Bosch "Invented for life" is a motto well known, but what about Bosch "Simply. Connected"? The new slogan highlights the new approach of the company that is presenting solutions for smart homes, smart cities, connected mobility, and Industry 4.0. While smart homes will offer greater convenience and safety, and smart cities will improve the quality of life, the Industry 4.0 initiative needs to make sure that the factory of the future is flexible, connected and smart and that it enables people, machines, and products to communicate with each other.
The Industry 4.0 term implies that versions 1.0, 2.0, and 3.0 must have previously existed, and, indeed, these refer to specific epochs in industrial history. Industry 1.0 refers to the steam-powered machines that, to varying degrees, required heavy physical work from people, for instance in transportation or mines. Version 2.0 is characterized by the assembly line, which sped up production by dividing processes into steps, making it faster, cheaper, and more efficient. This allowed mass markets to be served. Industry 3.0 refers to computers, and to the robots and machines they control, which allow an extensive degree of automation in production and a reduction in costs. This is the current status quo in industry.
The future of industry is built on all of these things, but is far more flexible. In addition, to some degree, it has the capacity to self-organize, since parts will tell machines how and when they "want" to be processed. The societal trend towards customization is a driving force for this. In such a factory, the lubricant is information, which tells the machines and robots how they should organize themselves for each project.
By connecting people, machines, and materials, a virtual image of the manufacturing process can be generated on the computer - in real time. The network of programs with mechanical and electrical components communicates via the internet. This makes constant coordination possible, even between locations around the world and beyond company boundaries.
Big Data is the ground on which the foundation of the I4.0 concept stands. The term "big data" is used to highlight scenarios where classical analytical architectures become inefficient for the analysis of the data that is generated. By definition, the "big data" concept is characterized by three key features known as the "3Vs": volume, variety and velocity. Referring strictly to the "3Vs", the data production landscape looks something like this: i) world's data production is doubling nearly every year, and this highly increasing trend affects all the economic/industrial sectors; ii) unstructured data grows much faster than structured data, and iii) the velocity at which data is generated becomes much higher, being driven by an increase in the number of devices/sensors that generate new data. The obvious fact is that the Big Data Age is here (it is already here for some time), and we have to adapt.
The manufacturing industry is one of the sectors producing big data. As one of the most important players in the production of automotive components, industrial and building products worldwide, our company is adapting to this natural evolution; and the first thing to adapt is the whole architecture.
The implementation of an end-to-end architecture for big data analytics is challenging due to various reasons: the necessity to fulfill customized requirements specific to different business sectors, the presence of not enough mature technologies for data integration/modeling/access on the market and also the need to adapt many of the analytical algorithms to the big data context are just some of the challenging issues. The new architecture has to cover the above mentioned challenges, and in our case to fulfill specific requirements such as (near) real time constraints, heterogeneous data sources or the analysis and visualization of large volumes of data. Our IT architects implement the generic solution presented in Figure 1, using primarily open source technologies. The solution is based on an Apache Hadoop Ecosystem and it is customized, to fulfill the requirements in our business sector.
Figure 1. Generic architecture for a big data analytics platform
Together with colleagues from Germany, US, and India, the Bosch associates in Cluj-Napoca develop analytical solutions based on the production data from the Bosch plants worldwide. We provide the means for real-time process monitoring, faster and better reaction against failure, prediction of events, data driven support for problem solving, optimization of risk management, reduction of reaction time of customer requests. From simple visual analytics to complex prediction models, the performed analytics need to fullfill the high quality standards imposed by the automotive industry, since they have a direct influence on the production processes.
Figure 2. The CRISP-DM model
The CRISP-DM model (Cross Industry Standard Process for Data Mining) is the standard we implement in our projects. The Bosch associates in Cluj-Napoca manage all the steps of a project, from business understanding to deployment. Understanding the industrial process, with the help of the process engineers in the plant, is a crucial part of the data understanding process, and this can only be achieved by maintaining close contact with the industrial process owners, in their respective centers of competence.
The development of a complete analytical solution relies, on the one hand, on a reliable IT infrastructure, and, on the other hand, on the interdisciplinary skills of our Bosch associates: DB expertise, programming skills and solid background in Mathematics, Statistics, Machine Learning and Data Mining. Thus, the solutions incubated in our analytics center are rolled out to specific Bosch plants worldwide. On a long-term basis, the target is to serve not only Bosch plants but also the external market.
In terms of benefits, looking only at Bosch's more than 250 plants worldwide, Bosch CEO, Dr. Volkmar Denner, estimates that Industry 4.0 will save the company hundreds of millions of euros annually in the years leading up to 2020.
by Ovidiu Mățan
by Mihai Câsu
