Laboratory of Advanced Technology and Intelligent Systems, National Engineering School of Sousse, BP 264, Sousse Erriadh 4023, University of Sousse, Tunisia
Auto-generated profile, based on publicly available information
Ali Lahouar
Laboratory of Advanced Technology and Intelligent Systems, National Engineering School of Sousse, BP 264, Sousse Erriadh 4023, University of Sousse, Tunisia
This paper proposes two new open-circuit fault detection methods suitable for the diagnosis of power electronics converters. The faulty semiconductor devices are identified using conducted and radiated electromagnetic signatures of the dc bus through nonintrusive measurements. The first method uses a low-cost electromagnetic interference filter to collect the common-mode emissions signature. The second one utilizes an external antenna to collect the emitted near-field signature. Both methods are tested on a three-level neutral point clamped inverter (NPC) inverter with the aim to identify the clamping diodes open-circuit faults. Indeed, each open-circuit fault affects the common-mode emission signature in the time-domain, while the fast Fourier transform of the emitted near-field showed substantial reduction of spectrum amplitude at a specific radio frequency range and the appearance of a new spectral rail. The effectiveness of these two methods has been tested through numerical simulations and validated by experimental results which confirmed its high performance in detecting single as well as multiple open-circuit faults for three-level NPC inverter.
Short term load forecasting is one of the most important tasks for power suppliers, and it is getting more important with deregulation of electricity market and emergence of smart grids. This paper proposes a load prediction model of one day ahead with resolution of one hour, using regression random forests. With information about season, temperature, type of the day and hourly load, a training process is performed to build the adopted model. A real load data set from Tunisian Power Company is used for test, and special attention is paid to the load profile which is specific to warm countries with excessive and unstable demand in summer. The results reflect accuracy and effectiveness of the proposed method, keeping low prediction error for long test periods.
With the great recent moves towards green energy exploitation worldwide, the solar photovoltaic (PV) power has gained much attention. Thanks to PV panels' cost drop and recent improvements in energy conversion systems, the PV installations are getting more and more integrated into power plants. Because of high correlation with weather conditions, accurate short-term PV output forecast is highly recommended. An accurate prediction is needed to assess the effective contribution of solar energy in the grid, and to overcome the problems of intermittence. This paper proposes a day-ahead prediction method of PV output, which estimates the power generated by solar panels with and without prior knowledge of solar irradiance. The proposed model is the random forest using bagging algorithm, characterized by built-in cross validation and immunity to irrelevant inputs. A special attention is paid to the choice of most influential weather conditions on future power. The proposed approach is validated through tests on real data from PV sites in Australia.
Predicting wind speed and direction is one of the most important and critic tasks in a wind farm, since wind turbine blades motion and thus energy production is closely related to wind behaviour. Machine learning techniques are often used to predict the non-linear wind evolution. In this context, this paper proposes a short term wind data prediction model based on support vector machines in their regression mode, which have the advantage of being simple, fast and well adapted for the short term. This research tries also to prove how wind direction may influence power generation, and why it is important to predict it. A real data set of wind speed and direction historical values is used, from Sidi Daoud wind farm, north-eastern Tunisia, in order to evaluate the proposed model. This forecasting system predicts wind speed and direction for the short term, from one to 10 hours in advance, using a set of past samples.