Predictive maintenance can reduce cost, increase productivity and streamline production. [shutterstock: 340585748, kenary820]
For industrial companies, predictive maintenance is the ideal point of entry into highly-automated production, based on internet technologies and the Internet of Things. Data gathered and processed in real time can improve production processes and bring down maintenance requirements.
The Internet of Things is made from a flow of status-data on machine components and ultimately this data is combined with information from third-party systems, such as ERP applications from SAP.
Conspicuous patterns pointing to malfunctions can be recognised at an early stage and measures introduced on a preventive basis. As a result, users avoid costs caused by production outages, contractual penalties and compensatory measures.
In addition, the period required for production-throughput is reduced and delivery dates can be determined and adhered to more precisely.
A typical application example of automation based on internet technologies is a service used in engine production, one that supervises production quality.
Accordingly, an automotive manufacturer produces thousands of cylinder heads in manufacturing process, from constructing the shape and then producing the metal, through to the post-production finishing.
During production a data-set is established for each cylinder-head, gathering details on the manufacturing process across all data-transmission points. Here, in real time, the automated production environment records process parameters such as times, measurements or temperatures.
Production-control then checks this data off against the setpoint values; that way it can supply first signals indicating possible deviations but cannot yet produce a solution for eliminating the error. In such a scenario, huge data quantities emerge, requiring experienced specialists to evaluate them in a manual process.
Only they can draw the correct conclusions based on the raw data supplied.
Making Data Work
A solution for data analytics creates greater transparency here, delivers clear statements on the progress of production, and highlights new interconnections that lead to an accelerated analysis of errors.
For this purpose, however, all relevant measurement data from the production process must be gathered on a continuous basis. These items of information can be used by an analytics solution for more extensive statistical analyses.
Those responsible for production can then recognise the situations in which the manufacturing process deviates from requirements and when a manual intervention is needed more quickly. Subsequently, it becomes easier to adhere both to customer requirements and also to technical tolerances.
On this basis, machine maintenance also becomes more predictable. Until real-time analysis is an option, companies need to wait to get the required data until a complete production batch has been manually evaluated for errors.
Because information about deviations from the norm is available quickly, there is less frequently a need to replace tools, for instance.
This means that companies boost their productivity on a lasting basis, also shortening the run-up phase in the production process. Such a scenario demands efficient data management, able to handle the huge data-flows that are recorded in real-time.
A Solution for SAP
Jointly with partners, NetApp supplies a platform for analyses, directed at forecasting. As the basis for storing the sensor data, applications such as Hana and the Big Data platform Hadoop come into operation.
The infrastructure consists of the FlexPod solution, a pre-validated and convergent infrastructure platform, with components from NetApp and Cisco. The modular solution includes a server, a network link-up via switches and storage with a flash memory.
This enables companies to accelerate their IT projects and to implement them with less risk, because all components are coordinated with one another.
Simultaneously, the solution brings the necessary stability and speed necessary for connecting SAP ERP with the analytics results produced by the Hana cluster or the Hadoop cluster respectively.
To be able to use the analytics results on each device, the solution can be combined with tools such as Lumira or Tableau. In addition, a central platform for the data-management is necessary in the scenario presented: this platform supports a high level of security against system outages.
NetApp provides a back-up solution for this, specially adapted to the Hana technology and based on snapshots, in order to secure production-related data provided at short intervals and to rapidly feed this in again if the need arises.
The NetApp flash-memory systems also have an integrated RAID function available to them, one that offers a data reduction of more than 30 per cent, compared to traditional Hadoop scenarios for production outage.
This way, in the event of data losses, companies can reduce the period needed for data-restoration by up to 500 per cent, compared to traditional back-up concepts.
The concept developed by NetApp for the internet of things is based upon an established reference architecture, one that enables a hardware environment to be built-up quickly.
Combined with solutions from SAP and the Hadoop distribution MapR, as well as other Hadoop technologies, companies establish analytics applications that are viable for the future, so as to make the internet of things a reality.
The application example of a manufacturing environment in the automotive sector shows that data, as a factor of production, is assuming an ever greater importance for companies as they set themselves up for future competitiveness.