| 摘要: | 本論文研究發展微電網之智慧型功率控制與最佳化經濟調度，提出以遞迴式模糊類神經網路為基礎之電壓回復控制器設計與基於拉格朗日乘數法之最佳化經濟調度，以智慧型控制來改善微電網電壓回復之暫態響應，及最佳化演算法直接求解實現經濟調度。在本論文的第一部分設計兩種控制策略來實現電網故障時的電壓調節：首先根據電網規範控制太陽光電系統(PV)以滿足低電壓穿越(LVRT)支撐的要求；另一方面，當電壓發生驟降時，控制儲能系統(BESS)以執行電壓回復控制(VRC)。當供電受到干擾時，電壓回復是微電網電力控制的重要任務，其中一種干擾為輸電線路短路所引起，這可能會導致微電網電壓驟降，甚至是停電，但傳統的比例積分控制器(PI)會受到微電網環境的非線性和不確定性所影響，導致性能下降。有鑒於此，本論文提出遞迴式小波派翠模糊類神經網路(RWPFNN)控制器用於儲能系統的電壓回復控制，來加快響應速度並降低暫態衝擊，因RWPFNN結合小波派翠模糊神經網路(WPFNN)及遞迴式網路結構的優點，是以大幅提高微電網電壓回復的控制性能。此外，為了檢驗太陽光電系統是否符合低電壓穿越的要求，及探討電壓回復控制的性能，而以建在台灣澎湖群島的七美島微電網進行研究，微電網中的太陽光電系統、風力發電機(WTG)和儲能系統分別透過獨立的升壓變壓器連接到同一責任分界點(PCC)，並且柴油發電機提供主要的電力來源並形成獨立的微電網。最後透過使用OPAL-RT即時模擬系統搭配兩部32位元浮點運算數位訊號處理器(DSPs)，除了所提出的RWPFNN外，也使用PI和FNN控制器來比較控制性能，驗證所提出的智慧型控制的有效性。本論文第二部分研究微電網之能源管理最佳化，除了考慮電力潮流限制和再生能源的不確定性，此外維持微電網營運商利潤的重要性及對主電網提供額外電力的需求亦不斷提升，因此將上述問題化為限制式，提出基於拉格朗日乘數的方法(Lagrange multipliers-based method)，可同時處理等式和不等式限制，並且直接精確地獲得最佳經濟調度解。為了驗證結果是否符合等式與不等式限制的條件，並探討基於拉格朗日乘數之方法的性能，而把本論文第一部分之七美島微電網模型進行改進，把原本的孤島狀態與主電網進行連接，形成一個與市電併網的電力系統，並且將所提出之方法與經驗式判斷法、牛頓-粒子群體法(Newton-PSO)及深度Q學習網路(DQN)比較，評估各種方法獲得的結果，所提出之基於拉格朗日乘數的方法花費較少的計算量，並且直接精確地獲得最佳經濟調度解。本論文並以OPAL-RT即時模擬系統搭配32位元浮點運算數位訊號處理器，驗證所提出基於拉格朗日乘數法之最佳化經濟調度的實用性。;The intelligent power control and optimal economic dispatch of microgrid are developed in this dissertation. A voltage restoration controller design based on recurrent fuzzy neural network to provide fast voltage restoration during grid faults and optimal economic dispatch using Lagrange multipliers-based method to directly obtain the optimal economic dispatch solution accurately are also proposed. In the first part of this dissertation, two voltage restoration strategies are designed: one is the regulatory grid code, which is the requirements of low voltage ride through (LVRT) to support grid operations, and the other is the voltage restoration control (VRC) during voltage sags. Voltage restoration is an important task for the power control of microgrid during utility disturbances. One of the disturbances is caused by short circuit on power line of the microgrid, which may lead to voltage sag and even blackout of the microgrid system. However, the control performance of the proportional-integral (PI) control system will be degraded in the microgrid owing to the nonlinearities and uncertainties of the controlled plants. To tackle this problem, the recurrent wavelet Petri fuzzy neural network (RWPFNN) controller is proposed for the VRC of battery energy storage system (BESS) to provide fast control response to mitigate the transient impact. Since the proposed RWPFNN combines the merits of wavelet Petri fuzzy neural network (WPFNN) and recurrent structure, the transient control performance of VRC in a microgrid can be much improved. Moreover, to examine the compliance with the requirements of LVRT of the photovoltaic (PV) plant and investigate the performance of the proposed VRC, the microgrid built in Cimei Island in Penghu Archipelago, Taiwan, is investigated. Furthermore, the PV system, the wind turbine generator (WTG) system and the BESS are connected to the same point of common coupling (PCC) with separated step-up transformers in the microgrid. In addition, the diesel generators provide the main power sources and form the isolated microgrid system. Through the experimental platform, which is built using OPAL-RT real-time simulator, with two floating-point digital signal processors (DSPs), the effectiveness of proposed intelligent controllers can be verified and demonstrated. Besides the proposed RWPFNN controller, both conventional PI controller and FNN controller are implemented for the comparison of the control performance. The second part of the dissertation deals with the optimal energy management of microgrid. In addition to considering power flow constraints and uncertainties of renewable energy, the importance of retaining profits of microgrid operators and the needs for providing extra supports to the main grid are rising. To meet all the requirements, a novel Lagrange multipliers-based method is proposed to deal with equality and inequality constraints simultaneously to directly obtain the optimal economic dispatch solution accurately. Moreover, the Cimei Island microgrid system used in the first part of the dissertation is modified to examine the compliance with the requirements of equality and inequality constraints and the performance of the proposed method. The main grid is designed to provide the utility power sources, which forms the grid-connected microgrid system. Furthermore, comparison of the proposed method with experience-based rule, Newton-particle swarm optimization (Newton-PSO) and deep Q-learning network (DQN) is provided to evaluate the obtained solutions. Compared with various methods, the proposed Lagrange multipliers-based method costs much lesser computing efforts and directly obtains the optimal economic dispatch solution accurately. Finally, through the experimental platform with OPAL-RT real-time simulator and floating-point DSP, the effectiveness of proposed Lagrange multipliers-based method is proven to be pragmatic. |
