Usually, the Internet of Things means “smart” household appliances, say, a kettle which notifies its owner that the water is boiling. But this concept has a global goal, within which engineers and scientists are planning to modernize and create enterprises where each machine will be equipped with multiple sensors for data collection and creation of a digital twin of manufacture. Together with scientists from South Ural State University, we are going to tell you about the goals set for businesses on the way to total digitalization of industry, and how contemporary Russian science helps implementing them.
Everything is tied up
For the beginning, let’s get a sense with the concept of the Internet of Things. It assumes establishment of a network which unites not computers or other customary electronic devices but any appliances at all. According to common understanding, the Internet of Things is intended so that ordinary things could be imparted with the ability to not only perform their direct function but also transmit data about it to other devices as well as receive such data. For example, a fridge, which determined that milk is about to expire based on signals sent by radio-frequency tags, can independently order a new portion from a shop without the owner’s participation.
While household “Internet of Things” remains at best a toy for techno geeks, the phenomenon is successfully winning over absolutely another territory – industrial enterprises. We surely are not talking about equipping factories with smart lamps or air conditioners. Instead, engineers propose to digitalize the manufacturing process itself, and in a perspective make factories to be fully independent from people.
It is nowadays already that almost all facilities at enterprises are equipped with digital control systems which allow shifting a part of responsibilities from personnel to machines, as well as monitor execution of commands. As a rule, all this equipment operates on its own, and after designing a factory, specialists can only obtain data regarding operation of each element separately, and this is not being done in a moment.
As engineers assume, a logical step that would allow enhancing automation and efficiency of production must be unification of disorganized equipment in a single network – the industrial Internet of Things. Due to the fact that each element of the production chain will contain sensors which will be monitoring state of the element itself as well as of released products, it is possible to not just understand what is happening with a certain machine or other equipment but also compose the overall picture of the happening at the factory, in other words, create its digital twin.
Application of a digital twin provides an enterprise and its personnel with several new possibilities at once. Due to the fact that data gets transmitted simultaneously from all the equipment, specialists can assess production efficiency, determine “narrow spots” and measure operating parameters of certain devices or the entire production chain in order to optimize manufacturing. This can be applied both at the beginning stage of production operation soon after it’s been launched, and further on. For example, this allows adjusting parameters of equipment for inevitable variations in quality of raw material.
Moreover, digital twin allows enhancing stability of production and reduce probability of equipment’s failure that can lead to total shutdown of the entire production line. Due to the fact that facilities will be equipped with multiple sensors and data about its operation will be constantly recording, it is possible to evaluate actual load and wear-off instead of the estimated ones, and carry out timely maintenance support, which is cheaper than unplanned overall repair.
“Having installed sensors at a turbine aggregate or a compressor and having constructed a model of vibrational field, we can beforehand predict occurrence of some failure. Not waiting until equipment becomes emergency, we can carry out preventive repair, which by time consumption and expenses requires less resources compared to emergency repair. This is the value of such technologies,” explained Aleksandr Shestakov, Rector of South Ural State University, where new methods of production control are being developed today.
Besides, introduction of digital twins and creation of factories with production lines which are totally controlled from a single centre will allow enhancing production flexibility. For example, this is how one can quickly rearrange production process and change the type of manufactured products. Provided an accurate enough model, digital twins can serve for checking new parameters of production process without putting real equipment and product release plans at risk.
In a word, development of digital twins is being related to radical changes in industry, which might seem far from reality. In fact, it is today already that there are not only plans and concepts in this sphere but real inventions; and some Russian factories are already implementing the digital twin concept in one way or another.
Digital twins in Russia and America
One of the most flawless implementations of the digital twin concept is carried out at an American steel mill called Big River Steel, located in the state of Arkansas. In 2017, the company launched its first production line at which various marks of steel intended for application in pipelines, automobile engineering and electronics are smelted. Its peculiarity is in the fact that originally the line was designed as completely computer-aided. Facilities used in the line are equipped with hundreds of thousands of sensors, data from which allow obtaining the overall picture of production in real-time mode and reduce human factor to a minimum.
“This steel mill is some kind of a ‘pilot ground’ for polishing the systems of Industry 4.0 till their expansion to larger industries. The company has been created in cooperation with SMS group company, a partner of SUSU. It unites modern automation technologies, methods of artificial intelligence in industry and many more,” explains Gleb Radchenko, Director of the School of Electrical Engineering and Computer Science of South Ural State University.
Digitalization is coming to Russia as well. For example, in 2015, engineers and inventors from SUSU created a control system of thermal electric station’s resources at Magnitogorsk Iron and Steel Works – this is the central thermal electric station which supplies the city and the plant with energy, and a more powerful central thermal power station. The system of model-forecasting control over steam boilers allows economizing fuel – natural gas – due to application of blast-furnace gas which gets produced during iron smelting in blast furnaces. Mathematical model uses data from sensors, and control algorithms, constructed on the basis of this model, regulate operational modes of boilers.
“Economic effect in the scale of one station reaches 50 million of rubles per year. It can be either higher or lower; everything depends on operational modes of the station and on equipment. The project is very profitable, and it is quickly paying off – in less than one year and a half,” says Head of the Laboratory for Power Saving in Social Sphere of the SUSU School of Electrical Engineering and Computer Science, Dmitry Shnaider.
Moreover, a team of engineers and programmers of South Ural State University created and innovative system of steel cast modeling in real-time mode. Specialists of SUSU invented it within a contest named SMS Group Data Challenge, held by SMS Group Company.
The task was to create a computer-aided system to prevent sticking of steel in the casting system. If this phenomenon is not prevented, the entire casting process must be stopped. More than 500 temperature sensors were installed in the casting system and evenly distributed along the entire casting form. Developers elaborated a neural network algorithm which analyzes temperature in different parts of the line, and predicts probability of sticking with high rate of accuracy.
Overall, the process seems to be simple – sensors are monitoring situation at the production line and transmit collected data to the controlling algorithms. But who is going to watch over sensors themselves? Because if collected data do not reflect real situation because of wear-off or some failures, data corruption can lead to even more hazardous consequences compared to failure of a sensor.
The simplest and most obvious solution is to ask employees periodically walk through the factory and check whether sensors are displaying correct data. This can become applicable if there are a few of several dozens of sensors. But when it comes to tens of thousands of sensors, this approach is simply impossible to use, and therefore automation is needed. Specialists of the SUSU International Laboratory for Self-Monitoring and Self-Validating Sensors and Systems together with the University of Oxford (Great Britain) are working on development of self-monitoring sensors that can independently determine their state and transmit data about it as well as about the main registered parameters. Besides, self-monitoring of sensors will be supplemented with data collection system which will be able to detect their “suspicious” behaviour.
Specialists of SUSU not only implement the invented systems at large industries; they also apply them right at their university. There are about 300 wireless temperature sensors at the university building, data from which get transmitted to the server. At that, the sensors have been developed by a company called Polytech-Automatic, a small innovative enterprise of the university, established by its alumni. Mathematical models analyze data from the sensors and control heat supply, switch it from one source to another and increase or reduce the load.
“Nowadays smart technologies allow bringing the process of production control up to a qualitatively new level. We can reduce expenses, make production to be more flexible and sustainable, predict wear-off of equipment – and all of this is thanks to the ‘digit’, virtual instruments. In fact, thanks to digitalization, we get principally new factories, the factories of the 21st century – and all this takes place without buying of new machines or production lines,” emphasizes Aleksandr Shestakov.