Siniša Fajt

This paper presents a Vibration-based Track Anomaly Detection (VTAD) system designed for real-time monitoring of urban tram infrastructure. The novelty of VTAD is that it converts existing public transport vehicles into distributed mobile sensor platforms, eliminating the need for specialized diagnostic trains. The system integrates low-cost micro-electro-mechanical system (MEMS) accelerometers, Global Positioning System (GPS) modules, and Espressif 32-bit microcontrollers (ESP32) with wireless data transmission via Message Queuing Telemetry Transport (MQTT), enabling scalable and continuous condition monitoring. A stringent ±6σ statistical threshold was applied to vertical vibration signals, minimizing false alarms while preserving sensitivity to critical faults. Field tests conducted on multiple tram routes in Zagreb, Croatia, confirmed that the VTAD system can reliably detect and locate anomalies with meter-level accuracy, validated by repeated measurements. These results show that VTAD provides a cost-effective, scalable, and operationally validated predictive maintenance solution that supports integration into intelligent transportation systems and smart city infrastructure.

Audio systems with unbalanced connections are susceptible to interference from ground loops, which manifests as hum and noise. This paper introduces and evaluates a novel passive Audio Interference Suppressor in Analog Audio Interface (AISAAI) designed to mitigate this problem. The AISAAI circuit is inserted between an audio device’s rectifier ground and its protective earth terminal, creating an optimized impedance path that reduces interference while ensuring safety. This approach is analyzed within a proposed Analog Audio Interconnection System (AAIS) framework. Experimental results show that common-mode voltages from protective earth potential differences are the primary source of interference. The optimized AISAAI suppressor achieves a consistent 15–30 dB reduction in measured audio interference across the audio band, regardless of the interconnect cable characteristics. This study confirms AISAAI as an effective solution for reducing ground loop noise in unbalanced audio systems and underlines the usefulness of the AAIS model for systemic analysis.

High frequency (HF) welding of polymer materials is increasingly used in modern manufacturing processes. The literature on HF welding process parameters was reviewed and it was found that 3–5 basic welding parameters were considered, which is insufficient for the scientific study of HF welding of polymeric materials. This article presents the mathematical expressions for the evaluation combining 17 influential parameters. For the first time, the specific and latent heat of the welded polymer material were used. The breaking forces of welds made by RF welding are investigated by varying the anode current, the coupling capacitor, and the exposure time of the HF electromagnetic fields. It was found that the amount of HF energy supplied depends on the breaking forces of the weld. A characteristic inflection point was also observed in the graph of the dependence of the breaking forces on the percentage of the coupling capacitor and the anode current. During elongation, it was observed that the weld is separated by peeling before the inflection point and breaks after the inflection point by tearing at the extruded edges of the weld. If the HF energy is applied to the weld for too long, there will be excessive melting of the material in the weld, thinning of the weld, unfavourable appearance of the extruded edges and electrical breakdown, and a drastic drop in the breaking force.

So far, scientific studies on high frequency welding (HF) have observed the three main parameters: anode current, welding time and pressure of upper electrode. Some authors have also observed and studied the welding frequency, the power, and the values of the dielectric constants of the materials or the angle of dielectric losses. Due to the complex effects that the parameters have on the quality of the welded joint, it is necessary to extend the research and study the mutual relationships between a large numbers of process parameters on the welded joint in order to achieve satisfactory strength of the welded joint. In this research, the number of considered parameters is increased to 17, each of which is relevant to the processes of HF welding, and their mathematical equations are presented. The research is carried out on a HF welding machine improved with a linear pneumatic actuator for lowering the upper electrode and regulating the compressive force, and a measuring machine with the appropriate measuring method for determining the breaking forces with a tension gage and performing the „T“ test is made. Numerous researches were carried out on the functional changes of the process parameters of the percentage share of the values of the coupling capacitor, the anode current and the welding time on the values of the breaking forces of the welded joints. Based on the conducted research, it was concluded that the optimization of the process parameters of HF machine must be performed before the welding process, which is determined by a certain amount of energy that must be introduced into the material. These parameters are the capacity of the output capacitor, i.e., the value of the electric current flowing through the material and the time for the development of the operating power to produce welded joints with a certain strength and quality of welded joints for a specific material used in the process.

The paper presents a new method for determining the machine-hand welding times of synthetic polymer materials using ultrasonic welding machines with rotary sonotrode. The method is based on the claims and observations of W. Möller intended for the clothing industry in the 1990s, according to which there is a spontaneous drop in sewing speed when strongly curved seams are joined, which is due to the possibility of human reactions. The method for determining machine-hand sewing times of curved seams was well accepted in garment production processes. It is used to standardize production time. Using ultrasonic welding machines with rotary sonotrode, the problem of determining the time of ultrasonic joining of curved seams on clothing or technical textiles remained unsolved. That is why is completely new and original model was created, which combines eight technical parameters of ultrasonic welds, eight technological parameters of the production process and seven ergonomic parameters depending on the psychophysical conditions of the workers. The systematic development of the mathematical relationship of all 23 parameters mentioned and the corresponding mathematical expressions for determining and calculating these parameters are presented. These results were also verified by experimental measurements, which show a favourable correlation between the calculated and measured machine hand times. The relationships between the recommended welding speeds, the critical radii of curvature and the number of reactions required according to Möller and the new model are also presented. An analysis of the success in the application of Möller’s and the new model was also performed.

Brojne provedene analize ne daju jedinstveni odgovor na pitanje ocjene akustičke kvalitete tonskih režija, budući da nisu uvijek uzete u obzir sve karakteristike specifične za tonske režije. U procesu opisanom u ovom radu sve su...

Modelling voice tract using an acoustical–mechanical analogy is shown in this article. The process of modelling voice tract is often reduced to one-dimensional acoustical model, i.e. planar waveguide model. With this acoustical-mechanical model, computer programs are utilized to analyze and synthesize the voice tract. Hence, it is possible to calculate models which describe the voice tract with satisfactory quality and detail level and these models can be used in practice and applied to particular cases. In this article the process of producing a voice tract model is presented. The process includes the usage of three different computer programs, each of them dedicated to one segment of modelling. Recordings for three speakers were made and the analysis of voice parameters was performed. After the analysis is completed, a two-dimensional model of the voice tract was made, which resulted in a mechanical model consisting of specific number of tubes. Finally, the evaluation procedure for the model is done by applying a software developed specifically for this purpose, in order to determine whether the resulting model truly represents the voice tract of each speaker.