV-Belastungen geschützt.

Die vorverarbeiteten Daten können je nach Umfeld drahtlos oder drahtgebunden über die Gateway-Funktionalität der MICA in Austauschformaten wie MQTT oder OPC UA an übergeordnete IT-Systeme übertragen werden. Für das Hängebahn-Transportsystem wurde die Wireless-Version der MICA verwendet, um die Sensorwerte in jedem mobilen Transportgestell direkt auszuwerten und per WLAN übertragen zu können.

Frühzeitige Warnmeldung verhindert den Anlagenstillstand

Für die Auswertung der Sensordaten werden Schwellwerte festgelegt, die der Betriebstechnik frühzeitig eine Meldung über den Wartungsbedarf geben. In der Praxis hat es sich als hilfreich erwiesen, wenn vor Ort ein Monitor installiert wird, der per Modbus RTU (Remote Terminal Unit) mit der MICA verbunden ist. Auf dem Monitor werden die aktuellen Kennwerte und kritische Systemzustände optisch in Form einer Ampelfunktion angezeigt. Zusätzlich erhalten ausgewählte Empfänger eine Alarmmeldung, wenn Schwellwerte überschritten werden. Ergänzend haben Mitarbeiter vor Ort die Möglichkeit, per Touch-Button Ereignisse wie Pausen, Störungen oder fehlendes Material an die MICA senden. So können die Sensorwerte mit echten Betriebsereignissen verknüpft werden.

Die Condition-Monitoring-Lösung kann durch Machine Learning zu Predictive Maintenance erweitert werden, um zeitliche Voraussagen über den nächsten Wartungseinsatz machen zu können. Die dafür notwendige Analyse historischer und aktueller Daten findet in der Regel mit Unterstützung von Cloud-Services statt. Hier werden Machine-Learning-Algorithmen entwickelt und mit den gesammelten Daten über die Vibrationen der Laufräder, den Schwankungen im Motorstrom sowie der Temperaturentwicklung der Antriebe trainiert. Predictive Maintenance ist damit in der Lage, schleichende Veränderungen in den Sensordaten als Anomalien zu erkennen und Schlussfolgerungen über den aktuellen und Voraussagen über den künftigen Zustand zu treffen.

Bei Themen wie Predictive Maintenance empfiehlt sich die Zusammenarbeit mit spezialisierten IT-Systemhäusern, die Erfahrung im Bereich Datenanalyse haben und Algorithmen, Regeln und Dashboards individuell für den Bedarf des Anwenders entwickeln und auch die Integration in bestehende Leit- oder ERP-Systeme übernehmen. HARTING hat dafür gemeinsam mit IT-Unternehmen das gegründet.

Schnelle Amortisation der Condition-Monitoring-Lösung

Die Einführung von Condition Monitoring rechnet sich relativ schnell. Die Planungs-, System- und Installationskosten sind durch das standardisierte Condition-Monitoring-Konzept gering und der Nutzen für die Anlagenbetreiber hoch. So lässt sich der Verschleiß von kritischen Bauteilen frühzeitig erkennen und ein Ausfall des Transportsystems vermeiden. Dadurch erhöhen sich die Anlagenverfügbarkeit und die Gesamtanlageneffektivität (OEE). Weniger Reparaturen und die bedarfsgerechte Wartung verringern außerdem die Instandhaltungskosten. Schließlich verbessert sich der Service für die betroffenen Kunden / Einsatzbereiche.

Ein weiterer Effekt von Condition Monitoring zeigt sich beim Retrofit. Durch die digitale Überwachung werden Altanlagen auf den aktuellen Stand der Technik gebracht. Die aufgrund des Product Lifecycle sonst häufigeren Wartungszyklen können auf den tatsächlichen Bedarf reduziert und die Nutzungszeit der Anlage verlängert werden.