考慮遮蔽情況之太陽光電系統發電功率模擬

建材一體型太陽光電發電系統的應用主要是在大樓帷幕外牆，停車場之遮陽棚、大樓天井、斜頂式屋頂之建築等。不僅可以增加建築物之美觀，更可以作為發電之運用。但其缺點為會受限於建築物設置方位而降低發電效率，也會受到週遭環境影響而降低輸出功率，例如建築物本身的陰影、樹木的陰影或者樹葉掉落至太陽能板上所產生的遮蔽效應，均會降低太陽能板效率或讓太陽能板受損。因此，在裝設太陽能模組之前，適當的發電量模擬技術以評估太陽光電發電系統整體的發電效能，是相當重要的前置作業。本論文的目的是利用微粒群演算法來求解太陽能模組在無遮蔽、有遮蔽無旁路和有遮蔽有旁路時的電壓-電流及電壓-功率特性曲線，並與傳統牛頓-拉佛森法的收斂效能、解答精確度及執行時間作比較。同時藉由上述研究方法模擬在不同日照條件下，太陽能模組的最大功率值。再搭配不同的變流器架構及型態(含集中式變流器、串式變流器、多階串式變流器、交流模組及交流電池等五類)，根據其輸入功率-效率曲線圖，計算整個太陽光電能發電系統ㄧ天的發電量。太陽能模組特性模擬結果顯示，微粒群演算法及牛頓-拉佛森法均能求解太陽光電能發電系統的特性曲線。但微粒群演算法對初始值較不敏感，不會發散，唯執行時間比牛頓-拉佛森法長。而微粒群演算法及牛頓-拉佛森法若收斂條件設定相同，則兩者在精確度上並無差異。另外，變流器發電量模擬結果顯示，在正常未遮蔽情況下，集中式變流器的發電量最大，大功率交流太陽能電池之變流器最小。但是在測試六個電池被遮蔽時，集中式變流器的發電量受影響最大，且發電量低於其它變流器類型，而以大功率交流太陽能電池之變流器受影響最小。若在測試中加入一組旁路二極體情況時，雖然集中式變流器輸出降低量可減緩，但仍為發電量最少者。

關鍵字

微粒群演算法；牛頓-拉佛森法；太陽光電發電

並列摘要

Building-integrated photovoltaic systems are usually applied on building facades, sun-shading canopy of parking area, building court and slop roof. It can be used to generate electrical power and enhance the appearance of the building. The disadvantages of building-integrated photovoltaic systems are low efficiency due to position of the building and low power generation due to the environment effects, such shadow of the building itself, shadow of the trees or leaves falling down on the photovoltaic panel, which can cause low efficiency or damage of the photovoltaic panel. Therefore, before installing photovoltaic modules, it is important to simulate the power generation of the photovoltaic systems. The purpose of this thesis is to solve the voltage-current and voltage-power characteristic curves by using particle swarm optimization algorithm under the conditions such as no-shading, shading without bypass diodes and shading with bypass diodes. The convergence performance, solution accuracy and execution time of particle swarm optimization algorithm are compared with those of the traditional Newton-Raphson method. At the same time, maximum power point is simulated for various irradiations by the previous mentioned methods. The photovoltaic modules are matched with various inverter configurations and types (including centralized inverter, string inverter, multi-string inverter, AC module and AC cells). Finally, the day energy outputs of whole the photovoltaic generation system are calculated according to the power-efficiency characteristic curves of the inverters. Test results of photovoltaic module simulations indicated that the characteristic curves of the photovoltaic systems can be solved by both particle swarm optimization algorithm and Newton-Raphson method. The particle swarm optimization algorithm is not sensitive with initial values, not diverged and the execution time is longer than that of Newton-Raphson method. The accuracies are not different if the convergence criteria are set to identical for both methods. Besides, test results of inverter simulations indicated that, under no-shading condition, centralized inverter has the biggest generation amount, however, AC cells has the smallest one. Under the shading condition (6 cells are shaded), the impact on generation amount is biggest for centralized inverter and it makes the centralized inverter has the smallest generation amount. However, the impact on generation amount is smallest for AC cells. If a bypass diode is shunted with a set of series photovoltaic cells, the impact on generation amount for centralized inverter can be reduced, but the generation amount of centralized inverter is still the smallest one when compared with those of other inverters.