The Impact of Participation Ratio and Bidding Strategies on New Energy’s Involvement in Electricity Spot Market Trading under Marketization Trends—An Empirical Analysis Based on Henan Province, China

Abstract

As new-energy electricity increasingly enters the post-subsidy era, traditional fixed feed-in tariffs and guaranteed purchase policies are not conducive to the optimal allocation of large-scale, high-proportion new-energy power due to the high pressure of subsidy funds and the fairness issues of power-generation grid connection. Encouraging new energy to participate in electricity market transactions is considered an effective solution. However, existing studies have presupposed the adverse effects of new energy in proposing market mechanism optimization designs for new-energy participation without quantitative results to support this, which is not conducive to a true assessment of the comprehensive impact of individual instances of new-energy participation in the market. To this end, this study, based on the actual experience and data cases of China’s electricity spot market pilot provinces, considers both unit commitment and economic dispatch in the electricity distribution process, and constructs a two-stage optimization model for electricity spot market clearing. According to the differences in grid connection time and the construction costs of new-energy power, differentiated proportions of new-energy participation in the market and bidding strategies are set. By analyzing the quantitative results of new energy participating in spot market transactions under multiple scenarios, using both typical daily data for normal loads and peak loads, the study provides theoretical support and a data basis for the optimized design of market mechanisms. The research results show that there is a non-linear relationship between the scale of new energy entering the market and its bidding strategies and market-clearing electricity prices. In the transition phase of the low-carbon transformation of the power sector, the impacts of thermal power technology with a certain generation capacity and changes in the relationship between power supply and demand on electricity prices are significant. From the perspective of the individual interests of new-energy providers, the analysis of their bidding strategies in the market is important.

1. Introduction

The global surge in new-energy electricity generation is fueled by a dual imperative: the external demand for environmental conservation and low-carbon growth, and the internal drive towards energy self-sufficiency and security. Advances in technology, coupled with economies of scale, have led to a dramatic decline in the investment and construction costs of wind and solar power. Even offshore wind technology is poised to reach grid parity with conventional energy sources. Thanks to their competitive and accessible pricing, as well as mature technology, renewable energies such as wind, solar, and biomass power have emerged as the energy methods of choice for nations undergoing energy and power transformations [1,2]. China is a case in point. According to the China National Energy Administration, by 2023, the country’s installed wind power capacity had reached approximately 441.34 GW, while solar power capacity stood at about 609.49 GW [3]. These two sources accounted for nearly 36% of the total installed capacity, marking increases of approximately 484.71% and 4020.96%, respectively, since 2013 [4]. The exponential growth of new energy in China has yielded impressive outcomes.

At the same time, we acknowledge that the growing penetration of new energy introduces structural risks to the energy and power system supply, and many countries have sequentially rolled out a suite of electricity marketization reforms. With a specific year as the cutoff, that is, the point at which wind and solar power achieve grid parity, a differentiated support policy is enforced through a combination of existing and incremental approaches. Specifically, China stipulates that new-energy power already connected to the grid before 2021 continues to receive the original fixed feed-in tariff subsidy. The incremental new-energy forms connected to the grid in 2021 and beyond will adopt a grid-parity policy on par with coal-fired power generation, and will be incentivized to participate in the electricity market through adjustments to the number of guaranteed purchase hours and the negotiation of market-based power contracts. The design of marketization mechanisms is not a one-off endeavor, but a process of progressive iteration and ongoing exploration. The market mechanisms for new-energy integration that are prevalent worldwide are chiefly aimed at medium-to-long-term horizons, utilizing instruments like contracts for difference to secure income for power generation. The crux of the issue lies in the significant impact of meteorological conditions on new-energy output, and the erratic nature of its output can lead to substantial deviation assessment costs, thus affecting generation efficiency. The spot market’s bidding mechanism, crafted on the premise of marginal cost pricing theory, holds the theoretical promise of reducing market electricity prices and societal energy costs when new energy participates in bidding, given its nearly negligible marginal cost of generation. However, this also poses a risk to the revenues of power-generation firms, which, in turn, dampens the enthusiasm of new-energy enterprises to enter the market. As the market is a critical vehicle for resource allocation, it is of profound practical significance to investigate how market-based approaches can wean new energy off its reliance on fiscal subsidies and foster its healthy and orderly development.

Currently, research on new energy’s participation in the electricity market has reached a substantial level, with a predominant emphasis on the architecture of market trading mechanisms, the dynamics of market bidding and clearing models, and the strategic decisions surrounding new-energy transactions. In the realm of trading mechanisms, scholars have predominantly delved into refining existing market structures to accommodate the unique requirements of new energy, which boasts low marginal costs but is subject to significant volatility. Du et al. [5] and Mu et al. [6] have advanced a trading mechanism that is based on energy blocks, accompanied by a corresponding clearing model that is particularly well-suited for microgrids that incorporate distributed new-energy sources. Moreover, new energy can engage in market activities by establishing virtual power plants in collaboration with other market participants, as conceptualized by Sadeghi et al. [6,7], who have devised models for virtual power plants to engage in both energy and ancillary service markets across various timeframes. A contingent of scholars posit that the entry of new energy into the competitive market landscape may disrupt the market’s resource-allocation mechanisms, particularly due to changes in clearing prices, which can distort market price signals and skew the interests of power-generation entities. To counteract this, research has explored differentiated pricing strategies for clearing and settlement, even extending to differentiated settlement prices for each generation entity, with the aim of optimizing trading mechanisms [8]. To bolster market efficacy, Shang et al. [9] have proposed a parallel trading model that aligns the demand side and the generation side, clearing markets. These studies, whether they are directed at addressing operational challenges or at analyzing foreign experiences to inform domestic mechanisms, have provided invaluable insights into the design of electricity market structures. However, in the reference process, it is essential to pay attention to multiple factors such as one’s own resource endowment characteristics and market positioning. Most Western countries focus on the development of distributed energy, and decentralized trading mechanisms may be more applicable [10].

Within the context of market bidding and clearing models, existing research primarily addresses the operational transformations introduced by new energy’s integration into market transactions, with a focus on enhancing model construction theory and optimizing solution algorithms. Li et al. [11] have integrated demand response into market transactions to counteract the volatility of new-energy grid integration, presenting a corresponding market-clearing model. Zhao et al. [12] and Wang et al. [13] have addressed the robustness challenges arising from the diversity of day-ahead market bids and the uncertainty of generation output by employing reinforcement learning methods for model resolution. Kardakos et al. [14] have proposed the use of diagonalization algorithms for model resolution. Essentially, the market trading model is consistent with the core of the classic generation scheduling model based on unit commitment and economic dispatch theory [15,16]. The market trading model further seeks the shadow price of the scheduled electricity quantity on top of the optimal scheduling results, which serves as the clearing price. These studies have offered novel perspectives on the construction and resolution of market bidding and clearing models.

In the realm of new-energy trading decisions, current research primarily concentrates on optimizing decision-making processes for the combined operation of new energy, thermal power, and energy storage, capitalizing on the temporal and spatial synergies among different power-generation entities to maximize profitability. However, the viability of such joint operations hinges on the efficacy of the profit-distribution mechanism, with pre-contractual agreements and post-contractual profit allocation being pivotal. Cooperative game theory is commonly utilized in the design of profit-distribution mechanisms [17,18]. Another strand of trading decision research is centered on the management of uncertainty, employing decision-theoretic optimization. Uncertainty is pervasive in new-energy output, consumer demand, and market pricing, which Liu et al. [19] have tackled through a data-driven approach. In addition, Sadeghi et al. [7] have employed deep learning techniques, specifically long short-term memory networks, to precisely model the demand for frequency regulation in joint bidding strategies across frequency regulation and energy markets.

Against the backdrop of the burgeoning trend of new energy’s expansion, existing research has made significant progress in strengthening the design of electricity market trading mechanisms, which has provided a solid foundation for this study in terms of market trading model construction and clearing mechanism design. However, in the face of the reality that large-scale new energy is about to participate in the electricity market, these studies, on the one hand, still follow traditional market trading methods, such as the guaranteed purchase of new energy, participation in long-term electricity transactions, and green electricity trading, to explore the design of safeguard mechanisms for new energy in long-term market transactions, rather than participating in spot market transactions. The new-energy characteristics of volatility and intermittency are better represented in the spot market. Market mechanism research at long time scales is less able to reflect the short-term uncertainty of new energy, and has weaker guiding significance for the construction of the electricity spot market. On the other hand, some studies have discussed the design of mechanisms related to new energy participating in spot market transactions, such as multi-market coupling research in the electricity energy market and ancillary service market. These studies often simplify new energy as a single generating entity, ignoring individual characteristics, and most of the research starts with the presupposition that new energy participating in the spot market will have a series of adverse effects on new-energy power-generation enterprises themselves. Therefore, more discussion is focused on how to ensure the income level of new energy through financial instruments such as price difference contracts or the marketized redistribution of benefits.

In traditional electricity market transactions, the participants on the power-generation side are mainly composed of units with stable power-generation capacity, such as coal-fired and gas-fired power. The usual practice is to deduct the predicted output of new energy from the total load demand to form a net-load curve, and then these conventional generating units compete for net-load demand to form clearing electricity prices. This method is easy to operate and can ensure that new-energy sources have priority on the grid. However, in the face of the current trend of the large-scale development of new-energy sources, continuing to follow traditional practices will rapidly reduce net-load demand, making it difficult for the system to consume a large amount of new-energy power. Moreover, this power does not form a market price through competition, making the spot market necessary. The price signal formed by the market seriously deviates from reality and there is a risk of market failure. Therefore, the research question of this study is how to promote new energy to participate in electricity spot market transactions in an orderly manner under the trend of marketization. This study adopts the perspectives of decision-makers in market mechanism design and investigates the trading decisions and policy effects of new-energy participation in electricity spot markets under different market-entry scales and bidding strategies. The key innovations of this study are encapsulated in the following: (1) investigating the influence of incremental and existing new energy on electricity spot market clearing prices under differentiated market participation scales and bidding strategies; (2) developing an electricity market transaction optimization model that accommodates the simultaneous bidding of wind, solar, and thermal power, as well as the multi-segmented bidding of generation units; (3) employing real-world data from the pilot provinces of China’s electricity spot market, in conjunction with power system simulation analysis, to conduct a quantitative analysis of the application outcomes of the electricity market clearing model.

The rest of the paper is organized as follows. Section 2 shows the research method, including the constructed two-stage power spot market clearing model and the corresponding solution process. Section 3 takes Henan, China, as a case study to conduct an empirical analysis of the optimization model. Including the basis for case selection, the sources of basic data, scenario setting and result analysis, and the effectiveness and applicability of the optimization model are verified by exploring case analysis results of different scenarios. The conclusions are presented in Section 4.

2. Methods

2.1. Construction of a Two-Stage Electricity Spot Market Clearing Model

The electricity spot market clearing model is essentially a joint dispatch model for multiple types of power-generation units. By matching supply and demand, it obtains the optimal dispatch sequence of generation resources at the unit level, and then obtains the shadow price of the marginal generator under the equilibrium condition of each time period, which is the market-clearing electricity price at this moment. Based on the theory of unit commitment and economic dispatch [20], the two-stage electricity spot market clearing model is constructed as follows.

2.1.1. First Stage: Unit Commitment Optimization

The central objective of this stage is to ascertain the startup and shutdown schedules for thermal power units, thereby establishing the essential boundary conditions for the subsequent stage of economic dispatch optimization.

(1) Objective function.

The establishment of the startup and shutdown schedules for thermal power units is a pivotal precondition for day-ahead generation planning, which is essential for maintaining the stability and resilience of the power system. Therefore, the present study seeks to minimize the overall grid dispatch cost [21], encompassing operational expenses such as unit dispatch costs, thermal power startup and shutdown expenses, and the costs associated with energy storage system calls, as outlined in Equation (1).

$$MinC={\displaystyle \sum _{t=1}^T\begin{array}{l}{\displaystyle \sum _{i=1}^N{O}_{i,t}^{th}{P}_{i,t}^{th}+u_{i,t}S{U}_{i,t}+\left(1-u_{i,t}\right)S{D}_{i,t}}+{\displaystyle \sum _{j=1}^J{O}_{j,t}^{re\text{-}p}{P}_{j,t}^{re\text{-}p}}\\ +{\displaystyle \sum _{k=1}^K{\alpha }_{k,t}^{re\text{-}s}{O}_{k,t}^{re\text{-}s}{P}_{k,t}^{re\text{-}s}}+{\displaystyle \sum _{m=1}^M{O}_{m,t}{P}_{m,t}}\end{array}}$$

(1)

where C is the overall grid dispatch cost; t and T represent the minimum trading period and the total trading period, usually 1 h and 24 h, respectively; i, j, k, and m represent the number of thermal power units, affordable new energy, subsidized new energy, and energy storage, respectively; N, J, K, and M are the corresponding total numbers of units and $O_{i,t}^{th}$, $O_{j,t}^{re\text{-}p}$, $O_{k,t}^{re\text{-}s}$, and $O_{m,t}$ are the marginal power-generation costs of these power technologies; $P_{i,t}^{th}$, $P_{j,t}^{re\text{-}p}$, $P_{k,t}^{re\text{-}s}$, and $P_{m,t}$, respectively, indicate the corresponding power outputs of these technologies; $u_{i,t}$ is a Boolean variable to represent the startup and shutdown state of thermal power unit i at time t; 1 and 0 indicate startup and shutdown, respectively; $S{U}_{i,t}$ and $S{D}_{i,t}$ are the costs of startup and shutdown; and ${\alpha }_{k,t}^{re\text{-}s}$ is an exogenous parameter that controls the market entry proportion. By adjusting its value, the impact of different market entry proportions can be analyzed.

(2) Constraints.

① Electricity equilibrium constraint.

For any given time t, the electricity supply and demand must be balanced.

$${\displaystyle \sum _{i=1}^N{P}_{i,t}^{th}}+{\displaystyle \sum _{j=1}^J{P}_{j,t}^{re\text{-}p}}+{\displaystyle \sum _{k=1}^K{P}_{k,t}^{re\text{-}s}}+{\displaystyle \sum _{m=1}^M{P}_{m,t}}=D_t+{\displaystyle \sum _{m=1}^M{F}_{m,t}}$$

(2)

where $D_t$ represents the electricity demand and $F_{m,t}$ represents the energy storage charging demand.

② Output upper- and lower-limit constraints of generation technologies.

$$\left\{\begin{array}{l}{g}_{i,t}^{\min }{G}_i^{th}\le P_{i,t}^{th}\le G_i^{th}\\ 0\le P_{j,t}^{re-p}\le g_{j,t}^{\max }{G}_j^{re-p}\\ 0\le P_{k,t}^{re-s}\le g_{k,t}^{\max }{G}_k^{re-s}\\ 0\le P_{m,t}\le G_m\end{array}\right.$$

(3)

where $g_{i,t}^{\min }$ is the technical parameter of the minimum output of the thermal power units. Because the hidden danger of boiler operation will occur when the output of the thermal power unit is too low, and the startup and shutdown cost of thermal power is high, the minimum output power is restricted, and it is regarded as shutdown when the actual power is less than the minimum power. $g_{j,t}^{\max }$ and $g_{k,t}^{\max }$, respectively, represent the maximum output technical parameters of the two new-energy power-generation technologies, and mainly depend on the endowment conditions of natural resources such as wind and solar energy in the region where the case study is located; $G_i^{th}$, $G_j^{re-p}$, $G_k^{re-s}$, and $G_m$, respectively, represents the installed capacity of the units i, j, k, and m, that is, the rated power.

③ Technical constraints on the operation of power technologies.

In addition to the aforementioned constraints, the safety-constrained unit commitment model also needs to consider technical constraints of various types of power-generation units, such as the ramping constraint of thermal power units, the startup and shutdown time constraint of thermal power units, the output conversion constraint of new energy units, the daily balance constraint of energy storage equipment, and so on. Since these technical constraints already have mature equation expressions, they will not be repeated here; readers can refer to the literature [22,23].

2.1.2. Second Stage: Economic Dispatch Optimization

Through the calculation of the first-stage model, the startup and shutdown state variable values of each thermal power unit can be obtained, which can then be used to formulate the startup and shutdown schedules for thermal power units under safety constraints. This result is introduced into the second-stage model as a parameter input for conducting economic dispatch optimization, thereby allocating the generation sequences and each time’s market-clearing electricity price.

(1) Objective function.

The second-stage model still aims to minimize the grid dispatch cost, but it should be noted that the startup and shutdown costs of thermal power have been excluded. This is specifically outlined in Equation (4).

$$MinC={\displaystyle \sum _{t=1}^T{\displaystyle \sum _{i=1}^N{O}_{i,t}^{th}{P}_{i,t}^{th}}+{\displaystyle \sum _{j=1}^J{O}_{j,t}^{re\text{-}p}{P}_{j,t}^{re\text{-}p}}+{\displaystyle \sum _{k=1}^K{\alpha }_{k,t}^{re\text{-}s}{O}_{k,t}^{re\text{-}s}{P}_{k,t}^{re\text{-}s}}+{\displaystyle \sum _{m=1}^M{O}_{m,t}{P}_{m,t}}}$$

(4)

(2) Constraints.

In the second stage, except for the startup and shutdown state and time constraints of the thermal power units, which can be cancelled due to the direct use of the corresponding optimization results of the first stage, the other constraints are consistent with the first stage.

(3) Bidding method of power units participating in the market.

Currently, the mainstream pricing theory of electricity spot markets is based on the marginal cost pricing theory of microeconomics, which is also applied by countries worldwide for market transactions and the design of generator bidding strategies. For new energy units, the marginal cost of generation is zero, thus allowing for more flexibility in their bidding strategies. To ensure priority in clearing, they can bid at a very low price; to maximize generation revenue, they can reference the bidding prices of thermal units and historical market-clearing electricity price data when bidding, as shown in Equation (5). For thermal units, market designers often use a segmented and incremental approach to guide their bidding, both to encourage them to bid as close to their marginal cost as possible and to ensure the efficiency of the clearing model’s solution, as depicted in Equation (6). On the one hand, by quoting the unit’s power generation in sections, it can encourage the unit to quote according to the real marginal cost of power generation in different operating intervals, which helps to ensure the minimization of system operating costs. On the other hand, through incremental bidding, it is possible to ensure that the function of power generation and bidding price is a convex function, which ensures the efficiency of model solving.

$$O_{j,t}^{re-p},O_{k,t}^{re-s}\in \left[p^{q-\min },p^{q-\max }\right]$$

(5)

where $p^{q-\min }$ and $p^{q-\max }$, respectively, represent the lower and upper limits of the bidding prices for power units in the spot market. The units can decide on their bidding behavior within the price limits according to their operational strategies.

$$\left\{\begin{array}{l}{O}_{i,t}^{th}=p^{q-\min },0\le P_{i,t}^{th}\le g_{i,t}^{\min }{G}_t^{th}\\ O_{i,t}^{th}=p^{q-\min }+{\alpha }^1,g_{i,t}^{\min }{G}_t^{th}<P_{i,t}^{th}\le g_{i,t}^1{G}_t^{th}\\ O_{i,t}^{th}=p^{q-\min }+{\alpha }^2,g_{i,t}^1{G}_t^{th}<P_{i,t}^{th}\le g_{i,t}^2{G}_t^{th}\\ \dots \\ O_{i,t}^{th}=p^{q-\min }+{\alpha }^n\le p^{q-\max },g_{i,t}^n{G}_t^{th}<P_{i,t}^{th}\le g_{i,t}^{\max }{G}_t^{th}\end{array}\right.$$

(6)

wherein thermal power units adopt a segmented and incremental bidding method, with corresponding bidding prices set according to different power output segments. ${\alpha }^n$ represents the increment in the bidding price compared to the price floor for the nth segment of output; $g_{i,t}^n$ represents the output regulation parameter for the nth segment of declaration. Market designers can adjust the value of n based on practical conditions to determine the level of detail in the declared output, aiming to guide thermal power units to bid closer to their actual marginal costs.

(4) Market-clearing prices and settlement prices.

By solving the shadow prices of the power balance constraints in the second stage, the market-clearing prices for the spot market at different trading times t can be obtained. Furthermore, based on the winning bid quantities of each unit, the average settlement prices for each power technology can be calculated. Taking thermal power as an example, Equation (7) illustrates the calculation of the average settlement price.

$$p_i^{spot}={\displaystyle \frac{{\displaystyle \sum _{t=1}^T{p}_t^{\lambda }{P}_{i,t}^{th}}}{{\displaystyle \sum _{t=1}^T{P}_{i,t}^{th}}}}$$

(7)

wherein $p_i^{spot}$ is the weighted-average settlement price for thermal power unit I; $p_t^{\lambda }$ is the real-time clearing price at time t.

2.2. Model Solution

Figure 1 illustrates the steps for solving the two-stage model constructed in this study. The two stages of the optimization model developed in this study are unit commitment and economic dispatch. By determining the startup and shutdown plans for thermal power units in the first stage and using them as input data for the second stage, the difficulty of solving the clearing price problem due to the nonlinear constraints can be reduced. Specifically, the first step is model data input, which includes market-related data, such as the types, quantities, and installed capacities of units participating in the market, as well as boundary condition data, such as typical daily load demand, the output characteristic functions of wind and photovoltaic power, and the application of scenario generation methods based on historical data and scenario reduction methods based on clustering. The second step is model construction, as described in Section 2.1. Due to the two-stage optimization approach, the efficiency of the model solution can be improved, and the model at each stage can be solved using the Gurobi 11.0 solver [24,25] on the Python 3.9.12 platform. Moreover, it should be noted that during the modeling and solution process, this study applied the integrated platform PyPSA developed based on Python, which provides a fundamental power system code to ensure the scientificity of power system operation simulation analysis while reducing the difficulty of modeling [26]. This platform has been widely used in academia and engineering fields.

3. Empirical Analysis

3.1. Case Overview

Henan Province is located in the middle and lower reaches of the Yellow River in central China, with a typical temperate monsoon climate. It is well connected by transportation with a developed railway network, and is one of the important starting points for the China–Europe freight trains in the “Belt and Road” initiative. It is a representative strong economic and large agricultural province in China. Henan’s wind and solar resources are not at the leading level overall compared to the northwestern provinces of China due to differences in wind speed and solar radiation caused by meteorological conditions. But due to declines in investment and construction costs brought about by technological progress, especially the government’s in-depth promotion of the energy revolution and the acceleration of the planning and construction of a new energy system, Henan has made significant progress in new-energy development in recent years. According to the statistical data obtained from the field research of State Grid Henan Electric Power Company, Zhengzhou, China, by the end of 2023, the province’s renewable energy-generation capacity reached nearly 100 billion kilowatt-hours, with a total installed capacity of 67.76 million kilowatts. Of this capacity, wind power accounted for 32.1%, photovoltaic power accounted for 55.1%, hydropower accounted for 7.9%, and other renewable energy sources accounted for 4.9%. As one of China’s pilot provinces for the construction of the second batch of electricity spot markets, Henan started long-term trial operations in the spot market as of December 2023, providing a relatively sufficient foundation for power-generation entities to participate in market transactions and cultivate a competitive mindset. The rapid rise and development of new energy in Henan Province have posed certain challenges to the assurance of local new-energy absorption levels and grid security, leading to a significant increase in the cost of low-carbon transformation in the energy sector. In terms of managing the cost of low-carbon transformation, market trading is currently a widely recognized approach. However, the impact of important parameters during the implementation process, such as the mode, scale, and progress of market entry, still requires quantitative analysis. Consequently, conducting an impact analysis of high-proportion new-energy market entry in Henan Province is both data-feasible and necessary for research.

3.2. Market Trading Mechanism Design and Basic Data

During the trial operation of the Henan electricity spot market, a centralized trading method was adopted, with full-volume declarations and centralized optimization clearing. Coal-fired thermal power units on the generation side participated in the market with a quantity-based bidding method, while new-energy and electricity users participated in the market with a quantity-based method, acting as market price takers and not participating in the bidding. Based on this, the basic trading mechanism of the spot market after new-energy entry was designed, as shown in Figure 2.

Table 1. Basic data on generating units participating in the spot market in Henan.

Type of Generating Unit		Installed Capacity (MW)	Mode
Thermal power		67,835	All
Wind power	Subsidized	15,183	10% of electricity
	Grid parity	6596	All
PV	Subsidized	6039	10% of electricity
	Grid parity	349	All

shows the installed capacities of various types of power-generation technologies participating in the spot market, obtained through field research. It can be seen that, under the basic trading mechanism, public coal-fired power-generation enterprises participate in the spot market with their entire electricity output. Centralized new energy is divided into two categories according to the difference in its grid connection time (before and after 31 December 2020): subsidized new energy and grid-parity new energy. Among them, the subsidized new energy, due to being connected to the grid before 31 December 2020, still implements the national benchmark feed-in tariff policy, so during the trial operation phase, only 10% of the electricity output participates in market trading, while the rest is acquired through government-authorized contracts for guaranteed purchase. Centralized new-energy stations connected to the grid after 31 December 2020, due to the implementation of the national subsidy reduction policy, therefore, like public coal-fired thermal power, implement full electricity participation in market trading.

Figure 3 shows the typical daily load demand in Henan Province. This includes spring and summer, and is used for subsequent comparative analysis. Figure 4 displays the predicted power-generation values of wind and photovoltaic power in Henan Province corresponding to two typical days, obtained through the clustering analysis of historical data collected from the Henan power grid.

3.3. Scenario Setting

After clarifying the boundary conditions with the types of unit and corresponding installed capacities, typical daily load curves, and the output curves of wind and photovoltaic power, scenarios are set as shown in

Table 2. Scenario settings for new-energy participation in market trading.

Seasons	Methods	Strategy	Subsidized New Energy	Scenarios
Spring/Summer	Declare quantity without quotation	None	10%	Passive business as usual (P-BAU)
			50%	Passive large-scale new energy (P-LSN)
	Declare quantity and quotation	Price floor	10%	Active business as usual (A-BAU)
			50%	Active large-scale new energy (A-LSN)
		Average price of thermal power	10%	Positive active business as usual (PA-BAU)
			50%	Positive active large-scale new energy (PA-LSN)
		Slightly higher than thermal power	10%	Aggressive active business as usual (AA-BAU)
			50%	Aggressive active large-scale new energy (PA-LSN)

.

Key discussions are focused on (a) the impact of new-energy market entry in different periods (regular and peak summer) on the simulated settlement average price of the spot market; (b) the impact of the scale of subsidized new-energy market entry on the simulated settlement average price of the market when new energy declares quantity without quotation; (c) and the impact of different bidding strategies and market-entry proportions of new energy on the simulated settlement average price of the market when new energy declares quantity and quotation.

3.4. Results Analysis

Through comparative analysis of the basic scenarios and in conjunction with the actual situation in Henan Province, the two-stage market-clearing model constructed in this study can be verified to have certain applicability. Figure 5 shows the clearing prices of the Henan spot market under typical days in scenario P-BAU, with an average clearing price of 125.49 CNY/MWh in spring and 255.93 CNY/MWh in summer. Combined with the output allocation results of the various types of power units shown in Figure 6, it can be seen that the main reasons for the lower average clearing price in the spring typical day are threefold: first, the overall load demand in spring is lower than in summer, leading to an excess supply situation in the electricity market; second, the output level of new energy, especially wind power, is significantly higher in spring than in summer; and third, the transaction method of new energy declaring quantity without quotation and prioritizing clearing lowers the average clearing price by further reducing the output level of thermal power with higher marginal generation costs. In contrast, in the summer typical day, due to the overall higher level of load demand, thermal power needs to generate at peak times to ensure the safety and stability of the power system, thereby raising the market-clearing price. Therefore, it can be inferred that this model can effectively simulate the reality of electricity trading in Henan Province and can be used as a basis for further research.

On the basis of the P-BAU scenario, when the proportion of new energy entering the market is further increased, the impact on the clearing price is nonlinear. Figure 7 shows the market-clearing results in spring and summer in the P-LSN scenario. It reveals that as the electricity output of new energy participating in the market increases, the average price in spring decreases by about 1.26 CNY/MWh, while the average price in summer increases by 72.16 CNY/MWh. The reason may lie in the fact that the power supply and demand in Henan Province are already in a tight balance during the summer, and when the electricity output of new energy participating in the market increases, the original new-energy priority trading electricity output authorized by the government also needs to be traded through the market, further increasing the load level on the demand side of the electricity market. This leads to a greater trend of supply being insufficient in peak load periods, disrupting the tight balance of the summer and causing electricity prices to rise.

The roles of new-energy providers in reducing the clearing average price after participating in the market are different, and the differentiation effect is mainly attributed to their respective generation capabilities and output characteristics.

Table 3. Differentiated settlement average price results of different units under P-BAU and P-LSN scenarios.

Seasons	Scenarios	Overall Average Price	Average Price of Thermal Power	Average Price of Wind Power	Average Price of PV
Spring	P-BAU	125.49	127.73	104.69	100.60
	P-LSN	124.23	127.73	104.69	100.60
Summer	P-BAU	255.93	256.21	245.88	256.38
	P-LSN	328.09	329.32	302.31	330.10

shows the settlement average price results of different units under the P-BAU and P-LSN scenarios. It can be seen from the weighted-average clearing prices of electricity output from different generation technologies that the decrease in the spring market-clearing average price is the result of the joint efforts of wind power and photovoltaic power, with photovoltaic power having the lowest average price as it becomes the marginal unit during the trough of noon load. In contrast, the decrease in the summer mainly relies on wind power, while the weighted-average transaction price of photovoltaic and coal-fired power is slightly higher than the overall average price. Combining the changes in load demand and new-energy output, it can be inferred that the reason lies in the greater synergy between new-energy output and load demand in spring, that is, the peak load demand is also the peak of new-energy output, leading to a significant reduction in the market-clearing electricity price. In summer, load demand increases significantly by about 50%, and power supply becomes tight during some periods. There is a more pronounced anti-peak characteristic, with the peak load demand coinciding with the low point of new-energy output, resulting in a higher frequency of peak electricity prices. However, wind power, due to its relatively high overall generation efficiency and the ability to generate electricity throughout the day, helps to lower the market electricity price to some extent.

3.5. Impact of Bidding Strategies for New Energy

The market participation strategy of declaring quantity without quotation makes new-energy enterprises price takers, making it difficult for them to play their role in market transactions. To address this, by comparing six scenarios under the strategy of declaring quantity and quotation, the results of new energy’s active trading decisions when participating in the market are explored.

Table 4. Typical daily trading simulation results of new energy actively participating in the market.

Method	Scenarios	Average Clearing Price (CNY/MWh)
		Spring	Summer
Declare quantity and quotation	A-BAU	125.49	302.99
	A-LSN	124.23	302.55
	PA-BAU	126.85	261.05
	PA-LSN	126.79	261.05
	AA-BAU	130.10	271.80
	AA-LSN	144.52	287.48

shows the simulation results of new energy under different bidding strategies on typical spring and summer days. Overall, the average clearing price in the spring market is lower than that in the summer, which is consistent with the basic analysis results in the previous section, further validating the effectiveness of the model. Specifically, in terms of the impact of increasing the proportion of new energy trading under different bidding strategies, whether in spring or summer, when the bidding price is less than or equal to the average price of thermal power, the increase in proportion will lead to a decrease in price, and the magnitude of the decrease is closely related to the bidding price and the market’s supply-and-demand status. The lower the bidding price or the more relaxed the market’s supply-and-demand status, the greater the decrease. However, when the new-energy bidding price is higher than that of thermal power, regardless of the scenario, the increase in the proportion of new energy will instead raise the market price. Combined with the dispatching results shown in Figure 8, it can be inferred that the main reason is that in the early stage of the low-carbon transformation of the power system, thermal power units still occupy a certain proportion of the power-generation capacity structure. After losing the legal guarantee of priority clearing, the behavior of new energy to bid at a high price will actually allow thermal power units to obtain a greater generation share, making new energy the marginal unit. In terms of the impact of new-energy bidding strategy changes on different market trading scales, on the typical spring day when the new-energy trading scale remains unchanged, at this time, the market demand is relatively low, and the supply-and-demand relationship is relatively relaxed. The higher the new-energy bidding price, the higher the market-clearing average price, and new energy can obtain excess profits through free-riding behavior. However, on the typical summer day, when the new-energy trading scale remains unchanged, there is a “U”-shaped curve relationship between the new-energy bidding price and the market-clearing price, with the left side being higher than the right side, as shown in Figure 9. The closer the new-energy bidding price is to the average price of thermal power, the lower the clearing price. The reason may be that, at this time, the overall demand is relatively high, and the pressure to ensure power supply is large. New energy bidding at a low price to pursue priority clearing can maximize its own generation profit, but it may increase the operational risk of the power system, leading to a surge in spot prices. While new energy bidding at a high price can also raise the market average price, it may also face the risk of not being awarded.

Overall, compared to the government-authorized contract electricity price (377.9 CNY/MWh) currently set in Henan, the clearing prices obtained in the various trading scenarios in this study are lower. From a societal perspective, this is beneficial to electricity users but detrimental to power-generation entities. For subsidized new energy subject to fixed feed-in tariffs, the generation income will significantly decrease, which will greatly suppress the willingness of new energy to participate in market transactions. The spot market trading mechanism based on marginal cost pricing theory, facing the inevitable trend of the vigorous development of new energy, can bring some cost relief to electricity users, but this is also based on the premise that system stability operation costs have not been effectively diverted. If the system costs added due to new-energy grid integration are included, both the power generation and electricity consumption sides may suffer from interest damage. Based on this, to promote the healthy and orderly development of the electricity market and encourage new energy to participate in market transactions, during the transition period of spot market construction, it is still necessary to study the corresponding new-energy participation in market transactions and the subsequent income guarantee mechanism to stimulate participation enthusiasm.

4. Conclusions

In response to the need for energy transformation, the construction of a new power system characterized by the large-scale development of new energy makes the traditional fixed feed-in tariff policy for new energy unsustainable due to excessive fiscal subsidy pressure. It has become an inevitable trend to promote new energy’s participation in the electricity market and to divert system costs through market-oriented means. To guide the trading strategies and mechanism design of new-energy participation in the market, this study constructs a two-stage clearing optimization model for the electricity market, with new energy competing alongside thermal power based on classical unit commitment and economic dispatch theories. Taking Henan Province, China, as an example, this study compares and discusses the differential impact of different trading proportions and bidding strategies of new energy on market prices and draws the following conclusions:

(1) The impact of new-energy participation in the spot market on clearing prices cannot be generalized. It is greatly influenced by the level of load demand and new energy bidding strategies. The better the time synergy between new-energy output and load demand, the lower the clearing price. When the anti-peak characteristics of new-energy output are more pronounced, even with a large scale of new-energy priority grid connection, the market-clearing price will fluctuate significantly due to the rise in system regulation costs, leading to frequent peak electricity price risks.

(2) Different market players have different preferences for new-energy bidding, and this difference is not linear. For new energy, when the power supply-and-demand situation is relatively stable, the higher the bidding price, the easier it is to obtain excess profits. When supply and demand are tight, the lower the bidding price, and the more excess profits can be obtained through a larger amount of electricity, winning the bidding. For electricity users, when the supply-and-demand situation is relatively stable, the lower the new-energy bidding price, the larger the trading scale, and the lower the clearing electricity price. When supply and demand are tight, the new-energy bidding price aligning with that of thermal power can actually lower the clearing price.

(3) In an environment where thermal power units still account for a large proportion of the industry and have not yet been phased out on a large scale, even in the spot market based on the mainstream marginal cost pricing theory, new energy cannot intentionally raise its bidding price due to its low marginal generation characteristics. Otherwise, the loss of market share due to the bidding risk may be even greater than the increase in income brought by the rise in electricity prices.

We have explored the impacts of different new-energy entry methods and offer strategies in relation to the spot market; although we have set up as many typical scenarios as possible for comparative demonstration, the Henan provincial electricity spot market is still in the stage of short-cycle trial operation, and has not been able to accumulate additional historical data. So, the conclusions of this study have certain limitations due to data availability, mainly reflected in the fact that the impacts of medium- and long-term electricity transactions on the spot market have not yet been taken into account, and the simulation of the operation of the spot market is still based on the basic data of a typical day. Therefore, the conclusion of this study has some limitations, mainly reflected in the fact that the impact of long-term power trading on the spot market has not been considered, and the operation simulation of the spot market is still based on the basic data of a typical day. In the future, we will further extend the simulation period and expand the case objects to deepen the study after obtaining more actual operation data.