Research of the possibility of using neural networks in the tests of locomotive hydraulic transmissions

Abstract

The possibility of developing a self-diagnostics system of the diesel locomotives hydraulic transmissions information-measuring test system is researched. The use of neural networks and fuzzy logic for the development of a self-diagnostics system of the diesel locomotives hydraulic transmissions information-measuring tests system is proposed. As the initial stage of developing a diagnostic system using neural networks, a neural network is presented which predicts the rotational speed of the hydraulic turbine wheel during run-time tests. For this purpose, a three-layered, fully permeable perceptron is constructed. This perceptron allows to predict the last 24 seconds of rotation of the hydraulic transmission turbine wheel shaft in run-out mode based on the previous 48 seconds with a fairly high accuracy. Such a developed neural network can be used as part of a self-diagnostic subsystem of the diesel locomotives hydraulic transmissions information-measuring test system and in tests of transmissions in run-out mode when the sensor D2-MMU-2 is used.

Authors and Affiliations

I. A. Kliushnyk

Keywords

Related Articles

Investigation of electromagnetic transients in a frequency controlled electric drive of switching devices of traction substation transformers

The study proposes the boundary conditions for selecting the capacitance of the converter to ensure the exclusion of the mode of self-oscillation and dangerous overvoltages in controlling the frequency of rotation of the...

Creating an effective technology for the formation railway routes with grain cargoes using models on the Petri net

The method of formalization technology of formation railway routes with grain cargoes with the models on Petri's nets is proposed. For the first time, modeling of railway technological processes of transportation in the...

Methods for automating the design of easily tested computer systems and devices based on digital automata

The relevance of the work is to provide minimal additional hardware costs during design automation of easy-tested digital devices, which are represented by models of control finite state machines on hardware description...

Experimental researches of computer-oriented methods of consumed electricity payment minimization

The purpose of the work is the experimental research of practical efficiency of the developed complex of computer-oriented methods of payment minimization for electricity consumption of railway transport traction electri...

Reception of information signals of systems of railway automatic equipment in the conditions of action of three-component interference

Typical situation in work of paths of signaling of systems of railway automatic equipment is need of reception of signal against the background of additive multicomponent interference. In quality interference component i...

Download PDF file
  • EP ID EP497647
  • DOI 10.18664/ikszt.v0i5.115370
  • Views 138
  • Downloads 0

How To Cite

I. A. Kliushnyk (2017). Research of the possibility of using neural networks in the tests of locomotive hydraulic transmissions. Інформаційно-керуючі системи на залізничному транспорті , 22(5), 8-15. https://europub.co.uk/articles/-A-497647