Classic and intelligent techniques aim to locate and track the maximum power point of photovoltaic (PV) systems, such as perturb and observe (P&O), fuzzy logic (FL), artificial neural networks (ANNs), and adaptive neuro-fuzzy inference systems (ANFISs). This paper proposes and compares three intelligent algorithms for maximum power point tracking (MPPT) control, specifically fuzzy, ANN, and ANFIS. The modeling of a single-diode equivalent circuit-based 3 kWp PV plant was developed and validated to achieve this purpose. Then, the MPPT techniques were designed and applied to control the buck–boost converter’s switching device of the PV plant. All three methods use the ambient conditions as input variables: solar irradiance and ambient temperature. The proposed methodology comprises the study of the dynamic response for tracking the maximum power point and the power generated of the PV systems, and it was compared to the classic P&O technique under varying ambient conditions. We observed that the intelligent techniques outperformed the classic P&O method in tracking speed, tracking accuracy, and reducing oscillation around the maximum power point (MPP). The ANN technique was the better control algorithm in energy gain, managing to recover up to 9.9% power.

A simple technique for obtaining the parameters of a linearized model for a didactic magnetic levitation system is introduced. The proposed procedure only involves basic concepts of dynamic systems and is suitable for use in undergraduate control experiments. The project and validation of a digital lead compensator based on the linearized model are also presented.

This work introduces a new fault detection method for photovoltaic systems. The method identifies short-circuited modules and disconnected strings on photovoltaic systems combining two machine learning techniques. The first algorithm is a multilayer feedforward neural network, which uses irradiance, ambient temperature, and power at the maximum power point as input variables. The neural network output enters a Sugeno type fuzzy logic system that precisely determines how many faulty modules are occurring on the power plant. The proposed method was trained using a simulated dataset and validated using experimental data. The obtained results showed 99.28% accuracy on detecting short-circuited photovoltaic modules and 99.43% on detecting disconnected strings.

Soft sensors based on deep learning have been growing in industrial process applications, inferring hard-to-measure but crucial quality-related variables. However, applications may present strong non-linearity, dynamicity, and a lack of labeled data. To deal with the above-cited problems, the extraction of relevant features is becoming a field of interest in soft-sensing. A novel deep representative learning soft-sensor modeling approach is proposed based on stacked autoencoder (SAE), mutual information (MI), and long-short term memory (LSTM). SAE is trained layer by layer with MI evaluation performed between extracted features and targeted output to evaluate the relevance of learned representation in each layer. This approach highlights relevant information and eliminates irrelevant information from the current layer. Thus, deep output-related representative features are retrieved. In the supervised fine-tuning stage, an LSTM is coupled to the tail of the SAE to address system inherent dynamic behavior. Also, a k-fold cross-validation ensemble strategy is applied to enhance the soft-sensor reliability. Two real-world industrial non-linear processes are employed to evaluate the proposed method performance. The obtained results show improved prediction performance in comparison to other traditional and state-of-art methods. Compared to the other methods, the proposed model can generate more than 38.6% and 39.4% improvement of RMSE for the two analyzed industrial cases.