According to scientists, automation and digitalization of equipment in the mining and processing industry lags far behind the advanced industries. This is especially acute for the issues of intelligent control. For example, control of vibrating machines which mix various bulk materials due to vibrations of the working surface. To achieve the desired effect, the surface should move chaotically, but the necessary parameters, such as speed, should be maintained. This is needed to avoid material particles separation into different layers.

To do this, it is necessary to ensure synchronous operation of the rotors of the vibrating machine with a phase difference equal to the number of Pi (3.14). The optimum state in this case will be comparable with a Huygens pendulum clock. The pendulums of these clocks attached to a wooden wall begin to move synchronously as time passes, despite the fact that they could be started asynchronously. At present, the optimum operating parameters of the vibrating machine are set manually.

“ Within the framework of the project supported by the megagrant, co-jointly with Alexander Lvovich Fradkov, Head of the Complex Systems Control Laboratory of IPMash RAS, and with Boris Rostislavich Andrievsky, Dr Eng, my scientific adviser, an algorithm based on reinforcement learning has been developed which «teaches» the machine's control system to bring into effect any given speed in real time. With that, nothing else is required: the operator can set the target speed in the interface window, and the algorithm adjusts the controller parameters all by itself so that the task should be performed as required," said Yulia Zaitseva, an IPMash RAS research scientist.

In addition, based on adaptive control, a new algorithm has been developed which allows you to adjust another important parameter – the phase shift between the rotors. It is implemented with minimal human involvement also, you only need to set the required phase shift and the program will do the rest of the work.

The use of these algorithms helps to control the machine through a computer, in the way as it is done in smart manufactures. A great success is, that for such a complex mechatronic object under control, these algorithms turned out to be workable.

Scientists state that the technical modernization of such manufacturing processes is comparable with the transition between industrial waves and requires great effort.

«But the result is worth the effort, this is proved in the works on the economic justification of the costs of industry branch development. The benefits would be enormous: not only for the beneficiaries in the face of investors, but also for the country's economy and the environment,» Yulia Zaitseva said.

The Laboratory forDigitalization, analysis and synthesis of complex mechanical systems, networks and mediums, where these studies are carried out, was established at IPMash RAS within the framework of the megagrant of the RF Government. The Laboratory is headed by Emilia Fridman, Professor of Electrical Engineering at the Faculty of Engineering at Tel Aviv University.

Areas of work of the Laboratory: digital intelligent control of multi-rotor vibration installations; autonomous navigation and control of multi-agent robotic and mechatronic complexes; adaptive control and evaluation for aircrafts at high delays and limitations on communication channels; development of digital control methods for nonlinear spatially distributed systems and strength properties of nonlinear acoustic metamaterials and development of new methods and algorithms for digital adaptive control of continuous distributed systems under conditions of significant disturbances.

The results of the study have been published in the journals Electronics, Symmetry and a series of articles of the IFAC and IEEE conferences.