Convolutional neural networks (CNNs) are widely used in image classification tasks, but their increasing model size and computation make them challenging to implement on embedded systems with constrained hardware reso...Convolutional neural networks (CNNs) are widely used in image classification tasks, but their increasing model size and computation make them challenging to implement on embedded systems with constrained hardware resources. To address this issue, the MobileNetV1 network was developed, which employs depthwise convolution to reduce network complexity. MobileNetV1 employs a stride of 2 in several convolutional layers to decrease the spatial resolution of feature maps, thereby lowering computational costs. However, this stride setting can lead to a loss of spatial information, particularly affecting the detection and representation of smaller objects or finer details in images. To maintain the trade-off between complexity and model performance, a lightweight convolutional neural network with hierarchical multi-scale feature fusion based on the MobileNetV1 network is proposed. The network consists of two main subnetworks. The first subnetwork uses a depthwise dilated separable convolution (DDSC) layer to learn imaging features with fewer parameters, which results in a lightweight and computationally inexpensive network. Furthermore, depthwise dilated convolution in DDSC layer effectively expands the field of view of filters, allowing them to incorporate a larger context. The second subnetwork is a hierarchical multi-scale feature fusion (HMFF) module that uses parallel multi-resolution branches architecture to process the input feature map in order to extract the multi-scale feature information of the input image. Experimental results on the CIFAR-10, Malaria, and KvasirV1 datasets demonstrate that the proposed method is efficient, reducing the network parameters and computational cost by 65.02% and 39.78%, respectively, while maintaining the network performance compared to the MobileNetV1 baseline.展开更多
This research entails the study of heat and mass transfer of nanofluid flow in a fluidized bed dryer used in tea drying processes in presence of induced magnetic field. A mathematical model describing the fluid flow i...This research entails the study of heat and mass transfer of nanofluid flow in a fluidized bed dryer used in tea drying processes in presence of induced magnetic field. A mathematical model describing the fluid flow in a Fluidized bed dryer was developed using the nonlinear partial differential equations. Due to their non-linearity, the equations were solved numerically by use of the finite difference method. The effects of physical flow parameters on velocity, temperature, concentration and magnetic induction profiles were studied and results were presented graphically. From the mathematical analysis, it was deduced that addition of silver nanoparticles into the fluid flow enhanced velocity and temperature profiles. This led to improved heat transfer in the fluidized bed dryer, hence amplifying the tea drying process. Furthermore, it was noted that induced magnetic field tends to decrease the fluid velocity, which results in uniform distribution of heat leading to efficient heat transfer between the tea particles and the fluid, thus improving the drying process. The research findings provide information to industries on ways to optimize thermal performance of fluidized bed dryers.展开更多
Currently,the Internet of Things(IoT)is revolutionizing communi-cation technology by facilitating the sharing of information between different physical devices connected to a network.To improve control,customization,f...Currently,the Internet of Things(IoT)is revolutionizing communi-cation technology by facilitating the sharing of information between different physical devices connected to a network.To improve control,customization,flexibility,and reduce network maintenance costs,a new Software-Defined Network(SDN)technology must be used in this infrastructure.Despite the various advantages of combining SDN and IoT,this environment is more vulnerable to various attacks due to the centralization of control.Most methods to ensure IoT security are designed to detect Distributed Denial-of-Service(DDoS)attacks,but they often lack mechanisms to mitigate their severity.This paper proposes a Multi-Attack Intrusion Detection System(MAIDS)for Software-Defined IoT Networks(SDN-IoT).The proposed scheme uses two machine-learning algorithms to improve detection efficiency and provide a mechanism to prevent false alarms.First,a comparative analysis of the most commonly used machine-learning algorithms to secure the SDN was performed on two datasets:the Network Security Laboratory Knowledge Discovery in Databases(NSL-KDD)and the Canadian Institute for Cyberse-curity Intrusion Detection Systems(CICIDS2017),to select the most suitable algorithms for the proposed scheme and for securing SDN-IoT systems.The algorithms evaluated include Extreme Gradient Boosting(XGBoost),K-Nearest Neighbor(KNN),Random Forest(RF),Support Vector Machine(SVM),and Logistic Regression(LR).Second,an algorithm for selecting the best dataset for machine learning in Intrusion Detection Systems(IDS)was developed to enable effective comparison between the datasets used in the development of the security scheme.The results showed that XGBoost and RF are the best algorithms to ensure the security of SDN-IoT and to be applied in the proposed security system,with average accuracies of 99.88%and 99.89%,respectively.Furthermore,the proposed security scheme reduced the false alarm rate by 33.23%,which is a significant improvement over prevalent schemes.Finally,tests of the algorithm for dataset selection showed that the rates of false positives and false negatives were reduced when the XGBoost and RF algorithms were trained on the CICIDS2017 dataset,making it the best for IDS compared to the NSL-KDD dataset.展开更多
Reinforced concrete (RC) constructions are the innovation of sustainable constructions replacing masonry constructions. Despite this, the use of concrete and steel to improve the performance of structural members in s...Reinforced concrete (RC) constructions are the innovation of sustainable constructions replacing masonry constructions. Despite this, the use of concrete and steel to improve the performance of structural members in service is a recurring problem due to the immediate or overtime appearance of cracks. The objective of this work was therefore to assess the damage phenomena of the steel-concrete interface in order to assess the performance of an RC structure. Samples of approximately 30 cm of reinforcement attacked by rust were taken from broken reinforced concrete columns and beams in order to determine the impact of corrosion on high adhesion steel (HA) and therefore on its ability to resist. The experimental results have shown that the corrosion degradation rates of reinforcing bars of different diameters increase as the diameter of the reinforcing bars decreases: 5% for HA12;23.75% for HA8 and 50% for HA6. Using the approach proposed by Mangat and Elgalf on the bearing capacity as a function of the progress of the corrosion phenomenon, these rates made it possible to assess the new fracture limits of corroded HA steels. For HA6 respectively HA8 and HA12, their initial limit resistances will decrease by 4/4, 3/4 and 1/4. Based on the results of this study and in order to guarantee their durability, an RC structure can be dimensioned by taking into account the effects of reinforcement corrosion.展开更多
Welded mild steel is used in different applications in engineering. To strengthen the joint, the weld can be reinforced by adding titanium alloy powder to the joint. This results in the formation of incomplete martens...Welded mild steel is used in different applications in engineering. To strengthen the joint, the weld can be reinforced by adding titanium alloy powder to the joint. This results in the formation of incomplete martensite in a welded joint. The incomplete martensite affects mechanical properties. Therefore, this study aims to predict the volume fraction of martensite in reinforced butt welded joints to understand complex phenomena during microstructure formation. To do so, a combination of the finite element method to predict temperature history, and the Koistinen and Marburger equation, were used to predict the volume fraction of martensite. The martensite start temperature was calculated using chemical elements obtained from the dilution-based mixture rule. The curve shape of martensite evolution was observed to be relatively linear due to the small quantity of martensite volume fraction. The simulated result correlated with experimental work documented in the literature. The model can be used in other powder addition techniques where the martensite can be observed in the final microstructure.展开更多
Approximately 450 million tons of plastic and agricultural waste are produced each year in the world. Only a small portion of this plastic waste is recycled, and a small portion of this agricultural waste is used as f...Approximately 450 million tons of plastic and agricultural waste are produced each year in the world. Only a small portion of this plastic waste is recycled, and a small portion of this agricultural waste is used as fuel or fertilizer, and the rest of this waste is left in the environment or is burned, resulting in environmental and air pollution. For proper disposal, plastic and agricultural waste can be used in the manufacture of composites as raw materials. In this study, we had evaluated the use of bean pod powder (BPp) was used as natural reinforcing filler in recycled polypropylene (rPP) based composites. BPp/rPP composite filaments were developed using the extrusion method and the samples were printed by Fused Filament Fabrication (FFF). Composites with rPP matrix containing different weight fractions of BPp (5%, 10% and 15%) were fabricated to observe and compare the mechanical properties (tensile, flexural, and compressive strength) of the filament composites. In addition, the filament surface was analyzed for roughness and particle size of bean pod powder. The results established that BPp/rPP composites exhibited better tensile, flexural, and compressive strength than rPP and pure PP. By adding 5 wt% BPp, the tensile strength of rPP increased from 20.4 MPa to 22.8 MPa. The highest flexural strength (15.05 MPa) was obtained at 5 wt% BPp among all composites and the highest compressive strength (24.5 MPa), was obtained at 10 wt% BPp. Therefore, it can be concluded that by carefully selecting the ratio of BPp to bean pod powder, it is therefore possible to positively influence the mechanical properties of the resulting composite.展开更多
Depression is a major public health problem around the world and contributes significantly to poor health and poverty. The rate of the number of people being affected is very high compared to the rate of medical treat...Depression is a major public health problem around the world and contributes significantly to poor health and poverty. The rate of the number of people being affected is very high compared to the rate of medical treatment of the disease. Thus, the disease often remains untreated and suffering continues. Machine learning has been widely used in many studies in detecting depressive individuals from their contents on online social networks. From the related reviews, it is apparent that the application of stacking for diagnosing depression has been minimal. The study implements stacking based on Extra Tree, Extreme Gradient Boosting, Light Gradient Boosting and Multi-layer perceptron and compares its performance to state of the art bagging and boosting ensemble learners. To better evaluate the effectiveness of the proposed stacking approach, three pretrain word embeddings techniques including: Word2vec, Global Vectors and Embeddings from language models were employed with two datasets. Also, a corrected resampled paired t-test was applied to test the significance of the stacked accuracy against the baseline accuracy. The experimental results shows that the stacking approach yields favourable results with a best accuracy of 99.54%.展开更多
The forecasting research literature has developed greatly in recent years as a result of advances in information technology. Financial time-series tasks have made substantial use of machine learning and deep neural ne...The forecasting research literature has developed greatly in recent years as a result of advances in information technology. Financial time-series tasks have made substantial use of machine learning and deep neural networks, but building a prediction model from scratch takes time and computational resources. Transfer learning is growing popular in tackling these constraints of training time and computational resources in several disciplines. This study proposes a hybrid base model for the financial time series prediction employing the recurrent neural network (RNN) and long-short term memory (LSTM) called RNN-LSTM. We used random search to fine-tune the hyperparameters and compared our proposed model to the RNN and LSTM base models and evaluate using the RMSE, MAE, and MAPE metrics. When forecasting Forex currency pairs GBP/USD, USD/ZAR, and AUD/NZD our proposed base model for transfer learning outperforms RNN and LSTM base model with root mean squared errors of 0.007656, 0.165250, and 0.001730 respectively.展开更多
The majority of spatial data reveal some degree of spatial dependence. The term “spatial dependence” refers to the tendency for phenomena to be more similar when they occur close together than when they occur far ap...The majority of spatial data reveal some degree of spatial dependence. The term “spatial dependence” refers to the tendency for phenomena to be more similar when they occur close together than when they occur far apart in space. This property is ignored in machine learning (ML) for spatial domains of application. Most classical machine learning algorithms are generally inappropriate unless modified in some way to account for it. In this study, we proposed an approach that aimed to improve a ML model to detect the dependence without incorporating any spatial features in the learning process. To detect this dependence while also improving performance, a hybrid model was used based on two representative algorithms. In addition, cross-validation method was used to make the model stable. Furthermore, global moran’s I and local moran were used to capture the spatial dependence in the residuals. The results show that the HM has significant with a R2 of 99.91% performance compared to RBFNN and RF that have 74.22% and 82.26% as R2 respectively. With lower errors, the HM was able to achieve an average test error of 0.033% and a positive global moran’s of 0.12. We concluded that as the R2 value increases, the models become weaker in terms of capturing the dependence.展开更多
Pneumonia ranks as a leading cause of mortality, particularly in children aged five and under. Detecting this disease typically requires radiologists to examine chest X-rays and report their findings to physicians, a ...Pneumonia ranks as a leading cause of mortality, particularly in children aged five and under. Detecting this disease typically requires radiologists to examine chest X-rays and report their findings to physicians, a task susceptible to human error. The application of Deep Transfer Learning (DTL) for the identification of pneumonia through chest X-rays is hindered by a shortage of available images, which has led to less than optimal DTL performance and issues with overfitting. Overfitting is characterized by a model’s learning that is too closely fitted to the training data, reducing its effectiveness on unseen data. The problem of overfitting is especially prevalent in medical image processing due to the high costs and extensive time required for image annotation, as well as the challenge of collecting substantial datasets that also respect patient privacy concerning infectious diseases such as pneumonia. To mitigate these challenges, this paper introduces the use of conditional generative adversarial networks (CGAN) to enrich the pneumonia dataset with 2690 synthesized X-ray images of the minority class, aiming to even out the dataset distribution for improved diagnostic performance. Subsequently, we applied four modified lightweight deep transfer learning models such as Xception, MobileNetV2, MobileNet, and EfficientNetB0. These models have been fine-tuned and evaluated, demonstrating remarkable detection accuracies of 99.26%, 98.23%, 97.06%, and 94.55%, respectively, across fifty epochs. The experimental results validate that the models we have proposed achieve high detection accuracy rates, with the best model reaching up to 99.26% effectiveness, outperforming other models in the diagnosis of pneumonia from X-ray images.展开更多
Financial Time Series Forecasting is an important tool to support both individual and organizational decisions. Periodic phenomena are very popular in econometrics. Many models have been built aiding capture of these ...Financial Time Series Forecasting is an important tool to support both individual and organizational decisions. Periodic phenomena are very popular in econometrics. Many models have been built aiding capture of these periodic trends as a way of enhancing forecasting of future events as well as guiding business and social activities. The nature of real-world systems </span><span style="font-family:Verdana;">is</span><span style="font-family:Verdana;"> characterized by many uncertain fluctuations which makes prediction difficult. In situations when randomness is mixed with periodicity, prediction is even much harder. We therefore constructed an ANN Time Varying Garch model with both linear and non-linear attributes and specific for processes with fixed and random periodicity. To eliminate the need for time series linear component filtering</span><span style="font-family:Verdana;">,</span><span style="font-family:Verdana;"> we incorporated the use of Artificial Neural Networks (ANN) and constructed Time Varying GARCH model on its disturbances. We developed the estimation procedure of the ANN time varying GARCH model parameters using non parametric techniques.展开更多
Recycled polypropylene filaments for fused filament fabrication were investigated with and without 14 wt% short fibre carbon reinforcements. The microstructure and mechanical properties of the filaments and 3D printed...Recycled polypropylene filaments for fused filament fabrication were investigated with and without 14 wt% short fibre carbon reinforcements. The microstructure and mechanical properties of the filaments and 3D printed specimens were characterized using scanning electron microscopy and standard tensile testing. It was observed that recycled polypropylene filaments with 14 wt% short carbon fibre reinforcement contained pores that were dispersed throughout the microstructure of the filament. A two-stage filament extrusion process was observed to improve the spatial distribution of carbon fibre reinforcement but did not reduce the pores. Recycled polypropylene filaments without reinforcement extruded at high screw speeds above 20 rpm contained a centreline cavity but no spatially distributed pores. However, this cavity is eliminated when extrusion is carried out at screw speeds below 20 rpm. For 3D printed specimens, interlayer cavities were observed larger for specimens printed from 14 wt% carbon fibre reinforced recycled polypropylene than those printed from unreinforced filaments. The values of tensile strength for the filaments were 21.82</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa and 24.22</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa, which reduced to 19.72</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa and 22.70</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa, respectively, for 3D printed samples using the filaments. Likewise, the young’s modulus of the filaments was 1208.6</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa and 1412.7</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa, which reduced to 961.5</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa and 1352.3</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa, respectively, for the 3D printed samples. The percentage elongation at failure for the recycled polypropylene filament was 9.83% but reduced to 3.84% for the samples printed with 14 wt% carbon fiber reinforced polypropylene filaments whose elongation to failure was 6.58%. The SEM observations on the fractured tensile test samples showed interlayer gaps between the printed and the adjacent raster layers. These gaps accounted for the reduction in the mechanical properties of the printed parts.展开更多
Objective: To bioprospect optimal phenological phases as source of novel molecules from native golden yellow Pleurotus citrinopileatus across four phenologies in both aqueous and ethanol extracts, and identify novel m...Objective: To bioprospect optimal phenological phases as source of novel molecules from native golden yellow Pleurotus citrinopileatus across four phenologies in both aqueous and ethanol extracts, and identify novel molecules responsible for these activities.Methods: Standard qualitative assay, Folin–Ciocalteu assay; aluminium chloride spectrophotometric, 2, 2-diphenyl-1-picrylhydrazyl, 2, 20-azinobis(3-ethylbenzothiazoline-6-suslfonic acid, ferricyanide reducing antioxidant power were used to determine total flavonoid, polyphenols, radical scavenging, and reducing power. Spectrophotometric methods were used for lycopene, b-carotene, and total carotenoids, while liquid chromatography quadrupole time of flight mass spectrometry was used for identification and comparative quantitation of polyphenols and flavonoids across the four phenological states. Chem Spider? database was used for the identification of compounds based on their empirical formula, accurate mass and literature review of previously reported compounds in mushroom.Results: Primordial phases exhibited higher contents of secondary metabolites than mature basidiocarps. Polyphenols content differed across physiological phases with primordials exhibiting significant high contents(P < 0.05) [(13.803 ± 0.797) mg gallic acid equivalent/g dry weight]. Distribution of total flavonoids was significantly different(P < 0.05) across physiological states and ranged from(3.311 ± 0.730) to(14.824 ± 0.890) mg quercetin equivalent g dry weight. Ten polyphenol acids and seven flavonoids compounds identified varied across these phases with primordials exhibiting relatively high peak areas. Total antioxidant activities showed a positive correlation with total polyphenols(r = 0.969;P < 0.05) and total flavonoids(r = 0.960; P < 0.05) across these phenologies.Conclusions: These findings provide evidence that primordials of golden yellow mushroom as opposed to their fruiting bodies are potent sources of bioactive health molecules.展开更多
Diesel fuel has been known as the most effective fuel but it is known as a fuel which produces harmful emissions. Later, emulsified diesel fuel was introduced as a better solution but there is no sufficient research d...Diesel fuel has been known as the most effective fuel but it is known as a fuel which produces harmful emissions. Later, emulsified diesel fuel was introduced as a better solution but there is no sufficient research data concerning combustion of emulsified fuel. The present work carried out a simulation of non-surfactant emulsified diesel fuel where composition of water in emulsion varied from 0% to 20% to determine the ratio of water to diesel which is more effective in reducing the exhaust emissions especially NOx. For this simulation,5% of water in diesel without surfactant was able to reduce NOx?up to 35%. It was shown that as the percentage of water increases, the power from that fuel combustion reduces.展开更多
Water scarcity in developing countries has forced farmers to use sewage as an alternative source of irrigation water. However, the usage of sewage for vegetable production has been known to cause excessive and often-u...Water scarcity in developing countries has forced farmers to use sewage as an alternative source of irrigation water. However, the usage of sewage for vegetable production has been known to cause excessive and often-unbalanced addition of nutrients hence posing a threat to food safety. The objective of this study was to determine the efficacy of slow sand filter and wetland plant in domestic wastewater treatment. To achieve this objective, samples were collected from the domestic wastewater collection pond within Jomo Kenyatta University of Agriculture and Technology (JKUAT). Laboratory tests were conducted on the collected samples and they revealed the presence of BOD, DO, pH, TDS, Sulfates, Chloride, Turbidity, Salinity, Conductivity, Alkalinity?and Coliform;whose values varied when compared with that of the parameters for standard irrigation water. This gave insight to the kind of treatments and filtration medium that were required to transform domestic wastewater into water fit for irrigation. A slow sand filter bed was designed and constructed using precisely six samples materials;sand, sand and wetland plants, gravel, gravel and wetland plants, mixture of gravel and sand, mixture of gravel and sand with wetland plants. These materials were used to identify the chemical and biological changes in domestic wastewater within a seven-day period. The water collected from the slow sand filter was tested, results showed that, of all six samples, slow sand filter using the mixture of gravel, sand with wetland plants had an average percentage efficient of 90% in removing all impurities from domestic wastewater thereby turning it into water suitable for irrigation. It is hoped that this study will provide a safe, easy, eco-friendly and cheap method of wastewater treatment while ensuring the sustainability of wastewater for irrigation and the expansion of green spaces in urban and peri-urban areas.展开更多
The global energy demand has continued to skyrocket, exacerbating the already severe energy problem and environmental pollution, prompting researchers to look for alternative energy sources. Exploration of waste lubri...The global energy demand has continued to skyrocket, exacerbating the already severe energy problem and environmental pollution, prompting researchers to look for alternative energy sources. Exploration of waste lubricating oil (WLO) as an alternative source of fuel has gained prominence among researchers due to its availability at low cost and the potential to generate energy while providing a safer means of disposal. The main challenge with WLO combustion is proper regulation of fuel and oxidizer during combustion to realize a near stoichiometric result. Additionally, WLO has high viscosity, hence preheating of the oil is necessary to lower the viscosity and enhance atomization, for a more efficient combustion process. This paper presents the optimization of flow parameters for combustion of WLO in a burner system by use of response surface methodology (RSM). The effects of air flow rate, injection pressure and fuel flow rate on combustion performance of a WLO burner were investigated. The highest flame temperature recorded was 1200°C at an air flow rate of 1 m3</sup>/min, fuel flow rate of 0.08 m3</sup>/hr and injection pressure of 20 bar. Tests on physical and chemical properties of WLO were conducted and characterized according to ASTM standard to ascertain its potential as an alternative fuel. The calorific values of WLO from petrol and diesel engines were found to be 41.23 MJ/kg and 42.65 MJ/kg respectively. Therefore, recycling of WLO by utilizing it as a fuel for burners has double benefits of mitigating environmental pollution and harnessing energy for process heating and power generation.展开更多
Wastewater management and purification remain one of the greatest problems of mankind. The biological wastewater treatment technique uses a biofilm media carrier where microorganisms attach themselves to the surface. ...Wastewater management and purification remain one of the greatest problems of mankind. The biological wastewater treatment technique uses a biofilm media carrier where microorganisms attach themselves to the surface. This biofilter is usually made from virgin plastic pellets and can also be produced from recycled waste plastic and used in wastewater treatment. The need to treat water using low-cost carrier media has led to finding alternative sources of materials for biofilter manufacturing. Therefore, this work is centered on the recycling of waste plastic to make filaments which are then used for 3D printing of a high specific surface area (SSA) less clogging biofilm carrier through the parametric redesign. In the current study, the polypropylene material was recycled to make a 2.85 mm diameter filament compatible with the Ultimaker S3. Moreover, analytical models and governing equations were developed for the design of the K3 Kaldnes and MB3 media. Empirical surface area (SA), specific surface area, and volume of the respective carriers were determined using the model developed. SolidWorks was used to design and evaluate the same parameters which were then compared to model results. The errors in SSA obtained from the model with respect to the SolidWorks results for both the K3 Kaldnes and MB3 media were 0.34% and 0.76% respectively. With these small error margins, the model can serve as a tool and guideline for the designing of cylindrically shaped carriers. By transforming plastics into biofilters, waste plastics are mopped up reducing pollutions. Consequently, the deployment of such biofilters will enhance efficient wastewater treatment for a cleaner environment and the wellbeing of human race.展开更多
The Bermudan option pricing problem with variable transaction costs is considered for a risky asset whose price process is derived under the information-based model. The price is formulated as the value function of an...The Bermudan option pricing problem with variable transaction costs is considered for a risky asset whose price process is derived under the information-based model. The price is formulated as the value function of an optimal stopping problem, which is the value function of a stochastic control problem given by a non-linear second order partial differential equation. The theory of viscosity solutions is applied to solve the stochastic control problem such that the value function is also the solution of the corresponding Bellman equation. Under some regularity assumptions, the existence and uniqueness of the solution of the pricing equation are derived by the application of the Perron method and Banach Fixed Point theorem.展开更多
Solar water heaters which provide a cost-effective and environmental friendly approach to hot water generation are in widespread application. Evacuated tube solar water heaters perform better than flat plate solar wat...Solar water heaters which provide a cost-effective and environmental friendly approach to hot water generation are in widespread application. Evacuated tube solar water heaters perform better than flat plate solar water heaters as a result of their greater surface area exposed for sunlight absorption. Water-in-glass evacuated tube solar water heaters are widely used as compared to heat-pipe solar water heaters due to their short payback periods. In this study, the performance of water-in-glass evacuated tube solar water heater is investigated through experiments under the climatic conditions in Kenya. The results revealed a daily efficiency range of 0.58 - 0.65 and a daily final outlet temperature greater than 55<span style="white-space:normal;">°</span>C given an initial temperature of 25°C.展开更多
Geothermal energy can be effectively utilized for grain drying to reduce carbon emissions and also cut operational costs associated with conventional methods. The main challenges encountered in the use of the geotherm...Geothermal energy can be effectively utilized for grain drying to reduce carbon emissions and also cut operational costs associated with conventional methods. The main challenges encountered in the use of the geothermal grain dryer, such as in Menengai, Kenya, include uneven grain drying and long throughput times. Grains near the hot air inlet dry at a faster rate compared to those near the exhaust end. Therefore, the grains must be recirculated within the dryer to achieve uniform moisture distribution. Grain recirculation is energy-intensive as it utilizes electricity running the elevator motors in addition to the suction pump. A Computational Fluid Dynamics (CFD) model was developed to study the airflow pattern and its impact on drying of maize. The model was simulated in ANSYS 21 and validated using experimental data. Finite volume discretization method was employed for meshing. Pressure-based segregated solver was used in the Computational Fluid Dynamics (CFD) simulation. Also, K-Omega turbulent model was used for enhancing wall treatment. The findings indicate that non-uniform hot air distribution across the grain buffer section causes uneven drying. Introducing filleted flow-guides results in a relatively uniform velocity, temperature, and turbulence kinetic energy distribution across the dryer. The average velocity and temperature magnitudes in lower compartments increased by 153.3% and 0.25% respectively for the improved dryer. In the upper compartments, the velocity and temperature increase were 176.5% and 0.22% respectively.展开更多
文摘Convolutional neural networks (CNNs) are widely used in image classification tasks, but their increasing model size and computation make them challenging to implement on embedded systems with constrained hardware resources. To address this issue, the MobileNetV1 network was developed, which employs depthwise convolution to reduce network complexity. MobileNetV1 employs a stride of 2 in several convolutional layers to decrease the spatial resolution of feature maps, thereby lowering computational costs. However, this stride setting can lead to a loss of spatial information, particularly affecting the detection and representation of smaller objects or finer details in images. To maintain the trade-off between complexity and model performance, a lightweight convolutional neural network with hierarchical multi-scale feature fusion based on the MobileNetV1 network is proposed. The network consists of two main subnetworks. The first subnetwork uses a depthwise dilated separable convolution (DDSC) layer to learn imaging features with fewer parameters, which results in a lightweight and computationally inexpensive network. Furthermore, depthwise dilated convolution in DDSC layer effectively expands the field of view of filters, allowing them to incorporate a larger context. The second subnetwork is a hierarchical multi-scale feature fusion (HMFF) module that uses parallel multi-resolution branches architecture to process the input feature map in order to extract the multi-scale feature information of the input image. Experimental results on the CIFAR-10, Malaria, and KvasirV1 datasets demonstrate that the proposed method is efficient, reducing the network parameters and computational cost by 65.02% and 39.78%, respectively, while maintaining the network performance compared to the MobileNetV1 baseline.
文摘This research entails the study of heat and mass transfer of nanofluid flow in a fluidized bed dryer used in tea drying processes in presence of induced magnetic field. A mathematical model describing the fluid flow in a Fluidized bed dryer was developed using the nonlinear partial differential equations. Due to their non-linearity, the equations were solved numerically by use of the finite difference method. The effects of physical flow parameters on velocity, temperature, concentration and magnetic induction profiles were studied and results were presented graphically. From the mathematical analysis, it was deduced that addition of silver nanoparticles into the fluid flow enhanced velocity and temperature profiles. This led to improved heat transfer in the fluidized bed dryer, hence amplifying the tea drying process. Furthermore, it was noted that induced magnetic field tends to decrease the fluid velocity, which results in uniform distribution of heat leading to efficient heat transfer between the tea particles and the fluid, thus improving the drying process. The research findings provide information to industries on ways to optimize thermal performance of fluidized bed dryers.
文摘Currently,the Internet of Things(IoT)is revolutionizing communi-cation technology by facilitating the sharing of information between different physical devices connected to a network.To improve control,customization,flexibility,and reduce network maintenance costs,a new Software-Defined Network(SDN)technology must be used in this infrastructure.Despite the various advantages of combining SDN and IoT,this environment is more vulnerable to various attacks due to the centralization of control.Most methods to ensure IoT security are designed to detect Distributed Denial-of-Service(DDoS)attacks,but they often lack mechanisms to mitigate their severity.This paper proposes a Multi-Attack Intrusion Detection System(MAIDS)for Software-Defined IoT Networks(SDN-IoT).The proposed scheme uses two machine-learning algorithms to improve detection efficiency and provide a mechanism to prevent false alarms.First,a comparative analysis of the most commonly used machine-learning algorithms to secure the SDN was performed on two datasets:the Network Security Laboratory Knowledge Discovery in Databases(NSL-KDD)and the Canadian Institute for Cyberse-curity Intrusion Detection Systems(CICIDS2017),to select the most suitable algorithms for the proposed scheme and for securing SDN-IoT systems.The algorithms evaluated include Extreme Gradient Boosting(XGBoost),K-Nearest Neighbor(KNN),Random Forest(RF),Support Vector Machine(SVM),and Logistic Regression(LR).Second,an algorithm for selecting the best dataset for machine learning in Intrusion Detection Systems(IDS)was developed to enable effective comparison between the datasets used in the development of the security scheme.The results showed that XGBoost and RF are the best algorithms to ensure the security of SDN-IoT and to be applied in the proposed security system,with average accuracies of 99.88%and 99.89%,respectively.Furthermore,the proposed security scheme reduced the false alarm rate by 33.23%,which is a significant improvement over prevalent schemes.Finally,tests of the algorithm for dataset selection showed that the rates of false positives and false negatives were reduced when the XGBoost and RF algorithms were trained on the CICIDS2017 dataset,making it the best for IDS compared to the NSL-KDD dataset.
文摘Reinforced concrete (RC) constructions are the innovation of sustainable constructions replacing masonry constructions. Despite this, the use of concrete and steel to improve the performance of structural members in service is a recurring problem due to the immediate or overtime appearance of cracks. The objective of this work was therefore to assess the damage phenomena of the steel-concrete interface in order to assess the performance of an RC structure. Samples of approximately 30 cm of reinforcement attacked by rust were taken from broken reinforced concrete columns and beams in order to determine the impact of corrosion on high adhesion steel (HA) and therefore on its ability to resist. The experimental results have shown that the corrosion degradation rates of reinforcing bars of different diameters increase as the diameter of the reinforcing bars decreases: 5% for HA12;23.75% for HA8 and 50% for HA6. Using the approach proposed by Mangat and Elgalf on the bearing capacity as a function of the progress of the corrosion phenomenon, these rates made it possible to assess the new fracture limits of corroded HA steels. For HA6 respectively HA8 and HA12, their initial limit resistances will decrease by 4/4, 3/4 and 1/4. Based on the results of this study and in order to guarantee their durability, an RC structure can be dimensioned by taking into account the effects of reinforcement corrosion.
文摘Welded mild steel is used in different applications in engineering. To strengthen the joint, the weld can be reinforced by adding titanium alloy powder to the joint. This results in the formation of incomplete martensite in a welded joint. The incomplete martensite affects mechanical properties. Therefore, this study aims to predict the volume fraction of martensite in reinforced butt welded joints to understand complex phenomena during microstructure formation. To do so, a combination of the finite element method to predict temperature history, and the Koistinen and Marburger equation, were used to predict the volume fraction of martensite. The martensite start temperature was calculated using chemical elements obtained from the dilution-based mixture rule. The curve shape of martensite evolution was observed to be relatively linear due to the small quantity of martensite volume fraction. The simulated result correlated with experimental work documented in the literature. The model can be used in other powder addition techniques where the martensite can be observed in the final microstructure.
文摘Approximately 450 million tons of plastic and agricultural waste are produced each year in the world. Only a small portion of this plastic waste is recycled, and a small portion of this agricultural waste is used as fuel or fertilizer, and the rest of this waste is left in the environment or is burned, resulting in environmental and air pollution. For proper disposal, plastic and agricultural waste can be used in the manufacture of composites as raw materials. In this study, we had evaluated the use of bean pod powder (BPp) was used as natural reinforcing filler in recycled polypropylene (rPP) based composites. BPp/rPP composite filaments were developed using the extrusion method and the samples were printed by Fused Filament Fabrication (FFF). Composites with rPP matrix containing different weight fractions of BPp (5%, 10% and 15%) were fabricated to observe and compare the mechanical properties (tensile, flexural, and compressive strength) of the filament composites. In addition, the filament surface was analyzed for roughness and particle size of bean pod powder. The results established that BPp/rPP composites exhibited better tensile, flexural, and compressive strength than rPP and pure PP. By adding 5 wt% BPp, the tensile strength of rPP increased from 20.4 MPa to 22.8 MPa. The highest flexural strength (15.05 MPa) was obtained at 5 wt% BPp among all composites and the highest compressive strength (24.5 MPa), was obtained at 10 wt% BPp. Therefore, it can be concluded that by carefully selecting the ratio of BPp to bean pod powder, it is therefore possible to positively influence the mechanical properties of the resulting composite.
文摘Depression is a major public health problem around the world and contributes significantly to poor health and poverty. The rate of the number of people being affected is very high compared to the rate of medical treatment of the disease. Thus, the disease often remains untreated and suffering continues. Machine learning has been widely used in many studies in detecting depressive individuals from their contents on online social networks. From the related reviews, it is apparent that the application of stacking for diagnosing depression has been minimal. The study implements stacking based on Extra Tree, Extreme Gradient Boosting, Light Gradient Boosting and Multi-layer perceptron and compares its performance to state of the art bagging and boosting ensemble learners. To better evaluate the effectiveness of the proposed stacking approach, three pretrain word embeddings techniques including: Word2vec, Global Vectors and Embeddings from language models were employed with two datasets. Also, a corrected resampled paired t-test was applied to test the significance of the stacked accuracy against the baseline accuracy. The experimental results shows that the stacking approach yields favourable results with a best accuracy of 99.54%.
文摘The forecasting research literature has developed greatly in recent years as a result of advances in information technology. Financial time-series tasks have made substantial use of machine learning and deep neural networks, but building a prediction model from scratch takes time and computational resources. Transfer learning is growing popular in tackling these constraints of training time and computational resources in several disciplines. This study proposes a hybrid base model for the financial time series prediction employing the recurrent neural network (RNN) and long-short term memory (LSTM) called RNN-LSTM. We used random search to fine-tune the hyperparameters and compared our proposed model to the RNN and LSTM base models and evaluate using the RMSE, MAE, and MAPE metrics. When forecasting Forex currency pairs GBP/USD, USD/ZAR, and AUD/NZD our proposed base model for transfer learning outperforms RNN and LSTM base model with root mean squared errors of 0.007656, 0.165250, and 0.001730 respectively.
文摘The majority of spatial data reveal some degree of spatial dependence. The term “spatial dependence” refers to the tendency for phenomena to be more similar when they occur close together than when they occur far apart in space. This property is ignored in machine learning (ML) for spatial domains of application. Most classical machine learning algorithms are generally inappropriate unless modified in some way to account for it. In this study, we proposed an approach that aimed to improve a ML model to detect the dependence without incorporating any spatial features in the learning process. To detect this dependence while also improving performance, a hybrid model was used based on two representative algorithms. In addition, cross-validation method was used to make the model stable. Furthermore, global moran’s I and local moran were used to capture the spatial dependence in the residuals. The results show that the HM has significant with a R2 of 99.91% performance compared to RBFNN and RF that have 74.22% and 82.26% as R2 respectively. With lower errors, the HM was able to achieve an average test error of 0.033% and a positive global moran’s of 0.12. We concluded that as the R2 value increases, the models become weaker in terms of capturing the dependence.
文摘Pneumonia ranks as a leading cause of mortality, particularly in children aged five and under. Detecting this disease typically requires radiologists to examine chest X-rays and report their findings to physicians, a task susceptible to human error. The application of Deep Transfer Learning (DTL) for the identification of pneumonia through chest X-rays is hindered by a shortage of available images, which has led to less than optimal DTL performance and issues with overfitting. Overfitting is characterized by a model’s learning that is too closely fitted to the training data, reducing its effectiveness on unseen data. The problem of overfitting is especially prevalent in medical image processing due to the high costs and extensive time required for image annotation, as well as the challenge of collecting substantial datasets that also respect patient privacy concerning infectious diseases such as pneumonia. To mitigate these challenges, this paper introduces the use of conditional generative adversarial networks (CGAN) to enrich the pneumonia dataset with 2690 synthesized X-ray images of the minority class, aiming to even out the dataset distribution for improved diagnostic performance. Subsequently, we applied four modified lightweight deep transfer learning models such as Xception, MobileNetV2, MobileNet, and EfficientNetB0. These models have been fine-tuned and evaluated, demonstrating remarkable detection accuracies of 99.26%, 98.23%, 97.06%, and 94.55%, respectively, across fifty epochs. The experimental results validate that the models we have proposed achieve high detection accuracy rates, with the best model reaching up to 99.26% effectiveness, outperforming other models in the diagnosis of pneumonia from X-ray images.
文摘Financial Time Series Forecasting is an important tool to support both individual and organizational decisions. Periodic phenomena are very popular in econometrics. Many models have been built aiding capture of these periodic trends as a way of enhancing forecasting of future events as well as guiding business and social activities. The nature of real-world systems </span><span style="font-family:Verdana;">is</span><span style="font-family:Verdana;"> characterized by many uncertain fluctuations which makes prediction difficult. In situations when randomness is mixed with periodicity, prediction is even much harder. We therefore constructed an ANN Time Varying Garch model with both linear and non-linear attributes and specific for processes with fixed and random periodicity. To eliminate the need for time series linear component filtering</span><span style="font-family:Verdana;">,</span><span style="font-family:Verdana;"> we incorporated the use of Artificial Neural Networks (ANN) and constructed Time Varying GARCH model on its disturbances. We developed the estimation procedure of the ANN time varying GARCH model parameters using non parametric techniques.
文摘Recycled polypropylene filaments for fused filament fabrication were investigated with and without 14 wt% short fibre carbon reinforcements. The microstructure and mechanical properties of the filaments and 3D printed specimens were characterized using scanning electron microscopy and standard tensile testing. It was observed that recycled polypropylene filaments with 14 wt% short carbon fibre reinforcement contained pores that were dispersed throughout the microstructure of the filament. A two-stage filament extrusion process was observed to improve the spatial distribution of carbon fibre reinforcement but did not reduce the pores. Recycled polypropylene filaments without reinforcement extruded at high screw speeds above 20 rpm contained a centreline cavity but no spatially distributed pores. However, this cavity is eliminated when extrusion is carried out at screw speeds below 20 rpm. For 3D printed specimens, interlayer cavities were observed larger for specimens printed from 14 wt% carbon fibre reinforced recycled polypropylene than those printed from unreinforced filaments. The values of tensile strength for the filaments were 21.82</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa and 24.22</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa, which reduced to 19.72</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa and 22.70</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa, respectively, for 3D printed samples using the filaments. Likewise, the young’s modulus of the filaments was 1208.6</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa and 1412.7</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa, which reduced to 961.5</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa and 1352.3</span><span style="font-size:10pt;font-family:""> </span><span style="font-family:Verdana;">MPa, respectively, for the 3D printed samples. The percentage elongation at failure for the recycled polypropylene filament was 9.83% but reduced to 3.84% for the samples printed with 14 wt% carbon fiber reinforced polypropylene filaments whose elongation to failure was 6.58%. The SEM observations on the fractured tensile test samples showed interlayer gaps between the printed and the adjacent raster layers. These gaps accounted for the reduction in the mechanical properties of the printed parts.
基金Supported by the African Union(African UnionNo.17/2014)
文摘Objective: To bioprospect optimal phenological phases as source of novel molecules from native golden yellow Pleurotus citrinopileatus across four phenologies in both aqueous and ethanol extracts, and identify novel molecules responsible for these activities.Methods: Standard qualitative assay, Folin–Ciocalteu assay; aluminium chloride spectrophotometric, 2, 2-diphenyl-1-picrylhydrazyl, 2, 20-azinobis(3-ethylbenzothiazoline-6-suslfonic acid, ferricyanide reducing antioxidant power were used to determine total flavonoid, polyphenols, radical scavenging, and reducing power. Spectrophotometric methods were used for lycopene, b-carotene, and total carotenoids, while liquid chromatography quadrupole time of flight mass spectrometry was used for identification and comparative quantitation of polyphenols and flavonoids across the four phenological states. Chem Spider? database was used for the identification of compounds based on their empirical formula, accurate mass and literature review of previously reported compounds in mushroom.Results: Primordial phases exhibited higher contents of secondary metabolites than mature basidiocarps. Polyphenols content differed across physiological phases with primordials exhibiting significant high contents(P < 0.05) [(13.803 ± 0.797) mg gallic acid equivalent/g dry weight]. Distribution of total flavonoids was significantly different(P < 0.05) across physiological states and ranged from(3.311 ± 0.730) to(14.824 ± 0.890) mg quercetin equivalent g dry weight. Ten polyphenol acids and seven flavonoids compounds identified varied across these phases with primordials exhibiting relatively high peak areas. Total antioxidant activities showed a positive correlation with total polyphenols(r = 0.969;P < 0.05) and total flavonoids(r = 0.960; P < 0.05) across these phenologies.Conclusions: These findings provide evidence that primordials of golden yellow mushroom as opposed to their fruiting bodies are potent sources of bioactive health molecules.
文摘Diesel fuel has been known as the most effective fuel but it is known as a fuel which produces harmful emissions. Later, emulsified diesel fuel was introduced as a better solution but there is no sufficient research data concerning combustion of emulsified fuel. The present work carried out a simulation of non-surfactant emulsified diesel fuel where composition of water in emulsion varied from 0% to 20% to determine the ratio of water to diesel which is more effective in reducing the exhaust emissions especially NOx. For this simulation,5% of water in diesel without surfactant was able to reduce NOx?up to 35%. It was shown that as the percentage of water increases, the power from that fuel combustion reduces.
文摘Water scarcity in developing countries has forced farmers to use sewage as an alternative source of irrigation water. However, the usage of sewage for vegetable production has been known to cause excessive and often-unbalanced addition of nutrients hence posing a threat to food safety. The objective of this study was to determine the efficacy of slow sand filter and wetland plant in domestic wastewater treatment. To achieve this objective, samples were collected from the domestic wastewater collection pond within Jomo Kenyatta University of Agriculture and Technology (JKUAT). Laboratory tests were conducted on the collected samples and they revealed the presence of BOD, DO, pH, TDS, Sulfates, Chloride, Turbidity, Salinity, Conductivity, Alkalinity?and Coliform;whose values varied when compared with that of the parameters for standard irrigation water. This gave insight to the kind of treatments and filtration medium that were required to transform domestic wastewater into water fit for irrigation. A slow sand filter bed was designed and constructed using precisely six samples materials;sand, sand and wetland plants, gravel, gravel and wetland plants, mixture of gravel and sand, mixture of gravel and sand with wetland plants. These materials were used to identify the chemical and biological changes in domestic wastewater within a seven-day period. The water collected from the slow sand filter was tested, results showed that, of all six samples, slow sand filter using the mixture of gravel, sand with wetland plants had an average percentage efficient of 90% in removing all impurities from domestic wastewater thereby turning it into water suitable for irrigation. It is hoped that this study will provide a safe, easy, eco-friendly and cheap method of wastewater treatment while ensuring the sustainability of wastewater for irrigation and the expansion of green spaces in urban and peri-urban areas.
文摘The global energy demand has continued to skyrocket, exacerbating the already severe energy problem and environmental pollution, prompting researchers to look for alternative energy sources. Exploration of waste lubricating oil (WLO) as an alternative source of fuel has gained prominence among researchers due to its availability at low cost and the potential to generate energy while providing a safer means of disposal. The main challenge with WLO combustion is proper regulation of fuel and oxidizer during combustion to realize a near stoichiometric result. Additionally, WLO has high viscosity, hence preheating of the oil is necessary to lower the viscosity and enhance atomization, for a more efficient combustion process. This paper presents the optimization of flow parameters for combustion of WLO in a burner system by use of response surface methodology (RSM). The effects of air flow rate, injection pressure and fuel flow rate on combustion performance of a WLO burner were investigated. The highest flame temperature recorded was 1200°C at an air flow rate of 1 m3</sup>/min, fuel flow rate of 0.08 m3</sup>/hr and injection pressure of 20 bar. Tests on physical and chemical properties of WLO were conducted and characterized according to ASTM standard to ascertain its potential as an alternative fuel. The calorific values of WLO from petrol and diesel engines were found to be 41.23 MJ/kg and 42.65 MJ/kg respectively. Therefore, recycling of WLO by utilizing it as a fuel for burners has double benefits of mitigating environmental pollution and harnessing energy for process heating and power generation.
文摘Wastewater management and purification remain one of the greatest problems of mankind. The biological wastewater treatment technique uses a biofilm media carrier where microorganisms attach themselves to the surface. This biofilter is usually made from virgin plastic pellets and can also be produced from recycled waste plastic and used in wastewater treatment. The need to treat water using low-cost carrier media has led to finding alternative sources of materials for biofilter manufacturing. Therefore, this work is centered on the recycling of waste plastic to make filaments which are then used for 3D printing of a high specific surface area (SSA) less clogging biofilm carrier through the parametric redesign. In the current study, the polypropylene material was recycled to make a 2.85 mm diameter filament compatible with the Ultimaker S3. Moreover, analytical models and governing equations were developed for the design of the K3 Kaldnes and MB3 media. Empirical surface area (SA), specific surface area, and volume of the respective carriers were determined using the model developed. SolidWorks was used to design and evaluate the same parameters which were then compared to model results. The errors in SSA obtained from the model with respect to the SolidWorks results for both the K3 Kaldnes and MB3 media were 0.34% and 0.76% respectively. With these small error margins, the model can serve as a tool and guideline for the designing of cylindrically shaped carriers. By transforming plastics into biofilters, waste plastics are mopped up reducing pollutions. Consequently, the deployment of such biofilters will enhance efficient wastewater treatment for a cleaner environment and the wellbeing of human race.
文摘The Bermudan option pricing problem with variable transaction costs is considered for a risky asset whose price process is derived under the information-based model. The price is formulated as the value function of an optimal stopping problem, which is the value function of a stochastic control problem given by a non-linear second order partial differential equation. The theory of viscosity solutions is applied to solve the stochastic control problem such that the value function is also the solution of the corresponding Bellman equation. Under some regularity assumptions, the existence and uniqueness of the solution of the pricing equation are derived by the application of the Perron method and Banach Fixed Point theorem.
文摘Solar water heaters which provide a cost-effective and environmental friendly approach to hot water generation are in widespread application. Evacuated tube solar water heaters perform better than flat plate solar water heaters as a result of their greater surface area exposed for sunlight absorption. Water-in-glass evacuated tube solar water heaters are widely used as compared to heat-pipe solar water heaters due to their short payback periods. In this study, the performance of water-in-glass evacuated tube solar water heater is investigated through experiments under the climatic conditions in Kenya. The results revealed a daily efficiency range of 0.58 - 0.65 and a daily final outlet temperature greater than 55<span style="white-space:normal;">°</span>C given an initial temperature of 25°C.
文摘Geothermal energy can be effectively utilized for grain drying to reduce carbon emissions and also cut operational costs associated with conventional methods. The main challenges encountered in the use of the geothermal grain dryer, such as in Menengai, Kenya, include uneven grain drying and long throughput times. Grains near the hot air inlet dry at a faster rate compared to those near the exhaust end. Therefore, the grains must be recirculated within the dryer to achieve uniform moisture distribution. Grain recirculation is energy-intensive as it utilizes electricity running the elevator motors in addition to the suction pump. A Computational Fluid Dynamics (CFD) model was developed to study the airflow pattern and its impact on drying of maize. The model was simulated in ANSYS 21 and validated using experimental data. Finite volume discretization method was employed for meshing. Pressure-based segregated solver was used in the Computational Fluid Dynamics (CFD) simulation. Also, K-Omega turbulent model was used for enhancing wall treatment. The findings indicate that non-uniform hot air distribution across the grain buffer section causes uneven drying. Introducing filleted flow-guides results in a relatively uniform velocity, temperature, and turbulence kinetic energy distribution across the dryer. The average velocity and temperature magnitudes in lower compartments increased by 153.3% and 0.25% respectively for the improved dryer. In the upper compartments, the velocity and temperature increase were 176.5% and 0.22% respectively.