頻率品質是電力系統中的一個重要議題。頻率品質的好壞取決於備轉容量（OR）是否充足。本研究分為兩部分，第一部份是考慮正常運轉情況，第二部分是考慮偶發事件情況。在正常運轉情境下，許多運營機構採用北美電力可靠性公司（NERC）制定的控制性能標準（CPS1），以規範頻率控制的品質，因此運營機構對於與CPS1分數有高度相關的頻率敏感因子和如何找出更有效的機組調度策略以確保CPS1分數達標的議題都有很大的興趣。由於頻率敏感因子通常具有多變量和高度相關的特點，本文採用集成學習技術（梯度提升決策樹演算法，GBDT）來構建台電系統的頻率響應模型（FRM），以評估CPS1得分。然後，將所提出的CPS1模型與機組排程（UC）相結合，以確保排程結果可以滿足CPS1分數目標。然而，在事故情境下，因為再生能源高度滲透的關係，使得頻率最低點變得更難預測。此外，大量的再生能源占比也改變了原本電力系統的機組排程結果。傳統上，頻率最低點是透過負載頻率控制模型(LFC)來評估，其中關鍵因子有：負載阻尼、系統慣性和調速機響應。然而，這些關鍵因子並不容易獲得。因此，本研究提出一個監督式學習方案，以追踪這些關鍵因子。此外，本研究提出一個新的關鍵因子—電能缺口率，它比負載阻尼更能反映負載對系統頻率的影響。利用所提出的監督式學習方案，實現關鍵因子的重要性排序和構建一個頻率最低點預測模型（FNM）。本研究提出的方案與知名的方法相比，所提出的模型（FNM）可準確預測系統頻率，並規劃出可靠的輔助服務的需求量。此外，在兩個情境的案例分析中，都能將高滲透再生能源發電情況列入考慮。案例結果顯示，本研究提出的方法可以有效地規劃輔助服務需求量，以確保電力系統的安全性和可靠性。

Frequency quality is an important issue in the power system. The quality of frequency depends on the preparation of operating reserve (OR). This study is divided into two parts, one is non-event OR preparation and the other is for contingency case. For normal operation, North American Electric Reliability Corporation’s (NERC) Control Performance Standard 1 (CPS1) is widely adopted by many operating authorities to examine the quality of the frequency control. The operating authority would have a strong interest in knowing how the frequency-sensitive features affect the CPS1 score and then come up with more effective unit-dispatch schedules for reaching the CPS1 goal. As frequency-sensitive features usually possess multi-variable and high-correlated characteristics, this research employed an ensemble learning technique (the Gradient Boosting Decision Tree algorithm, GBDT) to construct Frequency Response Model (FRM) of the Taiwan power company’s (TPC) system in Taiwan to evaluate CPS1 score. The proposed CPS1 model was then integrated with Unit Commitment (UC) program to determine the non-event OR that achieves the targeted CPS1 score. For the contingency case, frequency nadir prediction is the key to evaluate sufficiency of OR to ensure frequency security. However, the prediction becomes more challenging because renewables have significantly changed the generation portfolio within the system. Conventionally, the frequency nadir is determined by frequency sensitive features—load damping, system inertia, and effective governor response—which are assumed to be known. However, these key features are not easily obtained in a power system as it continuously changes during daily operation. To overcome the obstacle, this study proposes a supervised learning scheme that continuously traces these key features. A new feature—that better reflects the influence of the load on the system frequency than that of the load damping and system inertia is recognized by feature importance process. The construction of the frequency nadir model (FNM) was realized using the proposed supervised learning scheme. In the testing scenarios, both FRM and FNM models present their superiorities to predict the system frequency in comparison to the other prior arts. The resultant OR preparation is more reliable to ensure system security and frequency quality.

[1] E. Ela, M. Milligan and B. Kirby, “Operating reserves and variable generation,” National Renewable Energy Laboratory, Golden, Colorado, TP-5500-51978, Aug. 2011.
[2] “NERC Reliability Standard BAL-001-1-Real Power Balancing Control Performance,” Accessed on: Apr. 10, 2021. [Online]. Available: http://www.nerc.com/pa/Stand/Reliability%20Standards/BAL-001-1.pdf
[3] “NERC Reliability Standard BAL-002-0-Disturbance Control Performance,” Apr. 1, 2005. Accessed on: Apr. 10, 2021. [Online]. Available: https://www. nerc. com/files/BAL-002-0.pdf
[4] P. B. Luh et al., “Grid integration of intermittent wind generation: A Markovian approach,” IEEE Trans. Smart Grid, vol. 5, no. 2, pp. 732–741, Mar. 2014.
[5] W. Su, J. Wang, and J. Roh, “Stochastic energy scheduling in microgrids with intermittent renewable energy resources,” IEEE Trans. Smart Grid, vol. 5, no. 4, pp. 1876–1883, Jul. 2014.
[6] Q.P. Zheng, J. Wang, A.L. Liu, “Stochastic optimization for unit commitment—A review,” IEEE Trans. Power Syst., vol. 30, no 4, pp. 1913-1924, 2015.
[7] N. Zhang, Z. Hu, D. Dai, S. Dang, M. Yao, and Y. Zhou, “Unit commitment model in smart grid environment considering carbon emissions trading,” IEEE Trans. Smart Grid, vol. 7, no. 1, pp. 420–427, Jan. 2016.
[8] T. Mitani et al., “Annual Assessment of Large-Scale Introduction of Renewable Energy: Modeling of Unit Commitment Schedule for Thermal Power Generators and Pumped Storages", Energies, vol. 10, no. 6, pp. 738, May 2017.
[9] W. Chung, “Anticipatory fuzzy automatic generation control scheme using very short-term ANN ACE forecasting,” Ph.D. dissertation, Dept. Elect. Comput. Eng., Purdue Univ., West Lafayette, IN, USA, 2005.
[10] H. Ahmadi and H. Ghasemi, “Security-constrained unit commitment with linearized system frequency limit constraint,” IEEE Trans. Power Syst., vol. 29, no. 4, 2014.
[11] F. Pérez-Illanes, E. Álvarez-Miranda, C. Rahmann, and C. Campos-Valdés, “Robust Unit Commitment Including Frequency Stability Constraints,” Energies, vol. 9, no. 11, p. 957, Nov. 2016.
[12] M. Farrokhabadi, C. Canizares, and K. Bhattacharya, “Unit commitment for isolated microgrids considering frequency control,” IEEE Trans. Smart Grid, vol. 9, pp. 3270–3280, July 2018.
[13] Y.-Q. Bao, Y. Li, B. Wang, M. Hu and Y. Zhou, “Day-ahead scheduling considering demand response as a frequency control resource", Energies, vol. 10, no. 1, pp. 82, Jan. 2017.
[14] G. Zhang and J. D. McCalley, “Estimation of regulation reserve requirement based on control performance standard,” IEEE Trans. Power Syst., vol. 33, no. 2, pp. 1173-1183, March 2018.
[15] Y. Yang, C. Shen, C. Wu and C. Lu, “Control performance based dynamic regulation reserve allocation for renewable integrations,” IEEE Trans. Sustainable Energy, vol. 10, no. 3, pp. 1271-1279, July 2019.
[16] A. Ahmad, N. Javaid, N. Alrajeh, Z. A. Khan, U. Qasim and A. Khan, "A modified feature selection and artificial neural network-based day-ahead load forecasting model for a smart grid", Appl. Sci., vol. 5, no. 4, pp. 1756-1772, 2015.
[17] S. Mujeeb, T. A. Alghamdi, S. Ullah, A. Fatima, N. Javaid and T. Saba, "Exploiting deep learning for wind power forecasting based on big data analytics", Appl. Sci., vol. 9, no. 20, Oct. 2019.
[18] A. Mellit, A. Massi Pavan, E. Ogliari, S. Leva and V. Lughi, "Advanced methods for photovoltaic output power forecasting: A review", Appl. Sci., vol. 10, no. 2, pp. 487, Jan. 2020.
[19] A. Moradzadeh, S. Zakeri, M. Shoaran, B. Mohammadi-Ivatloo and F. Mohamamdi, "Short-term load forecasting of microgird via hybrid support vector regression and long short-term memory algorithms" Sustainability, Switzerland, vol. 12, no. 17, pp. 7076, Aug. 2020.
[20] J. López Gómez, A. Ogando Martínez, F. Troncoso Pastoriza, L. Febrero Garrido, E. Granada Álvarez, and J. A. Orosa García, “Photovoltaic Power Prediction Using Artificial Neural Networks and Numerical Weather Data,” Sustainability, vol. 12, no. 24, p. 10295, Dec. 2020.
[21] C. Seneviratne and C. Ozansoy, “Frequency response due to a large generator loss with the increasing penetration of wind/PV generation–A literature review,” Ren. Sustain. Energy Rev., vol. 57, pp. 659-668, 2016.
[22] P. Du, N. V. Mago, W. Li, S. Sharma, Q. Hu and T. Ding, “New ancillary service market for ERCOT,” IEEE Access, vol. 8, pp. 178391-178401, 2020.
[23] H. Gu, R. Yan and T. Saha, “Review of system strength and inertia requirements for the national electricity market of Australia,” CSEE J. Power Energy Syst., vol. 5, no. 3, pp. 295-305, Sep. 2019.
[24] D. Fernández-Muñoz, J. I. Pérez-íaz, I. Guisández, M. Chazarra, Á. Fernández-Espina, “Fast frequency control ancillary services: an international review,” Renew. Sustain. Energy Rev., vol. 120, p. 109662, 2020.
[25] L. Su, X. Qin, S. Zhang, Y. Zhang, Y. Jiang, Y. Han, “Fast frequency response of inverter-based resources and its impact on system frequency characteristics,” Global Energy Interconnection, vol. 3, no. 5, 2020.
[26] E. Pusceddu, B. Zakeri, G. C. Gissey, Synergies between energy arbitrage and fast frequency response for battery energy storage systems,” Appl. Energy, vol. 283, 2021.
[27] S. P. Melo, U. Brand, T. Vogt, J. S. Telle, F. Schuldt, K. V. Maydell, “Primary frequency control provided by hybrid battery storage and power-to-heat system,” Appl. Energy, vol. 233-234, 2019.
[28] P. Babahajiani, Q. Shafiee and H. Bevrani, “Intelligent demand response contribution in frequency control of multi-area power systems,” IEEE Trans. Smart Grid, vol. 9, no. 2, pp. 1282-1291, Mar. 2018.
[29] C.-Y. Lai and C.-W. Liu, “A scheme to mitigate generation trip events by ancillary services considering minimal actions of UFLS,” IEEE Trans. Power Syst., vol. 35, no. 6, pp. 4815-4823, Nov. 2020.
[30] M. Peydayesh and R. Baldick, “Simplified model of ERCOT frequency response validated and tuned using PMUs data,” IEEE Trans. Smart Grid, vol. 9, no. 6, pp. 6666-6673, Nov. 2018.
[31] W. Wang, W. Yao, C. Chen, X. Deng and Y. Liu, “Fast and accurate frequency response estimation for large power system disturbances using second derivative of frequency data,” IEEE Trans. Power Syst., vol. 35, no. 3, pp. 2483-2486, May 2020.
[32] “NERC Reliability Guideline Primary Frequency Control,” Jun. 4, 2019. Accessed on: Apr. 10, 2021. [Online]. Available: https://www.nerc.com/comm/OC/RS_GOP_Survey_DL/PFC_Reliability_Guideline_rev20190501_v2_final.pdf
[33] H. Huang and F. Li, “Sensitivity analysis of load-damping characteristic in power system frequency regulation,” IEEE Trans. Power Syst., vol. 28, no. 2, pp. 1324-1335, May 2013.
[34] D. D. Giudice, A. Brambilla, S. Grillo, F. Bizzarri, “Effects of inertia, load damping and dead-bands on frequency histograms and frequency control of power systems,” Int. J. Elect. Power Energy Syst., vol. 129, 2021.
[35] C. C. Wu and N. Chen, “Frequency-based method for fast-response reserve dispatch in isolated power systems,” Proc. Inst. Elect. Eng. Gener. Transm. Distrib., vol. 151, no. 1, pp. 73-77, Jan. 2004.
[36] R.-F. Chang, C.-N. Lu and T.-Y. Hsiao, “Prediction of frequency response after generator outage using regression tree,” IEEE Trans. Power Syst., vol. 20, no. 4, pp. 2146-2147, Nov. 2005.
[37] L. Wu, S. You, J. Dong, Y. Liu and T. Bilke, “Multiple linear regression based disturbance magnitude estimations for bulk power systems,” 2018 IEEE Power & Energy Society General Meeting (PESGM), Portland, OR, USA, pp. 1-5, 2018
[38] D. Zografos, T. Rabuzin, M. Ghandhari and R. Eriksson, “Prediction of frequency nadir by employing a neural network approach,” 2018 IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT-Europe), Sarajevo, Bosnia and Herzegovina, pp. 1-6, 2018
[39] M. S. Ibrahim, W. Dong, Q. Yang, “Machine learning driven smart electric power systems: Current trends and new perspectives,” Appl. Energy, vol. 272, 2020.
[40] O. A. Alimi, K. Ouahada and A. M. Abu-Mahfouz, “A review of machine learning approaches to power system security and stability,” IEEE ACCESS, vol. 8, pp. 113512-113531, 2020.
[41] L. Duchesne, E. Karangelos and L. Wehenkel, “Recent developments in machine learning for energy systems reliability management,” Proc. IEEE, vol. 108, no. 9, pp. 1656-1676, 2020.
[42] H.-W. Chiu, L.-R. Chang-Chien, C.-C. Wu, “Construction of a frequency compliant unit commitment framework using an ensemble learning technique,” Energies, vol. 14, no. 2, pp. 310, 2021.
[43] IEA (2016), Re-powering Markets: Market design and regulation during the transition to low-carbon power systems, Electricity Market Series, IEA, Paris,
[44] Power System Operation Standards and Manual. Taiwan Power Company: Taipei, 2000.
[45] The Architecture of Power Supply for TPC. Available online: http://www.moeaboe.gov.tw (accessed on 12 August 2021).
[46] X. Guan, P. B. Luh, H. Yen and P. Rogan, "Optimization-based scheduling of hydrothermal power systems with pumped-storage units," IEEE Trans. Power Syst., vol. 9, no. 2, pp. 1023-1031, May 1994.
[47] J. Alemany, D. Moitre, H. Pinto and F. Magnago, "Short-Term Scheduling of Combined Cycle Units Using Mixed Integer Linear Programming Solution," Energy and Power Engineering, vol. 5 no. 2, pp. 161-170., 2013.
[48] W. Huang, T. Hu, K. Hsu, C. Tsai and C. Wu, "Analysis of Simulated Bidding Results for Ancillary Services of Taiwan Power System: Using Thermal Power Plants as an Example", Monthly Journal of TPC's Engineering, vol. 795, no. 11, Nov. 2014.
[49] G. E. Alvarez, M. G. Marcovecchio and P. A. Aguirre, "Security-constrained unit commitment problem including thermal and pumped storage units: An MILP formulation by the application of linear approximations techniques", Elect. Power Syst. Res., vol. 154, pp. 67-74, Jan. 2018.
[50] R. Domingues, M. Filippone, P. Michiardi, J. Zouaoui, “A comparative evaluation of outlier detection algorithms: Experiments and analyses,” Pattern Recognition, vol. 74, pp. 406-421, 2018.
[51] F. T. Liu, K. M. Ting and Z.-H. Zhou, "Isolation forest", Proc. 8th IEEE Int. Conf. Data Mining (ICDM), pp. 413-422, Dec. 2008.
[52] S. Badillo, B Banfai, Birzele F, Davydov II, Hutchinson L, Kam-Thong T, et al, “An introduction to machine learning,” Clin Pharmacol Ther., vol. 107, pp. 871-885, 2020.
[53] S. Hochreiter and J. Schmidhuber, "Long Short-Term Memory", Neural Computation, Vol. 9, no. 8 pp. 1735–1780, 1997.
[54] A. Sherstinsky, “Fundamentals of Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) Network,” Phys. D Nonlinear Phenomena, vol. 404, 2020.
[55] R. K. Agrawal, F. Muchahary, M. M. Tripathi, "Long Term Load Forecasting with Hourly Predictions Based on Long-short-term-memory Networks", IEEE Texas Power and Energy Conference (TPEC), 2018.
[56] O. Sagi, L. Rokach, “Ensemble learning: a survey,” Wiley Interdisc. Rew., vol. 8, pp. 1-18, 2018.
[57] J. Friedman, “Greedy function approximation: a gradient boosting machine,” Ann. Stat., vol. 29, no. 5, pp. 1189–1232, 2001.
[58] "Scikit-learn, " Feb. 1, 2010. Accessed on: Apr. 10, 2021. [Online]. Available: https://scikit-learn.org/stable/
[59] T. Chen, T. He and M. Benesty, “Xgboost: Extreme gradient boosting,” R Package Version 0. 4-2, pp. 1-4, 2015.
[60] Y. Yu, X. Si, C. Hu and J. Zhang, “A review of recurrent neural networks: LSTM cells and network architectures,” Neural Comput., vol. 31, no. 7, pp. 1235-1270, Jul. 2019.
[61] C. S. Bojer and J. P. Meldgaard, “Kaggle forecasting competitions: An overlooked learning opportunity,” Int. J. Forecast, vol. 37, no. 2, pp. 587603, 2020.
[62] T. Hastie, R. Tibshirani, and J. Friedman, The Elements of Statistical Learning, 2nd ed. New York: Springer, 2009.
[63] H. Chiu, Y. Lin, L. Chang-Chien and C. Wu, "Modeling of the CPS1 for the Frequency Constrained Unit Commitment, “2019 IEEE PES Asia-Pacific Power and Energy Engineering Conference (APPEEC), 2019, pp. 1-5.
[64] H. H. Alhelou, M. E. Golshan, T. Njenda, N. Hatziargyriou, "An overview of UFLS in conventional, modern, and future smart power systems: challenges and opportunities, " Elect. Power Syst. Res., Feb. 2020.
[65] S.-J. Huang and C.-C. Huang, “An automatic load shedding scheme including pumped-storage units, " IEEE Trans. Energy Conversion, vol. 15, no. 4, pp. 427-432, Dec. 2000.
[66] C. Wu, W. Lee, C. Cheng and H. Lan, “Role and D-value of pumped storage units in an ancillary services market for isolated power systems- simulation in the Taiwan power system,” 2007 IEEE/IAS Industrial & Commercial Power Systems Technical Conference, Edmonton, AB, Canada, pp. 1-6, 2007.
[67] B. Ratner, “The correlation coefficient: Its values range between +1/1 or do they?,” J. Targeting Meas. Anal. Marketing, vol. 17, pp. 139-142, May 2009.