| DC 欄位 |
值 |
語言 |
| DC.contributor | 電機工程學系 | zh_TW |
| DC.creator | 呂光欽 | zh_TW |
| DC.creator | Kuang-Chin Lu | en_US |
| dc.date.accessioned | 2016-7-14T07:39:07Z | |
| dc.date.available | 2016-7-14T07:39:07Z | |
| dc.date.issued | 2016 | |
| dc.identifier.uri | http://ir.lib.ncu.edu.tw:444/thesis/view_etd.asp?URN=101581009 | |
| dc.contributor.department | 電機工程學系 | zh_TW |
| DC.description | 國立中央大學 | zh_TW |
| DC.description | National Central University | en_US |
| dc.description.abstract | 本論文研究發展以PC為基礎之智慧型控制系統應用於三相倂網型太陽光電(PV)系統在電網故障時之實虛功的控制。提出兩種以模糊類神經網路為基礎之智慧型控制器以實現倂網型太陽光電系統的智慧型控制系統，達成在電網故障期間調控PV系統實虛功以符合低壓穿越(LVRT)的規範並確保三相變流器的輸出電流不會超越電流安全值。首先提出結合了機率模糊類神經網路(PFNN)以及小波模糊類神經網路(WFNN)優點的機率小波模糊類神經網路(PWFNN)控制器；接著則是非對稱歸屬函數之TSK 型機率模糊類神經網路(TSKPFNN-AMF)控制器，為同時結合TSK型式的模糊類神經網路(TSKFNN)、非對稱歸屬函數(AMF)與機率模糊類神經網路(PFNN)等特點而成的控制器；是以，PWFNN與TSKPFNN-AMF控制器對於具有不確定性及高度非線性特性之複雜的非線性系統都有良好的處理能力。由於三相併網型太陽光電系統包含了太陽能板、升壓轉換器、三相變流器以及配電系統，因此是一種具有非確定性之非線性系統，所以此系統真正的非線性模型難以建立。像比例積分(PI)型這類的傳統型控制器因本身之線性特性之故，當受控體的參數改變以及未知的外來干擾等不確定性的影響，PI控制器將難以使該非線性系統達到預期的控制性能。因此，可採用本論文提出之PWFNN與TSKPFNN-AMF控制器來發展有學習能力的控制系統以處理具不確定性之非線性系統，如被用來調控併網型太陽光電系統在電網故障期間的實虛功控制。此外，本論文根據LVRT的規範提出計算電壓下降百分比與決定注入虛功電流比例的公式，同時為了避免於LVRT期間發生過電流，此公式在計算注入虛功電流大小時已加入預定的電流限制條件。還有，本論文針對三相併網型太陽光電系統之升壓轉換器與三相變流器研擬一種雙模式操作法，以使太陽光電系統之太陽能板與三相變流器在電網故障期間的實功流動維持平衡，並藉由注入額定電流以最大化三相變流器的功率耐受力。為達此目標，控制系統調整操作方式為模式II以降低實功輸出來確保不會超出最大額定電流。有關所提控制器的網路架構設計、線上學習流程以及收斂性分析等，文中皆有詳細的描述。另建立了1kW的三相併網型光電系統模擬器以測試本論文所提智慧型控制系統的性能，並設計多種的電壓故障類型以及測試情境以檢驗併網型太陽光電系統的LVRT能力。實驗結果顯示，雖然所提之控制器的控制性能比其他控制器如PI、FNN、WFNN等優異，但控制器的高複雜度以及故障期間注入電流含有較高的總諧波失真(THD)成為美中不足的缺點。當然本文也建立一些控制器性能評估的基準，以便客觀的評估控制器性能。 | zh_TW |
| dc.description.abstract | A PC-based intelligent power control system of the three-phase grid-connected photovoltaic (PV) system for active and reactive power control during grid faults is developed in this dissertation. Two fuzzy-neural-network (FNN) based intelligent controllers are proposed to perform the intelligent power control system to regulate the active and reactive power of the grid-connected PV system satisfying the low voltage ride through (LVRT) requirements and ensuring the injected currents within the safety value of the three-phase inverter. The first proposed intelligent controller is the probabilistic wavelet fuzzy neural network (PWFNN) controller, which combines the merits of probabilistic FNN (PFNN) and wavelet FNN (WFNN). The other is the Takagi-Sugeno-Kang type probabilistic fuzzy neural network with asymmetric membership function (TSKPFNN-AMF) controller, which is the combination of the characteristics of the Takagi-Sugeno-Kang type fuzzy neural network (TSKFNN), asymmetric membership function (AMF) and probabilistic neural network (PNN). Both PWFNN and TSKPFNN-AMF controllers are good at dealing with a complex nonlinear system with uncertainty, complexity, and strong nonlinearities. Since the three-phase grid-connected PV system, which includes PV panel, boost converter, three-phase inverter and grid, can be regarded as a nonlinear system with uncertainty, it is very difficult to develop a physical nonlinear model for the system. Traditional controller such as proportional-integral (PI) controller is difficult to guarantee the desired control performance with the presence of plant parameter variations and unknown external disturbances due to the linearity characteristic of a PI controller. Therefore, the proposed PWFNN and TSKPFNN-AMF controllers can be adopted to develop a system with learning capability for nonlinear systems with uncertainties and be adopted to control the active and reactive power of the grid-connected PV system during grid faults. Moreover, a formula for evaluating the percentage of voltage sags is derived to determine the ratio of the injected reactive current to satisfy the regulations. To reduce the risk of over-current during LVRT operation, a current limit is predefined for the injection of reactive current. Furthermore, a dual mode operation control method of the converter and inverter of the three-phase grid-connected PV system is proposed to balance the active power between the PV panel and the three-phase inverter and maximize the three-phase inverter power capability via injecting the full-rated current during grid fault. To achieve this objective, the control system applies the mode II operation strategy to curtail the active power to ensure the maximum rated current is not surpassed. In addition, the network structure, online learning algorithm, and convergence analysis of the proposed intelligent controllers are described in detail. An emulated 1kW grid-connected three-phase PV system is settled and tested to verify the performance of the proposed intelligent power control system. Various types of voltage sags and test scenarios are designed to investigate the LVRT capability of the grid-connected PV system. The experimental results show that although the control performances of the proposed controllers are superior to other controllers such as PI, FNN and WFNN, higher complexity of structure and current harmonic distortion of injected current during grid faults are the main defects. Additionally, some benchmarks of these controllers are also provided to evaluate the control performances. | en_US |
| DC.subject | 非對稱歸屬函數 | zh_TW |
| DC.subject | 電網故障 | zh_TW |
| DC.subject | 低壓穿越 | zh_TW |
| DC.subject | 太陽光電 | zh_TW |
| DC.subject | 機率小波模糊類神經網路 | zh_TW |
| DC.subject | 虛功注入 | zh_TW |
| DC.subject | 非對稱歸屬函數之TSK型機率模糊類神經網路 | zh_TW |
| DC.subject | asymmetric membership function (AMF) | en_US |
| DC.subject | grid faults | en_US |
| DC.subject | low voltage ride through (LVRT) | en_US |
| DC.subject | photovoltaic (PV) | en_US |
| DC.subject | probabilistic wavelet fuzzy neural network (PWFNN) | en_US |
| DC.subject | reactive power injection | en_US |
| DC.subject | Takagi-Sugeno-Kang type probabilistic fuzzy neural network with asymmetric membership function (TSKPFNN-AMF) | en_US |
| DC.title | 三相式併網型太陽光電系統之智慧型功率控制系統 | zh_TW |
| dc.language.iso | zh-TW | zh-TW |
| DC.title | Intelligent Power Control System of Three-Phase Grid-Connected PV System | en_US |
| DC.type | 博碩士論文 | zh_TW |
| DC.type | thesis | en_US |
| DC.publisher | National Central University | en_US |