Based on the results of experimental studies concerning the filtration coefficient, the Darcianity of the observed flows for eight cohesive soils at four hydraulic gradients was analyzed. It is observed that linear dependence of flow velocity on hydraulic gradient is an approximation only, and it is the worse the more cohesive a given soil is. Despite this, Darcy’s law can be a correct approximation of the empirical relationship between hydraulic gradient and the flow velocity, also in very cohesive soils. A statistical analysis was carried out to identify correlation between soil properties and permeability coefficient. For each soil, 109 parameters were analyzed, among others applying mercury intrusion porosimetry, scanning electron microscopy, dynamic image analysis, and laser diffraction. Ultimately, three single-variable models best fitted to the experimental data were found, using the plasticity index I_P as the independent variable, the average pore diameter D_P, and the convexity of silt fraction particles. All model parameters are statistically significant at p < 0.05. Comparison with reference multi-variable models showed that the best fit for experimental data is observed by the model with the plasticity index, while the results suggest low usability of single-variable models with structural parameters.

Electrical cables are a potential source of ignition and fire hazards in various types of buildings and industrial installations, as well as in all modes of transportation, including aircraft. Fires in buildings pose the greatest threat to human life and health. The composition of thermal degradation products depends mainly on the type of combustible materials and the type of combustion process—flaming or smoldering. The purpose of this paper was to determine, based on experimental studies, the effects of flaming and smoldering combustion on the response times of fire smoke detectors. In addition, the concentrations of fire gases formed in the process of duct combustion, including CO, SO₂, NO₂, NO, HCN, HCl, HBr and HF, were measured using an FTIR spectrometer. The results presented confirm the significant effect of the way the cable samples burned on detector tripping time. The highest concentration of smoke (Y) in the test chamber was obtained during flame combustion. It was further found that the characteristics of the cable insulation material used, such as the type of PVC, had a significant effect on the toxicity of the emitted gases. The largest amounts of toxic gases were emitted during the smoldering combustion of a cable with a plasticized PVC sheath.

Photovoltaic modules should pass series of tests and examinations, verifying the electrical and thermal characteristics of the module, before being released to the market, examining the properties of the modules should undergo appropriate durability and efficiency tests under real outdoor conditions with exposure to climatic conditions and under laboratory conditions. The performance of photovoltaic cells depends on many factors, such as solar irradiance, module operating temperature, installation location, weather conditions and module shading. In this paper, three selected photovoltaic modules were examined using a large-scale steady-state solar simulator. The current-voltage (I-V) characteristics of the photovoltaic modules were experimentally tested and analyzed. The next step was to implement a three-layer artificial neural network model (MLP). The experimental data obtained in the previous step coupled with output data from MLP were then used in global sensitivity analysis (GSA). Experiments carried out on a large-scale stationary solar simulator showed differences between the values declared by the manufacturer and the values obtained from measurements of PV modules. The first module tested achieved the maximum power point greater than that specified by the manufacturer, while the other two showed power drops; it was 85-87% for the second module, and 95-98% for the third, respectively. The performed global sensitivity analysis (GSA) for the MLP model showed that the parameters: eff (22.9) and Voc/V (14.19) have the largest effect on the power-voltage relationship, while U (7.29) has the smallest effect. The usefulness of machine learning (ML) methods in the comparative analysis of PV modules has been proved.