In cooperation with the Slovenian factory of electromotors and household devices, Domel d.d. from Železniki, we optimized the lamination of the universal motor for vacuum cleaners by genetic algorithm. We implemented a new procedure of designing rotor and stator of the universal motor for home appliances with the evolutionary approach. In order to find the optimal values of the independent geometric parameters of the rotor and the stator lamination a genetic algorithm was employed with the aim to reduce the main motor's power losses and to improve the efficiency of the motor. The geometry optimization gave about 30% lower power losses and about 5% higher efficiency. The optimization process took a few hours, which is much faster than the former practice performed manually by a skilled engineer. Using this procedure, the motor's technical quality, expressed as efficiency, was significantly improved. The genetic algorithm proved to be a simple and efficient search and optimization method for solving the common design problem in the industry. The results were reported in papers in journals and on international conferences. In collaboration with Engineering Ceramics department we investigated the rheological properties of ceramic suspensions in the frame of an application project for AET Tolmin. The main activities were targeted in revealing the factors affecting dynamic viscosity of ceramic suspensions. The relationship between these factors and the resulting viscosity is highly non-linear. Therefore, we decided to model it with artificial neural networks (ANN). The experimental plan was made to ensure the experiments would properly cover the space of input variables. Experiments were carried out and the software for simulation of multilayered perceptron with backpropagation learning rule was written. Together with Reactor Engineering department, we also worked on a simulation of a low-pressure injection moulding of ceramic suspensions. Special concern was devoted to solidification process, which is a key point for successful simulation of this type. In analog design, we developed a frequency dependant temperature sensor structure realized in a single serial loop. We also developed a method for temperature measurement in a system compliant with IEEE 1149.4 standard where 1149.4 infrastructure is used for processing temperature sensor data. Both solutions have been patented.