2023 - Repozytorium publikacji - Politechnika Gdańska

Pokaż publikacje z roku

Publikacje z roku 2023

Pokaż wszystkie

Practical Approach to IP Scheduled Throughput Measurements in Dual Connectivity Systems
- Arkadiusz Zięba
- Martin Kollar
- Krzysztof Tatarczyk
- Jarosław Sadowski
2023 Pełny tekst International Journal of Electronics and Telecommunications
IP scheduled throughput defined according to 3GPP TS 36.314 reflects user throughput regardless of traffic characteristics, and therefore has become one of the most important indicators for monitoring Quality of Service (QoS) of the end user in Evolved Universal Terrestrial Radio Access Network (E-UTRAN). However, networks built on a distributed architecture make the above definition impossible to be applied directly due to the implementation challenges. This paper gives an overview of the classical Long Term Evolution (LTE) architecture as opposed to Dual Connectivity (DC) topology and focuses on a novel method of solving the calculation issue with the IP scheduled throughput measurement in edge computing environment. Experimental results show a good agreement with the real end user perception.
Practical aspects of testing superconducting electromagnets by the capacitor discharge method taking into account the non-linearity of circuit parameters
- Michał Michna
- Andrzej Wilk
- Marek Wołoszyk
- Michał Ziółko
- Stanisław Galla
- Piotr Szwangruber
2023 Pełny tekst Archives of Electrical Engineering
The article presents selected issues related to the development and testing of the diagnostics systems dedicated for superconducting electromagnets. The systems were constructed to assess the production quality of superconducting electromagnets of the SIS100 synchrotron, a new accelerator being built as part of the Facility of Antiproton and Ion Research (FAIR). One of the systems is used for automatic checking of electrical connection parameters and the continuity of electric circuits. The role of the second device is to assess the quality of winding insulation and to estimate circuit parameters of electromagnet coils using the capacitor discharge method. The work presents measurements and analysis of current and voltage waveforms acquired during discharges on a magnet coil simulator and on the SIS100 main dipole electromagnet
Praktyczne aspekty fotowoltaiki
- Ewa Klugmann-Radziemska
2023
Wydajność instalacji fotowoltaicznej zależy od wielu czynników. Niektóre z nich są niezależne od właściciela modułów, inne pozwalają na podejmowanie świadomych działań, by lepiej wykorzystać potencjał instalacji i zwiększyć ilość generowanej energii, a tym samym poprawić efekt ekologiczny i ekonomiczny. Ograniczenie każdego z tych czynników ma istotne znaczenie ekonomiczne ze względu na możliwość uzyskania – przez zwiększenie sprawności konwersji fotowoltaicznej – większej mocy z tej samej powierzchni czynnej ogniw przy tym samym natężeniu promieniowania padającego. Wyżej wymienione czynniki powinny być analizowane na etapie wyboru modułów przez inwestora. Z kolei decyzja o miejscu i sposobie zainstalowania modułów i ich usytuowaniu jest podejmowana na etapie budowania instalacji i jest kluczowa dla uzyskania jak najbardziej korzystnych efektów jej pracy.
Pre-analytical aspects in metabolomics of human biofluids – sample collection, handling, transport, and storage
- Dorota Garwolińska
- Agata Kot-Wasik
- Weronika Hewelt-Belka
2023 Molecular Omics
Metabolomics is the field of omics research that offers valuable insights into the complex composition of biological samples. It has found wide application in clinical diagnostics, disease investigation, therapy prediction, monitoring of treatment efficiency, drug discovery, or in-depth analysis of sample composition. A suitable study design constitutes the fundamental requirements to ensure robust and reliable results from the study data. The study design process should include a careful selection of conditions for each experimental step, from sample collection to data analysis. The pre-analytical variability that can introduce bias to the subsequent analytical process, decrease the outcome reliability, and confuse the final results of the metabolomics research, should also be considered. Herein, we provide key information regarding the pre-analytical variables affecting the metabolomics studies of biological fluids that are the most desirable type of biological samples. Our work offers a practical review that can serve and guide metabolomics pre-analytical design. It indicates pre-analytical factors, which can introduce artificial data variation and should be identified and understood during experimental design (through literature overview or analytical experiments).
Prebiotics and probiotics in food
- Edyta Malinowska-Pańczyk
2023
Nowadays, food is not only used to satisfy hunger and is a source of nutrients, but is also considered a factor that directly affects human health. Consumers are looking for high-quality products that contain bioactive ingredients that affect the proper functioning of the body and good mood. Functional foods, which must exhibit health benefits when consumed as part of a balanced diet, are becoming increasingly popular among health-conscious individuals. Probiotics and prebiotics may be functional components of such foods.
Predicting emotion from color present in images and video excerpts by machine learning
- Aleksandra Wędołowska
- Dawid Weber
- Bożena Kostek
2023 Pełny tekst IEEE Access
This work aims at predicting emotion based on the colors present in images and video excerpts using a machine-learning approach. The purpose of this paper is threefold: (a) to develop a machine-learning algorithm that classifies emotions based on the color present in an image, (b) to select the best-performing algorithm from the first phase and apply it to film excerpt emotion analysis based on colors, (c) to design an online survey to check the accuracy of the annotations of the collected film data. In the first, three approaches to color extraction are tested, namely clustering colors into a palette of predefined colors, assigning colors to the RYB (Red, Yellow, Blue) model, and extracting a histogram of colors present in an image. This is based on image datasets containing color and emotion annotations. Classification is conducted using several algorithms, both deep learning and baseline artificial intelligence algorithms. The obtained results, under different configurations of parameters and training sets, are then presented. In the second part, the bestperforming algorithm from the first phase is applied to film excerpt emotion analysis based on colors. This is followed by the third part, which is an online survey created to check the accuracy of the algorithm’s annotations to the collected film data by the questionnaire respondents. Further, a discussion of the results achieved is presented. Conclusions contain a summary of the results and further direction for improving the performance of the created algorithm.
Predicting seismic response of SMRFs founded on different soil types using machine learning techniques
- Farzin Kazemi
- Neda Asgarkhani
- Robert Jankowski
2023 ENGINEERING STRUCTURES
Predicting the Maximum Interstory Drift Ratio (M-IDR) of Steel Moment-Resisting Frames (SMRFs) is a useful tool for designers to approximately evaluate the vulnerability of SMRFs. This study aims to explore supervised Machine Learning (ML) algorithms to build a surrogate prediction model for SMRFs to reduce the need for complex modeling. For this purpose, twenty well-known ML algorithms implemented in Python software are trained based on the dataset achieved from Incremental Dynamic Analyses (IDAs) performed on the 2-Story, 3- Story, 4-Story, 5-Story, 6-Story, 7-Story, 8-Story, and 9-Story SMRFs modeled in Opensees. Then, important features of weight, fundamental period of structure (T1), the RSN number of record subsets, the direction of the record subsets, soil types, and Sa(T1) of analysis, which affect the results of predictions, were selected by trial and error. Having these important features, data-driven techniques developed in Python software were compared for predicting the M-IDR of SMRFs as target. Results showed that ML algorithms of GPReg, PLSReg, SReg, LReg, GReg, MLPReg, SVM, and LSVR had lower values of coefficient of determination (R2 lower than 0.5) in both train and test datasets. In addition, XGBoost, BReg, HistGBR, and ERTReg algorithms achieved higher values of R2 (i.e. upper than 0.95 in the 5-Story SMRF) with low Mean Squared Error (MSE) for prediction of M-IDR. Therefore, using these algorithms mitigates the need for computationally expensive, time-consuming, and complex analysis, while preliminary prediction of M-IDR can be considered a low computational and efficient tool for seismic vulnerability assessment.
Prediction of maximum tensile stress in plain-weave composite laminates with interacting holes via stacked machine learning algorithms: A comparative study
- Faramarz Bagherzadeh
- Torkan Shafighfard
- Raja Muhammad Awais Khan
- Piotr Szczuko
- Magdalena Mieloszyk
2023 MECHANICAL SYSTEMS AND SIGNAL PROCESSING
Plain weave composite is a long-lasting type of fabric composite that is stable enough when being handled. Open-hole composites have been widely used in industry, though they have weak structural performance and complex design processes. An extensive number of material/geometry parameters have been utilized for designing these composites, thereby an efficient computational tool is essential for that purpose. Different Machine Learning (ML) tools were integrated to obtain the model with the highest accuracy considering the maximum tensile stress of composite plates with two interacting notches while comparing the effectiveness of each technique. Finite Element (FE) simulations were carried out inside the ABAQUS software by employing python macro code to provide a data-rich framework (8960 data). The predictions given by ML methods were compared with the data given by the numerical simulations. An evolutionary algorithm (TPOT) and automatic neural network search (AuoKeras) were utilized for that purpose. An automatic grid search technique was employed to select the best method which could predict the material attribute target values (maximum stress) for different tests. 1% of the data was given as training while 99% was for testing to ensure the robustness of the model. It was concluded that the model containing the Gradient Boosting Regression (GBR), PolyFeatures, and LassoLarsCV algorithms outperformed other ML combinations and Artificial Neural Networks (ANN) for predicting the target value. The coefficient of determination (2 ) and root mean square error (RMSE) of the proposed model were 0.97 and 253 respectively. Hence, this model could be utilized for prospective predictions in this type of materials and geometry by providing further reduction of the computational time and labor cost with high accuracy.
Pre‐exascale HPC approaches for molecular dynamics simulations. Covid‐19 research: A use case
- Miłosz Wieczór
- Vito Genna
- Juan Aranda
- Rosa M. Badia
- Josep Lluís Gelpí
- Vytautas Gapsys
- Bert L. de Groot
- Erik Lindahl
- Martí Municoy
- Adam Hospital
- Modesto Orozco
2023 Pełny tekst Wiley Interdisciplinary Reviews-Computational Molecular Science
Exascale computing has been a dream for ages and is close to becoming a reality that will impact how molecular simulations are being performed, as well as the quantity and quality of the information derived for them. We review how the biomolecular simulations field is anticipating these new architectures, making emphasis on recent work from groups in the BioExcel Center of Excellence for High Performance Computing. We exemplified the power of these simulation strategies with the work done by the HPC simulation community to fight Covid-19 pandemics.
Preparation and Characterization of Diamond-like Carbon Coatings for Biomedical Applications—A Review
- Klaudia Malisz
- Beata Świeczko-Żurek
- Alina Sionkowska
2023 Pełny tekst Materials
Diamond-like carbon (DLC) films are generally used in biomedical applications, mainly because of their tribological and chemical properties that prevent the release of substrate ions, extend the life cycle of the material, and promote cell growth. The unique properties of the coating depend on the ratio of the sp3/sp2 phases, where the sp2 phase provides coatings with a low coefficient of friction and good electrical conductivity, while the share of the sp3 phase determines the chemical inertness, high hardness, and resistance to tribological wear. DLC coatings are characterized by high hardness, low coefficient of friction, high corrosion resistance, and biocompatibility. These properties make them attractive as potential wear-resistant coatings in many compelling applications, including optical, mechanical, microelectronic, and biomedical applications. Another great advantage of DLC coatings is that they can be deposited at low temperatures on a variety of substrates and can thus be used to coat heat-sensitive materials, such as polymers. Coating deposition techniques are constantly being improved; techniques based on vacuum environment reactions are mainly used, such as physical vapor deposition (PVD) and chemical vapor deposition (CVD). This review summarizes the current knowledge and research regarding diamond-like carbon coatings.
Preparation, characterization, and manufacturing of new polymeric materials for 3D printing for medical applications
- Agnieszka Haryńska
2023 Pełny tekst
This work concerns the synthesis, formation, and characteristics of new filaments for 3D printing in FDM™/FFF technology for medical purposes. Two types of filaments were developed, i.e. degradable polyurethane and biodegradable polylactide-starch. The influence of the 3D printing process on selected filament properties was investigated. A detailed analysis of the filament formation process by the extrusion method was carried out, thus complementing the current state of knowledge. Porous structures and anatomical models were designed and 3D printed using, among others, developed filaments. The obtained details were subjected to a series of preliminary biological (in vitro) tests to determine their suitability for medical applications. The research results showed that the developed polyurethane filaments are biocompatible and susceptible to degradation, and the forming process does not affect their structural, thermal, and biological properties. Whereas, modification of polylactide with the addition of thermoplastic starch increased the hydrophilicity and susceptibility to hydrolytic degradation of the developed bio-filament. It has been shown that 3D printed polyurethane porous structures meet the prerequisites of cancellous bone tissue scaffolds. In turn, personalized anatomical models printed with the use of the developed bio-filament are characterized by increased compostability while maintaining the quality of printouts from commercial polylactide (PLA) filament. The basis of the presented doctoral dissertation is a series of six scientific articles published in journals indexed in the JCR list.
Previous Opinions is All You Need - Legal Information Retrieval System
- Maciej Osowski
- Katarzyna Lorenc
- Paweł Drozda
- Rafał Scherer
- Konrad Szałapak
- Kajetan Komar-Komarowski
- Julian Szymański
- Andrzej Sobecki
2023
We present a system for retrieving the most relevant legal opinions to a given legal case or question. To this end, we checked several state-of-the-art neural language models. As a training and testing data, we use tens of thousands of legal cases as question-opinion pairs. Text data has been subjected to advanced pre-processing adapted to the specifics of the legal domain. We empirically chose the BERT-based HerBERT model to perform the best in the considered scenario.
Price bubbles in commodity market – A single time series and panel data analysis
- Marcin Potrykus
2023 Pełny tekst QUARTERLY REVIEW OF ECONOMICS AND FINANCE
This paper examines thirty-five commodities, grouped into three market sectors (energy, metals, agriculture & livestock) in terms of the occurrence of price bubbles. The study was based on monthly data for each commodity separately and, in a panel approach, for selected sectors and for all commodities combined. The GSADF test and its version for panel data – panel GSADF – were used to identify bubbles. The beginning and end of the detected price bubbles were also determined. No price bubbles were found for commodities such as Bananas, Cocoa or Orange juice, while tin, tobacco and gold were identified as the commodities most prone to bubbles. Also, a distinction was made between those commodities characterized by short and infrequent periods of price bubbles (five commodities) and those characterized by frequent and usually lasting for at least six months periods of bubbles (eighteen commodities). The panel confirmed that the energy and metals sectors are exposed to periods of bubbles more frequently and for longer than the agriculture & livestock sector. For all identified panels, a clear impact of the financial crisis of 2008 and the European debt crisis on the emergence of commodity bubbles was found.
Probabilistic assessment of SMRFs with infill masonry walls incorporating nonlinear soil-structure interaction
- Farzin Kazemi
- Neda Asgarkhani
- Robert Jankowski
2023 Pełny tekst Bulletin of Earthquake Engineering
Infill Masonry Walls (IMWs) are used in the perimeter of a building to separate the inner and outer space. IMWs may affect the lateral behavior of buildings, while they are different from those partition walls that separate two inner spaces. This study focused on the seismic vulnerability assessment of Steel Moment-Resisting Frames (SMRFs) assuming different placement of IMWs incorporating nonlinear Soil-Structure Interaction (SSI). The aim is to explore the damage states of IMWs and use their ability for improving the vulnerability of SMRFs. For this purpose, the three, five, seven, and nine story levels (3-Story, 5-Story, 7-Story, and 9-Story) SMRFs were modeled considering four soil types. Incremental Dynamic Analyses (IDAs) were performed to determine the seismic performance limit-state capacities of SMRFs considering the Far-Fault (FF) record subset suggested by FEMA P695. To accurately model the influence of IMWs on the seismic response of SMRFs, a Tcl programming algorithm was developed to intelligently monitor the damage states of IMWs in each floor level. Results of the analysis show that assuming different placement of IMWs can significantly increase the seismic limit-state capacities of SMRFs with and without considering SSI effects. In addition, IMWs can play a crucial role to improve the seismic performances as well as the seismic collapse probability, which may be suggested for retrofitting purposes.
Probabilistic estimation of diverse soil condition impact on vertical axis tank deformation
- Kamil Żyliński
- Jarosław Górski
2023 Pełny tekst Bulletin of the Polish Academy of Sciences-Technical Sciences
The calculations of fuel tanks should take into account the geometric imperfections of the structure as well as the variability of the material parameters of the foundation. The deformation of the tank shell can have a significant impact on the limit state of the structure and its operating conditions. The paper presents a probabilistic analysis of a vertical-axis, floating-roof cylindrical shell of a tank with a capacity of 50000 m3 placed on stratified soil with heterogeneous material parameters. The impact of a random subsoil description was estimated using the Point Estimated Method (PEM). In this way, the number of analyzed FEM models was significantly reduced. This approach also makes it possible to assess the sensitivity of tank settlement and deformation to the changing foundation conditions.
Probability distribution of flicker noise in AuNPdecorated graphene–Si Schottky barrier diode
- Janusz Smulko
- Andrzej Kwiatkowski
- Katarzyna Drozdowska
- Lars Österlund
- Tesfalem Welearegay
- Adil Rehman
- Sergey Rumyantsev
2023
We present results of the probability distribution analysis of flicker noise generated in Au nanoparticle (AuNP) decorated graphene–Si Schottky barrier diodes with and without yellow light illumination (592 nm), close to the localized surface plasmon resonance in the AuNPs (586 nm). The AuNPs occupy imperfections in the single-layer graphene and reduce the flicker noise intensity generated in the graphene layer. The estimated probability distribution exhibited an asymmetry shift when the sample was irradiated by yellow light (a 10-fold increase of the skewness coefficient). This effect is attributed to AuNPs collecting low-frequency fluctuations in the graphene layer and reducing 1/f noise.
Probiotic potential of Bacillus Isolates from Polish Bee Pollen and Bee Bread
- Karolina Pełka
- Ahmer Hafeez
- Randy Worobo
- Piotr Szweda
2023 Pełny tekst Probiotics and Antimicrobial Proteins
The main goal of this study was the evaluation of the probiotic potential of 10 Bacillus spp. strains isolated from 5 bee bread and 3 bee pollen samples. The antagonistic interaction with Staphylococcus aureus and Escherichia coli was a primary criterion for the preliminary selection of the isolates. Three out of ten strains—PY2.3 (isolated from pollen), BP20.15 and BB10.1 (both isolated from bee bread)—were found to be possible probiotic strains. All these strains are safe for humans (exhibiting -hemolytic activity) and meet all essential requirements for probiotics in terms of viability in the presence of bile salts and acid conditions, hydrophobicity, auto-aggregation, and co-aggregation with the cells of important human pathogenic bacteria. They also assimilate more than 30% of cholesterol after 24 h of incubation. These three isolates are resistant to penicillin but sensitive (or exhibit moderate resistance) to the other nine antibiotics tested herein. On the basis of whole-genome sequencing, BP20.15 and BB10.1 were classified as B. subtilis and PY2.3 as B. velezensis. Moreover, genomic analyses revealed that all these isolates are potential producers of different antimicrobial compounds, including bacteriocins and secondary metabolites. The outcomes of this study have proven that some of the Bacillus strains isolated from bee pollen or bee bread are potential probiotics.
Problem jakości powietrza wewnętrznego na przykładzie placówek opieki nad dziećmi
- Małgorzata Rutkowska
- Mariusz Marć
- Natalia Jatkowska
- Bożena Zabiegała
2023
Zgodnie z informacjami udostępnionymi przez Światową Organizację Zdrowia, zanieczyszczone powietrze, jest przyczyną siedmiu milionów przedwczesnych zgonów na całym świecie. Wciąż jednak w społeczeństwie panuje przeświadczenie, że największe ryzyko dla zdrowia człowieka płynie z oddychania powietrzem atmosferycznym, pomijając istotność jakości powietrza wewnętrznego, które może być zanieczyszczone w podobnym stopniu lub nawet w większym niż powietrze zewnętrzne. Jest to o tyle istotne, iż szacuje się, że człowiek przeciętnie spędza ok. 90% swojego czasu w pomieszczeniach wewnętrznych. W społeczeństwie można wyróżnić dwie podgrupy szczególnie podatne na niekorzystny wpływ, jaki wywiera zanieczyszczone powietrze na organizm człowieka, są to dzieci, oraz osoby starsze. Dzieci w wieku przedszkolnym są szczególnie narażone na działanie zanieczyszczeń obecnych w fazie gazowej nie tylko ze względu na nie w pełni rozwinięty układ oddechowy ale również ze względu na istotny stopień narażenia na kontakt z kurzem i materią zawieszoną czy długi czas przebywania w pomieszczeniach zamkniętych takich jak żłobki czy przedszkola. Głównym celem przeprowadzonych badań jest kompleksowa ocena stopnia narażenia dzieci na obecność cząstek materii zawieszonej, wybranych lotnych związków organicznych, rtęci i dwutlenku węgla w powietrzu wewnętrznym. Badania prowadzone były dwutorowo: w czasie rzeczywistym, bezpośrednio w placówkach, za pomocą dedykowanego przenośnego sprzętu pomiarowego, oraz w laboratorium, gdzie pobrane próbki powietrza zostały poddane wieloczynnikowej analizie. Placówki będące obiektem badań zlokalizowane były w różnych regionach Polski i zostały wyselekcjonowane na podstawie stopnia zurbanizowania, zalesienia i zanieczyszczenia powietrza atmosferycznego.
Problemy niepewności i integracji w przetwarzaniu danych o stanie emocjonalnym użytkownika komputera
- Grzegorz Brodny
2023 Pełny tekst
Istnieje wiele programów automatycznego rozpoznawania emocji z pojedynczych modalności. Wykorzystują one różnorodne modele reprezentacji emocji, a także metody przetwarzania. Obecnie wyraźnie widoczny jest trend do wykorzystania analizy wielomodalnej oraz wielokanałowej, mającej na celu poprawę skuteczności rozpoznawania emocji oraz poprawę niezawodności. W dziedzinie automatycznego rozpoznawania emocji nie określono typowego modelu reprezentacji emocji ani typowej metody analizy wielomodalnej czy wielokanałowej. Celem pracy doktorskiej było zaproponowanie metody integracji danych o stanie emocjonalnym użytkowników komputerów, która uwzględnia różne programy, kanały dostępu oraz metody reprezentacji emocji, a także kontekstową niepewność pomiarów symptomów emocji. Autor pracy zaproponował nową metodę integracji składającą się z wzorca architektonicznego Scoreboard, metody kwantyfikacji niepewności oszacowania emocji oraz algorytmu integracji informacji o stanie emocjonalnym, opartego o miarę niepewności. Zaproponowane rozwiązanie zapewnia modularność i adaptowalność poprzez zastosowane paradygmaty obiektowe, pozwala na łatwą integrację umożliwiającą wykorzystanie rozwiązań napisanych w różnych technologiach, cechuje się odpornością na nieprawidłowe dane wejściowe oraz nieprawidłowe działania komponentów, dzięki zastosowaniu ich dynamicznej oceny w czasie działania programu. Zaproponowana metoda kwantyfikacji pozwala na liczbową reprezentację niepewności. Z kolei zaproponowana metoda integracji informacji o stanie emocjonalnym wyznacza poziom wiarygodności odpowiadający jakości danych wejściowych, a także zgodności danych z poszczególnych kanałów obserwacji takich jak zgodności informacji o stanie emocjonalnym z dwóch kamer. W ramach ewaluacji zaproponowanego rozwiązania wykorzystano trzy metody badawcze: eksperymenty na zaetykietowanym emocjonalnie zbiorze danych SEMAINE, symulacje przypadków brzegowych przygotowane przez autora pracy, eksperymenty oraz studium przypadku z wykorzystaniem danych z badania przeprowadzonego na Politechnice Gdańskiej na stanowisku Monitora Emocji. Przeprowadzono dziewięć eksperymentów, pięć symulacji oraz studium przypadku Monitora Emocji, co pozwoliło na walidację postawionych tez. Wykorzystywanymi w badaniach modalnościami były analiza mimiki twarzy, analiza prozodii głosu oraz analiza ładunku emocjonalnego tekstu. Za pomocą zaproponowanej wielomodalnej analizy uzyskano poprawę niezawodności.
Process control of air stream deodorization from vapors of VOCs using a gas sensor matrix conducted in the biotrickling filter (BTF)
- Dominik Dobrzyniewski
- Bartosz Szulczyński
- Piotr Rybarczyk
- Jacek Gębicki
2023 Pełny tekst Archives of Environmental Protection
This article presents the validity, advisability and purposefulness of using a gas sensor matrix to monitor air deodorization processes carried out in a peat-perlite-polyurethane foam-packed biotrickling filter. The aim of the conducted research was to control the effectiveness of air stream purification from vapors of hydrophobic compounds, i.e., n-hexane and cyclohexane. The effectiveness of hydrophobic n-hexane and cyclohexane removal from air was evaluated using gas chromatography as the reference method and a custom built gas sensor matrix consisting of seven commercially available sensors. The influence of inlet loading (IL) of n-hexane and cyclohexane on the biotrickling filtration performance was investigated. The prepared sensor matrix was calibrated with use of two statistical techniques: Multiple Linear Regression (MLR) and Principal Component Regression (PCR). The developed mathematical models allowed us to correlate the multidimensional signal from the sensor array with the concentration of the removed substances. The results based on gas chromatography analyses indicated that the elimination efficiencies of n-hexane and cyclohexane reached about 40 and 30 g m-3 h-1, respectively. The results obtained using a gas sensor matrix revealed that it was possible not only to determine concentration reliably of investigated hydrophobic volatile organic compounds in the gas samples, but also to obtain results of a similar high level of quality as the chromatographic ones. A gas sensor matrix proposed in this work can be used for on-line real-time monitoring of biofiltration process performance of air polluted with n-hexane and cyclohexane.