Algoritem za celostno vrednotenje fotovoltaičnega in vetrnega potenciala večjih geografskih območij

doktorska disertacija

Niko Lukač (Author), Borut Žalik (Mentor), Gorazd Štumberger (Co-mentor)

Abstract

V doktorski disertaciji predlagamo nov algoritem za izračun fotovoltaičnega in vetrnega potenciala nad večjim geografskim območjem. Pri tem veliko geografsko območje predstavimo v topološki strukturi mreže, ki jo skonstruiramo iz visokoločljivostnih podatkov laserskega snemanja LiDAR. Najprej relativno, glede na lokacijo, izračunamo položaj Sonca, senčenje ter anizotropno obsevanje na površino, z upoštevanjem večletnih meritev direktnega in difuznega obsevanja. Izračunano trenutno vrednost globalnega obsevanja integriramo po času, pri čemer upoštevamo tudi od globalnega obsevanja nelinearno odvisno karakteristiko izkoristka (učinkovitosti) obravnavanega fotovoltaičnega sistema. V drugem delu doktorske disertacije predstavimo novo metodologijo za izračun vetrnega potenciala nad dano mrežo, kjer z računalniško simulacijo določimo vetrni tok kot skupek zračnih molekul z Lagrangeovo metodo hidrodinamike zglajenih delcev ter modelom turbulence temelječim na Reynoldsovim povprečenjem. Za izvedbo izračuna potencialnega gibanja vetra upoštevamo še večletne podatke meteoroloških meritev hitrosti in smeri vetra, s katerimi zgradimo logaritmični vetrni profil za določitev začetnih pogojev delcev vetra. Za izračun izhodne moči uporabimo nelinearno karakteristiko moči vetrne elektrarne, ki je odvisna od hitrosti vetra. Integracijo slednje po času uporabimo za napoved proizvodnje električne energije. Obe metodologiji združimo v predlagan algoritem, ki ga implementiramo na grafični procesni enoti s tehnologijo CUDA, s čimer dosežemo učinkovito paralelno izvajanje v doglednem času. V eksperimentalnem delu doktorske disertacije s primerjavo izračunanih vrednosti in neodvisnih meritev ovrednotimo natančnost in točnost izračuna glede na vhodne podatke. Z algoritmom tudi vrednotimo fotovoltaični in vetrni potencial večjega geografskega območja mesta Maribor. Pri fotovoltaičnem potencialu v povprečju dosežemo točnost glede na podatke meritev do 97%, pri vetrnem potencialu pa do 92%.

Keywords

modeliranje;računalniška simulacija;daljinsko zaznavanje;meteorološke meritve;podatki LiDAR;paralelno računanje;GPU;GPGPU;CUDA;hidrodinamika zglajenih delcev;

Data

Language:	Slovenian
Year of publishing:	2016
Typology:	2.08 - Doctoral Dissertation
Organization:	UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher:	[N. Lukač]
UDC:	004.925:551.521.1(043.3)
COBISS:	285053696
Views:	1546
Downloads:	270
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Algorithm for the determination of photovoltaic and wind potential over large geographic areas
Secondary abstract:	In this doctoral dissertation, a new algorithm is proposed for the estimation of photovoltaic and wind potential over large area, which is represented with a topological grid structure that is constructed from high-resolution laser-scanned LiDAR data. At first, the position of the Sun relative to the geographic location is calculated, as well as shadowing and anisotropic irradiance, by considering multiannual measurements of direct and diffuse irradiance. Calculated instantaneous global irradiance is integrated with respect to time using the global irradiance-dependent nonlinear efficiency characteristics of a given photovoltaic system. In the second part of the dissertation, a new methodology is presented for the estimation of wind potential over the constructed topological grid, in which the computer simulation of wind flow represented as clusters of air molecules is performed by using the Lagrangian method of smoothed particle hydrodynamics and Reynolds averaged turbulence model. In order to implement the estimation of wind flow, multiannual meteorological measurements of the wind velocity and direction are also considered, which are then used for constructing logarithmic wind profile that defines the initial conditions of the wind particles. In order to calculate the output power the nonlinear characteristics of wind systems are used, which depend on the calculated wind velocities. The calculated power values can then be integrated with respect to time in order to forecast electrical energy production. Both methodologies are merged into the proposed algorithm, which we implement on a graphics processing unit using CUDA technology, in order to perform the calculations in an acceptable time. Within the experimental part of the doctoral thesis, the accuracy of the calculation based on the input data is evaluated by comparing the calculated values with independent measurements. The proposed algorithm was also applied over the large geographical area of Maribor, in order to estimate its photovoltaic and wind potentials. The estimated photovoltaic potential has an agreement In this doctoral dissertation, a new algorithm is proposed for the estimation of photovoltaic and wind potential over large area, which is represented with a topological grid structure that is constructed from high-resolution laser-scanned LiDAR data. At first, the position of the Sun relative to the geographic location is calculated, as well as shadowing and anisotropic irradiance, by considering multiannual measurements of direct and diffuse irradiance. Calculated instantaneous global irradiance is integrated with respect to time using the global irradiance-dependent nonlinear efficiency characteristics of a given photovoltaic system. In the second part of the dissertation, a new methodology is presented for the estimation of wind potential over the constructed topological grid, in which the computer simulation of wind flow represented as clusters of air molecules is performed by using the Lagrangian method of smoothed particle hydrodynamics and Reynolds averaged turbulence model. In order to implement the estimation of wind flow, multiannual meteorological measurements of the wind velocity and direction are also considered, which are then used for constructing logarithmic wind profile that defines the initial conditions of the wind particles. In order to calculate the output power the nonlinear characteristics of wind systems are used, which depend on the calculated wind velocities. The calculated power values can then be integrated with respect to time in order to forecast electrical energy production. Both methodologies are merged into the proposed algorithm, which we implement on a graphics processing unit using CUDA technology, in order to perform the calculations in an acceptable time. Within the experimental part of the doctoral thesis, the accuracy of the calculation based on the input data is evaluated by comparing the calculated values with independent measurements. The proposed algorithm was also applied over the large geographical area of Maribor, in order to estimate its photovoltaic and wind potentials. The estimated photovoltaic potential has an agreement of 97%, while the agreement of the estimated wind potential is 92%.of the estimated wind potential is 92%.
Secondary keywords:	Fotonapetostni potencial;Disertacije;Algoritmi;Vetrni potencial;
URN:	URN:SI:UM:
Type (COBISS):	Dissertation
Thesis comment:	Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko
Pages:	XII, 119 str.
ID:	9139761