A Clustering-Based Method for Quantifying the Effects of Large On-Grid PV Systems

Abstract

Analyzing the impacts of large on-grid photovoltaic (PV) systems on the performance of the electric network is an essential task prior to the installation of these systems. To quantify these impacts, a method based on chronological simulations can be used. The main advantage of this method is its ability to provide information about the impacts of the fluctuation of the power generated from the PV systems. However, this method requires performing extensive analysis and simulations, making it impractical for utility studies, especially if long historical data with subhourly time resolution is used. In this paper, a new method that utilizes the data efficiently while preserving the temporal information of the generated PV power is proposed. The method takes advantage of the clustering techniques to group together segments of the output PV power having similar features. Hence, a representative segment for each group can be chosen and used in the analysis and simulations. This representative segment can provide information about the expected performance of other segments of the group. The validity and usefulness of the proposed method are demonstrated by identifying the suitable size and site of a large PV system.

Integrating short term variations of the power system into integrated energy system models: A methodological review

“…More advanced approaches make use of clustering algorithms to cluster days with similar load, wind speed and solar irradiance patterns. Different clustering algorithms, such as Ward's hierarchical clustering algorithm [75], the kmedoids [86], k-means [87][88][89] and fuzzy C-means algorithm [86] have been applied in this regard. Once all days are grouped into a number of clusters, a single representative day is selected from each cluster.…”

Grey-box identification for photovoltaic power systems via particle-swarm algorithm

“…Furthermore, the aforementioned equations of PV modules require numerical solution and thus are sometimes replaced with simplified equations that relate the power output with the efficiency of the system and variation in radiation and temperature [19], [20]. These equations are basically a translation of performance measurement from standard test measurements (STC; Air Mass 1.5 spectrum with global irradiance (G=1000W/m 2 and module temperature = 25 o C).…”

“…Another version of equation (7) is given below [20]: (4) G t is the global irradiance on the titled surface in W/m 2 , K T is thermal derating coefficient of the PV module in %/ o C, A a area of the PV array in m 2 , m is the module efficiency, dust is 1-the fractional power loss due to dust on the PV array, mis is 1-the fractional power loss due module mismatch, DCloss is 1-the fractional power loss in the dc side, MPPT is 1-fractional power loss due to the MPPT algorithm, T C is the cell temperature in o C, T ao is the ambient temperature at STC conditions in o C. The ac power of the PV system is then estimated by using manufacturer's efficiency curve of three phase inverter. The simplified PV equation adopted for this work is given below [21]:…”

Grey-Box Modeling for Photo-Voltaic Power Systems Using Dynamic Neural-Networks

“…However, these equations require numerical solution and thus are sometimes replaced with simplified equations that relate the power output with the efficiency of the system and variation in radiation and temperature [9], [10]. These equations are basically a translation of performance measurement from standard test measurements (STC; Air Mass 1.5 spectrum with global irradiance (G=1000W/m 2 and module temperature = 25 o C).…”