Solutions to get Big Data into production do not have to be expensive. But implementation requires the combined know-how of skilled workers and data specialists.
Big Data and Analytics already consider 94 percent of OEMs and Tier 1 suppliers in the automotive industry to be relevant. But the technology is only fully operational at seven percent. This was the result of a study conducted by BearingPoint a few months ago. A quarter of the 120 decision-makers in the survey had just introduced corresponding solutions, often in the areas of marketing and sales.
It is obvious that Big Data Analytics is also an important instrument for quality and efficiency in production. This is especially relevant for small and medium-sized suppliers. “Many companies are aware that they have to deal with Big Data. At the same time, there is a fear of investment. If we want to make it work, the processes have to be adapted - and that costs money," says Dr. Frank Breitenbach, senior expert in planning methodology at EDAG Production Solutions, an engineering service provider located at the interface between product and production for the automotive industry. According to Breitenbach's experience, the effort required to retrieve data from the plants is rather limited: “It is essential to think in detail beforehand and to make hypotheses about how things can be connected. A lot of data can be obtained from the machine control system and complemented by a few sensors ". However, the already-" mass-existing" data from the PLC would not be used. On the contrary, there is no reason why I should measure each nonsense and throw it into the cloud - this will confuse the people who process the data," says Frank Breitenbach. At the same time, the range of sensor and raspberry Pi solutions (small single-board computer) is huge. In any case, it is important to consider what really makes sense.
Breitenbach also sees the numerous promises of analytics providers insinuating that it is sufficient to simply collect data in the cloud as a further problem. “A company that has no control over its processes does not benefit from a cloud solution that is supposed to make everything better. There are many solutions where you have to make a clear distinction: Is this charlatanism or is the provider sincere and honest." warns the expert. It is of special importance to integrate expertise into data analytics projects. It is crucial to know the process and what happens in the plant. This will result in the parameters. Therefore, the team must include the people who are maintaining a plant." Data Scientists by themselves can't solve the tasks, but to play a double pass users are required. IT is getting very close to the process, which was not possible before." explains Frank Breitenbach. The challenge for companies lies not in the analysis systems themselves, which are often available as an affordable open source solution, but in deciding which data is to be recorded and what flows into the algorithms.
High benefits are achieved by focusing on quality. For example, if a tube is to be sawn into 200-millimetre pieces, the saw blade and thus the sales department will change, but the customer specification is within a certain limit value range. If the data analysis shows where you have to interfere in order to get the Gauss curve into the delta, I have a typical, very simple application that brings a lot of benefits", says EDAG man Breitenbach, citing an example. Another proven way to take advantage of the collected data is to use self-developed apps – no rocket science nowadays with appropriate development platforms. You can show users exactly those data that is relevant for their work.
Would you like to talk about the efficient use of Big Data in your production? Our colleague Dr. Frank Breitenbach will be pleased to help you.
The interview was conducted by Daniela Hoffmann from automotiveIT.