• 作者： 鄭林合; 許源浴
• 作者服務機構： 台灣大學電機工程學研究所
• 中文摘要： 本文提出一個利用於直流傳輸系統的整流電流控制器（rectifier current regulator, RCR )之自調式次同步共振控制器（supplemental subsynchronous damping conrol, SSDC）以抑制高壓交/直流並聯傳輸系統中的轉軸振盪。此整流電流控制器之參數，係利用一個經過監督式訓練後的多層結構類神經網路加以調度，以期能在擾動狀態下依然保持良好的阻尼特性。神經網路的輸入為系統之工作點，輸出則為控制器之參數。從時域的模擬測試結果可知，本文所提出之類神經網路調變之自調式次同步共振控制器，確實較傳統之定參數式控制器可以維持系統運轉於較寬廣的工作範圍，且無須繁瑣的即時系統參數估測，同時在多機系統上本方法依然具有良好之結果。
• 英文摘要： Torsional oscillations in an AC/DC transmission system are damped by the supplementalsubsynchronous damping control (SSDC), which is, in principle, a rectifier current regulator (RCR). Toprovide good damping characteristics of torsional oscillations under disturbance conditions, it is desirableto adapt the parameters of SSDC according to generator load changes. To do this, a multilayer feedforwardneural network is used to achieve the desired control action based on generator loading conditions. Afterhaving been trained by a supervised training process with the patterns in the training set, the artificialneural network (ANN) can be used to tune the control parameters. When a pair of generator outputs Pand Q, which are used to describe the operating point of a generator, are fed to the input terminal of theANN, a suitable set of parameters for SSDC will be generated by the ANN as its outputs. To demonstratethe effectiveness of the proposed adaptation scheme, both single-machine and multi-machine systems wereused as test systems. Time domain simulations were performed on these systems for comparison purposes.It is concluded from the results that the proposed adaptive SSDC can provide better damping effects oftorsional oscillations than other control schemes can over a wide operating range of generators.
• 中文關鍵字： subsynchronous resonance; artificial neural network; supplemental subsynchronous damping control
• 英文關鍵字： --