本研究在於設計等電壓不同電流源電路，電路板輸入端滿足最大功率追蹤，輸出端使用類神經網路實現能量自鎖榨取功能，隨著附載及光源強弱調整工作週期，用以改善現行太陽能板之技術，在串接多塊太陽能板時當某塊太陽能板故障或遮蔭時，造成因電位下降而導致集電時整體輸出功率降低；亦或當太陽能板陣列某一片太陽能板損壞時，因太陽能板前端串接蕭特基二極體，導致該片太陽能板電壓無法輸出而失效。經實驗結果顯示，在同樣光照強度之下，針對有使用本裝置之太陽能組合板較傳統之太陽能組合板做比較，在不同負載之下改善之太陽能板能夠增加9%~65%的效率。在並聯輸出時假如某一片太陽能板被遮蔭或損壞時，整體轉換效率仍能維持在90%以上並使時輸出電壓一致。;Our research using with common voltage in the non-common current circuit which can fulfill Maximum power point tracking in the input side, and self-locking output power with Neural network learning method. That can modify the duty cycle along the load resistors to provide the output energy more efficiently. Within this technology, we can improve current application in solar cell domain. Typically when solar cells in series but one of them failed or broken, it will cause the voltage to fail to output so that to total power efficiency very low. In other condition, if we add stocky diodes in solar cells but one of them broken, the leakage current will break all the solar cells. According to our test result, the solar cells embedded our circuit which can provide the solar system no matter solar cells in series or parallel. Further, it can improve the system more powerful and efficient. In the same condition, we can provide up to 60% efficiency than the traditional devices. And we can still provide up to 90% with equal voltage values even the shade effect.
