Electricity remains one of the most vital resources for industrial, domestic, and agricultural applications. However, electricity theft has emerged as a significant challenge, contributing to substantial power losses and severe economic repercussions for utility companies. This study examines the role of smart meters (SMs) in minimizing electricity theft and reducing energy losses by transitioning from traditional analogue meters to advanced SMs equipped with automated billing and metering systems. Data collected from the SM system in the Akre energy distribution network reveal that, following the implementation of SMs, overall electrical power losses were reduced by 17.1%, while theft incidents decreased by 96.4%. These results demonstrate that the deployment of SMs significantly contributes to lowering total power losses and yields considerable financial benefits for both utility providers (UPs) and consumers. Moreover, the system enhances the ability to remotely monitor and control customer meters, allowing continuous oversight of meter readings without requiring physical visits. This remote functionality strengthens theft prevention measures, improves grid reliability, and reduces operational costs. The findings highlight the potential of the SM system in advancing power efficiency and promoting a more secure and cost-effective energy distribution network.

The environmental conditions in large-scale, intensive poultry farming systems require high precision, and accurate prediction of environmental factors is critical for effective control. Existing control methods generally focus on the prediction and control of individual environmental factors without considering the interdependencies among these factors, leading to low prediction and control accuracy. To address the complex nature of the environmental system in poultry houses, characterised by multi-factor dependencies, an adaptive environmental control system based on Multi-feature Long Short-Term Memory (Multif-LSTM) was proposed. The Multif-LSTM model within the system calculates the dependencies between environmental factors using correlation coefficients and establishes a multi-input, multi-output neural network architecture. External climate factors are also incorporated during the input phase. Experimental comparisons conducted in a duck house environment, with Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) models, show that the Multif-LSTM model outperforms others in terms of prediction accuracy. For NH₃ concentration, the Root Mean Square Error (RMSE), the Mean Absolute Percentage Error (MAPE), and Coefficient of Determination (R²) values are 1.34, 8.3, and 0.55, respectively; for temperature, they are 0.29, 2.83, and 0.98; and for relative humidity, they are 1.73, 2.46, and 0.95, respectively. Compared to the average performance of the RNN and LSTM models, the RMSE is reduced by 2.5, MAPE by 4.6, and R² increased by 0.32. The results demonstrate that the Multif-LSTM model achieves higher prediction accuracy and is suitable for high-precision adaptive environmental control in poultry houses.

Indonesia, a significant exporter of coffee, faces persistent challenges in accurately identifying and classifying coffee varieties based on aromatic characteristics, primarily due to the subjective variability of human sensory evaluation. To address these limitations, an electronic nose (e-nose) system was developed for the classification of coffee varieties through the analysis of aromatic profiles. The system integrates a DHT-22 sensor and four gas sensors (MQ-5, MQ-4, MQ-3, and MQ-135) to measure humidity, temperature, and gas concentrations from coffee vapor. Data acquisition was facilitated by the Arduino Uno platform, while classification was conducted using the Naive Bayes Classifier (NBC) algorithm. The e-nose achieved a classification accuracy of 82.2%, as validated through a confusion matrix and performance metrics, including precision, recall, and F1-score. Among the gas sensors employed, the MQ-4 sensor, which detects methane, demonstrated the highest response sensitivity, whereas the MQ-3 sensor, designed to detect alcohol, exhibited the lowest. This system significantly mitigates the inherent subjectivity associated with traditional aroma assessment methods and offers considerable potential for enhancing quality control protocols in coffee production processes. Future work will focus on integrating advanced machine-learning algorithms, optimizing sensor array performance, and expanding the dataset to include a broader diversity of coffee samples. These advancements are expected to further refine the system's classification capabilities and contribute to more robust quality assurance in the coffee industry.