• PC Simulation and Sizing of PV Systems, ISES Solar World Congress, Budapest 1993
• PVDIM: PC Program for PV Simulation and Sizing, 12th EPVSEC, Amsterdam, 1994
• Simulation et Dimensionnement Photovoltaïque: Critères pour la Garantie des Résultats, COMPLES Symposium on Calculation methods for solar energy systems, Perpignan (F), 9/1994
• PV and controlled nuclear fusion: a case study in European energy research, 13th EPVSEC, Nice, 1995
• On PV simulation tools and sizing techniques: a comparative analysis towards a reference procedure, 13th EPVSEC, Nice, 1995
• Stand-alone PV guarantee of results: a validated statistical methodology, ISES EuroSun Conference, Freiburg, 1996
• Preliminary validation of synthetic radiation time series by means of PV simulation, 14th EPVSEC, Barcelona, 1997
• Dal sole elettricità senza fine, AAM Terra Nuova, 11/1997

Les méthodes de dimensionnement photovoltaïque utilisées couramment sont basées principalement sur des bilans énergétiques des systèmes PV, dont le caractère intrinsèquement non linéaire est négligé. De plus ces techniques ne permettent pas l'analyse de la fiabilité des installations, nécessaire pour étendre aux systèmes PV la garantie des résultats.

Une procédure de simulation à pas de temps fin a été proposée [4] dans le but d'optimiser les coûts des installations et d'accroître le niveau de confiance dans les performances des systèmes PV, aussi bien autonomes que raccordés au réseau.

Comme la production d'énergie d'un système PV est le résultat d'une chaîne de processus stochastiques, nous avons analysé la sensibilité des résultats de la simulation aux divers paramètres caractéristiques.

Nous avons évalué le poids des fluctuations météorologiques autour des années moyennes, de la répartition diurne/nocturne et des fluctuations statistiques de la charge, de la taille du générateur et du stockage sur les performances attendues, c'est à dire sur le taux de couverture des besoins envisagé lors du projet d'un système autonome ou du coût du kWh pour les systèmes raccordés au réseau.

Nous avons ainsi montré quels sont les points clés d'un système PV pour des conditions limites données et quelles sont les mesures à prendre dans les buts : a) d'augmenter la disponibilité du système ; b) d'optimiser les bilans énergétiques, soit l'adéquation entre les composants; c) d'optimiser les coûts d'investissement.

European Union has been sponsoring researches in the domains of nuclear fusion and PV, with the aim of providing a secure and environmentally benign energy source. We have analysed the EU energy requirements and compared the present status, the expectations, the environmental impact, the energy pay-back time and the costs of the two energy technologies, taking into account the different status of maturity. A set of recommendations has been proposed to address the energy policy of EU and to promote the widest dissemination of RE.

Sizing techinques that have been proposed so far don’t allow any precise estimation of the system yield. We have showed how to combine the rules of thumb techniques with a precise hourly simulation to assess on the sensitivity on the local climatology of some leading parameters of a stand-alone PV system. We have also analysed the optimal matching of the array size and the inverter nominal power to enhance the average efficiency of grid connected systems in various European sites.

Stand-alone PV systems performances and reliability are dramatically affected by the sizing rules used for their design. The common attitude of systems conceivers to tackle the problem is the system oversizing, which actually reduces the risk of energy shortage but increases the cost of the system and precludes a quantitative assessment on the probability of energy shortage (Loss Of Load Probability). 
Our work is intended to help solving the problem of LOLP estimation by means of a thorough analysis of the reliability of a stand-alone PV system, carried out with Monte Carlo simulation techniques. We have deduced some empirical rules to assess on the LOLP that a system is likely to meet during its operations for 2 French sites, covering thus very different climatologic areas (continental and Mediterranean climates).
An innovative approach we have implemented in our simulation is the use of a gaussian distribution of the expected load, rather than a constant value all the year around. We have thus shown the derating of reliability when taking into account just a constant load value.
We expect that the proposed methodology will help PV designers to evaluate the correct parameters for broad geographical areas (but homogeneous by a climatologic point of view), to be adjusted precisely when dealing with specific sites.

Detailed simulation is a mandatory requirement for the effectiveness of RE systems and RUE. In Southern European countries the availability of multivariate solar radiation data, necessary for an accurate simulation, is scarce. The JOULE III CLIMED Project is intended to provide methodologies for synthetic data generation. In this paper are presented the main results of the preliminary validation of some statistical features characterising solar radiation, by means of PV simulation. Monitored radiation and temperature data have been remixed and sorted so as to highlight one feature at a time and to find which parameters are relevant in the synthetic generation of data when applied to PV systems. Similar analysis will be performed to find out the relevant parameters for DHW systems and buildings thermal behaviour.