IRANIAN JOURNAL OF BIOSYSTEMS ENGINEERING
Year:2024 | Volume:55 | Issue:1|Papers Count:6

Machine learning modeling can help overcome some of the limitations of gas sensors, such as high operational conditions, drift errors, limited selectivity, the need for a large amount of labeled data, and cost and fabrication challenges. In this research, an electronic nose system was developed for the detection of sulfur dioxide and acetic acid. three treatments, including sunny, acetic, and sulfuric, were prepared in three repetitions, and each was exposed to olfactory sensors for 60 minutes to record the sensor responses to each treatment. Then, the data obtained from the sensor responses were examined by machine learning models to determine the modeling accuracy of each method. The results showed that the utilized Gradient Boost Regression model with a determination coefficient of 0.9972, root mean square error of 0.0209, mean absolute error of 0.0026, and relative root mean square error of 0.0209 was able to model the gas sensor responses well for the introduced treatments. Furthermore, by analyzing the results, the type and degree of correlation between the sensor responses to each other and over time were determined to evaluate their behavior prediction. Then, based on the conducted modeling, it was revealed that MQ9, MQ3, MQ5, and TGS2620 sensors, with determination coefficients of 0.8668, 0.8786, 0.9458, and 0.9074, and root mean square errors of 0.0163, 0.0168, 0.0083, and 0.0227, respectively, provided more accurate and predictable responses compared to MQ135, TGS822, TGS810, and MQ4 sensors.

Cinnamon is one of the most important spices that has medicinal properties. Detecting adulteration in cinnamon powder using laboratory methods is expensive, time-consuming, and requires expert. Hyperspectral imaging is specifically used in the assessment of food safety and quality. The purpose of the present research is to detect adulteration in cinnamon powder using hyperspectral imaging. In the present study, 15 samples of cinnamon were prepared with 0, 5, 15, 30 and 50% adulteration levels. The adulterants were chickpea flour, wheat flour, and sea foam that were used separately. The hyperspectral imaging system received the light emitted from the samples in the visible and near-infrared ranges from 400 to 950 nm wavelength and recorded their hyperspectral images in the computer. After selecting the effective wavelengths and extracting the features from the images, the efficient features were selected and then classified using the support vector machine method. The correct classification rates of the classifier with one-against-one strategy in classification of the efficient features selected from the hyperspectral images related to the light emitted from the visible and infrared ranges to detect adulteration of wheat flour, chickpea flour, and sea foam powder in cinnamon were 95.55, 85.56, and 96.66%, respectively. Its correct classification rates with one-against-all strategy were equal to 78.88, 77.77, and 94.44%, respectively.

Curcumin as a natural hydrophobic compound has anti-microbial and anti-cancer properties, but its low stability and high sensitivity have limited the bioavailability of this compound. The purpose of this research is to design chia seed protein and mucilage hydrogel for curcumin encapsulation. The encapsulation of this compound in the hydrogel structure can be an effective way to protect this compound during digestion in the digestive tract. For this purpose, first, hydrogel protein and mucilage 12.5% , which was optimized in terms of texture characteristics. The release behavior of curcumin in two conditions of stomach and intestine simulation for protein and mucilage 7.5% and 12.5% hydrogels were evaluated. The results showed that curcumin loaded in hydrogel protein and mucilage 12.5% has a better stability to free curcumin during heat and optical operations. The results related to the controlled release in laboratory conditions indicated that the total amount of curcumin release during gastric-intestinal digestion was 60.71% for the sample containing 7.5% protein and mucilage and 27.30% for the sample containing 12.5% protein and mucilage. As a result, the release rate of curcumin decreased with the increase of mucilage concentration in the simulated conditions of the stomach and intestine, which can show the good ability of combined hydrogels to protect curcumin in gastrointestinal conditions and deliver it to the colon. The release behavior of curcumin in the gastrointestinal tract was of the Fickian release type.

The specific detection of the type and severity of plant abiotic stresses to take timely measures helps prevent yield reduction. This study introduces a new method to detect the type and severity of stress in rice plants under salinity, drought, and heat conditions by investigating microRNAs. The concentration of eight microRNAs in the tissue of plants subjected to salinity, drought, and heat conditions was measured with the help of an optical biosensor based on gold nanoparticles. The biosensor worked based on probe-target hybridization, in which the mixture of probe/citrate-capped gold nanoparticles (compound 1) and microRNA/polyethyleneimine-capped nanoparticles (compound 2) resulted in the aggregation of nanoparticles and changes in their spectroscopic properties. In the following, machine learning methods were used to predict the type and severity of stress using such concentrations. The results showed that the support vector machine optimized by the genetic algorithm was able to detect the severity of salinity, drought, and heat stress applied to rice plants with appropriate performance and with coefficients of determination of 0.94, 0.91, and 0.86, respectively. Then, the results of feature selection based on the cooperative game theory showed that among the microRNAs studied, miRNA-156, miRNA-393, and miRNA-159 had the largest contribution in predicting drought, salinity, and heat stresses in the rice plants, respectively. The findings of the research show that the examination of plant microRNAs with the help of optical biosensors can lead to reliable features for determining plant growth conditions and plant stresses in the early stage.

Conversion of biomass resources such as agricultural residues can enhance the energy supply and improve the diversify of the energy portfolio. One of the primary methods for preparing solid fuel for energy uses involves densifying shredded residues, a process known as briquetting. This study investigated the modeling of ozone pretreatment and its impacts on the production of bagasse fuel briquettes, utilizing a hydraulic press and employing response surface methodology (RSM). The effect of independent variables including the particle size of bagasse (≤1.18, 1.18-2.36, and 2.36-4.75 mm), brequetting temperature (280, 320, and 360 °C), moisture content of bagasse during ozonation (20, 35, and 50%), and ozonation time (15, 25, and 35 min.) was evaluated on on the physical, mechanical, and chemical properties of the briquettes. The results indicated that the moisture content of bagasse played a significant role in the ozonation pretreatment. Increasing the moisture content during the ozonation process to approximately 30% enhanced the toughness of the produced briquettes; however, further increasing it to 50% resulted in a decrease in toughness. The calorific value of the briquettes was calculated using validated experimental equations based on the results of proximate analysis. Findings revealed that ozone pretreatment boosts the calorific value of the briquettes compared to those without pretreatment. The optimal conditions for briquette production were identified as 15 minutes of ozonation, a bagasse particle size smaller than 1.18 mm, a moisture content of 33.448%, and a briquetting temperature of 280 °C. The optimum briquette density and toughness were found to be 982.310 kg/m³ and 249.934 kPa, respectively. Overall, ozonation pretreatment significantly enhanced the properties of sugarcane bagasse fuel briquettes.

The depletion of fossil fuels, environmental issues, and climate change make the development of renewable energy, especially wind energy, essential. The main challenge in developing wind energy is selecting suitable locations for power plants, where land roughness plays a significant role. This study aimed to prioritize suitable areas based on land roughness using remote sensing in Kiashahr. Land use classification results by the SVM algorithm from 2000 to 2020 showed changes in 2,957.66 hectares of the region. Predicted maps from Markov Cellular Automata models for 2030 were used to ensure practical application of results for future years. The simulated map was gridded based on Wieringa roughness length data to generate maps of roughness length and classes. Results showed that 84 cells, equivalent to 1,363.98 hectares, in the first and second classes have the best potential for wind power plants. Additionally, land use maps for 2030 indicated that a large part of the region is used for agriculture, mostly rice cultivation. These areas have a roughness length of 0.25 m for only two months of the year, and for the rest of the year, they have a roughness length of 0.1 m (class 4) and 0.03 m (class 3). Overall, considering a roughness length of up to 0.25 meters, 552 cells, equivalent to 8,963.36 hectares, were identified as suitable for wind power plants. The findings of this research can help identify suitable areas for wind power plant construction and assist in modeling wind speed near the hub of tall wind turbines.