Möchtest du unsere Inhalte auf Deutsch sehen?

x
x

Intelligent pump monitoring using artificial intelligence

““
IoT Use Case - ifm
6 minutes Reading time
6 minutes Reading time

Monitor process variables and detect impending failures early on

Goal of the project

The supply pump is one of the central components of many plants. It ensures fluid circulation through the entire system. Damage to this pump will result in the failure of the entire plant – this must be avoided at all costs.

In this specific case, the pump is operated at various speeds, which makes static monitoring more difficult. Artificial intelligence can be used in a targeted manner here: AI simplifies pump monitoring while increasing the efficiency of the entire process. Process deviations (anomalies) are detected early and downtime is minimised.

Business Case

Avoidance of unplanned
downtimes

Early detection
of incipient damage

Setup possible
without specialist

On average, customers achieve:

8 h

savings in labour hours
thanks to simple configuration

€ 1,300

lower
integration costs

€ 15,000

cost savings through prevention
of complete damage

Successful digitalisation of a supply pump

Without having to make any changes or interventions on the PLC or software side, a supply pump can be successfully digitised with the moneo software solution. Thanks to intelligent pump monitoring, critical process variables are recorded during the ongoing production process.

By collecting and structuring the data, the process becomes more transparent and the associated potential for optimisation is shown. If there are any deviations from the specified process, moneo issues warnings and alarms at an early stage. This enables a fast response to changing parameters. Downtimes are reduced and maintenance can be planned and carried out in a more focused way. All measures also increase process and product quality while simultaneously boosting plant availability.

Furthermore, the tools in moneo are easy to use and offer a detailed visualisation of the entire process, so that all machines and plants can be kept track of.

  • Simple can be used without a data science expert
  • Convenient – through automated data processing
  • Intelligent –through targeted use of AI
  • Reliable –condition-based monitoring
  • Customised – customisable thresholds for warnings and alarms.

Value proposition

  • Machine availability

How to intelligently monitor a pump

To enable intelligent monitoring of a supply pump, critical process variables must first be identified. These measured variables can be recorded using the corresponding sensors on the pump. These include measured variables such as flow, speed, pressure, temperature and vibration. Previously, pumps were not monitored – or if so, only to a limited extent – for vibration and speedbecause at best only static process variables could be recorded. Monitoring and visualisation systems like moneo that could be used to monitor the process were often not installed either.

The result:no transparency, no monitoring, no alarms. And this, in case of doubt, resulted in complete plant failure, even when damage to the pump could have been foreseen.

This is precisely where AI comes in:the recording of critical measured variables forms the database that is used to define the normal state in the software. Ai-supported models are then trained on this basis to enable monitoring regardless of the operating state – in this case: regardless of the speed range. Deviations from the normal state are detected by the software and signalled as an alarm.

moneo simulation video

Benefits of AI-supported monitoring

Artificial intelligence facilitates dynamic monitoring of complex machine and production processes. The AI-powered assistant moneo SmartLimitWatcher monitors critical process values and sends alarms at an early stage if there are deviations from threshold values. Depending on the requirements, individual parameters and sensitivity to deviations can be set. By accurately recording the resulting data, dynamic threshold values can be observed and evaluated in the pump’s different operating states.

The tool is simple and intuitive to operate, so no data science expertise is necessary.

Automated pump maintenance

The complete solution including ifm sensors and moneo software offers numerous options and interfaces that help automate processes and thereby improve the entire supply chain process. Data processed in moneo can be exported using different protocols. It is possible to use data via MQTT or OPC UA in a third-party system or to transmit data directly to AWS, Azure or SAP via a specific connector.

By integrating ifm’s own SFI (shop floor integration) interface, it is also possible to have a direct connection to SAP PM. With the on-premises solution, the interface between the production and business levels enables the corresponding follow-up processes to be triggered automatically in the event of warning messages.

For example, if the dynamic threshold values of a pump are exceeded, the maintenance engineer receives a notification via SAP. Specific steps, such as maintenance and replacement as well as orders, can also be triggered automatically.

Condition-based maintenance of your pump and the automated processes allow you to reduce production costs considerably and increase machine availability.

  • ✓ moneo
    MQTT or OPC UA → AWS, Azure, SAP
  • ✓ SAP Integration
    SFI (shop floor integration) → SAP PM
  • ✓ Alert
    email notification (threshold)
  • ✓ Spare parts
    automatic ordering through SAP
  • Cloud
  • pump monitoring possible in the cloud (if artificial intelligence is not used)

This solution is currently possible on-premises if artificial intelligence is used. Connection to the moneo Cloud is also an option if artificial intelligence is not used.

How to install condition-based pump monitoring

To record the critical process variables on a supply pump, they must first be identified. With the help of the appropriate IO-Link sensors from ifm, this data can be recorded at the pump. The IO-Link masters are connected to the server via an internal, secured network (VLAN). The sensors used are each connected to an IO-Link master.

Measured variables such as flow, speed, pressure, temperature and vibration data are recorded. The data recording of the normal condition is used to create a model that enables monitoring irrespective of the operating status and can thus identify any deviations from the normal condition. The IIoT platform moneo takes care of recording and visualising the data. The moneo SmartLimitWatcher is used to analyse the dynamic process values and to automatically detect deviations in a critical process variable at an early stage. The stored data is analysed and a corresponding model is calculated.

Pumps can be operated in different states (e.g. under load or idle). Different thresholds are permissible in each of these states. The moneo SmartLimitWatcher sets dynamic thresholds for this purpose. If process values are outside a specific range, warnings or alarms are issued.

The moneo SmartLimitWatcher continuously monitors the target variable (e.g. temperature, flow, vibration or current consumption) with regard to the production quality, efficiency or plant condition. After a teach-in phase, this model takes over the monitoring of the pump and reports any deviations from the normal condition.

The target variable, in this case the flow, is monitored using the moneo SmartLimitWatcher. For this purpose, the “support variables” (speed, pump pressure, vibration data) are used for monitoring. They describe the flow characteristics in different operating states. For example, with increasing flow, the speed and the pump pressure increase as well.

System structure

  1. IO-Link master (VLAN)
  2. IO-Link flow meter
  3. IO-Link speed sensor
  4. IO-Link pressure sensor
  5. IO-Link temperature sensor
  6. IO-Link vibration sensor
moneo Software (SmartLimitWatcher)
  • ✓ Analysis and model calculation
  • ✓ Historical data
  • ✓ Visualisation
  • ✓ Monitoring
  • ✓ Alarm function
  • ✓Automatic detection of deviations from normal conditions
  • ✓ Threshold violations via SFI to SAP

Text taken from the original – ifm

Get our IoT Use Case Update now

Get exclusive monthly insights into our use cases, activities and news from the network - Register now for free.