The purposes of this study are to analyze the power efficiency of BIPV system under actual operating condition and to reveal the factors of performance decrease through annual monitoring of the south-facing vertical BIPV window system (the capacity of 10.6 kWp) whose performance was affected by partial shading. The study first investigated the monthly PR of each inverter to verify what caused the decrease in power performance by analyzing PR variation. Then, the shading analysis was performed using a simulation to analyze the capture loss of measured power performance and ultimately to find out the effect by partial shading components such as small overhang louvers of each floor, obstructions of adjacent building, and nearby hillock. The one-year measurement of the power performance showed that the annual average reference yield was 2.15 h/day and the final yield was 1.52 h/day; moreover, it was analyzed that the annual average capture loss was 0.49 h/day and the system loss was 0.14 h/day. Monthly average reference yield in summer (between April and August) showed only 1.28 h/day which was 54% lower than the other seasons with average 2.77 h/day. The annual average of PR was 0.69 (the monthly maximum of 0.77 and minimum of 0.58); the average PR in the summer with low reference yield was 0.64, while the average PR for other seasons was 0.73. Whereas monthly variation of LS (system loss) did not show big difference of under 9%, monthly LC (capture loss) highly varied with fluctuation. These primarily arise from the effect of the partial shading and vertically installed angle of south-facing BIPV modules. The BIPV system consisted of four arrays and installed on the same south facade. However, the power performance analysis per array showed that each array had different performances. The arrays located above the facade had the highest PR, from 0.74, 0.75, 0.66, and 0.62 respectively in order. In addition, LS and LC analysis of each array showed that there was performance difference mainly by LC, due to the influence of an adjacent building located in the south. The analysis using the simulation showed the adjacent building did not cast any shading but affected on decrease in diffuse solar radiation whose amount was depending on the array location. When it comes to the lowest array, nearly 7% of annual insolation was blocked by front building. The shading analysis, furthermore, revealed that the overhang louvers caused a partial shading problem in spite of short extrusion of 40 mm depth all through the year, and particularly, showed a big effect during summer because of high solar altitude. The extruded louvers caused decrease in 4.5% of average insolation loss over the year. (C) 2016 Elsevier Ltd. All rights reserved.