This paper demonstrates the practical implementation of the State of Charge (SoC) measurement of lithium-ion batteries employing the Coulomb Counting Method (CCM). A low-priced microcontroller is used to implement the measurement process. The measured SoC for four different battery discharge profiles are obtained and plotted. The measured results are corroborated with computed values using Python. Further, the performance indices of the system are computed and provided and found to be satisfactory.

This research article proposes a distributed maximum power point tracking (MPPT) strategy in a dual-faced open cubic photovoltaic (DFOC-PV) system using K-means clustering. The system is designed to harvest solar energy continuously throughout the day, independent of the sun's direction, thereby eliminating the need for mechanical tracking mechanisms. Due to its compact and direction-free design, the proposed architecture is highly suitable for space-constrained environments. In contrast, conventional flat-panel PV systems are limited by large area requirements for comparable power output. This work comprehensively investigates all feasible array configurations of the DFOC-PV system under natural partial shading, which inherently occurs due to its structural design. The results indicate that the integration of a distributed maximum power extraction (DMPE) circuit significantly improves the energy yield across all configurations. The hardware experiments demonstrate that, for an identical land area, the proposed DFOC-PV system delivers higher energy output than a conventional flat-panel PV system under sunny, partly cloudy, and cloudy conditions. To further enhance performance under nonuniform irradiance conditions, the 28 PV panels in the proposed DFOC-PV system are grouped using K-means clustering algorithm, based on irradiance similarity. Clustering reduces current mismatch losses and improves overall energy harvesting by up to 2.1% under partial shading. Finally, a techno-economic analysis confirms the feasibility of the proposed system, yielding a lower levelized cost of energy ($\NewRupeeSymbol$13.28/kWh), a higher internal rate of return (14%), and a decreased energy payback period compared to the conventional PV configuration.

Dynamic line rating (DLR) is a smart-grid technology used to enhance the power transmission capacity of a transmission line. Time series forecasting, such as DLR forecasting, faces several challenges, including capturing complex relationships among multiple weather variables and modelling long-term spatiotemporal interactions in a faster and computationally efficient manner. Additionally, delivering interpretable predictions and adhering to physical standards complicate the process. This study investigates the feasibility of a novel DLR forecast model that utilizes self-attention in conjunction with a temporal convolutional neural network (TCN), ensuring an accurate and computationally efficient DLR forecast. The proposed model's DLR forecast is validated using historical weather data and line data collected from four locations along a 765 kV transmission line in the southern grid of India. Furthermore, the performance of the proposed model is compared with the state-of-the-art machine learning (ML) and deep learning (DL) models. The results demonstrate that the proposed model achieves up to 40% reduction in root mean square error (RMSE) and symmetric mean absolute percentage error (SMAPE) compared to bi-directional gated recurrent unit with attention (BIGRU-Attention), and approximately 28% improvement in coefficient of determination (R2) compared to gated recurrent unit with attention (GRUAttention), while reducing training time by up to 18% compared to hybrid convolutional neural network and gated recurrent unit (CNN-GRU), which represents the closest-performing baseline in terms of computational efficiency. Overall, the proposed hybrid model, which incorporates self-attention layers with TCN, yields more accurate, interpretable, and computationally efficient DLR forecasts.

The Minimization of Drive Test (MDT), an integral feature of the Long Term Evolution (LTE) network, facilitates coverage estimation of eNodeBs using measurement reports collected from User Equipment (UE). It eliminates the need for traditional drive tests and becomes a crucial feature for detecting possible network issues. Despite its potential, the implementation of MDT is hindered by challenges such as the sparse availability of user reports and inaccuracies in user positioning. This study employs a Convolutional Long Short-Term Memory (CNN2D-LSTM) model to estimate cellular signal coverage, identify network issues, and provide solutions for LTE signal optimization. The proposed methodology utilizes MDT data collected from UE through a custom-made Android app, referred to as iMDT (informed MDT). The acquired data undergo meticulous filtering based on predefined rules to eliminate inaccuracies. To demonstrate the efficacy of the proposed CNN2D-LSTM hybrid model, a comparative analysis is conducted against five other deep learning models: CNN1D-LSTM, CNN1D-BiLSTM, CNN2D-BiLSTM, LSTM, and CNN2D. The proposed iMDT measurement is performed in a live 2.1 GHz LTE network. In contrast to conventional MDT, which often produces a large volume of data with substantial positioning errors, iMDT notably minimizes data generation and achieves 41.36% of valid data in the resulting dataset.

The Extended Kalman Filter (EKF) is one of the most widely adopted model-based estimation technique employed for accurate estimation of State of Charge (SOC) of lithium-ion batteries. While EKF performs exceedingly well, its performance largely depends on judicial choice of process and measurement noise covariance matrices represented as Q and R respectively. Manual tuning of these covariance matrices is a challenging task due to the nonlinear characteristics of batteries and varying operating conditions, often leading to degraded estimation accuracy and slower convergence. To address this limitation, this paper proposes a Golden Jackal Optimization-based Extended Kalman Filter (GJO-EKF) for optimal tuning of the EKF covariance matrices. The tuning process is formulated as an optimization problem of minimizing the SOC estimation error with Q and R as variables to be optimized and then the steps of GJO method are followed. The proposed method is first evaluated under standard dynamic driving cycles, namely US06, PDMHC, and HWFET and the results are presented. The simulation results illustrate that the proposed GJO-EKF consistently outperforms the conventional EKF achieving RMSE and MAE values below 0.93% and 0.85% respectively, together with faster convergence and improved SOC tracking accuracy under dynamic operating conditions. Subsequently, the proposed algorithm is experimentally validated and measured results are presented in this paper. Experimental results further confirm the superiority of the proposed approach over the conventional EKF with RMSE and MAE values below 0.15% and 0.13% respectively, for all tested discharge profiles. The simulation and experimental investigations demonstrate that the proposed GJO EKF framework provides accurate, robust, and computationally efficient SOC estimation, making it suitable for real-time battery management system applications.

This work reports the practical implementation of ant colony optimized (ACO) extended Kalman filter (EKF) for state of charge (SoC) estimation of lithium-ion batteries (LiBs). The scheme is realized by employing a low-priced microcontroller. Two prominent parameters of EKF, namely, process (Q) and measurement (R) error covariances, are fine-tuned using the ACO algorithm. The problem of finding optimum values of Q and R together with SoC estimation is reformulated as an optimization problem. The proposed method has several advantages, such as computational simplicity, derivative-free operation, and near-optimal convergence with random initial guess. The performance, effectiveness, and robustness of the new SoC estimation are demonstrated for four different operating points of the tested battery. Measured mean square, mean absolute, and mean absolute percentage errors of 0.0165%, 1.2%, and 2%, respectively, with 99.34% R-squared value indicates that the proposed method is a promising candidate for the estimation of SoC of LiBs. Furthermore, the effectiveness of the ACO-tuned EKF approach is experimentally validated in comparison to the conventional EKF method.

This research focuses on the State of Charge (SoC) estimation of Lithium-ion batteries employing the Support Vector Machine (SVM) algorithm. The five hyperparameters of SVM are optimally identified using Particle Swarm Optimization (PSO). Four publicly available driving cycles, namely US06, LA92, CSHVC, and HWFET are employed for the optimization and subsequent evaluation of the proposed method. The first 70% of each dataset is used for the optimization of hyperparameters and the remaining 30% is used for the testing of the algorithm. The SVM with optimized hyperparameters is then employed for SoC estimation of each driving cycle at different temperatures and the results are presented. It is observed that the performance indices, namely RMSE, MAE, and R-2 values are far superior to the values in the available literature. Further, the percentage error in SoC is found to be below 2.5%, indicating improved performance of the proposed method.

This paper proposes a Golden Jackel Optimized (GJO) Support Vector Machine (SVM) algorithm towards the State of Charge (SoC) estimation of Li-ion batteries. All five hyperparameters of the SVM algorithm are optimally tuned using Golden Jackal Optimization. The proposed approach is evaluated for various publicly available driving cycles namely, LA92, HWFET, and PDMHC. The first 70% of each dataset is employed for the optimization of hyperparameters and the remaining 30% is used for the testing. Then the SVM algorithm with optimally tuned hyperparameters is then employed for the SoC estimation of same driving cycle at different temperatures. The performance indices namely RMSE, MAE, MSE, MAPE, and $R^{2}$ are computed and presented. Further, the percentage error in SoC is found to be below 2.5% for all test situations. The presented results demonstrate that the proposed approach is a promising tool for SoC estimation of Li-ion batteries.

The wireless power transfer (WPT) has gained popularity in the recent days for charging of various portable electronic gadgets such as robotic vacuum cleaners, pacemakers and also mobile phones. WPT is also largely employed for charging of electric vehicles. For a WPT system, two major components are transmitter and receiver which are mutually coupled together for power transfer. Here, the transmitter coil is positioned statically while the receiver coil is placed in the electronic gadget or in the electric vehicle. The structures of coupling coils are generally of circular, square and rectangular arrangements. The transmitter and receiver coils could be ideally completely aligned or misaligned depending upon the designed structures. The mutual inductance (MI) between the transmitter and receiver coils is a vital parameter which decides the wireless power transfer efficiency and for different coil structures. Hence, this parameter needed to be investigated for different coil structures with and without perfect alignment. This paper makes an attempt to compute the MI for different coil structures with and without perfect alignments employing Finite Element Method (FEM). For comparison, a prototype system was developed in the laboratory and the computed MI values are compared with measured values and there is a good correspondence between the two are observed. Further, the MI is measured for different arrangements and presented in this paper. It is observed that the MI of WPT system decreases with displacement between the two coils as well as the quantum of misalignment. Furthermore, square coil arrangement with perfect and various misalignment leads to higher MI.

Dynamic Line Rating (DLR) has gained significant attention in recent years due to its potential to optimize the utilization of transmission lines, improve system reliability, and reduce renewable energy curtailment. Therefore, forecasting the DLR accurately is essential. This research article proposes a novel approach for forecasting the DLR using the Temporal Convolutional Network (TCN). Here, TCN forecasts the weather parameters at the selected locations along the line. These forecasted weather parameters are utilized to compute the day-ahead DLR. The proposed method is applied to a transmission line in the southern Indian grid. TCN-based DLR computation method is compared with benchmark deep learning and machine learning models widely used in forecasting applications. The comparison shows that the proposed DLR computation method performs better than the other models. Incorporating weather forecasting using TCN into the line rating calculations enables a more accurate assessment of available transmission capacity under varying climatic conditions. DLR ensures 120% more ampacity than static line rating over 80% of the time.

The proposed work compares machine learning algorithms for fruit classification using apples, mandarins, oranges, and lemons. The goal is to identify the most accurate and precision-score algorithm. To assemble a robust dataset, we purchased several dozen oranges, lemons, and apples of various subtypes and meticulously recorded their mass, width, height, and color score using nano semiconductor sensors. Using this recorded data statistical analysis is conducted for identifying the accurate fruit classification machine learning method. The accurate prediction rate is determined by subtracting the actual and anticipated values. K-Nearest Neighbors (KNN) shows superior performance, achieving accuracies of 0.989, 0981 and 0.979 on training, validation and testing sets. The performance of the KNN algorithm combined with the W-H-CS feature combination technique is highly dependent on the choice of k and relevance of the selected features.

Battery technologies, a crucial element of contemporary energy storage systems, have extensive use in several industries including electric cars, portable gadgets, and grid storage. The identification of the different problems associated with batteries is critical to ensure their reliability, performance, and safety. Traditionally, many techniques rely on hardware-based solutions and form the basis of early battery management systems (BMS), which have significant limitations concerning accuracy, flexibility, real-time operation, and scalability. These deficiencies call for more advanced, data-based methods, such as machine learning, which significantly raise the performance in fault detection and reliability while reducing costs. This paper reviews the progress in battery technology and fault detection techniques with emphasis on the transformational role of machine learning (ML) in enhancing the capabilities of the battery management system. The machine learning methods raise the accuracy of fault detection and provide a means for constructing a safe and dependable battery system for many applications.

This article presents a dual-faced open cubic photovoltaic model for maximizing sunlight capture from multiple angles. The proposed model is directionless and will be placed on a horizontal surface. The open cubic structure allows for more efficient use of space, making it ideal for areas with limited land availability. This is particularly beneficial in urban environments or other constrained spaces, such as the rooftops of high-rise buildings, where maximizing power generation per square meter is crucial. A prototype model of four cubical structures with relevant circuitry has been developed using squared photovoltaic (PV) panels. Partial shading and irradiation mismatches in the proposed model are addressed using two types of hardware circuitry, namely the multi-input single-output buck converter and the single-input single-output buck converter, along with the conventional perturb and observe maximum power point tracking algorithm. The system efficiency of the proposed model is benchmarked against that of a conventional PV system covering the same land area. A techno-economic analysis is carried out, and key indicators such as the levelized cost of energy, net present value, internal rate of return, and energy payback period are estimated using PVsyst software.

Standardization is essential for any technological advancement to have widespread adoption, ensure interoperability, enhance compatibility and create a consistent user experience across various platforms and sectors. In industries, Programmable Logic Controllers (PLCs) and Human-Machine Interfaces (HMIs) play a crucial role in automating processes, improving operational efficiency, ensuring safety and providing real-time monitoring and control. The evolution of the Internet of Things (IoT) has significantly improved the industrial automation process to a higher level by enhancing the connectivity and data-sharing capabilities of PLCs and HMIs, enabling predictive maintenance and remote management. To meet the increasing demand and seamless integration of diverse industrial devices, many modern PLCs have built-in support for IoT implementation (IoT-ready PLCs) using industrial communication protocols like Ethernet Industrial Protocol (Ethernet/IP), PROFINET, Modbus, etc., with a standard interface. On the other hand, many old PLCs (legacy PLCs) continue to remain influential in industrial automation, managing complex processes. However, they lack built-in support for IoT integration due to different communication interfaces, proprietary protocols, and obsolete technology. Interconnecting these legacy PLCs under a shared network for data exchange, diagnosis, and remote monitoring in the industrial setup is challenging. It demands a huge capital expenditure to replace the existing legacy PLC infrastructure with IoT-ready PLCs. This challenge needs to be addressed through standardization and middleware solutions. This paper presents the integration of different PLCs deployed at diverse locations for the Industrial Lighting Management System (LMS) with a centralized HMI through a shared network using IoT Technologies.

A large volume of solar energy dissemination in a supply grid originates extreme variations in the load, resulting in a duck-form load arc that can cause stability issues. Also, the cost of energy consumption is found to vary between the off-peak and peak loads observed in the duck-shaped load curve. Accurate load forecast and demand response program is a key task for duck curve management in a distribution structure. Hence, this work proposes a Demand Response (DR) program using deep learning neural networks namely Long Short-Term Memory (LSTM). The proposed DR program is implemented in a modified 12 bus radial distribution network for duck curve management, where voltage stability is taken care of simultaneously minimizing the electricity cost in an energetic pricing environment. LSTM is used for forecasting the load and linear programming is used for load shedding. Therefore, this paper resolves the dual aims of flattening the duck-shaped load arc and minimizing electricity costs by combining them into a single objective function.

In a dynamic pricing scheme at a secondary distribution network, the residential electricity cost minimization and effective distributed energy resources management with energy storage (ES) is challenging using either compressed air or pumped hydro. Since, each one has its inherent limitations along with merits like ES with compressed air requires a very deep air storage cavern which results in increased constructional cost. Whereas ES with pumped hydro suffers from gravitational issues and also it requires suitable construction to maintain height difference. In particular, with the benefits of energy storage and solar integration, all residential houses lumped in the distributed network are willing to draw much of their load demand at minimum pricing intervals. This situation may lead to crowding phenomena among the residential houses. In order to overcome the crowding phenomena, in this paper a novel energy routing technique with hybrid energy storage has been proposed. The hybrid energy storage is a combination of compressed air energy storage and pumped hydro storage. The proposed energy routing technique assigns the role of each house as a seller or buyer by identifying the surplus or paucity power along with the status of their energy storage. A priority factor is calculated, in this past contribution made by each residential house and load demand are considered to be key parameters. Further, the linear integer program is employed to schedule the buyers to buy the electrical energy at minimum price without any crowding phenomena. The performance of the proposed novel technique has been validated on12 bus distribution system with 100 MVA, 11 kV ratings and IEEE 33 bus test system with 100 MVA, 12.66 kV ratings. The result analysis shows that the proposed novel technique is robust and applicable for real-time smart grid environment.

Adequate illumination is integral to performing any activity in manufacturing industries. In large-scale industries, various capacities and quantities of high bay luminaires (roof lights) are installed and distributed over the factory roofs to achieve the required illumination. Such enormous quantum is traditionally controlled by grouping them under multiple digital time switches distributed at diverse locations. The arduous activity of frequent rescheduling of time switches on a need basis is addressed by adopting a centralised controller called the Lighting Management System (LMS). However, the LMS lacks individual control of each distribution point (point-based) of roof lights due to limited control wires, leading to substantial energy wastage. This paper presents the design and implementation of point-based switching control using the Modbus Remote Terminal Unit (RTU) protocol on the existing control cable.

This article proposes an improved control strategy for a multifunctional unified active power filter (UAPF) based hybrid AC/DC microgrid system. Here, a hybrid microgrid incorporates both AC/DC sources and loads through a UAPF. The associated control strategy for the converters enables to enhance the power quality (PQ) with a flexible bidirectional power flow. The shunt converter (STC) and series converter (SRC) will distribute the power in the controlled current source and controlled voltage source, respectively. The reference quantities for the converters are based on a computation of powers and inject both sinusoidal and non-sinusoidal quantities based on system conditions. In this, optimal phase angle control is included for optimal utilization of converters and their rating. Additionally, the automatic switching transition between gridconnected to islanded mode and vice versa is used for seamless power transfer and continuous operation of the system. For this, control logic is developed based on frequency, phase angle and voltage. During islanded mode, the system generates a reference phase angle and frequency that match the load demand. Finally, the proposed system is developed in a 1kVA prototype laboratory test bed and its controller is designed in the FPGA platform for parallel processing of the converters. Its performance and effectiveness are validated in different operating scenarios based on the type of load, grid voltage fluctuations, available power at the DC link, etc. To show the significant effectiveness of the proposed system, in different scenarios, the various parameters of each side are collected through a data logger and are presented.

The most versatile resource for storing energy is one that can rapidly charge or discharge while supporting the use of renewable energy. As renewable energy sources advance rapidly, batteries play a pivotal role in this progress. When integrating battery energy storage into a renewable energy system, it’s crucial to address the issue of battery degradation while implementing operational strategies. Furthermore, since solar irradiation varies due to changing cloud conditions, it can impact how batteries charge and discharge. This study focuses on investigating battery degradation and lifetime. Experimental work is being conducted with lead acid batteries connected to a solar photovoltaics system. The paper provides a detailed investigation of commonly used methods for predicting battery lifespan. It also analyzes aspects such as the effects of depth of discharge (DoD) and battery charge/discharge on temperature changes due to degradation. Using the coarse average approach, global battery aging, weighted Ah aging method, and RFC method, this paper estimates the DoD, temperature, life cycle loss (%), and lifespan and evaluates the extent of battery degradation. The battery lifespan is estimated using this method to be 8.42, 8.72, 8.33, and 8.93 years, respectively.

Solar tracking systems (STS) are essential to enhancing solar energy harvesting efficiency. This study investigates the effectiveness of STS for improving the energy output of Photovoltaic (PV) panels. Optimizing solar energy capture is crucial as the demand for renewable energy sources continues to rise. The research evaluates various types of STS, including passive, active, single-axis, dual-axis, hybrid, and models based solar tracker systems, and analyzes their performance under different environmental conditions. This paper explores the latest developments in STS, identifies challenges, and outlines potential advancements to promote the widespread adoption of solar tracking technologies. The comparison between STS and fixed solar panel systems shows a significant increase in energy production with STS. This paper underscores the potential of STS in advancing sustainable energy solutions and emphasizes the need for further research and development in this field.