Repository of the University of Rijeka
Not a member yet
40085 research outputs found
Sort by
Reconstruction of Railway Transport System in the Service of the Port of Rijeka Container Terminal Development
No full textIn this article, authors emphasize the significance of the reconstruction of the railway transport system
for development of the port of Rijeka container terminal.
The most important projects related to Rijeka’s rail transport system are the multimodal platform
development and interconnection to the Adriatic Gate Container Terminal, and the future construction
of a new high-efficiency railway Rijeka – Zagreb.
Reconstruction of the “Brajdica” railway station and building extension of the railway infrastructure
has increased the loading/unloading capacity of the port Rijeka for container delivery by rail.
The aim of this article is to show how the reconstruction of the “Brajdica” railway station and terminal
within the port of Rijeka container terminal has improved the efficiency of the dispatch/delivery of
containers by railway.
Construction of a new high-efficiency railway Rijeka - Zagreb will create a significant shortening of
railway connections of Southeast Europe with its central and western parts and will improve conditions
of traffic exploitation, i.e. future development of the port of Rijeka container terminal
Thermal load monitoring of HV cable lines
No full textU radu se analizira važnost termićkog nadzora u elektroenergetskim sustavima, posebno u kontekstu povećane potrošnje električne energije i bržeg razvoja infrastrukture. Visokonaponski kabeli su osnovni elementi elektroenergetskih mreža, ali su podložni različitim uvjetima koji mogu utjecati na njihovu funkcionalnost i sigurnost. Prekomjerno zagrijavanje kabela može dovesti do oštećenja izolacije, smanjenja životnog vijeka kabela i povećanja rizika od električnih kvarova ili požara.
Rad detaljno opisuje osnovne karakteristike VN kabelskih vodova, termičke utjecaje na njih i metode termičkog nadzora. Poseban naglasak stavljen je na DTS i RTTR sustave, koji omogućavaju kontinuirano praćenje temperature i optimizaciju kapaciteta prijenosa energije. DTS sustavi koriste optička vlakna za praćenje temperature duž cijelog kabela, dok RTTR sustavi koriste računarske modele za izračunavanje trenutnog dopuštenog strujnog opterećenja temeljenog na stvarnim uvjetima.
Implementacija ovih sustava ima značajan utjecaj na ekonomsku i ekološku održivost elektroenergetskih mreža. Praćenjem temperature i optimizacijom rada kabela smanjuju se gubici energije, poboljšava energetska učinkovitost i produžava se životni vijek infrastruktureThe thesis analyzes the importance of thermal monitoring in electrical power systems, particularly in the context of increased electricity consumption and rapid infrastructure development. High-voltage cables are fundamental components of electrical power grids, but they are susceptible to various conditions that can affect their functionality and safety. Overheating of cables can lead to insulation damage, reduced cable lifespan, and increased risk of electrical faults or fires.
The thesis provides a detailed description of the basic characteristics of high-voltage cable lines, thermal influences on them, and methods of thermal monitoring. Special emphasis is placed on DTS (Distributed Temperature Sensing) and RTTR (Real-Time Thermal Rating) systems, which enable continuous temperature monitoring and optimization of energy transmission capacity. DTS systems use optical fibers to monitor temperature along the entire cable length, while RTTR systems use computer models to calculate the current allowed current load based on real-time conditions.
The implementation of these systems has a significant impact on the economic and environmental sustainability of electrical power grids. By monitoring temperature and optimizing cable operation, energy losses are reduced, energy efficiency is improved, and the lifespan of infrastructure is extended
Klinički potencijal ekstracelularnih vezikula u patologiji središnjeg živčanog sustava
No full textThe number of neurological disorders has been increasing for the last 30 years, and they are now among top 10 causes of mortality. As for neurodegenerative diseases in which Alzheimer’s and Parkinson’s disease predominate, there is a very small number of available and mostly symptomatic drugs. It is therefore a priority to find appropriate therapies, as well as biomarkers for early detection and monitoring of such disorders. Extracellular vesicles are formed in almost all cells, they transfer macromolecules like DNA, RNA and proteins to give an image of the tissue from which they originate. This makes them an excellent diagnostic as well as therapeutic tools. Therefore, there is a substantial goal in the need to find an adequate procedure for isolation of vesicles, in order to find a good biomarker from body fluids that does not require invasive methods.Broj neuroloških poremećaja raste zadnjih 30 godina te su oni sada unutar deset vodećih uzroka smrtnosti u svijetu. Što se tiče neurodegenerativnih bolesti u kojima prevladavaju Alzheimerova i Parkinsonova bolest, postoji vrlo mali broj dostupnih, i to većinom simptomatskih lijekova. Prioritetno je stoga pronaći odgovarajuće terapije, ali i biomarkere za rano otkrivanje i praćenje takvih poremećaja. Izvanstanične vezikule stvaraju se u gotovo svim stanicama, prenose makromolekule, DNA i RNA te daju sliku tkiva od kojeg potiču. To ih čini potencijalno dobrim dijagnostičkim, ali i terapeutskim alatom. Stoga se javlja potreba za pronalaženjem adekvatnog postupka izolacije vezikula kako bi se iz tjelesnih tekućina pronašao dobar biomarker bez korištenja invazivnih metoda
Application of sliding and rolling bearings in marine system
No full textLežajevi su ključni strojni elementi koji omogućuju relativno gibanje dijelova uz minimalno trenje, podupiru osovine i prenose opterećenja. Dijele se na klizne i valjne. Klizni ležajevi koriste klizni kontakt i pogodni su za sustave s visokim opterećenjima i tihim radom, dok valjni smanjuju trenje kotrljajućim elementima, omogućuju veću energetsku učinkovitost pri višim brzinama. Njihova primjena ovisi o radnoj brzini, vrsti opterećenja i zahtjevima održavanja. U pomorstvu su ključni za pouzdan rad brodskih sustava, a njihov kvar može uzrokovati ozbiljne probleme, uključujući povećano trenje, pregrijavanje i oštećenja mehaničkih sklopova, što dovodi do zastoja i povećanja troškova održavanja.Bearings are essential machine elements that enable relative motion between parts with minimal friction, support shafts, and transmit loads. They are classified as sliding or rolling bearings. Sliding bearings use direct surface contact and are ideal for high-load, low-noise applications, while rolling bearings reduce friction with rolling elements, improving energy efficiency at higher speeds. Their selection depends on operational speed, load type, and maintenance requirements. In maritime applications, bearing ensure reliable operation of ship systems, and their failure can lead to friction increase, overheating, and mechanical damage, causing downtime and high maintenance cost
The Connection Between Socioeconomic Factors and Dietary Habits of Children with Down Syndrome in Croatia
No full textChildren with Down syndrome often face significant feeding difficulties and health comorbidities that may contribute to undernutrition or obesity. This study assessed dietary habits and nutritional status among 104 children with Down syndrome in Croatia, representing 11.5% of this population. Results showed that over 30% of children aged 1 to 15 were overweight. Over 60% never consumed whole grain bread, while more than 50% avoided fish, nuts, or seeds. Despite rural families more frequently producing their own food (meat p = 0.009; fruits/vegetables p = 0.035), no significant improvement was observed in the children’s diets compared to their urban counterparts. Urban children consumed milk (p = 0.008) and fermented dairy (p = 0.005) more often. Children of university-educated mothers had higher vegetable (p = 0.031), meat (p = 0.025), olive oil (p = 0.003), and nut (p = 0.029) consumption, and a lower intake of processed meats (p = 0.008) and salty snacks (p = 0.040). Families spending less than 50% of income on food also showed significantly healthier dietary patterns. Feeding difficulties in children with Down syndrome are commonly associated with sensory sensitivities, oral-motor impairments, and comorbid medical conditions. These challenges are often intensified by parental anxiety, delayed introduction of diverse foods, and inadequate professional support. Collectively, these factors contribute to selective eating, poor nutrient intake, and disordered eating behaviors. This study underscores the need for individualized nutritional interventions that address the unique physiological and sensory requirements of both children and adults with Down syndrome. Effective strategies should extend beyond general dietary recommendations to include early exposure to a variety of food textures, specialized feeding support, and the management of coexisting health conditions. Family education and engagement play a crucial role in achieving positive nutritional outcomes. Empowering parents and caregivers—especially those in socioeconomically disadvantaged or rural communities—can facilitate the alignment of food accessibility with healthy dietary practices. The findings of this research offer valuable guidance for the development and implementation of national strategies aimed at enhancing the nutrition and long-term health of individuals with Down syndrome
BETWEEN REASON AND IMPULSE: PRICING STRATEGIES AND CROATIAN CONSUMERS
No full textThe high competitiveness of the retail environment forces manufacturers and vendors to constantly try to draw attention to their products. This is especially true in the current climate of global uncertainty, growing economic pessimism and new technologies influencing consumer behaviour. Consumers make numerous decisions every day that are influenced not only by their preferences but also by emotional and cognitive factors. Therefore, psychological pricing is becoming an increasingly important strategy. This paper examines how Croatian consumers perceive and react to various forms of psychological pricing, including prices with the final digit 9 (referred to as "charm prices"), the anchoring effect, the lure effect, and other methods. Their preferences for key product attributes, such as quality, price, brand design, and others, are also analysed. The research was conducted with a sample of 456 respondents using an online questionnaire, and the data were analysed using Mann-Whitney U and Kruskal-Wallis H tests. The results show that while most respondents describe themselves as rational and responsible shoppers who plan their purchases, they still tend to engage in impulse buying, especially in situations involving special offers. Men find prices ending in a 9 more attractive and cheaper than women, and are more likely to round them up to a lower number. Older age groups and less educated respondents show a greater sensitivity to psychological prices, while younger respondents place more value on product design and brand. In summary, although rational behaviour is the ideal of most consumers, emotional and heuristic influences have a significant impact on their decisions. The results can help companies to optimise their pricing strategies, taking into account demographic differences among customers, especially in the context of market changes and increasing pressure on household budgets
Modeliranje procesa deformiranja hiperelastičnih materijala pomoću fizikalno proširenih neuronskih mreža : doktorski rad
No full textMachine learning has gained an important place both in daily interactions and in scientific
development. It has a wide range of applications and many different approaches which can
be used to solve various problems. Neural networks are one of these approaches and are
a potent regression tool which is applied in a large number of scientific fields. In recent
years it has gained popularity in the field of computational mechanics as an alternative
material model replacing (or complementing) existing analytical models. These neural
network models are useful for describing complex relationships that existing models can
not fully describe. They are also used as a general model that can be applied to a wide
range of materials. This thesis places a focus on the latter, with the aim of creating a
general material model for hyperelastic materials such as rubber. With analytical models
there is a need to fit different models (Neo-Hookean, Mooney-Rivlin, Ogden, Yeoh, Gent,
etc.) to experimental data and use the one that approximates the data best. Each analytical model has some advantages compared to the others as well as drawbacks, and the
choice of the model also depends on what the application of the model will be. By using
a neural network as a model the goal is to obtain a single model that has no drawbacks
and can function for any material. This also speeds up the material modelling process as
only one single model is generated instead of fitting multiple models and choosing between them. In this thesis a detailed investigation is made between different modelling
strategies. Firstly, a neural network that predicts stress from strain is developed and is a
representative of the most common approach in neural network modelling. Afterwards, a
more general model using the same neural network architecture is created that is similar
to conventional invariant models so that it predicts the energy while taking the invariants
of the right Cauchy-Green deformation tensor as inputs. It was demonstrated that it can be
used as a direct replacement for the invariant models by reusing the existing framework
for the numerical implementation of invariant based models. Secondly, a significant improvement to the neural network architecture is done through the implementation of a
custom activation function suited for modelling of hyperelastic behaviour. It is accompanied by the implementation of certain conditions from solid mechanics such as objectivity,
thermodynamic consistency, normalisation of energy, non-negativity of energy, normalisation of stress and polyconvexity. All of these improvements have created a neural
network model, referred to as LINEXP-PANN, that can capture a wider range of material behaviours with a significantly reduced dataset size compared to the previous simpler
neural network models. The LINEXP-PANN model was used for modelling damage in
rubber-like materials which is known as the Mullins effect. A new modelling strategy was
developed for modelling the Mullins effect using neural networks where certain weights
are reused and shared within the neural network. The majority of the work is based on
the assumption of incompressibility, a common assumption when modelling rubber-like
materials. The LINEXP-PANN model was also extended to compressible behaviour confirming the general modelling capabilities of the model. Finally, a comparison is made
between the LINEXP-PANN model and another data-driven method called Data-Driven
Computational Mechanics (DDCM) where the advantages of the LINEXP-PANN model
are demonstrated.Strojno učenje je postalo dio naše svakodnevice i dio znanstvenoga razvoja. Ima široko
područje primjene i postoje mnogi različiti pristupi koje se može primijeniti za rješavanje
problema. Neuronske mreže su jedan od tih pristupa i moćni su regresijski alat koji se
primjenjuje u mnoštvu znanstvenih polja. Sa porastom primjene strojnog učenja u inženjerstvu, neuronske mreže postaju sve popularniji alternativni materijalni model koji
zamjenjuje (ili nadopunja) postojeće analitičke modele. Materijalni modeli temeljeni na
neuronskim mrežama su korisni za opisivanje složenih ponašanja koje postojeći analitički
modeli ne mogu u potpunosti opisati. Također se koriste i kao opći modeli koji mogu
opisati širi raspon materijala. U ovom doktorskome radu se pozornost stavlja na drugo
svojstvo primjene gdje se želi stvoriti opći materijalni model za hiperelastične materijale sa naglaskom na materijale slične gume. Kada se koriste analitički modeli potrebno
je prilagoditi više modela (Neo-Hookean, Mooney-Rivlin, Ogden, Yeoh, Gent, itd.) na
eksperimentalne podatke i odabrati onaj model koji najbolje opisuje podatke. Svaki analitički model ima neke prednosti u odnosu a druge modele, no i nedostatke u odnosu na njih,
te se mora odabrati model ovisno o tome koji najviše odgovara određenoj primjeni. Koristeći neuronsku mrežu kao model cilj je dobiti jedinstveni model koji nema nedostataka i
može se koristiti za bilo koji materijal. Ovo također ubrzava proces materijalnoga modeliranja jer je potrebno prilagoditi samo jedan model umjesto više njih između kojih se mora
birati. Prvi korak u radu je bio razvoj neuronske mreže koja predviđa naprezanje iz deformacije i predstavnik je uobičajenoga načina modeliranja neuronskim mrežama. Potom
je stvoren općenitiji model temeljen na istoj arhitekturi neuronske mreže koji je sličan invarijantnim modelima jer predviđa energiju i koristi invarijante desnog Cauchy-Greenova
tenzora deformiranja kao ulazne podatke. Pokazalo se da ovaj model može biti korišten i
kao jednostavna zamjena za klasične invarijantne modele jer koristi već postojeće rutine za
računarsku implementaciju invarijantnih modela. Drugi korak čine unaprijeđenja samoj
arhitekturi neuronske mreže od kojih je najznačajnija implementacija vlastite aktivacijske
funkcije primjerene modeliranju hiperelastičnoga ponašanja. Korištenje vlastite aktivacijske funkcije omogućava ispunjenje određenih uvjeta iz mehanike čvrstoga tijela poput
objektivnosti, termodinamičke konzistencije, normalizacije energije, ne-negativnost energije, normalizacija naprezanja i polikonveksnost. Promjene u arhitekturi i ispunjenje
ovih uvjeta su vodile stvaranju modela temeljenog na neuronskoj mreži kojem se nadjenulo ime LINEXP-PANN (Linear EXPonential Physics-Augmented Neural Network,
hrv. linearno eksponencijalna fizikalno proširena neuronska mreža). Ovaj model može
opisati široki raspon materijalnih ponašanja koristeći znatno manju količinu podataka za
treniranje neuronske mreže u usporedbi sa jednostavnijim modelima predstavljenim u prvome koraku. Prethodno spomenuti LINEXP-PANN model je korišten i za u modeliranju oštećenja kod gumenih materijala koji je poznat kao Mullinsov efekt. Razvijen je
novi pristup modeliranju Mullinsova efekta neuronskim mrežama koji je specifičan jer su
određene težine dijeljene u neuronskoj mreži među njenim dijelovima. Većina ovoga rada
je temeljena na pretpostavci nestlačivosti koja je česta u kontekstu modeliranja gumenih
materijala. Novorazvijeni LINEXP-PANN model je proširen i za modeliranje stlačivog
hiperelastičnog ponašanja čime se dodatno potvrđuju prilagodljivost i općenitost modela.
Naposljetku je napravljena usporedba između LINEXP-PANN modela i jedne alternative
metode temeljene na podacima Data-Driven Computational Mechanics (DDCM, hrv. podacima pogonjena računalna mehanika) te su prikazane prednosti LINEXP-PANN modela
The Association Between Women's Perception of Birth During the Pandemic, Companion of Choice and Support From Health Professionals: A Cross‐Sectional Study in 20 Countries in the WHO European Region
No full textBackground: Mitigation measures implemented in response to the COVID-19 pandemic led to significant changes in maternity care across Europe, including restrictions on companions during labor and birth. This cross-sectional study explores the association between the presence of a companion of choice and a positive perception of the birth experience. Additionally, it explores the association between health professionals' attention, assistance, and availability during labor and birth and a positive perception of birth. Methods: We utilized a structured, validated online questionnaire, available in 25 languages, to assess the quality of maternal care during the COVID-19 pandemic from women's perspectives. We conducted logistic regression to explore associations between variables related to the presence of a companion of choice, health professionals' attention, assistance, and availability, and positive perceptions of birth, when controlled for confounders, including birth mode and medical interventions. Results: Responses from 48,039 women across 20 countries in the WHO European Region were included. Always having a companion of choice during birth (aOR: 2.11) and always receiving adequate care from health professionals (assistance aOR: 2.12, attention aOR: 36.64, availability aOR: 2.12) were associated with positive birth perception. Instrumental births (aOR: 0.76), episiotomies (aOR: 0.74), fundal pressure (aOR: 0.52), and cesarean births (planned aOR: 0.80, unplanned prelabor aOR: 0.60, unplanned in-labor aOR: 0.52) were associated with less positive birth perceptions. Discussion: This study highlights the critical role of having a chosen companion and receiving adequate attention, assistance, and availability from health professionals in promoting positive birth perceptions, even in times of crisis such as the COVID-19 pandemic. Ensuring the presence of a companion of choice and comprehensive professional support is crucial for delivering high-quality, respectful maternity car
Association between Dental Variables and Hashimoto's Disease: A Retrospective Cohort Study
No full textAbstract Objectives The present study aimed to compare dental, endodontic, and periodontal status in patients with Hashimoto's disease and healthy patients, as well as to disclose the relation between dental variables and Hashimoto's disease. Materials and Methods The research included 85 patients affected by Hashimoto's thyroiditis (analyzed group) and 85 healthy patients (control group). The two groups were matched according to age and gender. Data regarding patients' health status was acquired from their medical records. Data regarding dental, endodontic, and periodontal status were acquired from patients' dental records and digital panoramic radiographs. Statistical Analysis Since a nonnormal distribution of data was detected, a median and interquartile range were used as a measure of central tendency and dispersion. Mann–Whitney U test and chi-squared test were used to test the differences between the groups. Multiple linear regression analysis and logistic regression analysis were used to test the association of the data. Results Healthy participants had significantly higher median number of teeth with secondary caries (median 2; interquartile range 1–3) when compared with participants with Hashimoto's disease (median 1; interquartile range 0–2; p
ASSESSMENT OF 10-YEAR CARDIOVASCULAR RISK IN THE ADULT POPULATION IN THE LIVNO AREA, BOSNIA AND HERZEGOVINA
No full textKardiovaskularne bolesti (KVB) vodeći su uzrok smrtnosti u svijetu, a prevalencija ovih bolesti raste, posebno u zemljama s niskim i srednjim prihodima. Procjena kardiovaskularnog rizika (KVR) ključna je za pravovremeno prepoznavanje rizika i intervenciju. Europsko kardiološko društvo preporučuje korištenje SCORE tablice za procjenu desetogodišnjeg KVR. Cilj istraživanja bio je procijeniti KVR u odrasloj populaciji Livna, Bosna i Hercegovina, korištenjem SCORE tablice. Istraživanje je obuhvatilo 50 ispitanika starijih od 40 godina, a rezultati su pokazali da muškarci imaju značajno viši prosječni KVR od žena (10,88% vs 7,06%). Također, stariji ispitanici (61-70 godina) imali su najviši KVR, dok mlađe skupine imaju niže vrijednosti. Obrazovna struktura nije imala značajan utjecaj na KVR. Zaključak naglašava potrebu za javnozdravstvenim preventivnim mjerama u Livnu.Cardiovascular diseases (CVD) are the leading cause of mortality worldwide, and the prevalence of these diseases is increasing, particularly in low- and middle-income countries. Cardiovascular risk (CVR) assessment is crucial for the timely identification of individuals at risk and intervention. The European Society of Cardiology recommends using the SCORE table to assess ten-year CVR. The aim of the study was to assess the CVR in the adult population of Livno, Bosnia and Herzegovina, using the SCORE table. The study included 50 participants aged over 40, and the results showed that men have significantly higher average CVR compared to women (10.88% vs. 7.06%). Furthermore, older participants (61-70 years) had the highest CVR, while younger age groups had lower values. Educational background did not significantly affect CVR. The conclusion highlights the need for public health preventive measures in Livno