Eye Feature Extraction with Calibration Model using Viola-Jones and Neural Network Algorithms

The field of mechanical fault diagnosis has entered the era of “big data”. However, existing diagnostic algorithms, relying on artificial feature extraction and expert knowledge are of poor extraction ability and lack self-adaptability in the mass data. In the fault diagnosis of rotating machinery, due to the accidental occurrence of equipment faults, the proportion of fault samples is small, the samples are imbalanced, and available data are scarce, which leads to the low accuracy rate of the intelligent diagnosis model trained to identify the equipment state. To solve the above problems, an end-to-end diagnosis model is first proposed, which is an intelligent fault diagnosis method based on one-dimensional convolutional neural network (1D-CNN). That is to say, the original vibration signal is directly input into the model for identification. After that, through combining the convolutional neural network with the generative adversarial networks, a data expansion method based on the one-dimensional deep convolutional generative adversarial networks (1D-DCGAN) is constructed to generate small sample size fault samples and construct the balanced data set. Meanwhile, in order to solve the problem that the network is difficult to optimize, gradient penalty and Wasserstein distance are introduced. Through the test of bearing database and hydraulic pump, it shows that the one-dimensional convolution operation has strong feature extraction ability for vibration signals. The proposed method is very accurate for fault diagnosis of the two kinds of equipment, and high-quality expansion of the original data can be achieved.

Download Full-text

Management of the Torch Structure with the New Methodological Approaches to Regulation Based on Neural Network Algorithms

Energies ◽

10.3390/en14071909 ◽

2021 ◽

Vol 14 (7) ◽

pp. 1909

Author(s):

Konstantin Osintsev ◽

Sergei Aliukov ◽

Yuri Prikhodko

Keyword(s):

Neural Network ◽

Coal Dust ◽

Combustion Process ◽

Analytical Calculation ◽

Continuous Distribution ◽

Initial Section ◽

Distribution Functions ◽

Thermophysical Characteristics ◽

Network Algorithms ◽

Furnace Volume

A method for evaluating the thermophysical characteristics of the torch is developed. Mathematically the temperature at the end of the zone of active combustion based on continuous distribution functions of particles of solid fuels, in particular coal dust. The particles have different average sizes, which are usually grouped and expressed as a fraction of the total mass of the fuel. The authors suggest taking into account the sequential nature of the entry into the chemical reactions of combustion of particles of different masses. In addition, for the application of the developed methodology, it is necessary to divide the furnace volume into zones and sections. In particular, the initial section of the torch, the zone of intense burning and the zone of afterburning. In this case, taking into account all the thermophysical characteristics of the torch, it is possible to make a thermal balance of the zone of intense burning. Then determines the rate of expiration of the fuel-air mixture, the time of combustion of particles of different masses and the temperature at the end of the zone of intensive combustion. The temperature of the torch, the speed of flame propagation, and the degree of particle burnout must be controlled. The authors propose an algorithm for controlling the thermophysical properties of the torch based on neural network algorithms. The system collects data for a certain time, transmits the information to the server. The data is processed and a forecast is made using neural network algorithms regarding the combustion modes. This allows to increase the reliability and efficiency of the combustion process. The authors present experimental data and compare them with the data of the analytical calculation. In addition, data for certain modes are given, taking into account the system’s operation based on neural network algorithms.

Download Full-text