ANALYSIS AND EXPERIMENTAL VALIDATION OF VARIOUS PHOTOVOLTAIC SYSTEM MODELS

1. ELECTRIMACS 2002, 7th International Conference on Modeling and Simulation of Electric Machines, Converters and Systems August 18th through 21st 2002 The polynomial model The  one-diode model The  two-diode model ANALYSIS AND EXPERIMENTAL VALIDATION OF VARIOUS PHOTOVOLTAIC SYSTEM MODELS O. GERGAUD, B. MULTON, H. BEN AHMED LÉSiR – Brittany Branch,ENS de Cachan Ker Lann campus – 35170 BRUZ – FRANCE Phone:(33)2 99 05 52 64 ; Fax: (33)2 99 05 93 28 gergaud@bretagne.ens-cachan.fr PHOTOVOLTAIC CONVERSION CHAIN ENERGY ANALYSIS Within the framework of the optimization of the dimensionnement and\or the management of the transfers of energy of a joint photovoltaic and wind production system hooked up to the network via a continuous bus bar and associated with an electrochemical storage device One wishes to determine the characteristic of the power supplied in the continuous bus, Pbat, according to the weather and electric conditions, that is the illumination in the plan of the sun panels Esd, their temperature Tj, as well as the level of the tension of the continuous bus, Vbat. The model should be at the same moment enough handbook and enough express. comparison of measurements with GROUP 1 " two-diode model" over one day production period (Dec. 17, 2001) comparison of measurements with GROUP 1 "one-diode model" over one day production period (Jan. 24, 2002) ELECTRICAL MODELS OF THE PANELS • photocurrent : Iph = P1. Es. [1+ P2. ( Es – Eref ) + P3 . (Tj – Tjref) ] • polarization current of the junction PN : CHARACTERISTICS OF THE MPPT CONVERTERS • saturation current : • Efficency model : • Ps converter output power, • P0 empty converter losses, • K1 a coefficient representing the losses proportional to the square of the current ; • K2 losses proportional to the current. • final equation of the model : 7 parameters P1, P2, P3, P4, A, Rs and Rsh to be determined • P0 being measured, calculated K1 and K2 two diodes are present for the PN junction polarization phenomena 8 parameters P1, P2, P3, P4, A, Rs and Rsh to be determined COMPLETE SYSTEM • Simulation of the complete system on sevral days : measurements 1000 800 600 400 200 0 Insolation (W/m²) • final equation of the model : March 31, 02 April 1, 02 April 2, 02 April 3, 02 April 4, 02 mesures / simulation 25 20 15 10 5 0 Energy produced (kWh) March 31, 02 April 1, 02 April 2, 02 April 3, 02 April 4, 02 simulation-mesures 3 parameters P1, P2, P3 to be determined 0.1 0 -0.1 -0.2 -0.3 -0.4 Simulation error (kWh) DETERMINATION OF PARAMETERS VALUES - POWER ANALYSIS • Identification of the parameters by genetic algorithm • For both first models, resolutiin of an implicit function : • IGthéo = f ( IGthéo, VGréelle, Es, Tj ) IGthéo = f ( IGreèlle, VGréelle, Es, Tj ) • Minimization of the error : • For the model polynomial determination of the parameters from the maxima of the characteristics power / tension real of the PV: March 31, 02 April 1, 02 April 2, 02 April 3, 02 April 4, 02 • Errors on the energy produced during the March 30 - April 4 2002 recording campaign, Measurement/Model comparison : • Very good energy depiction of troid presented models • The errors identified lie below the 2% level, they may just as easily stem from insolation, temperature and power production measurements as they may from our models • Shorter time of calculations in the cas of the polynomial model The polynomial model has therefore been selected for subsequent use in modeling the photovoltaic chain as part of our system optimization efforts