| 摘要: | 本文使用TRNSYS 軟體模擬中壢地區（緯度25oN，經度121oE）的聚光型太陽光電系統（concentrating photovoltaic, CPV）電力性能，並與實際量測的實驗數據做比較。 本實驗室開發的CPV 系統分別搭配四種追日控制方式（開迴路追日控制、閉迴路追日控制、混合式追日控制及指向誤差修正追日控制），CPV 系統搭配不同的追日控制方式產生不同的追日偏差，因此影響實驗和模擬結果二者的差異。 在TRNSYS 軟體的模擬中，由於軟體內建的標準元件是針對一般PV 模組，但CPV模組因為採用三五族三接面太陽電池和傳統的矽基太陽電池有相當差異，故本文參考相關文獻及電池溫度的實驗數據，推導CPV 模組模擬參數的設定值。 CPV 系統搭配開迴路和閉迴路追日控制產生的追日偏差角度大於其他二種追日控制方式，追日偏差角度變化範圍分別為0.04o~0.95o 和0.22o~0.77o，在晴朗的天氣下，每日平均輸出功率的模擬誤差分別為26%和47.1%。而搭配混合式和指向誤差修正追日控制具備回饋機制，因此它們的表現優於其他二種方法，不僅產生的追日偏差角度小，連帶著量測與模擬的輸出功率差亦小。搭配混合式和指向誤差修正追日控制對應的追日偏差角度變化範圍分別為0.01o~0.02o 和0.04o~0.2o，在晴朗的天氣下，二者每日平均輸出功率的模擬誤差為8.9%和4.35%。因此建議CPV 系統應搭配具有回饋修正的追日控制法像混合式或是指向誤差修正追日控制。 本文於TRNSYS 軟體中新增電力調節器的模擬元件，以2010 年 8、9 月（晴天且有足夠量測數據）長期的實驗數據在不考慮系統的追日偏差來驗證CPV 系統的模擬結果，其8、9 月累積發電量的模擬誤差分別為5.28%和9.53%。 In this study, predicted performance of concentrating photovoltaic (CPV) system using TRNSYS is compared with outdoor measurement in Jhong-Li area (latitude 25oN and longitude 121oE). The CPV system developed in our laboratory is implemented with four sun-tracking approaches (open loop tracking, closed loop tracking, hybrid tracking and pointing-error correction tracking). CPV system with different sun-tracking approaches generates different tracking offset-angle, thus affecting the deviations between measured and predicted results. In TRNSYS, the PV array component is used to simulate the performance of general PV module. However the characteristics of solar cell is different between III-V type and Si-based type. Therefore, modified parameters in the PV array component are derived based on available references. CPV system with open loop and closed loop sun-tracking approaches generated larger tracking offset-angle than the other two trackings, with range of offset-angle 0.04o~0.95o and 0.22o~0.77o, respectively. On sunny day, the daily average difference between measured and predicted power for open-loop and close-loop are 26% and 47.1%, respectively. Performance of our CPV system with hybrid and pointing-error correction tracking approaches, both embedded with feedback correction, are better than the other two sun-trackings, not only resulted in smaller tracking offset-angle but also better agreement between the measured and predicted power output. The range of offset-angle for the hybrid and pointing-error correction algorithms are 0.01o~0.02o and 0.04o~0.2o, respectively. The daily average difference between measured and predicted power were 8.9% and 4.35% on sunny day, respectively. Thus, it is suggested that CPV system should implemented with feedback correction such as the hybrid or pointing-error correction algorithms. This study adds an inverter component used in TRNSYS to perform inverter output simulation. Without considering the system's sun-tracking error, using long-term measured data for validating simulation of CPV system in August and September (sunny days with enough measurement), 2010. Errors of cumulated power output using TRNSYS are 5.28% and 9.53%, respectively. |
