Identifying defects in electrical power systems during field inspections is a difficult task, since faults are generally not visible and may be intermittent. To find possible adverse conditions, specific inspection equipment is used. The ultrasound detector is the equipment normally used to inspect outdoor insulating systems; however, using it demands operator experience. To improve the defect condition classification, artificial intelligence techniques are applied to assist the operator in the decision task and thereby facilitate the identification of faulty insulating devices in the grid. The training of artificial neural network (ANN) models is an important step in solving the classification problem. This study aims to evaluate the training capacity in terms of the performance of different optimisation methods for the calculation of the mean square error after convergence. Traditional methods such as Gradient Descent and its variations will be presented, as well as methods that employ high computational effort such as quasi‐Newton and Levenberg–Marquardt. In order to base these concepts, a review will be presented on the use of these algorithms and on the problem of classification of insulators in distribution networks. The results show that there is a considerable performance difference between the calculation methods.

The generation of electric energy by photovoltaic (PV) panels depends on many parameters, one of them is the sun’s angle of incidence. By using solar active trackers, it is possible to maximize generation capacity through real-time positioning. However, if the engines that update the position of the panels use more energy than the difference in efficiency, the solar tracker system becomes ineffective. In this way, a time series forecasting method can be assumed to determine the generation capacity in a pre-established horizon prediction to evaluate if a position update would provide efficient results. Among a wide range of algorithms that can be used in forecasting, this work considered a Neuro-Fuzzy Inference System due to its combined advantages such as smoothness property from Fuzzy systems and adaptability property from neural networks structures. Focusing on time series forecasting, this article presents a model and evaluates the solar prediction capacity using the Wavelet Neuro-Fuzzy algorithm, where Wavelets were included in the model for feature extraction. In this sense, this paper aims to evaluate whether it is possible to obtain reasonable accuracy using a hybrid model for electric power generation forecasting considering solar trackers. The main contributions of this work are related to the efficiency improvement of PV panels. By assuming a hybrid computational model, it is possible to make a forecast and determine if the use of solar tracking is interesting during certain periods. Finally, the proposed model showed promising results when compared to traditional Nonlinear autoregressive model structures.

Reliability in the electric power system is fundamental to the development of society, for which rapid and accurate methods of fault identification are required. Faults in distribution insulators are hardly visible and the fault behavior is often intermittent, which makes its diagnosis a difficult task. Fault diagnosis with the ultrasound equipment has been used efficiently since this equipment is directional and not influenced by sunlight. However, the interpretation of the signal generated by this equipment requires an experienced operator and they are also susceptible to provide false diagnostics. The use of advanced algorithms to classify electrical system conditions has been proven as a great alternative to automate operator decisions. This article proposes the use of artificial intelligence algorithms such as single-layer and multilayer Perceptron for classification of distribution insulators conditions. The use of artificial neural networks for insulator classification is an innovative subject. Some researchers have already worked on partial discharges however not specifically for fault classification in insulators of distribution networks. The application of this technique can make the inspection of the electrical system automated and, in this way, more accurate and efficient. The results of the analysis showed that the application of signal linearization technique joint with artificial intelligence is a good alternative to locate faults in insulators.

Purpose: This paper aims to propose a robust kinematic controller based on sliding mode theory designed to solve the trajectory tracking problem and also the formation control using the leader–follower strategy for nonholonomic differential-drive wheeled mobile robots with a PD dynamic controller.Design/methodology/approach: To deal with classical sliding mode control shortcomings, such as the chattering and the requirement of a priori knowledge of the limits of the effects of disturbances, an immune regulation mechanism-inspired approach is proposed to adjust the control effort magnitude adaptively. A simple fuzzy boundary layer method and an adaptation law for the immune portion gain online adjustment are also considered. An obstacle avoidance reactive strategy is proposed for the leader robot, given the importance of the leader in the formation control structure.Findings: To verify the adaptability of the controller, obstacles are distributed along the reference trajectory, and the simulation and experimental results show the effectiveness of the proposed controller, which was capable of generating control signals avoiding chattering, compensating for disturbances and avoiding the obstacles.Originality/value: The proposed design stands out for the ability to adapt in a case involving obstacle avoidance, trajectory tracking and leader–follower formation control by nonholonomic robots under the incidence of uncertainties and disturbances and also considering that the immune-based control provided chattering mitigation by adjusting the magnitude of the control effort, with adaptability improved by a simple integral-type adaptive law derived by Lyapunov stability analysis.

The control architecture for Connected and Automated Vehicles (CAVs) proposed in this work combines three features: Scalable, Distributed, and Reconfigurable. The term Scalable is used in the sense that the number of CAVs can change in the runtime. The term Distributed is used in the sense that the control is totally embedded in the CAVs, without a central manager/coordinator. And the term Reconfigurable is used in the sense that the CAVs may change their destinations and paths at any time. In this architecture, we assure the nonblocking and collision-free behavior of the system, and also the privacy of the path for each CAV. The control structure is composed of a local plant, a path controller, and a coordination controller, all modeled by automata. The dispute for road occupation between CAVs is handled to avoid collisions, based on a wireless communication where each CAV imposes disablings of events (actions) to other CAVs. We propose an online treatment for deadlock avoidance, with no need to compute a monolithic controller in any phase of the design. The path and coordination controllers are proposed to be embedded and synthesized in the CAV at runtime, which allows scalability and reconfiguration. The architecture is validated through a simulation environment in Robotarium and a practical experimental testbed with Lego Mindstorms robots. Two examples are given, as well as three simulations and three experiments are executed. One of the simulations was run for 54h, varying from 1 to 36 CAVs and the architecture is most efficient at 22.2% of the CAVs occupation rate on the map. Results show that all requirements were met. Note to Practitioners —This paper was motivated by the problem of controlling a large-scale system of CAVs, such as big cities. In these kinds of systems, vehicles are starting and stopping their routes at any time and it is desirable that the control system be scalable . Furthermore, it is unfeasible to implement on central controller/coordinator in such a large system, making it desirable to have a distributed control system, embedded in each vehicle. Finally, due to the dynamics of traffic, such as congestion, it is desirable that the control system be reconfigurable . Existing control techniques have some of these features but not all of them in one methodology. In this paper, we propose one control architecture for CAVs, which comprises all three features at runtime. With these features, a CAV may, for example, start its route at any time (scaling the system) and change it at any time (reconfiguring) without interfering in the control of other CAVs. The proposed architecture supports many solutions for CAVs that can be modeled jointly, such as intersection management, road merging, path planning, and roundabouts. In the article, we develop the mathematics of the control and apply it to simulations and experiments. Results show that the proposed architecture can be applied in practice. Our codes for the simulations and experiments are shared in a public repository.