About thirty people attended the day on digitization in the industrial sector that was held on Tuesday 27 June at the DayOne Business Center in Barcelona. A session jointly organized by the Digital Cluster and the MAV Cluster.

The event, which lasted for three hours, started with the presentation of the report "Edge Computing: State of the Art of Intelligent IoT" by Albert Rivero of Worsley Acceleration Services.

Edge computing offers data processing very close to the source where it is generated. With regard to processing in the cloud, it allows obtaining a much higher resolution and the detection of small anomalies reacting in real time while reducing network traffic. In addition, it is an important enabler for machine learning.

Rivero explained that there are many companies that are dedicating themselves to offering solutions to be able to optimize the processing capacity required by machine learning, such as HAILO, which manufactures microchips that allow machine learning; Deeplite, which designs new optimized algorithms to reduce energy consumption, and DataQube, which generates types of servers that can be placed very close to where the data is generated.

As Rivero explained, the future is to reach a hybrid cloud-edge model, where the more local processing offered by the edge will coexist with that of the cloud, which will be reserved for more demanding applications.

Rivera also presented some examples where edge computing will allow a paradigm shift in sectors such as mobility, retail, agriculture, industry and health. According to him, edge computing will drive the industrial IoT by offering new solutions to increase the productivity and quality of industrial processes.

Then, different successful cases of digitization of companies in the advanced materials sector were presented.

Xevi Sagué, CSMO - Director, Commercial and Industrial Marketing at AETECH

Sagué stated that in order for the industry to digitize its processes, the first thing that is needed is to have powerful communication networks (whether physical or wireless). As he explained, the integration of the processing part allows, through the application of predictive models, to reduce, for example, sudden plant stoppages and consequently the economic losses that this entails.

Sagué pointed out that the knowledge of the workers will be very valuable for companies, and in this sense he assured that having a digitized process is essential to allow the transfer of this specialized knowledge.

Oleguer Sagarra Pascual, co-CEO of Dribia Data Research

Sagarra emphasized that the first step to be able to improve a process is to digitize it in order to obtain all the information and improve it through prediction, automation and optimization tools.

In order to start this process, it is necessary to choose a project that does not "burn our shoes", but that has an impact on the company. It is very important to have someone inside who believes in it and accompanies the whole process.

In the industrial case, the goal of artificial intelligence applied to processes is not for the machine to make all the decisions and go alone, but rather a binomial machine + specialist technician will always be needed, who will be the one who ends up making the last decision

David Saldes, Data Science at R+D+i, Celsa Barcelona

Finally, Saldes assured that the application of AI has served them to improve energy efficiency in the steelmaking process while also improving the sustainability of the process, since the improvement of the accuracy of the assistance system in the loading of raw materials it has an impact on the optimization and reduction of their consumption.

Panel moderated by Marc Milian, Worsley

During the round table it was made clear that companies need to make their own way to be able to achieve the objectives in terms of digitization. Small changes must be made, start from the bottom and grow.

Likewise, it was emphasized that often the main limitation when implementing a digitization process is being able to show in numbers what it will entail as an economic improvement in the short and long term. Thus, it was concluded that although it may seem like an expensive solution, it is much more expensive to do nothing.

Finally, the speakers pointed out that there are three types of companies: the innovative ones, which will be the leaders in implementing this type of strategy; followers, and those who will do nothing and are destined to close.

Work tables and sharing of ideas and conclusions

The day ended with the work tables and a sharing of ideas where it was emphasized that the first step to start implementing digitization strategies is to identify the problem and trust the appropriate specialists.

At the same time, it was agreed that companies that do not get on this train, in the short or long term, will end up disappearing because they will no longer be profitable. In this context, it was emphasized that small and medium-sized companies are much more agile and can apply this type of solution in less time.

However, companies need to invest in innovation strategies and have key stakeholders and users identified. In this way, if they want to opt for a certain aid or apply for a call, such as an AEI or an IRC, they will benefit much more.

To finish, it was concluded that collaboration between companies in digitization projects is key, as alliances make it possible to find synergies and face a challenge with more strength. At this point, clusters play a key role in helping to bridge the gap between their members.