本文利用粒子群演算法來設計被動式濾波器，找到最佳調諧點及電容電感值，使系統電流諧波失真率合乎台電所訂之諧波管制標準，改善低壓配電系統的諧波問題。當今有很多領域在研究最佳化問題時採用了仿生物智慧的概念，而粒子群演算法即屬於其中一種較新的技術。粒子群演算法是導源於觀察群體性動物覓食的行為，每隻個體藉由自身的經驗及群體的互動最後得到食物。因此，應用至最佳化問題時，每個粒子都是所求問題的解，經過自身的認知及群體的指引而最後收尋到最佳解。本研究亦考慮建設成本及系統功因提升等成本因素，以達到低成本、高效率的目標。最後以一低壓配電系統為例，並和其它演算法的結果作比較，證實粒子群演算法於抑制諧波效果和成本上優於其它演算法。 This thesis uses Particle Swarm Algorithm to design a passive filter in order to find the turning point and the capacitor value enabling the generated electric harmonics within the limit set by Taiwan Power Company and to improve the harmonic problems of low voltage electric system. Nowadays, many fields take the concept of simulating biological wisdom for doing researches on most optimization problems, and Particle Swarm Optimization is one of the latest technologies. Particle Swarm Optimization is resulted from the observation of birds’ food-seeking behaviors; that is, each bird gets his food by his self-experiences and interactions among the group. Every particle having its own solution to the questions goes through self- recognition and group guide up to the best solution. In addition, this thesis considers the budget of construction and the improvement of power factors to meet the goal of low budget and high efficiency. Finally, for a system of low voltage distribution system, comparison of the results of the proposed method with those of the other method shows that the proposed method indeed suppresses the effects of harmonics better than the other method.
