Diabetic retinopathy (DR) is one of the diseases that cause blindness globally. Untreated accumulation of fat and cholesterol may trigger atherosclerosis in the diabetic patient, which may obstruct blood vessels. Retinal fundus images are used as diagnostic tools to screen abnormalities linked to diseases that affect the eye. Blurriness and low contrast are major problems when segmenting retinal fundus images. This article proposes an algorithm to segment and detect hemorrhages in retinal fundus images. The proposed method first performs preprocessing on retinal fundus images. Then a novel smart windowing-based adaptive threshold is utilized to segment hemorrhages. Finally, conventional and hand-crafted features are extracted from each candidate and classified by a support vector machine. Two datasets are used to evaluate the algorithms. Precision rate (P), recall rate (R), and F1 score are used for quantitative evaluation of segmentation methods. Mean square error, peak signal to noise ratio, information entropy, and contrast are also used to evaluate preprocessing method. The proposed method achieves a high F1 score with 83.85% for the DIARETDB1 image dataset and 72.25% for the DIARETDB0 image dataset. The proposed algorithm adequately adapts when compared with conventional algorithms, hence will act as a tool for segmentation.

The rapid advancement in technology and rise in energy consumption have motivated research addressing Demand-Side Management (DSM). In this research, a novel design for Home Energy Management (HEM) is proposed that seamlessly integrates Battery Energy Storage Systems (BESSs), Photovoltaic (PV) installations, and Electric Vehicles (EVs). Leveraging a Mixed-Integer Linear Programming (MILP) approach, the proposed system aims to minimize electricity costs. The optimization model takes into account Real-Time Pricing (RTP) tariffs, facilitating the efficient scheduling of household appliances and optimizing patterns for BESS charging and discharging, as well as EV charging and discharging. Both individual and multiple Smart Home (SH) case studies showcase noteworthy reductions in electricity costs. In the case of multiple SHs, a remarkable cost reduction of $46.38\%$ was achieved compared to a traditional SH scenario lacking integration of a PV, BESS, and EV.

In today’s power systems, electric vehicles (EVs) constitute a significant factor influencing electricity dynamics, with their important role anticipated in future smart grid systems. An important feature of electric vehicles is their dual capability to both charge and discharge energy to/from their battery storage. Notably, the discharge capability enables them to offer vehicle-to-grid (V2G) services. However, most V2G research focuses on passenger cars, which typically already have their own specific usage purposes and various traveling schedules. This situation may pose practical challenges in providing ancillary services to the grid. Conversely, electric school buses (ESBs) exhibit a more predictable usage pattern, often deployed at specific times and remaining idle for extended periods. This makes ESBs more practical for delivering V2G services, especially when prompted by incentive price signals from grid or utility companies (UC) requesting peak shaving services. In this paper, we introduce a V2G energy sharing model focusing on ESBs in various schools in a single community by formulating the problem as a leader–follower game. In this model, the UC assumes the role of the leader, determining the optimal incentive price to offer followers for discharging energy from their battery storage. The UC aims to minimize additional costs from generating energy during peak demand. On the other hand, schools in a community possessing multiple ESBs act as followers, seeking the optimal quantity of discharged energy from their battery storage. They aim to maximize utility by responding to the UC’s incentive price. The results demonstrate that the proposed model and algorithm significantly aid the UC in reducing the additional cost of energy generation during peak periods by 36% compared to solely generating all electricity independently. Furthermore, they substantially reduce the utility bills for schools by up to 22.6% and lower the peak-to-average ratio of the system by up to 9.5%.

Accurate electricity demand forecasting serves as a vital planning tool, enhancing the reliability of management decisions. Apart from that, achieving these aims, particularly in managing peak demand, faces challenges due to the industry’s volatility and the ongoing increase in residential energy use. Our research suggests that employing deep learning algorithms, such as recurrent neural networks (RNN), long short-term memory (LSTM), and gated recurrent units (GRU), holds promise for the accurate forecasting of electrical energy demand in time series data. This paper presents the construction and testing of three deep learning models across three separate scenarios. Scenario 1 involves utilizing data from all-day demand. In Scenario 2, only weekday data are considered. Scenario 3 uses data from non-working days (Saturdays, Sundays, and holidays). The models underwent training and testing across a wide range of alternative hyperparameters to determine the optimal configuration. The proposed model’s validation involved utilizing a dataset comprising half-hourly electrical energy demand data spanning seven years from the Electricity Generating Authority of Thailand (EGAT). In terms of model performance, we determined that the RNN-GRU model performed better when the dataset was substantial, especially in scenarios 1 and 2. On the other hand, the RNN-LSTM model is excellent in Scenario 3. Specifically, the RNN-GRU model achieved an MAE (mean absolute error) of 214.79 MW and an MAPE (mean absolute percentage error) of 2.08% for Scenario 1, and an MAE of 181.63 MW and MAPE of 1.89% for Scenario 2. Conversely, the RNN-LSTM model obtained an MAE of 226.76 MW and an MAPE of 2.13% for Scenario 3. Furthermore, given the expanded dataset in Scenario 3, we can anticipate even higher precision in the results.

The penetration level of renewable energy sources is increasing worldwide with incentives and subsidies for declining greenhouse gas emissions. Nevertheless, determining the optimal location and size of renewable distributed generators (RDGs) remains a challenging task, owing to the uncontrollable reactance that dominates power distribution networks in voltage control and its sensitivity to weather conditions. Hence, without considering the reactive compensation of generators, RDG integration incurs undesired total power losses and puts the system at risk for voltage instability and collapse. This research proposes Load Disabling Nodal Analysis for Robust Voltage Stability (LDNA-RVS), a method that determines the optimal location and size of RDGs and aims to improve robust voltage stability by considering reactive compensation while enhancing the loss reduction efficiency (LRE) of the RDG integration. The proposed LDNA-RVS method has been applied to the IEEE 33-bus and IEEE 69-bus test distribution systems, and the results are compared with other techniques to prove its effectiveness in terms of robust voltage stability and LRE against the uncontrollable reactive compensation.

Electronic noses (E-Noses) are becoming popular for food and fruit quality assessment due to their robustness and repeated usability without fatigue, unlike human experts. An E-Nose equipped with classification algorithms and having open ended classification boundaries such as the k-nearest neighbor (k-NN), support vector machine (SVM), and multilayer perceptron neural network (MLPNN), are found to suffer from false classification errors of irrelevant odor data. To reduce false classification and misclassification errors, and to improve correct rejection performance; algorithms with a hyperspheric boundary, such as a radial basis function neural network (RBFNN) and generalized regression neural network (GRNN) with a Gaussian activation function in the hidden layer should be used. The simulation results presented in this paper show that GRNN has more correct classification efficiency and false alarm reduction capability compared to RBFNN. As the design of a GRNN and RBFNN is complex and expensive due to large numbers of neuron requirements, a simple hyperspheric classification method based on minimum, maximum, and mean (MMM) values of each class of the training dataset was presented. The MMM algorithm was simple and found to be fast and efficient in correctly classifying data of training classes, and correctly rejecting data of extraneous odors, and thereby reduced false alarms.