應用類神經網路於高占比風能電力系統中發電機跳機之最低頻率預測

Frequency Nadir Prediction of Power System with High Wind Penetration after Generators Tripping Using an Artificial Neural Networks

張銘仁(Ming-Jen Chang)

指導教授 ： 陳昭榮 ； 李清吟

國立臺北科技大學/電資學院/電機工程系研究所/碩士(2013年)

https://doi.org/10.6841/NTUT.2013.00327

摘要

在電力系統運轉中維持頻率穩定是動態安全需求的重要指標。隨著變速型風力發電機逐漸增加下，風力機組將逐步取代傳統發電機組，若風力發電機無法對系統提供慣性響應與頻率控制，加上風力發電的不確定性，將會導致系統頻率變動率增高。當電力系統發生頻率事件時，而造成系統頻率快速下降，影響系統頻率穩定度，因此含高占比再生能源下之電力系統頻率響應成為相當重要的一項研究課題。
本論文以倒傳遞類神經網路估測能力應用於預測電力系統之暫態頻率最低點與頻率變化率電驛之最小卸載量。研究中以澎湖系統實際運轉資料為例，利用PSS/E批次模擬系統於不同運轉下，柴油引擎發電機組跳機事故之系統暫態頻率響應。將頻率低於系統安全限制57.4Hz之案例，以二分法求解最小卸載量。最後利用系統損失發電量、風力發電量、柴油機組總發電量、頻率變化率以及備轉容量等資料分別建立頻率最低點與卸載量預測模型。
由模擬結果顯示，本文提出之類神經網路可準確地估測頻率最低點以及在維持系統穩定運轉前提下，卸除最小負載量，可提供調度人員作預防性調度及卸載規劃。

關鍵字

頻率響應 ； 風力發電機 ； 類神經網路 ； 頻率變化率 ； 低頻卸載

並列摘要

Maintaining frequency stability is dynamic security requirements in power system operations. With the increase in operation of variable-speed wind turbine, the traditional generator will be replaced gradually. Due to the intermittent characteristic of wind power generation, if the wind turbine generators (WTGs) provide neither inertial response nor frequency control, the rate of system frequency variation will be obvious. When generators trip in a hybrid system with the connection of WTGs and diesel generators, the drop in system frequency is apparent; this will influence the stability of system frequency. Therefore, the issue of frequency response in a power system with high penetration of renewable energy will be imperative.
The aim of the thesis is to predict the frequency nadir and minimum load shedding using back propagation artificial neuron network (BPANN) in transient condition. Firstly, the transient frequency response in Penghu power system is studied when diesel generator is tripped, in which the different patterns of load demand, wind power generation and diesel generation in Penghu power system is taken into consideration. The software, namely PSS/E, is utilized to perform the simulation. According to the simulation results, the cases of system frequency blow the security setting (57.4Hz) can be obtained. Then, the bisection method is used to figure out the minimum load shedding. Finally, the data of total power generation of diesel generators, WTGs, power loss of the tripping, spinning reserve and the rate of system frequency variation is used to establish the model of predicting minimum load shedding and frequency nadir.
The simulation results show that proposed BPANN can accurately estimate that on premise of maintaining system frequency stability the frequency nadir and minimum load shedding. The results of the thesis can provide the system operator with a reference relative to the preventive dispatch and the plan of load shedding.

並列關鍵字

Frequency Response ； Wind Turbine Generators (WTGs) ； Artificial Neural Networks ； Rate of Change of Frequency ； Under-frequency Load Shedding

參考文獻

[33]張顯洋，DOS批次檔案設計技巧，臺北:第三波文化事業股份有限公司，1995。

[34]潘建宏，電力系統與離岸風場併接方法之研究，碩士論文，臺北科技大學電機工程系，臺北，2011。

[3]J. Kabouris, and F. D. Kanellos, "Impacts of large-scale wind penetration on designing and operation of electric power systems," IEEE Transactions on Sustainable Energy, vol. 1, no. 2, 2010, pp. 107-144.

[4]T. Ackermann, Wind Power in Power Systems, New York: John Wiley & Sons, 2005.

[5]G. Lalor, A. Mullane and M. O’Malley, "Frequency control and wind turbine technologies," IEEE Transactions on Power Systems, vol. 20, no. 4, 2005, pp. 1905-1913.