Leïla-Hayet Mouss

A smart algorithm for the diagnosis of short-circuit faults in a photovoltaic generator

This paper deals with a smart algorithm allowing short-circuit faults detection and diagnosis of PV generators. The proposed algorithm is based on the hybridization of a support vector machines (SVM) technique optimized by a k-NN tool for the classification of observations on the classifier itself or located in its margin. To test the proposed algorithm performance, a PV generator database containing observations distributed over classes is used for simulation purposes.

A regression algorithm for the smart prognosis of a reversed polarity fault in a photovoltaic generator

This paper deals with a smart algorithm allowing reversed polarity fault diagnosis and prognosis in PV generators. The proposed prognosis (prediction) approach is based on the hybridization of a support vector regression (SVR) technique optimized by a k-NN regression tool (K-NNR) for undetermined outputs. To test the proposed algorithm performance, a PV generator database containing sample data is used for simulation purposes.

Modeling the PV generator behavior submit to the open-circuit and the short-circuit faults

In this paper, we proposed a new mathematical model of a faulty photovoltaic generator operation. It presents its behavior, when its subjected to the open-circuit and the short-circuit faults at its basic components as: cells, bypass diodes and blocking diodes. Such kind of modeling will allow developing fault detection and diagnosis methods. Indeed, the proposed model will be used to set normal and fault operation conditions database, which will facilitate learning and classifications phases.

Faults modeling of the impedance and reversed polarity types within the PV generator operation

In this paper, we proposed a new mathematical model of the I-V characteristic of a faulty photovoltaic generator. It presents its behavior in normal and faulty operations. In particular, when its basic components such as cells, bypass and blocking diodes are subjected to the impedance or reversed polarity faults. The developed model of the faulty PV generator will allow studying of the I-V characteristic, measures the tolerances of the technical functions, avoids numerous experiments, and ensure better assessment of fault consequences.

Fuzzy Modeling of Single Machine Scheduling Problems Including the Learning Effect

In this chapter, we consider the single machine scheduling problem including uncertain parameters and position based learning effect with the aim to minimize the weighted sum of jobs completion times. Due to the ill-known quantities within the model, the determination procedures of optimal solutions in the conventional way is not an affordable task and more elaborated frameworks should be developed. In this context, we introduce two solution approaches for the proposed fuzzy scheduling problem in order to obtain an exact or a satisfactory near optimal solution. The first approach is based on the extension of the well-known Smith’s rule resulting in a polynomial algorithm with a complexity O(n l o g(n)). However, a severe constraint on jobs (fuzzy agreeability concept) should be satisfied in this case. The second approach based on optimization methods is built upon the assumption of unequal fuzzy release dates in addition to the absence of fuzzy agreeability constraint. Three trajectory based metaheuristics (Simulated annealing, taboo search and kangaroo search) are implemented and applied to solve the resulting problem. For the proposed methods throughout the chapter, several numerical experimentations jointly with statistical deductions are provided.

Smart diagnosis algorithm of the open-circuit fault in a photovoltaic generator

This article proposed a new smart diagnosis algorithm of the open-circuit fault in a PV generator. For the faults conventional diagnosis, it used the analysis of the actual operation parameters of the PV generator. For the faults smart diagnosis, it based on the optimization of SVM technique by the neural network for the classification of observations located on its margin. The resulting algorithm can ensure a better monitoring function of the open-circuit fault within the PV generator, with a high classification rate and a low error rate.