International Scientific Journals
of Scientific Technical Union of Mechanical Engineering "Industry 4.0"

In this study, the effect of production parameters of B4C + ZrO2 composites on density was modelled by using Artificial Neural Network (ANN). The composites were produced by using powder injection molding method (PIM). In the sintering stage, pressureless sintering method under argon atmosphere was used. As the production parameters, amount of additional (A, wt.%) and sintering temperature (T, ◦C) were defined. The main aim of the study is to obtain the experimental conditions giving maximum density. As a results of this study, the production parameters of hard sintered materials like B4C + ZrO2 could be modelled by using ANN method to optimize and predict because the prediction error is blow percentage of 10%. Therefore, the research and development time and cost can be reduced by using this method.

The aim of the article is to investigate the optical properties of Bulgarian honey in regard to the potential of honey discrimination on the base of its botanical origin. Samples from three types of honey (acacia, linden, and honeydew) are measured by a fluorescence spectrometer recording emission from 350 to 800 nm with excitation at 370, 395 and 405 nm. A combination of fluorescence emission spectra with some colorimetric parameters (CIELab) is used as input data of three types of honey classifiers: the first two are based on linear and quadratic discriminant analysis, and the third one uses an artificial neural network. The neural classifier is realized as a multilayered perceptron with backpropagation learning algorithm. Principal components analysis (PCA) is used for reducing the number of inputs and for a proper visualization of the experimental results. The comparative analysis of the three classifiers is based on leave-one-out-cross validation test carried out in MATLAB environment.

In this study boron carbide based sandblasting nozzles were produced by Low Pressure Powder Injection Molding (LPPIM) method, and wear behaviors of the nozzles were examined. The addition powder, addition ratio and sintering temperature were used as input parameters while density, micro hardness and wear rate were used as output parameters in the experimental design. This study consists of 3 steps: 1) production of standard samples and characterization, 2) modeling of proses parameters using Artificial Neural Network (ANN) method, 3) selection of nozzle material and production of nozzle, and testing. As a results of this study, ANN method can be used for modeling of process parameters of powder injection molding since the average value of the prediction error is below 7%, and boron carbide based products can be produced by using LPPIM method.