Analysis of a large database to optimise products and production processes
Implementation of Business Intelligence systems, data analysis in order to optimise the processes and the resulting products.
Manufacturing Big Data or Industrial Analytics is the specialisation of methods and tools that are required to handle and process large amounts of data on the manufacturing and Supply Chain Management fields.
The data can therefore come from IoT systems that are connected to the production layer or from the exchange between the IT systems for the planning and synchronization of production and logistics flows.
Manufacturing Big Data includes the application of new techniques and tools of Data Analytics & Visualization, simulation and forecasting, to highlight the information that is hidden in the data and its effective use to support quick decisions.
IoT and BigData are often together considered, two of the enabling technologies for Industry 4.0, and in particular for its fundamental element, namely the “Cyber Physical Convergence”. Following a logical “bottom-up” approach, the essential elements of these technologies can be illustrated as follows.
Starting from IoT technology, the following elements are highlighted:
the miniaturization of sensing and wireless communication technologies now makes it possible to include devices (embedded systems) in almost every physical object for the collection of information on the physical environment and their communication on the network. For objects that are born without such devices, it is now possible for them to be added at a later time; therefore, practically every physical object has (or may have) the ability to generate data on its state and the state of the physical environment that surrounds it;
the availability of pervasive high-capacity wireless networks (eg., lte, wifi, zigbee, bluetooth) allows to collect this data and to connect practically any physical object on the internet. This allows for, on the one hand, the collection and sharing of data that is generated by the devices embedded in the physical objects, and on the other hand configurate them and also remotely act on the physical objects and their surrounding environment;
users’ personal devices (smartphones, tablets, wearable objects such as smartwatches) can also generate data, communicate them on the internet and receive commands remotely, exactly according to the same paradigm;
the set of these three elements (pervasiveness of embedded systems, pervasiveness of networks, pervasiveness of personal devices connected to the network) constitutes the technological basis of the internet of things.
From a technological point of view, the Internet of Things is therefore the extension of the traditional Internet, designed to make specific objects, computers, communicate without a particular link with the surrounding physical world – to a network that allows physical objects to communicate directly with each other and to people to interact with both near and remote physical objects. With the widespread use of IoT devices, we are witnessing the so-called “data deluge”, ie the availability of an enormous amount of “raw” data generated by the devices in the physical environment.
If on the one hand the presence of such a large quantity of data provides such exceptional opportunities, on the other hand, it becomes even more necessary to couple with technology-specific technologies. These are needed for the management, data integration and the knowledge extraction of the raw data that is generated by IoT devices, commonly known as BigData Analytics (which, in turn, generally require high capacity computing infrastructure to both store and analyse such quantities of data, this is currently provided through cloud storage and computing platforms).
The synergy between IoT and Big Data technologies is one of the foundations of Cyber-Physical Convergence (and of the corresponding Cyber-Physical Production Systems – CPPS). Cyber Physical Convergence is defined by a circular process (Information Value Loop) between the physical world and the cyber world (Internet). Thanks to IoT technologies, objects and people constantly generate data that passes from the physical world to the cyber world via pervasive networks. In the cyber world, Big Data technologies make it possible to analyse collected data by extracting knowledge.
It should also be noted that the same focus on the aspects of IoT and Big Data is the basis of the initiative (IDS) led by Fraunhofer in Germany, the main German research institution oriented towards industrial innovation. In particular, IDS is seen as the enabling factor for all Industry 4.0 solutions and is focused on the collection, management and analysis of data on the entire production chain, both within the different units of the same company, and within the various companies of the same production chain.
Contextualized in the world of Industry 4.0, Cyber-Physical Convergence allows a continuous interaction between things, data, people and services, which is the basis of many of the fundamental concepts of Industry 4.0.
In particular, it is thanks to this interaction that the continuous circular process of:
maintenance and reconfiguration of production processes.
There may be different areas of application of the circular process illustrated above. If applied to a single production process, for example, this approach allows you to monitor the process accurately and continuously in order to control it effectively, constantly improve it over time and readjust it with respect to the variability of the external context (with benefits in terms of costs. , timing and flexibility of the process in question from a “zero-defect” perspective). If extended to different departments and lines of the same company, it allows for optimal internal integration, improving performance at the company level (vertical integration). If extended outside the companies (i.e. if other external companies also adopted IoT and Big Data technologies), the new technologies would offer the possibility, so far never been so powerful, to integrate into wider supply chains, dynamically identifying manufacturing partners in networks which would otherwise be impossible to connect, even if only for temporary productions (horizontal integration). Finally, if the IoT and Big Data approach were also extended to the world of customers, the flow of information on products during their life cycle and on the customers themselves would allow the company business models to evolve in the direction of offering high added value and circular economy services.
Given that IoT and Big Data technologies are currently mature (although always evolving in the research sector), it is possible to plan timely and immediate interventions for the adoption of IoT and Big Data technologies right away, with a view to a longer-term migration. towards Industry 4.0 models.
By way of example, we can indicate the following technologies that are currently already available on the market. Regarding IoT technologies:
ARDUINO and RASPBERRY PI.
They allow the creation of small integrated systems at a very low cost. They are essentially general hardware platforms, on which it is possible to “mount” sensors of various types (light, humidity, movement, etc.) and cards for wireless communication that implement the most common communication standards (wifi, zigbee, …). They also have local processors that can be programmed for the analysis of data collected “on site”.
It is the name of commercial solutions based on the ieee 802.15.4 standard for the construction of low energy consumption sensor networks (IoT). The zigbee devices can communicate with each other and receive commands remotely, to change their status. They are often used for control applications, such as home automation, but have applications in many sectors, such as retail, smart parking, intelligent lighting, smart metering, etc.
Bluetooth low energy.
It is the evolution of bluetooth technology for IoT devices with very low energy consumption. It allows wireless communication between physical objects in the vicinity, such as wearable devices (smart watches, smart badges, etc.) and with devices inserted in the surrounding physical environment (eg smart thermostats).
Sigfox, lora, cellular-iot.
They are technologies for wireless coverage of large geographical areas for ultra-low consumption iot devices. Typical applications are intelligent lighting or smart metering. For example, in the sigfox case, a network service operator (conceptually similar to a cellular operator) guarantees coverage in particular geographic areas. By purchasing IoT devices (typically sensors) that can communicate on the sigfox network, you automatically have the ability to create systems for monitoring and controlling physical objects. In particular, sigfox collects the data generated by the sensors and provides them to the user who installed them by subscribing to an access service via the cloud.
It is one of the leading manufacturers of physical IoT sensors and devices. The available devices allow to collect data on a very large variety of physical quantities. It is possible to associate network cards with these sensors to enable the transmission of data on the main IoT technologies.
Regarding Big Data technologies, we can mention:
- Hadoop. It is the reference open technological framework for programming analysis systems on large amounts of data.
- Hive. It allows the efficient execution of requests (queries) for the collection and analysis of data on distributed systems, compatible with Hadoop.
- Spark. Reference technology for the efficient programming of parallel systems for large-scale data analysis.
- HBase e Cassandra. Technologies for managing large amounts of data on distributed systems with particularly high performance.
By applying these technologies, it is possible, for example, from now on:
- integrate IoT devices for monitoring the various stages of production;
- analyse Big Data from the production process or from the use of products by customers.
- build circular processes according to the Information Value Loop scheme. At the moment it is reasonable to think of implementations in this sense within a single production reality or, in the case of larger companies, as a support for the “vertical” integration of various production units of the same company. One of the most immediate examples of this approach is the predictive maintenance of machinery.