Simple

Feeder’s load state recognition based on the analysis of the current oscillograms

Nikita Mukhlynin
Ural Power Engineering Institute, Ural Federal University named after the first President of Russia B.N. Yeltsin

Vladislav Samoylenko
Ural Power Engineering Institute, Ural Federal University named after the first President of Russia B.N. Yeltsin

Abstract
This paper provides a way for solving the problem of recognizing the composition of electrical loads, which are connected to a common feeder of the power supply node to identify the state of operation of each of them separately. These solutions can be used for remote control over especially critical electrical loads when it is impossible to equip them with measuring sensors directly, or if creation of a dispatching system for them is not economically feasible. The load state recognition method is based on a system that provides measuring highly discrete current oscillograms at the feeder head. This system includes three stages of processing measurements that determine its setting and the procedure for further operation. To process the measurements at different stages, the discrete wavelet transform algorithm, as well as the mathematical mechanism of the continuous wavelet transform are used. The key idea is the use of sets of non-standard mathematical mother wavelet functions in the algorithm, which make it possible to recognize the electrical load inside a complex current load signal. Description of each of the stages, as well as the algorithm, are given in the paper. On the example of mathematical processing of real high-discrete current measurements obtained during an active experiment and their further analysis, it is shown that the proposed recognition method provides a stable determination of the state of individual power receivers, including one or more drive mechanisms. More complex electrical receivers, including many servos and actuators, have less stability to be recognized. In addition, the paper presents the results of determining the boundaries of load recognition in the conditions when measurement point of high-discrete currents is remote for other uncontrolled loads and elements of the electrical grid. The possibility of the load recognizing not only individual electrical receivers from a group of different ones, but also electrical receivers from a group of identical ones, operating on a common load, is shown in the paper. The use of the described recognition system may be relevant for industrial companies with electrical grid, as well as for utility companies that collect data on the state of powerful electrical loads of large consumers. In the future, the system algorithm will be improved to be able to identify shadow generation sources installed at consumer’s side. The possibility of recognizing types of events in the power system is not excluded.