Abstract
As a dispatchable renewable energy technology, the fast ramping capability of concentrating solar power (CSP) can be exploited to provide regulation services. However, frequent adjustments in real-time power output of CSP, which stems out of strategies offered by ill-designed market, may affect the durability and the profitability of the CSP plant, especially when it provides fast regulation services in a real-time operation. We propose the coordinated operation of a CSP plant and wind farm by exploiting their complementarity in accuracy and durability for providing frequency regulation. The coordinated operation can respond to regulation signals effectively and achieve a better performance than conventional thermal generators. We further propose an optimal bidding strategy for both energy and frequency regulations for the coordinated operation of CSP plant and wind farm in day-ahead market (DAM). The validity of the coordinated operation model and the proposed bidding strategy is verified by a case study including a base case and sensitivity analyses on several impacting factors in electricity markets.
ENVIRONMENTAL aspects of thermal power generation has been the subject of more scrutiny since renewable energy sources such as wind and solar find their way and gain more clout as part of the power generation mix in many countries. The European Union (EU) has proposed a 26%-34% increase by 2030 in the utilization of renewable energy [
Several studies have shown that a wind turbine can control its active power output through rotor inertia [
Concentrating solar power (CSP) technology is one of the ways to efficiently harness solar energy. Unlike the photovoltaic (PV) technology, a CSP plant, which is usually equipped with a large-capacity thermal energy storage (TES), offers a valuable energy capacity [
When CSP and wind power both provide frequency regulation, they can exhibit complementary characteristics in terms of accuracy and durability. The accuracy refers to the regulation signal tracked in a timely and unbiased manner. The durability refers to the regulation signal tracked for a long time. A CSP plant with its inherent dispatchability excels a wind unit in frequency regulation accuracy, although a wind unit provides regulation services which can be adjusted easily without increasing its fatigue loads [
Considering the above issues, this paper proposes the coordinated operation of CSP plant and wind farm to provide frequency regulation according to the complementary characteristics of wind power and CSP, and formulates a performance-based optimal bidding model for the frequency regulation strategy of the coordinated system in day-ahead market (DAM). Additionally, the model ensures that the TES of CSP returns to its initial heat storage level at the end of the trading day. The case study results demonstrate the significant benefits of the coordinated system to provide regulation services. The contributions of this paper are summarized as follows.
1) We propose to provide frequency regulation by the cooperation of wind farm and CSP plant, and construct a two-stage stochastic model for the coordinated bidding of wind power and CSP in DAM (in the first stage) with the simulation of real-time operation that considers the variability of wind power in the second stage.
2) We propose a coordinated strategy which prioritizes the wind power durability to respond to regulation signals and exploits the accurate CSP regulation capability to compensate any inefficiencies in the coordinated offer.
3) We propose an optimal bidding model for a coordinated system with performance-based regulation and deviation penalty mechanism in the proposed real-time frequency regulation strategy.
The rest of the paper is organized as follows. The technical benefit of coordinated system for frequency regulation is introduced in Section II. The framework of the market and the optimal bidding model for the coordinated system is established in Section III. In Section IV, the validity of the model is verified by the study case. Conclusions are given in Section V.
In this section, we present the proposed coordinated operation of the CSP plant and wind farm and maximize its revenue by properly bidding in the energy and regulation markets based on corresponding market prices and the available wind energy. This is based on the observation that CSP plant and wind farm are highly complementary in terms of accuracy and durability for providing frequency regulation in the regulation market.
The output of a wind turbine is determined by its pitch angle and rotor speed, as the turbine works in its maximum power point tracking (MPPT) mode. The wind turbine can support frequency regulation through the de-loading control [
The frequency regulation accuracy of a wind turbine is affected by its available energy, reserved capacity for providing regulation service, and a performance score reflecting the frequency regulation accuracy which can reach 0.7 to 0.8 [
CSP is one of the non-fossil fuel technologies with promising applications. It does not increase the uncertainty of the system. Instead, it exploits the energy storage characteristics of TES to make solar energy a dispatchable resource [
Designing a reasonable frequency regulation operation strategy not only maintains frequency stability, but also brings potential benefits to the coordinated system. Based on this idea, a cooperation strategy of wind power and CSP for participating in the energy and regulation markets is proposed.

Fig. 1 Coordinated strategy of a coordinated system.
The frequency control error of the wind power in real-time operation is due to the intermittent or insufficient output of wind power, which is assumed and simulated as a random variable. The actual output of the CSP plant will be varied in real-time operation to compensate the control error of the wind, and the deficiency, if any, of the committed regulation capacity of the wind in DAM as compared with the required regulation capacity in real-time operation. The frequency regulation output of the CSP plant, denoted by , is shown in (1), in which the regulation-up/regulation-down capacity is considered half of the committed regulation capacity .
(1) |
The coordinated strategy takes advantage of the accuracy of CSP plant and the durability of wind turbine in providing regulation services, which can provide more regulation capacities, and accordingly higher profits as compared with the independent operations of the CSP plant and wind farm.
Note that the level of control for the cooperation strategy of the coordinated system in (1) is for the real-time operation in time resolution of k. Thus, k will not show up explicitly in the bidding model of the coordinated system in DAM (hourly) and real-time operation (15 min) as will be discussed in the next section. However, the improved performance score will indirectly affect the bids and the profitability of the joint system in DAM and real-time operation, which will be demonstrated in case study.
We assume that the coordinated system is treated like other market participants which submits hourly energy and regulation bids to the ISO in DAM. The DAM time index is t and its time resolution is 1 hour. The time index for the real-time operation is and its time resolution is 15 min. The regulation signal is a normalized signal which is updated every 2 s [
The compensation mechanism in regulation market is assumed to follow a two-part payment for regulation resources: capacity payment and performance payment [
(2) |
In the electricity market such as PJM, the performance score will be calculated by the ISO through a performance score calculation engine (PSCE) for each regulation resource after the operational hour and the score is reported to resource owners [
is compensated according to m and committed regulation capacity, which is expressed as:
(3) |
In this paper, the proposed operation of the coordinated system maximizes its revenue by properly bidding in the energy and regulation markets based on the corresponding market prices and the available wind energy.
The DAM is cleared according to hourly market bids. However, the available wind power can fluctuate drastically in real-time operation. Furthermore, the regulation performance of wind power is calculated using its real-time energy base point, which can also be set to be 1 hour. Accordingly, the wind power frequency regulation capability will be limited by its minimum available output within 1 hour. Otherwise, its frequency performance will be reduced significantly due to insufficient wind power. As shown in

Fig. 2 Illustration of bidding capacity in different time resolutions.
In this paper, the objective function is to maximize the income F of the coordinated system, which is equal to the income of each market minus the penalty cost. Note that real-time decision will vary in each real-time scenario, but the day-ahead decision will be the same.
(5) |
(6) |
(7) |
(8) |
(9) |
1) Operational constraints of CSP plant [
(10) |
(11) |
(12) |
(13) |
(14) |
(15) |
(16) |
(17) |
(18) |
(19) |
(20) |
One important CSP parameter for the participation in the electricity market is the TES level at the end of the trading day. We consider that the TES of the CSP plant returns to its initial level at the end of the trading day. The constraints are:
(21) |
(22) |
(23) |
(24) |
Considering the ramping constraints of CSP plant, its regulation-up/regulation-down capacity should not exceed a percentage of the maximum output, which is expressed as:
(25) |
(26) |
In addition, the regulation-up for the CSP plant is either half of the total regulation or the available regulation-up, whichever is smaller, as shown in (26). Similarly, the regulation-down for the CSP plant is either half of the total regulation or the real-time energy base point, whichever is smaller, as shown in (27).
(27) |
(28) |
2) Operational constraints of wind farm: wind power output cannot exceed its available capacity, which is expressed as:
(29) |
(30) |
Similarly, considering the wind farm ramping constraints, its regulation-up/regulation-down capacity should not exceed a percentage of the available capacity, which is expressed as:
(31) |
(32) |
3) Operational constraints of the coordinated system: CSP plant and wind farm should have sufficient power for the committed regulation capacity in DAM.
(33) |
(34) |
The aggregate energy base point submitted by the coordinated system is equal to the sum of the respective energy base points:
(35) |
The coordinated system in this case study consists of a wind farm with an installed capacity of 200 MW and a CSP plant with a maximum power output of 50 MW. The parameters of the CSP plant are shown in
The price scenarios are generated based on the PJM historical market data in June 2019, and historical wind power and DSI from a certain area of the Northwest China are used to generate the corresponding scenarios for the coordinated system. We implement the K-medoids parallel clustering algorithm to cluster the data into 10 scenarios [

Fig. 3 Energy and regulation market prices, and available wind power for scenario 5. (a) Hourly prices in energy market and available wind power. (b) Hourly prices in regulation market.

Fig. 4 Normalized regulation signal and wind power control error.
As shown in

Fig. 5 Day-ahead bidding strategy of coordinated system.
The real-time energy base points provided by the wind farm will change as the available wind power fluctuates, which results in the deviation between the energy base points provided in real-time operation and the energy bids provided in DAM, and accordingly a penalty. However, since the regulation performance of wind power is calculated based on real-time energy base points, the more flexible change of the energy base points compared with day-ahead energy bids enables the wind farm to commit and fulfill more regulation capacities. In

Fig. 6 Real-time bidding strategy of wind farm.

Fig. 7 Real-time bidding strategy of CSP plant.

Fig. 8 CSP plant behavior in base mode.
Four different operation modes are compared to demonstrate the benefit of coordinated operation of CSP plant and wind farm based on the proposed model. Mode 1 refers to the proposed coordinated operation of CSP plant and the wind farm; mode 2 refers the coordinated system without applying the coordinated strategy; mode 3 refers to the independent operation of the wind farm without applying the CSP plant; mode 4 refers to the independent operation of the CSP plant without applying the wind farm. We assume the performance score is 0.8 in mode 3 and 0.95 in mode 4, which considers the rapid ramping capability of the CSP plant.
Table II shows a comparison of the hourly average bidding capacity for different operation modes, where the sum of individual regulation capacities of CSP plant and wind farm is 10.2 MW, which is 1.8 MW (or 15.0%) and 7.4 MW (or 42%) less than that of the coordinated system in mode 1 and mode 2, respectively. This shows that the coordinated operation of CSP plant and wind farm can provide more regulation capacity to earn a higher profit in the regulation market. The coordinated CSP plant and the wind farm without applying the coordinated strategy tracks the regulation signal and compensates control errors together. Thus, it provides larger bids to the regulation market than the coordinated system in mode 1. Furthermore, it can be observed that the hourly energy bid of the coordinated system without applying the coordinated strategy is 85.4 MW, and the sum of hourly individual energy bids of CSP plant and wind farm is 90.1 MW, which is 15.4 MW (or 15.3%) and 10.7 MW (or 10.6%) less than the coordinated system with the proposed strategy, respectively. The wind farm tracks the regulation signal alone and compensates the wind power control error in the independent operation. Thus, it provides smaller bids to the energy market. The CSP plant with dispatchable characteristics in the coordinated system can choose to increase the energy bid in appropriate periods. Thus, the coordinated operation of CSP plant and wind farm provides energy bids to the energy market to earn a higher profit as well as regulation capacity in the regulation market while ensuring the regulation accuracy.
Table III compares the incomes of different operation modes. The coordinated operation can generate a higher extra income, which is mainly due to the increase in energy bids and the improvement in the regulation performance score through coordination. If half of this extra income in mode 1 is allocated to the CSP plant and the other half goes to the wind farm, the revenue of the CSP plant will increase by about 20.1% as compared with that of mode 4, and the revenue of the wind farm will increase by about 11.2% as compared with that of mode 3.
Table IV compares the incomes of the coordinated system for different performance scores. It can be observed that the total income of the coordinated system increases with the improvement in the performance score. The regulation market income shows an upward trend, indicating that the performance score improvement provides a strong economic incentive for regulation resources to improve their regulation accuracy. In addition, when the coordinated system has the same performance score as that of mode 3, the total incomes of the coordinated operation of CSP plant and wind power increase by 12.8% as compared with the sum of independent operations of the wind farm (model 3) and the CSP plant (model 4).
To study the impact of the installed capacity of the wind farm, we fix the installed capacity of the CSP plant and vary the installed capacity of the wind farm. We define a return on investment (ROI) metric, as shown in (36), which is the ratio of the annual income to the investment [
(36) |
and are set to be 1170 $/kW and 3200 $/kW, respectively [

Fig. 9 ROIs for different ratios of CSP plant capacity to wind farm capacity.
Similar to that in Section IV-E, the ROI is analyzed to study the impact of the installed capacity of the CSP plant for a fixed installed capacity of the wind farm.

Fig. 10 ROIs for different ratios of wind farm capacity to CSP plant capacity.
The total incomes of the coordinated system are related to the maximum regulation capacities of the CSP plant and the wind farm. We use the ratios and to characterize the maximum regulation capacities of the wind farm and the CSP plant.

Fig. 11 Total income of coordinated system.
In this paper, a coordinated strategy for CSP and wind power providing frequency regulation is proposed which is based on the complementary characteristics of CSP plant and wind farm for providing regulation services. We consider the regulation performance score and deviation penalty to establish an optimal bidding model for a coordinated operation of CSP plant and wind farm. The case study results point out that the coordinated system can provide more regulation services, alleviating the regulation pressure of the system while increasing the coordinated system income. Since the regulation error of wind has a strong relationship with the turbulence, it is determined that the gap in regulation accuracy for the wind farm could be filled by the CSP plant. The increased regulation performance score and energy bids, enabled by the CSP plant, will bring extra income, as illustrated in the results. Finally, we conduct sensitivity analysis to study the impact of various parameters such as the installed capacity of wind farm and CSP plant on the total income of the coordinated system, which can provide insights into deciding the configuration of the coordinated system.
Besides, the coordination of CSP and wind power to participate in peak load regulation is also an area of interest, which will be an interesting research area in the future.
Nomenclature
Symbol | —— | Definition |
---|---|---|
A. Indices and Sets | ||
—— | Time index in real-time operation | |
—— | Time resolution in real-time operation (15 min) | |
—— | Time resolution for regulation signal (2 s) | |
—— | Time resolution in day-ahead market (1 hour) | |
—— | Time index for regulation signal | |
—— | Index for scenario | |
—— | Set of scenarios | |
—— | Time index in day-ahead market | |
—— | Set of time | |
B. Parameters | ||
—— | Ratio of deviation penalty | |
—— | Probability of scenario | |
—— | Frequency control error of wind power at time | |
—— | Correction factor for solar incident angle | |
—— | Conversion efficiency for concentrator | |
—— | Thermal efficiency for rankine cycle | |
—— | Real-time energy market price at time in scenario | |
—— | Regulation market capacity price at time | |
—— | Regulation market performance price at time | |
—— | Unit investment cost of concentrating solar power (CSP) plant | |
—— | Unit investment cost of wind farm | |
—— | Total incomes | |
—— | Compensation fee for regulation capacity | |
—— | Penalty cost of real-time deviation at time t | |
—— | Compensation fee for regulation performance | |
—— | Normalized regulation signal at time | |
—— | Direct solar irradiation intensity at time in scenario | |
—— | Regulation performance score | |
—— | Mileage ratio for regulation resource | |
—— | The maximum/minimum output power of CSP | |
—— | The maximum available wind power at time in scenario | |
—— | Rated capacity of wind farm | |
—— | The maximum/minimum thermal storage capacity of thermal energy storage (TES) | |
—— | Initial heat storage level of TES | |
—— | The maximum charging/discharging rate of TES | |
—— | Upward/downward ramping capability of power block (PB) | |
—— | Area of solar mirror field | |
C. Decision Variables | ||
—— | Reserved regulation-down capacity for CSP plant at time in scenario | |
—— | Reserved regulation-down capacity for wind farm at time in scenario | |
—— | Capacity bid in day-ahead energy market at time | |
—— | Capacity bid in regulation market at time | |
—— | Aggregate real-time energy base point at time in scenario | |
—— | Real-time energy base point provided by CSP at time in scenario | |
—— | Real-time energy base point provided by wind power at time in scenario | |
—— | Output power of CSP plant at time in scenario | |
—— | Thermal energy absorbed from solar field (SF) at time in scenario | |
—— | Thermal energy of PB from SF at time in scenario | |
—— | Thermal energy of TES from SF at time in scenario | |
—— | Thermal energy curtailed by SF at time in scenario | |
—— | Heat storage level of TES at time in scenario | |
—— | Thermal energy of PB from TES at time in scenario | |
—— | Reserved regulation-up capacity for CSP plant at time in scenario | |
—— | Reserved regulation-up capacity for wind farm at time in scenario |
References
European Commission. (2010, Nov.). Energy 2020: a strategy for competitive, sustainable and secure energy. [Online]. Available: http://library.arcticportal.org/1536/1/LexUriServ.pdf [Baidu Scholar]
U.S. Department of Energy. (2008, Jul.). 20% wind energy by 2030: increasing wind energy’s contribution to US electricity supply. [Online]. Available: https://www.energy.gov/sites/prod/files/2013/12/f5/41869.pdf [Baidu Scholar]
J. Morren, S. W. H. de Haan, W. L. Kling et al., “Wind turbines emulating inertia and supporting primary frequency control,” IEEE Transactions on Power Systems, vol. 21, no. 1, pp. 433-434, Feb. 2006. [Baidu Scholar]
R. G. de Almeida and J. A. P. Lopes, “Participation of doubly fed induction wind generators in system frequency regulation,” IEEE Transactions on Power Systems, vol. 22, no. 3, pp. 944-950, Aug. 2007. [Baidu Scholar]
P. Cartwright, L. Holdsworth, J. B. Ekanayake et al., “Coordinated voltage control strategy for a doubly-fed induction generator (DFIG)-based wind farm,” IEE Proceedings:Generation, Transmission and Distribution, vol. 151, no. 4, pp. 495-502, Jul. 2004. [Baidu Scholar]
G. Ramtharan, J. B. Ekanayake, and N. Jenkins, “Frequency support from doubly fed induction generator wind turbines,” IET Renewable Power Generation, vol. 1, no. 1, pp. 3-9, Mar. 2007. [Baidu Scholar]
A. Zertek, G. Verbic, and M. Pantos, “Optimised control approach for frequency-control contribution of variable speed wind turbines,” IET Renewable Power Generation, vol. 6, no. 1, pp. 17-23, Jan. 2012. [Baidu Scholar]
M. Dreidy, H. Mokhlis, and S. Mekhilef, “Inertia response and frequency control techniques for renewable energy sources: a review,” Renewable and Sustainable Energy Reviews, vol. 69, pp. 1364-0321, Jul. 2017. [Baidu Scholar]
E. Saiz-Marin, J. Garcia-Gonzalez, J. Barquin et al., “Economic assessment of the participation of wind generation in the secondary regulation market,” IEEE Transactions on Power Systems, vol. 27, no. 2, pp. 866-874, May 2012. [Baidu Scholar]
Y. Chen, R. Leonard, M. Keyser et al., “Development of performance-based two-part regulating reserve compensation on MISO energy and ancillary service market,” IEEE Transactions on Power Systems, vol. 30, no. 1, pp. 142-155, Jan. 2015. [Baidu Scholar]
S. H. Madaeni, R. Sioshansi, and P. Denholm, “Estimating the capacity value of concentrating solar power plants with thermal energy storage: a case study of the Southwestern United States,” IEEE Transactions on Power Systems, vol. 28, no. 2, pp. 1205-1215, May 2013. [Baidu Scholar]
REN21. (2019, Jun.). Global status of renewable energy: REN21’s renewables 2019 global status report. [Online]. Available: https://www.ren21.net/gsr-2019/ [Baidu Scholar]
C. Philibert, P. Frankl, and Z. Dobrotkova, “Concentrating solar power: technology roadmap,” International Energy Agency, Paris, France, Tech. Rep., Jan. 2010. [Baidu Scholar]
K. Dallmer-Zerbe, M. A. Bucher, A. Ulbig et al., “Assessment of capacity factor and dispatch flexibility of concentrated solar power units,” in Proceedings of 2013 IEEE Grenoble Conference, Grenoble, France, Jun. 2013, pp. 1-6. [Baidu Scholar]
H. Sun, W. Wu, Q. Guo et al., “Reducing generation uncertainty by integrating CSP with wind power: an adaptive robust optimization-based analysis,” IEEE Transactions on Sustainable Energy, vol. 6, no. 2, pp. 583-594, Apr. 2015. [Baidu Scholar]
P. Denholm and M. Hummon, “Simulating the value of concentrating solar power with thermal energy storage in a production cost model,” NREL Tech. Rep. TP-6A20-56731, Nov. 2012. [Baidu Scholar]
G. He, Q. Chen, C. Kang et al., “Cooperation of wind power and battery storage to provide frequency regulation in power markets,” IEEE Transactions on Power Systems, vol. 32, no. 5, pp. 3559-3568, Sept. 2017. [Baidu Scholar]
R. Sioshansi and P. Denholm, “Benefits of colocating concentrating solar power and wind,” IEEE Transactions on Sustainable Energy, vol. 4, no. 4, pp. 877-885, Oct. 2013. [Baidu Scholar]
H. M. I. Pousinho, H. Silva, V. M. F. Mendes et al., “Self-scheduling for energy and spinning reserve of wind/CSP plants by a MILP approach,” Energy, vol. 78, pp. 524-534, Aug. 2014. [Baidu Scholar]
J. Aho, “Controlling wind turbines for secondary frequency regulation: an analysis of AGC capabilities under new performance based compensation policy,” in Proceedings of 13th International Workshop on Large-scale Integrated Wind Power into Power Systems, Berlin, Germany, Sept. 2014, p. 5. [Baidu Scholar]
S. Zhao, Y. Fang, and Z. Wei, “Stochastic optimal dispatch of integrating concentrating solar power plants with wind farms,” International Journal of Electrical Power & Energy Systems, vol. 109, pp. 575-583, Jul. 2019. [Baidu Scholar]
E. Du, N. Zhang, C. Kang et al., “Impact of wind power scenario reduction techniques on stochastic unit commitment,” in Proceedings of 2016 Second International Symposium on Stochastic Models in Reliability Engineering, Life Science and Operations Management (SMRLO), Beer-Sheva, Israel, Sept. 2016, pp. 202-210. [Baidu Scholar]
PJM. (2016, Jan.). PJM manual 12: balancing operations. [Online]. Available:https://www.pjm.com/-/media/documents/manuals/archive/m12/m12v37-balancing-operations-11-16-2017.ashx [Baidu Scholar]
FERC. (2011, Oct.). Frequency regulation compensation in the organized wholesale power markets. [Online]. Available:http://www.ferc.gov/whats-new/comm-meet/2011/102011/E-28.pdf [Baidu Scholar]
Spanish Electricity Market Operator. (2011, Jan.). Electricity market activity rules. [Online]. Available: http://www.omel.es [Baidu Scholar]
Y. Fang and S. Zhao, “Joint optimal operation and bidding strategies of concentrating solar power plants with wind farms,” Proceeding of the CSEE, vol. 40, no. 1, pp. 39-49, Jan. 2020. [Baidu Scholar]
L. Kotzur, P. Markewitz, M. Robinius et al., “Impact of different time series aggregation methods on optimal energy system design,” Renewable Energy, vol. 117, pp. 474-487, Sept. 2018. [Baidu Scholar]
R. Bakhshi and P. A. Sandborn, “A return on investment model for the implementation of new technologies on wind turbines,” IEEE Transactions on Sustainable Energy, vol. 9, no. 1, pp. 284-292, Jan. 2018. [Baidu Scholar]