Levelized Cost of Electricity Prediction and End-User Price Deduction Model for Power Systems with High Renewable Energy Penetration

Abstract

With the rapid growth in the scale of high-percentage new energy generation, the structure of the new power system is changing. Influenced by the uncertainty and zero marginal cost characteristics of new energy, the security cost required by the power system under the high proportion of new energy access has increased dramatically. How to accurately measure the cost of the power system and assess the trend of the system cost changes and the impact on its end-user price has become critical. Accordingly, this paper creatively proposes a levelized cost of electricity (LCOE) prediction and end-user price deduction model for power systems with high renewable energy penetration. Firstly, the power system factor cost prediction model is constructed from the three dimensions of power-side, grid-side, and system operation cost. Secondly, a levelized cost of electricity prediction model is constructed based on the above model. Again, based on the analysis of the end-user price composition, the end-user price deduction model is proposed. Finally, the data of Gansu Province is selected for example analysis, and the results show that, in 2060, the power LCOE will be 0.064 USD/kWh, the system LCOE will be 0.103 USD/kWh, and the end-user price will rise to 0.1 USD/kWh.

1. Introduction

As the energy revolution continues to advance, the scale of China’s high-proportion new energy power generation is growing at a rapid pace, and the structure of the new power system is undergoing fundamental changes. In July 2024, General Secretary Xi Jinping emphasized the need to speed up the construction of a new-type energy system and a new-type power system at the Third Plenary Session of the 20th CPC Central Committee, offering guidance and a pathway for advancing the energy transition. In the same month, the “Action Plan for Accelerating the Construction of a New Type of Electric Power System (2024–2027)” was published, which centered on the basic principles of “clean and low-carbon, safe and abundant, economically efficient, synergistic between supply and demand, and flexible and intelligent,” and which proposed nine special actions to accelerate the construction of a new type of power system to achieve results and to enhance the grid’s ability to consume, configure, and regulate new energy. Large-scale penetration of a high percentage of new energy will reshape the cost structure and features of the power system. Influenced by the zero marginal cost characteristics of new energy, the aggregate cost of the electrical energy within the system reduces, while due to the high randomness and uncertainty of new energy, the required regulation resources increase significantly, and the cost of security rises dramatically. Determining how to accurately measure the total system cost with high renewable energy penetration, and assessing how new energy access affects the cost characteristics of the power system and end-user prices, are difficult problems that must be addressed.

Precise cost prediction is critical for realizing the extensive application and sustainable growth of new energy. In terms of system cost prediction, Ref. [1] analyzes the new energy system cost. Ref. [2] predicts the future cost of offshore floating photovoltaics through historical economic data. Ref. [3] calculates the carbon emission and power generation cost prediction of a synergistic power plant using the life cycle assessment method. Ref. [4] predicts the cost of the upcoming tendency of solar power generation based the learning curve model. Ref. [5], by analyzing the cost structure of photovoltaic power generation, constructs a learning curve model to compute the power generation costs of photovoltaic power plants and predict future cost trajectories. In Ref. [6], starting from Wright’s basic learning curve model, a two-factor learning curve model is formulated to estimate the cost of solar photovoltaic power generation in China in the coming ten years. Ref. [7] developed learning curve models for wind power costs to analyze the economic feasibility of wind power costs. The current analysis of new energy system cost mainly focuses on the field of wind power and photovoltaic power generation, and the methods of prediction mainly adopt diverse technical means such as historical economic data prediction, analysis of specific production technology routes, the life cycle assessment method, and the learning curve model. These methods have played a crucial role in boosting the accuracy of new energy system cost prediction and revealing the cost change trends and influencing factors. However, most of the current studies still have the problems of focusing on the cost prediction of a single energy type or a specific technology route, relying too much on historical data in their models, and the limitations of historical data due to the rapid development of energy technology. In addition, in the study of levelized cost of electricity (LCOE), Ref. [8] proposes the calculation method of wattage cost and LCOE from the production technology route, material cost, production line cost, and depreciation, and predicts its LCOE. Ref. [9] measured the LCOE of electrochemical energy storage and analyzed the multiple factors affecting LCOE. Ref. [10], in order to assess the approximate change in cost over time, calculated the levelized cost of electricity using a life cycle assessment and life cycle cost approach that considered economic and environmental impacts. Current research on electricity prices is mainly focused on measurement, and there are relatively few studies on the long-term trend of electricity prices. Therefore, the structure and operating characteristics of the electricity market should be fully understood, and a system cost and LCOE prediction model suitable for high renewable energy penetration should be established to accurately predict and optimize the cost structure.

Accurate analysis of end-user price is a prerequisite for guiding the power grid and the government to rationally formulate diversion policies. Under the current market mechanism, the pace of investment in new energy installations is too fast, and the power system needs more regulating resources to maintain the balance and stability. However, the Notice on the Establishment of Capacity Tariff Mechanism for Coal-fired Power and the Notice on Transmission and Distribution Tariffs for Provincial Grids in the Third Regulatory Cycle and Related Matters stipulate that a two-part tariff mechanism is implemented for coal-fired power and pumped storage and that capacity tariffs are included in the system’s operating fees to channel to industrial and commercial users. Therefore, the increase in the installed scale of coal-fired power and pumped storage in the power system with high renewable energy penetration will lead to higher electricity prices for users. Ref. [11] analyzes the composition of the electricity price to the household and the impact of each component on the final price. Ref. [12] uses artificial neural network methodology to carry out electricity-price-level forecasting. Ref. [13] believes that the low adaptability of the traditional forecasting model and the difficulty of improving forecasting accuracy in the context of electricity market reform constitute a forecasting dilemma. Therefore, there is a need to explore the forecasting model adapted to the new situation in order to solve this problem. Ref. [14] believes that large-scale, high proportions of new energy access will have an impact on the level of electricity prices, combined with feed-in tariffs, investment, and other factors affecting the proposed system equivalent tariff calculation method. Ref. [15] proposes a tiered marginal pricing mechanism for electricity spot markets that takes into account the affordability of customer-side tariffs. Ref. [16] carries out a study on the calculation method of step tariffs in the electricity market based on user demand to meet the differentiated needs of different types of user tariff pricing. Ref. [17] establishes a pumped storage capacity tariff measurement model to quantify the impact of pumped storage power plant development on the terminal electricity price. Ref. [18] refines the pumped storage capacity tariff mechanism at the current stage and analyzes the mechanism of its impact on user electricity prices. Most of the current studies are on the impact of the large-scale development of a single power source on users. However, there are fewer studies on the derivation of the end-user price of a new type of power system with high renewable energy penetration, on the trend of changes in user tariffs from a multi-dimensional perspective, and on the consideration of the impact of the system cost on the end-user price after the increase in the cost of multiple regulating resources in the process of energy transition.

To tackle the above problems, this study puts forward a model for forecasting the LCOE of power systems and deriving end-user prices with high-penetration new energy access. By analyzing the new energy transition path, measuring the cost evolution trend of the power system under the high proportion of new energy access, and analyzing the trend of user electricity price change from the two dimensions of the full cost of the power system and the cost of regulating resources, we clarify and quantify the impact of the energy transition cost on the end-user electricity price.

2. Power System Factor Cost Prediction Model

2.1. Power System Cost Components

Power system cost refers to all the costs incurred during the design, construction, operation, and maintenance of the power system, mainly including power-side costs, grid-side costs, and system operation costs.

Among them, power-side costs denote the costs associated with power generation in the power system, encompassing the construction, operation, and maintenance costs of coal-fired power, hydropower, new energy storage, pumped storage, and new energy. Grid-side costs refer to the costs associated with the transmission and distribution networks in the power system, mainly transmission and distribution costs, including permitted costs, permitted returns, and taxes. System operation costs refer to the costs required to provide safe, reliable, and efficient operation of the power system, including coal-fired power flexibility retrofit costs, demand response costs, and ancillary services costs. Power system cost components are shown in Figure 1.

2.2. Model Building

2.2.1. Two-Factor Learning Curve Cost Prediction Model

The conventional learning curve model, containing merely a single explanatory variable, disregards the impact of technological research and development (R&D) on technological costs, frequently resulting in large learning efficiency estimates [19]. This study integrates the factor of technological advancement into the learning curve and utilizes a two-factor learning curve to forecast trends in the power-side costs of power systems. The expression is presented below.

$${\mathrm{C}}^t=C_0{\left(I^t\right)}^{-a}{\left(R^t\right)}^{-b}$$

(1)

$$L_b=2^{-b}$$

(2)

where $R^t$ stands for the variable of technological R&D progress, $b$ stands for the elasticity coefficient, and $L_b$ stands for the learning rate. This paper takes cumulative R&D investment as a variable to illustrate technological progress. The model expression is as follows:

$$C^t=C_0{\left(E_{total}^t\right)}^{-a}{\left(R^t\right)}^{-b}$$

(3)

The coefficients $a$ and $b$ in the formula are unknown, and the linear form is obtained by taking logarithms on each side of the formula:

$$\ln C^t=\ln C_0-a\ln E_{total}^t-b\ln R^t$$

(4)

We use the least squares method to obtain the optimal estimates of the elasticity coefficients a and b. After obtaining the coefficient estimates, we use the goodness-of-fit index to evaluate the fit.

2.2.2. Linear Regression Grid Cost Prediction Model

The increase in the new energy installed scale and load demand will cause the increase in grid-side cost. In order to explore the development trend of grid-side costs at key future time nodes, this paper refines the new energy installed scale and load as explanatory variables of the multiple regression model; the expression is shown below:

$${\mathrm{C}}_{grid}={\beta }_1{P}_t+{\beta }_2{L}_t+\epsilon$$

(5)

where $C_{grid}$ is the grid-side cost, $P_t$ is the new energy installed capacity, and $L_t$ is the load.

2.2.3. Production Simulation System Operating Cost Prediction Model

System operating costs are the costs of providing safe, reliable, and efficient operation of the electric power system, including coal-fired power flexibility retrofit costs, demand response costs, the cost of adding independent energy storage, and system frequency costs. Of these, system frequency costs primarily include frequency mileage costs and frequency capacity costs. Capacity costs and mileage costs in the actual settlement process are the accumulation of costs in each settlement cycle, because the operating effect of each frequency resource is different in practice, so the comprehensive frequency performance index of each frequency resource is considered in the calculation of costs, and the single-day capacity cost $C_{cap}$ and mileage cost $C_{mil}$ are, respectively:

$$C_{cap}={\displaystyle \sum _{t=1}^{96}{\displaystyle \sum _{i\in N}{C}_{i,t}{K}_{i,t}{p}_t^{cap}}}$$

(6)

$$C_{mil}={\displaystyle \sum _{t=1}^{96}{\displaystyle \sum _{i\in N}{M}_{i,t}{K}_{i,t}{p}_t^{mil}}}$$

(7)

The system operating cost measurement formula is:

$$C_{O\&M}=C_{str}+C_{tra}+C_{dev}+C_{cap}+C_{mil}$$

(8)

where $C_{i,t}$ indicates the winning capacity of resource $i$ in time period $t$. $M_{i,t}$ indicates the winning frequency mileage. $K_{i,t}$ indicates the composite frequency performance index. $p_t^{cap}$ and $p_t^{mil}$ indicate the capacity price and mileage price of the unified clearing of the market $i$. $C_{str}$ indicates the cost of independent storage. $C_{tra}$ indicates the cost of coal flexibility retrofits, and $C_{dev}$ indicates the demand response cost.

The system operating costs, other than frequency regulation costs, are measured based on production simulation using the power system with high renewable energy penetration, with the aim of minimizing system costs in the planning period; the specific objective functions are set out below.

$$\left\{\begin{array}{l}\min C_{sys}={\displaystyle \sum _{t=1}^T\left(C_{gene,t}+C_{str}+C_{tra}+C_{dev}+C_{uhv}+s_{j,t}^{on}\right)}+{\displaystyle \sum _{t=1}^T(C_w{P}_{wd}^t+V_d{D}_{wd}^t)}\\ C_{gene,t}=C_{TP}{P}_{TP,t}+C_{PV}{P}_{PV,t}+C_{WPP}{P}_{WPP,t}+C_W{P}_{W,t}+C_{ES}{P}_{ES,t}+C_P{P}_{P,t}\\ C_{str}={\displaystyle \sum _{k\in N_{str}}\frac{r{\left(1+r\right)}^{y_k}}{{\left(1+r\right)}^{y_k}-1}\times \left(c_{str,k}^p{P}_{k,\max }+c_{str,k}^e{E}_{k,\max }\right)}\\ C_{tra}={\displaystyle \frac{r{(1+r)}^{y_t}}{{(1+r)}^{y_t}-1}}\times c_{tra}{\alpha }_{tra}{P}_{coal}\\ C_{dev}={\displaystyle \frac{r{(1+r)}^{y_d}}{{(1+r)}^{y_d}-1}}\times c_{cl}{\alpha }_{cl}{P}_{load}\\ C_{uhv}={\displaystyle \frac{r{\left(1+r\right)}^{y_u}}{{\left(1+r\right)}^{y_u}-1}}\times c_{uhv}{\alpha }_{new}{P}_r^{uhv}\end{array}\right.$$

(9)

where $C_{sys}$ is the system cost; $C_{TP}$, $C_{ES}$, $C_W$, $C_{WPP}$, $C_{PV}$, and $C_P$ are the unit operating costs of coal-fired power, new energy storage, hydropower, wind power, photovoltaic, and pumped storage, respectively; $P_{TP,t}$, $P_{PV,t}$, $P_{WPP,t}$, $P_{W,t}$, $P_{ES,t}$, and $P_{P,t}$ are the power generation; $C_{str}$ is the cost of new energy storage for consuming new energy; $C_{tra}$ is the coal-fired power flexibility transformation cost; $C_{dev}$ is the demand response cost; $C_{uhv}$ is the export cost; $C_w$ is the cost of removing new energy; $P_{wd}^t$ is the new energy cutting power; $V_d$ is the loss of load cutting; $D_{wd}^t$ is the load cutting power; and $s_{j,t}^{on}$ is the startup and shutdown cost for each unit. The unit of the installed coal-fired power capacity $P_{coal}$ is MW; $P_r^{uhv}$ is the transmission capacity of the UHV line; ${\alpha }_{tra}$ and ${\alpha }_{cl}$ are the ratio of the flexibility reform capacity to the total installed coal-fired power capacity and the ratio of the controllable load to the total load; and $P_{load}$ is the total load. ${\alpha }_{cl}$ and ${\alpha }_{new}$ are parameters defined by the configuration model, which denote the ratio of controllable load to total load and the proportion of new energy power transmitted via UHV lines. $y_k$, $y_t$, $y_d$, and $y_u$ are the life cycles of new energy storage units, thermal power flexibility units, demand response, and UHV lines, respectively.

The constraints are given in references [20,21,22].

3. Levelized Cost of Electricity Prediction Model

The levelized cost of electricity (LCOE) refers to the ratio of the integrated full-life-cycle cost to the present value of electricity generated throughout the life cycle. It is mainly used to assess the electricity costs of different energy forms, with the calculation process outlined below.

(1)

Convert all the costs of each part of the project in each phase of the project life cycle to the present value.

(2)

Aggregate the project costs and deduct the value of the real options to obtain the whole life cost of the project.

(3)

Calculate the total power generation over the life cycle of the project and calculate its discounted value.

(4)

Compare the total project life cycle cost with the total generation capacity to obtain the cost of electricity.

The model for estimating the whole life cycle cost of electricity is shown below:

$$LCOE={\displaystyle \frac{P_{dyn}+{\displaystyle \sum _{n=1}^{T_{O\&M}}\frac{P_{O\&M}\times (1-R_{tax})}{{(1+R_{dis})}^n}-{\displaystyle \sum _{n=1}^{T_{O\&M}}\frac{P_{dep}\times R_{tax}}{{(1+R_{dis})}^n}-\frac{V_{res-value}}{{(1+R_{dis})}^{T_{O\&M}}}}}}{{\displaystyle \sum _{n=1}^{T_{O\&M}}\frac{E_{acc}}{{(1+R_{dis})}^n}}}}$$

(10)

where $P_{dyn}$ is the dynamic investment, $T_{O\&M}$ is the operation period of the project, $D_{dep}$ is the depreciation cost of fixed assets, $P_{O\&M}$ is the operation and maintenance cost, $R_{tax}$ is the income tax rate, $R_{dis}$ is the discount rate, $V_{res-value}$ is the residual value of fixed assets, and $E_{acc}$ is the annual power generation.

4. Modeling the Impact of High Renewable Energy Penetration on End-User Price

4.1. User Electricity Price Components

As clearly stipulated in the 14th Five-Year Plan for electric power industry development, the price structure for industrial and commercial users is composed of the feed-in tariff, transmission and distribution tariffs, system operation costs, and government funds and surcharges. Among them, the feed-in tariff refers to the contract price signed between market users and power generation enterprises or power selling companies. The feed-in tariff refers to the cost corresponding to the power loss incurred in the process of the transmission of electricity from the generation side to the user side, which is calculated on the basis of the actual feed-in tariff for purchased electricity and the comprehensive line loss rate. The deviation between the forecasted and actual value of the line loss rate of the feed-in tariff is apportioned or shared with all industrial and commercial users on a monthly basis as the gain or loss in line loss of the agency purchase. Transmission and distribution tariffs refer to the costs incurred in transmitting electricity from the power plant to the customer end, including grid construction investment, operation and maintenance costs, and loss costs. System operation costs include ancillary service costs, coal-fired power and pumped storage capacity tariffs, line loss agency procurement gains and losses in the feed-in chain, and new gains and losses from tariff cross-subsidization. They are apportioned or shared in equal proportion between the market users and the users of power purchased by agents of the grid enterprises. Governmental funds and surcharges include the National Major Water Conservancy Project Construction Fund, Reservoir Migration Support Fund, Renewable Energy Tariff Surcharge Fund, Agricultural Network Loan Repayment Fund, etc., as shown in Figure 2.

4.2. Impact of End-Costs on Electricity Prices

4.2.1. System Operating Cost

(1)

Capacity tariffs

The Notice on the Establishment of a Capacity Tariff Mechanism for Coal-fired Power and the Notice on Transmission and Distribution Tariffs for Provincial Grids and Related Matters for the Third Regulatory Cycle are provided for a two-part tariff mechanism for coal-fired power and pumped storage, with capacity tariffs incorporated into system operating fees to be channeled to industrial and commercial users. The calculation is as follows:

$$P_{coal}=k\times C_{fixed}$$

(11)

$$Q_{coal}=P_{coal}\times I_{capacity}^{\max }$$

(12)

where $P_{coal}$ denotes the price of electricity per kilowatt of capacity per year for coal-fired power, $C_{fixed}$ denotes the fixed cost per kilowatt per year for coal-fired power units, and $k$ indicates the capacity tariff recovery ratio. $Q_{coal}$ indicates the total capacity tariff of the coal-fired power unit and $I_{capacity}^{\max }$ indicates the declared maximum output of the coal-fired power unit.

$$\begin{array}{l}{\displaystyle \sum _{s=1}^{n-1}\left[C{I}_s-C{O}_s\right]{\left(1+IRR\right)}^{-s}}+{\displaystyle \sum _{s=n-3}^{n-1}0.8\left(C{I}_s^{Electricty}+C{I}_s^{Ancillary}\right)}\\ +{\displaystyle \sum _{t=n}^{40}\left(C{I}_t-C{O}_t\right){\left(1+IRR\right)}^{-t}}+C{I}_{40}{\left(1+IRR\right)}^{-40}=0\end{array}$$

(13)

where $C{I}_s$ is the cash inflow (capacity tariff revenue) in the year $s$ prior to verification; $C{O}_s$ is the actual cash outflow in the year $s$ prior to verification, $C{I}_s^{Electricty}$ is the tariff revenue from electricity in the last regulatory cycle prior to verification, $C{I}_s^{Ancillary}$ is the revenue from participation in the ancillary services market or the ancillary services compensation mechanism in the last regulatory cycle prior to verification, $C{I}_t$ is the average annual level of revenue from capacity tariffs after verification (excluding residual income from fixed assets in the last year of the period, which is shown separately later), $C{O}_t$ is the projected annual cash outflow, $IRR$ is the internal rate of return on capital, and $C{I}_{40}$ is the residual income from fixed assets in the last year of the operating period.

(2)

Ancillary service costs

System operating costs for evacuation to the customer side include some of the costs of ancillary services and capacity charges for coal-fired power, pumped storage, and independent energy storage. The cost of market compensation for frequency regulation ancillary services is apportioned between the unit operating on the entire network during the month and the market-based electricity consumers. The user is required to share the cost of running the system:

$$C_{total}=C_c+C_o^{coal}+C_o^{pump}$$

(14)

$$C_c={\displaystyle \frac{Q_c}{Q_{toal}}}{C}_{total}$$

(15)

where $C_{total}$ is the total cost of ancillary services, $C_c$ is the shared cost of ancillary services on the customer side, $Q_c$ is the electricity consumption of the customers in the province, and $Q_{total}$ is the total power generation in the provincial grid.

4.2.2. Transmission and Distribution Tariff

Transmission and distribution costs are authorized on the basis of the principle of “permitted costs + reasonable revenue”, which mainly includes permitted costs, permitted revenues, and taxes. The permitted revenues of the transmission and distribution network are stipulated in the Provincial Grid Transmission and Distribution Tariff Pricing Measures, and the permitted revenues of the power companies consist of permitted costs, permitted revenues, and taxes.

Therefore, the transmission and distribution network costs $C_m$ and transmission and distribution tariffs $C_p$ of the electricity consumers at each voltage level can be expressed as follows:

$$C_m=P{C}_m+P{R}_m+T_m-D{I}_m$$

(16)

$$C_p={\displaystyle \frac{C_m}{Q_s}}$$

(17)

where $P{C}_m$, $P{R}_m$, $T_m$, and $D{I}_m$ denote the permitted costs, permitted revenues, taxes, and deductions of the transmission and distribution network, respectively. And $Q_s$ denotes the amount of electricity sold.

$$P{C}_m=P{C}_m^{base}+P{C}_m^{incr}$$

(18)

$$P{C}_m^{base}=F{A}_m^{base}(\alpha +\beta +\gamma +\eta )$$

(19)

$$P{C}_m^{incr}=I_m\theta (\alpha +\beta +\gamma +\eta )$$

(20)

$$P{R}_m=E{A}_m\lambda$$

(21)

where $F{A}_m^{base}$ denotes the net fixed asset value of the transmission and distribution network in the base period; $\alpha$ denotes the depreciation rate; $\beta$, $\gamma$, and $\eta$ denote the material and repair rate, labor rate, and other operating expense rate, respectively; $I_m$ denotes the projected new investment in the regulatory cycle; and $\theta$ denotes the proportion of the new investment that is accounted for as fixed assets. $E{A}_m$ denotes the depreciable net effective asset value and λ denotes the permitted rate of return on the transmission and distribution network.

4.2.3. Line Loss Cost, Government Funds, and Surcharges

(1)

Line Loss Cost

The formula for calculating the line loss cost is as follows:

$$C_{line}=P_m\times (r_{loss}/(1-r_{loss}))\times Q_m$$

(22)

where $C_{line}$ donates line loss cost for users, $P_m$ donates users’ on-line tariff for the month, $r_{loss}$ donates line loss rate, and $Q_m$ donates users’ settled electricity volume for the month.

Among them, the user on-line tariff is the agency power purchase tariff of the household or the (weighted average) market power purchase tariff of the household.

(2)

Government funds and surcharges

Measured based on local policies. Taking Gansu as an example, according to the “Inventory List of Gansu Provincial Government Funds”, the details of government funds and surcharges are shown in

Table 1. Breakdown of government funds and surcharges in Gansu Province.

Project Name		Levy Criteria
National Fund for Construction of Major Water Conservancy Projects		1.566 × 10−4 USD/kWh
Central Reservoir Migrant Support Fund	Large and Medium-sized Reservoir Migrants Post-retirement Support Fund	3.654 × 10−4 USD/kWh
	Inter-provincial Large and Medium-sized Reservoir Fund	1.114 × 10−3 USD/kWh
Local Reservoir Migrant Support Fund	Provincial Large and Medium Reservoir Fund	1.114 × 10−3 USD/kWh
	Small Reservoir Migrant Support Fund	2.784 × 10−5 USD/kWh
Renewable Energy Tariff Surcharge Fund		1.114 × 10−3 USD/kWh
Agricultural Network Loan Repayment Funds		2.784 × 10−3 USD/kWh

.

5. Calculus Analysis

5.1. Power System Factor Cost Prediction Result

(1)

Power-side cost

This section collects data related to the development scale, R&D expenditure, and construction cost of coal-fired power, wind power, photovoltaic, conventional hydropower, pumped storage, and new types of energy storage in Gansu Province from 2017 to 2023. And it predicts the cost of each power source in Gansu Province from 2025 to 2060, in which hydropower has a lower potential for future development and is no longer predicted.

Using the simulation outcomes of the two-factor curve and the installed capacity scale of each power source between 2025 and 2060, the costs of various power sources in Gansu Province are computed, as displayed in

Table 2. The 2025–2060 costs by power source.

Cost (×108 USD)	2025	2030	2035	2040	2045	2050	2055	2060
Coal-fired power cost	8.49	14.48	21.30	30.90	47.47	71.83	102.73	119.99
Wind power cost	3.06	5.15	6.40	8.07	9.47	10.44	10.16	11.55
Photovoltaic cost	2.64	4.18	5.71	7.38	8.63	9.33	10.58	10.86
New energy storage cost	0.42	0.97	1.67	2.78	4.45	5.43	6.96	7.93
Pumped storage cost	-	4.45	6.68	10.16	12.39	14.20	17.82	20.88

.

(2)

Grid-side cost

The new energy installed scale, load, and grid investment in Gansu Province from 2014 to 2023 are selected as the base data to be fitted by multiple linear regression. Applying IBM SPSS Statistics (Version 26) analysis software to fit the basic data, ${\beta }_1$ is 0.018, ${\beta }_2$ is 0.014, $\epsilon$ is 17.34, and the specific grid cost prediction model is shown below.

$$Y=0{.018\mathrm{P}}_t+0{.014\mathrm{L}}_t+17.34$$

The results of the renewable energy share and grid investment forecast results are shown in Figure 3.

We calculate the transmission and distribution tariffs for Gansu Province for the period 2025–2060 based on the net fixed asset value of the transmission and distribution network for the base period of 2025–2060, projected new investment for the regulatory cycle, permitted returns, taxes, deduction of revenue data, and sales of electricity in Gansu Province.

From Figure 4, in 2025–2060, the overall transmission and distribution tariffs are basically stable and show a slight upward trend, with the transmission and distribution tariff in 2025 being 0.028 USD/kWh, in 2030 being 0.027 USD/kWh, and in 2060 being 0.031 USD/kWh.

(3)

System operation cost

In this example, the typical years of 2025–2060 are selected as the planning period, and the production simulation results in the allocation of regulating resources during the planning period are as shown in Figure 5, and the results of calculating the cost of regulating resources are as shown in

Table 3. Regulatory resource cost results.

Regulatory Resource Cost (×108 USD)	2025	2030	2035	2040	2045	2050	2055	2060
Flexibility retrofit	2.37	3.76	5.57	7.52	10.72	21.72	29.93	36.75
Demand response	3.34	4.32	5.99	10.44	19.35	26.17	18.65	25.20
Independent energy storage	0.14	0.28	0.56	1.25	2.37	3.06	4.32	5.01
Total	5.85	8.35	12.11	19.21	32.43	50.95	52.90	66.96

.

This section collects data related to energy storage and conventional units in terms of load, winning capacity, frequency mileage, and frequency hours from 2025 to 2060 in Gansu Province, calculating the cost of ancillary services in the system operation fee from 2025 to 2060 in Gansu Province. According to “Gansu Province Electricity Ancillary Service Market Operation Rules”, the frequency market comprehensive frequency performance index cap is 1.5. Each market participant can declare a daily frequency mileage offer for the coming week on the platform of the power operating organization; the offer cap is tentatively set at 1.6704 USD/MW, and the smallest unit of the declared price is 0.01392 USD/MW. And according to the Gansu Electricity Ancillary Service Market Operation Rules, there is no capacity compensation for the main participants in frequency in Gansu Province, so only mileage cost is considered in this section, and the basic data is shown in

Table 4. Frequency regulation cost results.

Frequency Regulation Cost (×108 USD)	2025	2030	2035	2040	2045	2050	2055	2060
Energy storage frequency regulation cost	0.14	0.19	0.25	0.30	0.37	0.44	0.51	0.58
Conventional unit frequency regulation cost	0.17	0.24	0.31	0.38	0.46	0.55	0.63	0.73
Total frequency regulation cost	0.30	0.43	0.56	0.68	0.83	0.99	1.14	1.31

.

From the results of the regulatory resource cost and frequency regulation cost calculations, the system operating costs can be calculated, as shown in

Table 5. Results of system operating cost calculations.

System Operating Cost (×108 USD)	2025	2030	2035	2040	2045	2050	2055	2060
Flexibility retrofit	2.37	3.76	5.57	7.52	10.72	21.72	29.93	36.75
Demand response	3.34	4.32	5.99	10.44	19.35	26.17	18.65	25.20
Independent energy storage	0.14	0.28	0.56	1.25	2.37	3.06	4.32	5.01
Frequency regulation cost	0.30	0.43	0.56	0.68	0.83	0.99	1.14	1.31
Total	6.15	8.78	12.67	19.89	33.26	51.94	54.03	68.26

.

5.2. Levelized Cost of Electricity (LCOE) Prediction Result

(1)

Power LCOE

Based on the basic data of Gansu Province coal-fired power, wind power, photovoltaic, pumped storage, and energy storage in the key years 2025–2060, the LCOE for these key years can be assessed via the levelized cost of electricity model, followed by an evaluation of the comprehensive power generation cost trend for each power source. Leveraging the results of power supply, grid, and system operation costs, the trend of the system LCOE can be computed. The findings are displayed in

Table 6. LCOE for key years.

LCOE (USD/kWh)	2025	2030	2035	2040	2045	2050	2055	2060
Coal-fired power	0.043	0.045	0.050	0.056	0.064	0.073	0.088	0.090
Wind power	0.027	0.026	0.026	0.026	0.025	0.025	0.024	0.024
Photovoltaic	0.022	0.022	0.022	0.022	0.022	0.022	0.021	0.021
Pumped storage		0.039	0.038	0.036	0.035	0.034	0.031	0.031
New energy storage	0.045	0.046	0.042	0.036	0.031	0.030	0.029	0.028
Power LCOE	0.051	0.053	0.054	0.056	0.058	0.060	0.063	0.064
System LCOE	0.079	0.081	0.084	0.087	0.091	0.097	0.098	0.103

.

As shown in Figure 6, in the process of realizing “carbon peak, carbon neutral”, the power LCOE will be on the rise. Over the long term, the development scale of new energy keeps expanding continuously, superimposed on the steady increase in coal prices which will cause the LCOE to rise. In 2025 the power LCOE is 0.051 USD/kWh, in 2030 it is 0.053 USD/kWh, and in 2060 it will rise to 0.064 USD/kWh.

As shown in Figure 7 through the comprehensive changes in the cost of electricity in all segments, the system LCOE will continue to rise. In the long term, although the scale of new energy development will stabilize, the rise in coal prices and the increase in grid costs will cause the system LCOE to rise. It is expected that the system LCOE in 2025, 2030, and 2060 will be 0.079 USD/kWh, 0.081 USD/kWh, and 0.103 USD/kWh, respectively.

5.3. End-User Price Prediction Result

According to the results of grid-side costs, system operation costs, line loss costs, and government fund and surcharge measurements, the impact of terminal costs on the end-user price is analyzed, and the results of the measurements are shown in

Table 7. Results of end-user price measurement.

End-User Price Component (USD/kWh)	2025	2030	2035	2040	2045	2050	2055	2060
Feed-in tariff	0.047	0.049	0.051	0.052	0.054	0.056	0.059	0.059
Transmission and distribution tariff	0.028	0.027	0.027	0.028	0.029	0.030	0.030	0.031
System operating costs	0.003	0.008	0.008	0.007	0.006	0.006	0.005	0.005
Line losses	0.001	0.001	0.001	0.001	0.001	0.001	0.001	0.001
Government funds and surcharges	0.003	0.003	0.003	0.003	0.003	0.003	0.003	0.003
Electricity price	0.082	0.088	0.090	0.091	0.094	0.096	0.098	0.100

, and the trend of changes in the electricity price of industrial and commercial users is shown in Figure 8.

There are comprehensive changes in the cost of electricity in all segments; the terminal cost will continue to rise. In 2030, the new energy installed capacity increases significantly, resulting in higher terminal costs; the end-user price during the flat period rises to 0.088 USD/kWh. The new energy installed capacity tends to stabilize, the rising trend of the end-user price tends to flatten out, and the terminal end-user price in 2060 is 0.1 USD/kWh.

6. Conclusions

This paper innovatively proposes a model for forecasting the LCOE and end-user price derivation for a high proportion of new energy sources connected to the power system. Unlike current research that primarily focuses on cost predictions for single energy types, the power-side cost prediction method developed in this paper based on learning curves can calculate the costs of multiple power sources such as coal-fired power, wind power, photovoltaic power, pumped storage hydropower, and new energy storage in Gansu Province from 2025 to 2060. Additionally, the grid-side cost prediction model proposed in this paper, which incorporates regression equations, fills a research gap in grid-side cost prediction. Furthermore, this paper conducts production simulation analysis targeting system cost minimization to analyze the development trend of the system operating cost. Based on the calculation results of the above model, an LCOE calculation model is constructed, and the power cost of coal-fired power, wind power, photovoltaic power, pumped storage, new energy storage, and the system is calculated by combining relevant data. The example results show that the power LCOE and system LCOE will be on the rise. In the long term, the new energy development scale continues to expand; superimposed on the steady increase in coal prices, it will cause the LCOE to rise. In 2025, 2030, and 2060, the power LCOE will be 0.051 USD/kWh, 0.053 USD/kWh, and 0.064 USD/kWh. System LCOE will be 0.079 USD/kWh, 0.081 USD/kWh, and 0.103 USD/kWh.

In addition, we analyze the components of user electricity price and measure the system operation cost, including the capacity tariffs and ancillary service costs of coal-fired power, pumped storage, and independent energy storage. And we measure the user electricity price from 2025 to 2060 based on the analysis of transmission and distribution tariffs and the analysis of line loss costs, government funds, and surcharges. The results of the calculations show that the overall trend of the end-user price is increasing. In 2025, the end-user price will be 0.082 USD/kWh. By 2030, the end-user price will rise to 0.088 USD/kWh. As the scale of the new energy installed capacity stabilizes, the upward trend in end-user prices will flatten out, with the end-user price reaching 0.1 USD/kWh by 2060. According to the announcement by State Grid Gansu Electric Power Company on the electricity purchase price for commercial and industrial users in January 2025, the electricity price for commercial and industrial users is 0.0823 USD/kWh. The end-user price predicted in this paper is consistent with the actual situation, which proves that the prediction model proposed in this paper is accurate and practical. In addition, the predictive results of this paper can provide an empirical basis for regulatory agencies to revise market rules, using price signals to guide the flattening of user-side load curves and improve the overall operational efficiency of the power system. At the macroeconomic level, stabilizing expectations for electricity prices can provide a fundamental foundation for high-quality economic development in the context of energy transition.