Energy demand forecasting is crucial for planning and optimizing the use of energy resources in building facilities. However, integrating digital solutions and learning techniques into legacy buildings presents significant challenges due to limited or outdated resources, hampering predictive analytics in these buildings and their circuits. To fill this gap, this article proposes an innovative demand forecasting strategy using an AIoT retrofit architecture based on the SmartLVGrid metamodel. This architecture allows remote monitoring of legacy building circuits, facilitating the collection, processing and storage of data in the cloud. We use several learning algorithms, including linear regression, support vector regressor, random forest regressor, XGBoost regressor, and long short-term memory (LSTM) neural network, to predict energy demand 15 min ahead, identifying potential overruns of contracted demand in accordance with Brazilian regulations. After Bayesian optimization, the LSTM neural network outperformed other models for most of the selected datasets and detected 32 out of 38 demand overruns on the test set. XGBoost and random forest followed closely, detecting 30 demand overruns. Overall, our cost-effective solution optimizes energy usage and efficiently mitigates potential demand exceedances in building installations. This is achieved through a step-by-step approach to upgrading existing aging facilities, which promotes energy efficiency and sustainability.

Este artigo apresenta um sistema elaborado para detectar e diagnosticar faltas em plantas industriais de tanque multivariável. Um modelo da planta foi desenvolvido em ambiente computaciona Matlab/Simulink. A identificação de uma falta é realizada por uma técnica multi-modelo, baseada em Neuro-Fuzzy. Nessa técnica os parâmetros do modelo da planta são estimados em tempo real e comparados com parâmetros de um banco de modelos identificados previamente em condições saudáveis e de falta. A partir dessa comparação foi gerado um resíduo, que, através do qual, o percentual da intensidade da falta foi diagnosticado. Com essa metodologia foi possível monitorar a ocorrência de faltas em sistemas industriais em tempo real, auxiliando a programação da manutenção desses sistemas.

The Internet of things (IoT) paradigm promotes the emergence of solutions to enable energy-management strategies. However, these solutions may favor the disposal or replacement of outdated but still necessary systems. Thus, a proposal that advocates the retrofit of pre-existing systems would be an alternative to implement energy monitoring. In this sense, this work presents a strategy for monitoring electrical parameters in real time by using IoT solutions, cloud-resident applications, and retrofitting of legacy building electrical systems. In this implementation, we adapted the SmartLVGrid metamodel to systematize the insertion of remote monitoring resources in low-voltage circuits. For this, we developed embedded platforms for monitoring the circuits of a building electrical panel and application for visualization and data storage in the cloud. With this, remote monitoring of the consumer unit was carried out in relation to energy demand, power factor, and events of variations of electrical parameters in the circuits of the legacy distribution board. We also carried out a case study with the proposed system, identifying events of excess demand in the consumer unit, mitigating the individual contribution of the installation circuits in this process. Therefore, our proposal presents an alternative to enable energy management and maximum use of existing resources.

Introduction: Electromagnetic interference caused by electric power lines adversely affects the signals of electronic instruments, especially those with low amplitude levels. This type of interference is known as common-mode interference. There are many methods and architectures used to minimize the influence of this kind of interference on electronic instruments, the most common of which is the use of band-reject filters. This paper presents the analysis, development, prototype and test of a new reconfigurable filter architecture for biomedical instruments, aiming to reduce the common-mode interference and preserve the useful signal components in the same frequency range as that of the noise, using the technique of dynamic impedance balancing. Methods: The circuit blocks were mathematically modeled and the overall closed-loop transfer function was derived. Then the project was described and simulated in the VHDL_AMS language and also in an electronics simulation software, using discrete component blocks, with and without feedback. After theoretical analysis and simulation results, a prototype circuit was built and tested using as input a signal obtained from ECG electrodes. Results: The results from the experimental circuit matched those from simulation: a 97.6% noise reduction was obtained in simulations using a sinusoidal signal, and an 86.66% reduction was achieved using ECG electrodes in experimental tests. In both cases, the useful signal was preserved. Conclusion: The method and its architecture can be applied to attenuate interferences which occur in the same frequency band as that of the useful signal components, while preserving these signals.