Best Practice

Reply’s Axulus is helping companies to deploy technology innovations using the IIoT and edge computing

In industrial production, technology deployments are tremendous and the sector is experiencing a continuous innovation. From the 1960s’ PLC-based automation to the development of hardware and software automation stacks, which evolved into digital solutions running on the cloud, many things have changed. However, it is not only a matter of technology, but also of time: what before needed months - if not years - to be developed, can now be deployed within weeks. However, technology, especially in industry, has great value as long as it can solve concrete challenges and generate measurable value on top of existing solutions.

Taking this perspective, technologies like cloud and edge are two new arrows in the quiver on the hunt for solutions to generate additional value within the industrial environment. Both technologies can be added to the industrial solution stack, and then interplay with other technologies like AI or novel sensor and automation equipment, thus contributing to solutions that drive concrete value for manufacturers and industrials. Of course, the deployment of innovative solutions is key for competitive advantages for companies in this sector. This is the area where Industrie Reply operates: the Reply Group company focused on solutions for the industrial market has a sound experience of use cases related to Industrial Iot (IIoT) and edge technology in manufacturing.

Today industrial customers are challenged by the ever-increasing speed and complexity of novel technologies. Although IIoT promises substantial value-add and nearly all industrial companies have implemented strategies and technical proof of concepts (PoCs), only a few of these customers can scale these solutions and realise their full business potential. The technological innovation cycle is considerably faster and easier than scaling the solution to an industrial set-up. In this direction, Reply presents a new way for customers to identify use cases and scale IIoT solutions at scale, and generate substantial value add for their business.

There are many potential solutions that depend on high-volume data being available for analysis and training of algorithms, such as predict failures of critical assets. Examples of this include mechanical assets like drives, where failure patterns are typically analysed based on vibration data, with high frequency and data volume. Such tailor-made methodologies and tools like Industrie Reply’s Axulus Value Scaling Accelerator are key to speed-up innovation and roll-out in an industrial environment based on use cases and solution templates as they allow clear workflows and scalable collaboration.

The challenge to apply these solutions at any production environment is that the capacity for analysing this data can’t be allocated, for example, at the PLC level (as the PLC ‘needs to steer production’), whereas the vast amount of raw data also cannot be sent to cloud environments for analysis, as the production network bandwidth needs to be reserved for production-related data communication. Therefore a dedicated edge component (like an industrial PC for instance) can be the answer. This way it is possible to pre-process data on a ‘local stage’ while leaving scaling and predictive capacities to the cloud. The process is quite simple: thanks to the edge component the system executes data cleansing and data compression tasks (like FFT) close to the asset itself, and then sends only the relevant information (e.g. the vibration spectrum) over the production network to the cloud for predictive maintenance applications.

This allows solutions to be implemented easily and smoothly with no latency concerns. For example, the vast amount of data generated by numerous electronic manufacturing lines can be used to train algorithms in the cloud (data and compute-heavy) in order to identify quality issues of the produced boards. Once these algorithms are trained to a level of accuracy and fulfil the required quality standards, the trained algorithms can be pushed to be executed on edge devices close to the line, thus avoiding large data traffic across the production network, while still being able to check the data of each board produced at high speed. Edge, however, allows also to simulate process outcomes, thus bringing an invaluable benefit to the whole industrial process optimisation.

You may also like

No contents here.