The Hydrogen Trade-Off: Optimizing Decarbonization Pathways for Urban Integrated Energy Systems

Abstract

Rapid socio-economic development has made energy application and environmental issues increasingly prominent. Hydrogen energy, clean, eco-friendly, and highly synergistic with renewable energy, has become a global research focus. This study, using the EnergyPLAN model that includes the electricity, transportation, and industrial sectors, takes Jinan City as the research object and explores how hydrogen penetration changes affect the decarbonization path of the urban integrated energy system under four scenarios. It evaluates the four hydrogen scenarios with the entropy weight method and technique, placing them in an order of preference according to their similarity to the ideal solution, considering comprehensive indicators like cost, carbon emissions, and sustainability. Results show the China Hydrogen Alliance potential scenario has better CO₂ emission reduction potential and unit emission reduction cost, reducing them by 7.98% and 29.39%, respectively. In a comprehensive evaluation, it ranks first with a score of 0.5961, meaning it is closest to the ideal scenario when cost, environmental, and sustainability indicators are comprehensively considered. The Climate Response Pioneer scenario follows with 0.4039, indicating that higher hydrogen penetration in terminal energy is not necessarily the most ideal solution. Instead, appropriate hydrogen penetration scenarios should be selected based on the actual situation of different energy systems.

1. Introduction

In recent years, climate disasters such as increased global warming, frequent occurrences of extreme weather, the oil crisis, and nuclear accidents have triggered a profound reflection on the transformation of the energy mix [1]. Under the dual drive of global climate change and energy structure transformation, realizing the deep decarbonization of the integrated energy system (IES) has become a core issue in addressing environmental challenges and promoting sustainable development [2]. Hydrogen energy, as a zero-carbon energy carrier, is becoming increasingly strategic in the energy transition [3].

Hydrogen energy is at the forefront of research in the field of integrated energy systems due to its advantages; it is clean, efficient and renewable [4]. In-depth exploration of the optimal configuration of an integrated electric–hydrogen energy system is crucial to achieving net-zero emissions and managing the volatility of renewable energies. Recent studies have focused on optimizing the configuration and operation of such systems from multiple perspectives. For example, to balance the interests of multiple stakeholders, Zhang et al. [5] developed a multi-criteria decision-making framework and identified an optimal configuration for an industrial park combining natural gas, photovoltaics, and hydrogen storage, which achieved a significant 76% reduction in carbon emissions. Similarly, for achieving net-zero demand matching, Wada et al. [6] assessed a fully renewable-powered hydrogen system and concluded that while technically feasible, such projects currently require subsidies or additional revenue streams to be economically profitable, highlighting a key economic challenge. To specifically address renewable energy volatility, the integration of hydrogen storage has been proven effective. Wu et al. [7] demonstrated that configuring a system with hydrogen storage can increase the renewable energy consumption rate by 6.54% and reduce grid interaction by 45.01% compared to systems without any storage. Building on this, Xu et al. [8] further explored hydrogen’s role in a comprehensive demand response framework, showing that a coordinated power–gas–heat–hydrogen strategy can cut energy costs by 12~27%. Beyond configuration, researchers are also refining operational strategies and integrating advanced technologies. Yang et al. [9] tackled long-term supply–demand imbalances by proposing a nested optimization method, finding that integrating waste heat recovery with hydrogen storage could reduce total annual operational costs by 1.16% while significantly improving computational efficiency. Meanwhile, Wu et al. [10] proposed a coordinated planning model featuring high-temperature electrolysis powered by concentrated solar plants, developing a detailed framework to improve the overall efficiency of the power-to-hydrogen process. From a market perspective, Li et al. [11] revealed the substantial economic potential of these systems; their multi-level market model showed that by including income from hydrogen sales, the economic benefits of an IES with offshore wind could increase by as much as 205.64%.

With the rapid development and widespread application of artificial intelligence technology [12,13], a substantial body of research has focused on enhancing system efficiency and facilitating low-carbon development through the formulation of diverse optimization models and the utilization of various algorithms and mechanisms [14]. A primary focus of this research has been to develop optimization models that simultaneously minimize costs and carbon emissions, with studies reporting significant achievements. For instance, Dong et al. [15] demonstrated that a coordinated hydrogen-based IES in an industrial park could achieve a remarkable 95.44% reduction in emissions. The benefits of collaboration were further quantified by Wang et al. [16], whose cooperative game theory model showed that joint operation could reduce total costs by 22.96% and emissions by 80.11% compared to independent operation scenarios. To achieve these reductions, many studies incorporate market mechanisms and advanced technologies. Yang et al. [17] employed a ladder-type carbon trading mechanism and demand response, resulting in a 3.56% operational cost reduction and a 5.44% increase in Photo Voltaic (PV) consumption. Similarly, Zhan et al. [18] integrated fuel cell waste heat recovery with a tiered carbon trading model, leading to a 30.7% total cost reduction and an 18.8% decrease in carbon emissions. Beyond general optimization, researchers are tackling specific operational complexities and uncertainties. Sohani et al. [19] highlighted the critical impact of component degradation, showing that accounting for performance loss could necessitate increasing PV and electrolyzer capacities by over 11% and 38%, respectively, to meet system targets. To manage temporal challenges, Liu et al. [20] introduced a two-layer model with seasonal hydrogen storage, demonstrating its effectiveness in leveraging seasonal complementarities to reduce natural gas purchase costs by 9.10% under high renewable penetration. To handle real-time uncertainties, Fang et al. [21] proposed a multi-stage, multi-timescale energy management framework, proving its economic benefits in a system with inter-microgrid hydrogen trading. Furthermore, the state of the art includes advanced algorithmic contributions and strategic market analysis. Judge et al. [22], for example, developed an improved metaheuristic algorithm that outperformed several existing methods and designed a hybrid system projected to avoid 72.05% of carbon emissions. From a market investment perspective, Joseph et al. [23] used a game-theoretic model to analyze a liberalized energy market, concluding that a round-trip efficiency of at least 50% is crucial for the profitability of vector-coupling storage and quantifying its advantages over Li-ion batteries.

At present, the research on hydrogen energy in integrated energy systems is mostly limited to green hydrogen power generation and energy storage scenarios, with insufficient depth of scenario design, a lack of quantitative analysis of sub-gradient penetration rate, and neglect of the decarbonization potentials of terminal areas such as industrial fuel substitution and transportation energy transformation. These matters have not been fully explored, which results in the serious underestimation of the overall contribution of hydrogen energy to the integrated energy system. The overall contribution of hydrogen energy to the integrated energy system has been seriously underestimated, and the focus is mostly on a single sector of power or transportation, ignoring the synergistic effect between industry, transportation, and power systems, which makes it difficult to demonstrate the coupling characteristics of the integrated energy system and restricts the exploration of the overall optimization path of the system. Therefore, based on EnergyPLAN, we constructed a model of an integrated energy system coupled with the electric power, transportation, and industrial sectors, constructed four end-use energy penetration scenarios, and quantified the demand for hydrogen under different scenarios, which provides a more comprehensive assessment path for an integrated energy system to realize deep decarbonization at a cross-regional scale.

2. Energy Development Modeling

2.1. Model Building

The integrated energy system encompasses coal, oil, natural gas, electricity, heat, wind, and solar energy, which support economic and social operations. It has been transformed from a “traditional structure centered on a single network” to an “integrated system centered on electricity”, and the key to the transformation lies in the coordination of multisectoral supply and demand activities and the synergistic promotion of the efficient integration of traditional fossil energy sources and new clean energy sources [24]. Due to the seasonality and volatility of renewable energy, it is necessary to rely on hourly or even smaller scale time data to accurately characterize the annual output and energy use behavior of the system.

2.1.1. Modeling Framework

EnergyPLAN is an energy system analysis model developed by Aalborg University in Denmark, specifically designed for technical–economic simulation of multi-sector coupled systems such as electricity, industry, and transportation [25]. The model employs a deterministic input–output framework with hourly resolution [26], simulating the system’s annual operation based on user-inputted energy demand, renewable energy availability, technical costs, and other parameters, and outputting key indicators such as fuel consumption, carbon emissions, and total system costs. Its primary function is to conduct scenario comparison analyses to evaluate the comprehensive benefits of different energy strategies. However, the model also has limitations, primarily that it is a simulation tool rather than an optimization model and cannot automatically optimize solutions. Additionally, it simplifies the power grid into a single-node model without considering network congestion, so its analysis focuses on overall system balance rather than specific grid operation details. The constructed reference model framework is shown in Figure 1.

2.1.2. Mathematical Expression

i.

Electricity sector

The electricity sector is the core of the integrated energy system. The annual electricity demand of a region is the sum of the hourly electricity demands of various sectors. In order to cover a longer time period in the analysis and consider the minor impact that leap years may have, this study selected leap years as typical analysis years, with a total duration of 8784 h [25], as shown in Equation (1).

$$P_{el}=\sum _{n=1}^{8784}{p}_{el}\left(n\right)$$

(1)

where P_el is the annual electricity demand; p_el is the hourly electricity demand; and 8784 is the total leap year hours.

For renewable energy sources, the calculation function for power generation is shown in Equation (2).

$$P_{res}={\displaystyle \frac{\epsilon C_{res}{Q}_{res}}{max\left(Q_{res}\right)}}$$

(2)

where P_res is the renewable energy generation capacity; ε is the correction factor between capacity and production; C_res is the installed renewable energy generation capacity; and Q_res is the distribution of electricity production under the 8784 h benchmark.

ii.

Transportation sector

The fuel demand in the transportation sector is intended to describe the potential changes in the sector, and there is no hourly distribution data for fuel demand within the model, which is calculated as a functional equation as shown in Equation (3).

$$P_{trans}=P_{di}+P_{pe}+P_{jp}+P_{hy}$$

(3)

where P_total is the total fuel demand in the transportation sector, P_di is the diesel demand, P_pe is the gasoline demand, P_jp is the jet fuel demand, and P_hy is the hydrogen demand.

iii.

Industry sector

Fuel consumption in the industrial sector takes into account the demand for coal, oil, and natural gas as well as biomass, which is calculated as a function of Equation (4) [27].

$$P_{in}=P_{co}+P_{oil}+P_{na}+P_{bi}$$

(4)

where P_in is the industry sector fuel demand, P_co is the coal demand, P_oil is the fuel oil demand, P_na is the natural gas demand, and P_bi is the biomass demand.

2.2. Energy Forecasting Model

2.2.1. Energy Demand Projections

This paper employs a time series-based multiple linear regression (MLR) method. MLR is a classic statistical analysis technique that aims to quantify the relationship between a dependent variable and multiple independent variables by establishing a linear equation. This method not only explains historical data but also makes predictions based on future scenarios of the independent variables. First, based on economic and social development theory, we selected macroeconomic and social development indicators closely related to energy consumption as potential independent variables, constructing the following initial theoretical model (5) [28].

$$ln{P}_i=a+b_{1}lnGDP+b_{2}lnPop+b_{3}lnSec+b_{4}lnUrb+b_{5}lnEte$$

(5)

where P_i is the energy demand, GDP is the GDP per capita, Pop is the population size, Sec is the share of secondary industry, Urb is the rate of urbanization, Ete is the consumer price index, and a, b₁, b₂, b₃, b₄, and b₅ are the model parameters.

We collected historical indicator data for Jinan City, sourced from the “Jinan City Statistical Yearbook.” Subsequently, we used the statistical analysis software Statistical Package for the Social Sciences Statistics 27, which is distributed by IBM Corp (IBM SPSS Statistics 27) to solve the model parameters. Through regression analysis, we tested the explanatory power of each independent variable on the dependent variable and its statistical significance (typically measured by the p-value) and excluded non-significant variables to ultimately determine the optimal predictive model [29]. The detailed regression analysis process, data sources, and results verification are presented in Appendix A.

2.2.2. Renewable Energy Supply Projections

To make a prediction of the future renewable energy supply, mainly for Jinan City, we combined the potential of wind power, photovoltaic, and biomass with policy planning data and our scenario analysis method to predict the future installed capacity, assuming that the installed capacity from 2025 to the policy planning year will follow a path of linear growth. The resulting prediction model is shown in Equation (6).

$$\left\{\begin{array}{c}{C}_y=C_{2025}+{\displaystyle \frac{Q_5-C_{2025}}{5}}\times y, 0\le y\le 5\\ C_y=C_{2025}+{\displaystyle \frac{Q_{10}-C_{2025}}{10}}\times y, 5\le y\le 10\end{array}\right.$$

(6)

where C_y is the installed capacity after year y, C₂₀₂₅ is the installed capacity in the base year, Q₅ is the policy planning installed capacity target after 5 years, Q₁₀ is the policy planning installed capacity target after 10 years, and y is the forecast year.

3. Hydrogen Energy Scenario Penetration Modeling

3.1. Modeling of Hydrogen End-Use Energy Penetration

Hydrogen energy needs more than ten years or even decades to be deeply integrated into the existing energy structure [30]. The sectors with high hydrogen energy penetration such as electric power, industry, and transportation were selected to be analyzed, and the scenario application analysis model was constructed based on the end-use energy penetration rate of hydrogen energy. Based on the energy demand and renewable energy supply prediction model, the development path of an integrated energy system in Jinan City in the coming 10 years under the basic scenario was simulated, as well as its end-use energy consumption. According to the output of the BEIS scenario simulation, the terminal energy consumption is shown by Equation (7).

$$FTC=E_y-{Exp}_y+{Imp}_y+\sum _i\sum _j{f}_{FTC,y,i,j}$$

(7)

where FTC is the final energy consumption, E_y is the total electricity production, Exp_y is the exported electricity, Imp_y is the imported electricity, f_FTC is the final fuel consumption, with subscript i for the energy-using sector and subscript j for the fuel type, and y is the year of simulation.

The amount of hydrogen used for each hydrogen energy scenario is shown in Equation (8).

$$\begin{gathered} H_{total,n}={\gamma }_n\times {FTC}_y \\ n=\{CRPS,QPDS,CHAPS\} \end{gathered}$$

(8)

where H_total,n is the total annual hydrogen use and γ_y,n is the hydrogen end-use energy penetration rate.

i.

Modeling hydrogen penetration in the electricity sector

Hydrogen penetration in the electricity sector is mainly realized through gas turbine hydrogen doping [31] and coal power doping. The amount of fuel replaced by hydrogen is shown in Equations (10) and (11). The proportion of hydrogen in each of the demand–supply distribution application scenarios is shown in Figure 2.

$$H_{elec}={\alpha }_{elec}\times H_{total,n}$$

(9)

$$M_{elec,ng}=H_{elec}\times {\beta }_{elec}\times {\mu }_{ng}$$

(10)

$$M_{elec,co}=H_{elec}\times {\beta }_{elec}\times {\mu }_{co}$$

(11)

where H_elec is the total hydrogen usage in the electricity sector, M_elec,ng is the hydrogen blending ratio for gas turbines, M_elec,co is the hydrogen blending ratio for coal-fired power generation, α_elec is the percentage of hydrogen demand in the electricity sector, β_elec is the percentage of hydrogen supply in the electricity sector, μ_ng is the hydrogen substitution capacity for gas, which takes the value of 50%, and μ_co is the hydrogen substitution capacity for coal, which takes the value of 50%.

ii.

Modeling hydrogen penetration in the industry sector

Hydrogen penetration in the industry sector is mainly realized by substituting hydrogen for coal and natural gas. In order to simplify the analysis process, the substitution of natural gas in the industry sector is realized through hydrogen injection into the gas network. The amount of fuel that is replaced by hydrogen is shown in Equations (13)–(15).

$$H_{Ind}={\alpha }_{Ind}\times H_{total,n}$$

(12)

$$M_{Ind}=H_{Ind}\times {\beta }_{Ind}$$

(13)

$$H_{inj}={\alpha }_{inj}\times H_{total,n}$$

(14)

$$M_{inj}=H_{inj}\times {\beta }_{inj}$$

(15)

where H_Ind is the total hydrogen use in the industry sector, α_Ind is the hydrogen demand share in the industry sector, and β_Ind is the hydrogen supply share in the industry sector. H_inj represents the total hydrogen consumption in the natural gas network, α_inj represents the proportion of hydrogen demand in the gas network, β_inj represents the proportion of hydrogen supply in the gas network, M_Ind represents the amount of hydrogen-based alternative fuels in the industrial sector, and M_inj represents the amount of hydrogen-based alternative fuels replacing natural gas in the gas network.

iii.

Modeling hydrogen penetration in the transportation sector

Based on the difficulty of substituting hydrogen for different fuels in the transportation sector, the study considers only the substitution of diesel and gasoline. The amount of diesel and gasoline that is replaced by hydrogen in this sector is shown in Equations (17) and (18).

$$H_{trans}={\alpha }_{tr}\times H_{total,n}$$

(16)

$$M_{tr,di}=H_{trans}\times {\beta }_{tr}\times {\theta }_{di}$$

(17)

$$M_{tr,pe}=H_{trans}\times {\beta }_{tr}\times {\theta }_{pe}$$

(18)

where M_tr,di is the amount of hydrogen replacing diesel fuel, M_tr,pe is the amount of hydrogen replacing gasoline, H_trans is the total hydrogen use in the transportation sector, α_tr is the percentage of hydrogen demand in the transportation sector, β_tr is the percentage of hydrogen supply in the transportation sector, θ_di is the hydrogen substitution capacity for diesel fuel, which takes the value of 60%, and θ_pe is the hydrogen substitution capacity for gasoline, which takes the value of 40%.

3.2. Hydrogen Energy Application Scenario Construction

i.

Basic energy inertia scenario (BEIS)

As a comparison to the “inertia development” characteristic of the benchmark, assuming that Jinan maintains its current energy structure (coal-based power generation, zero application of hydrogen), the development path of the integrated energy system for the next 10 years was simulated, and the carbon emission intensity of the outgoing electricity remained unchanged. This was used to highlight the marginal improvement effect of hydrogen on the peak carbon pathway in the other scenarios.

ii.

Climate response pioneer scenario (CRPS)

The International Energy Agency (IEA) “Zero Emissions 2050: A Roadmap for the Global Energy Sector” [32] proposes that hydrogen energy should account for 2% of end-use energy consumption in 2030. This percentage will need to be further increased in 2035 to support the target of a 1.5 °C temperature rise. As a core city involved in the ecological protection of the Yellow River Basin, Jinan should strictly control the production of hydrogen from fossil energy sources and focus on promoting the integration of “wind, light, hydrogen, and storage” projects, which would set the penetration rate of hydrogen energy in terminal energy at 3%.

iii.

Policy-driven scenario (QPDS)

Jinan’s hydrogen industry development plan for the pilot zone for conversion of old and new dynamics explicitly proposes to build a 100-billion hydrogen industry cluster by 2035 and to construct a 20-square-kilometer “China Hydrogen Valley”. This goal should be supported by a significant increase in the penetration rate of hydrogen energy, which is set at 5% in terminal energy.

iv.

China Hydrogen Alliance Potential Scenario (CHAPS)

Directly related to the China Hydrogen Energy Alliance’s national forecast (5% of end-use energy by 2030), and in light of Shandong Province’s position as the core area of the “Northern Hydrogen Valley”, the penetration rate of hydrogen energy in this scenario was set at 6.5%. The input parameters for hydrogen energy penetration in 2035 for different scenarios are shown in

Table 1. 2035 hydrogen energy terminal energy penetration model parameters.

Hydrogen Energy in Different Sectors		CRPS (TWh)	QPDS (TWh)	CHAPS (TWh)
FTC		403.530
Htotal,n		12.106	20.177	26.229
Electricity	Helec	0.726	1.211	1.574
	Melec,ng	0.218	0.363	0.472
	Melec,co	0.218	0.363	0.472
Transport	Htrans	4.116	6.860	8.918
	Mtr,di	0.988	1.646	2.140
	Mtr,pe	1.482	2.469	3.210
Industry	HInd	6.658	11.097	14.426
	MInd	3.995	6.658	8.656
	Hinj	0.605	1.009	1.311
	Minj	0.363	0.605	0.787

.

4. Energy Evaluation Modeling

4.1. Construction of Evaluation Indicators

i.

Economic indicators

The total cost is used as the core economic indicator in integrated energy systems [33], which directly reflects the economics of system planning and operation and is a key constraint in balancing environmental and energy objectives. It is calculated as shown in Equation (19).

$$AT{C}^{’}=\sum _{y}ATC=\sum _y(C_{inv}+C_{o\mathcal{s}m}+C_e)$$

(19)

where ATC′ is the cumulative cost, ATC is the total annual cost, C_inv is the initial investment cost, C_o&m is the fixed O&M cost; C_e is the energy cost, and y is the year.

Furthermore, to quantitatively assess the magnitude of the annual cost difference between the hydrogen penetration scenarios and the baseline scenario, especially when visual comparisons may not be remarkable, this study employs the root mean square error (RMSE). RMSE provides a single, aggregate measure of the deviation between two time-series datasets [34]. It is calculated as shown in Equation (20).

$$RMSE=\sqrt{{\displaystyle \frac{1}{n}}\sum _{t=1}^n{({ATC}_{hydro,t}-{ATC}_{BEIS,t})}^2}$$

(20)

where RMSE is the root mean square error of the cost, ATC_hydro,t is the annual total cost for a given hydrogen scenario in year t, ATC_BEIS,t is the annual total cost for the BEIS scenario in year t, and n is the total number of years in the simulation period.

ii.

Environmental indicators

The environmental objective of the integrated energy system is to reduce the emission of greenhouse gases, especially carbon dioxide, which is evaluated in terms of the total annual carbon emissions of the system. These emissions include those from fuel combustion and from hydrogen production. Notably, our model incorporates emissions from coal-to-hydrogen production coupled with Carbon Capture and Storage [35]. This strategy is designed to scale up hydrogen supply in the initial decades, thereby avoiding exclusive reliance on more expensive green hydrogen produced via electrolysis. It is calculated as shown in Equation (21). The CO₂ emission factors for fuel combustion, as listed in

Table 2. CO₂ emission factors for different fuels.

Fuel Types	Coal	Fuel Oil	Natural Gas
CO2 emission factors (kg/GJ)	90.9	75.8	61.8

, were sourced from internationally recognized guidelines to ensure accuracy [36].

$$CEV’=\sum _y{CEV}_y=\sum _y[\sum _i{fuel}_i\times {\sigma }_i+({CO}_{2,pro}-{CO}_{2,CCS})]$$

(21)

where CEV′ is the cumulative carbon emissions, CEV is the annual carbon emissions, fuel is the fuel consumption, σ is the fuel carbon emission factor, CO_2,pro is the carbon dioxide produced by hydrogen production from coal, CO_2,ccs is the carbon dioxide captured by the combination of effective carbon reduction technologies (CCS), y denotes the year, and i denotes the fuel type.

iii.

Technical indicators

With the increase of installed renewable energy capacity in the integrated energy system, the energy consumption capacity is insufficient, resulting in the phenomenon of wind and light abandonment. The cumulative Critical Excess Electricity Production (CEEP) [37] and cumulative Renewable Energy Sources (RES) are used as the technical evaluation indexes. RES is used to measure the penetration rate of renewable energy sources and CEEP is used to measure the capacity of renewable energy consumption, and the calculation method is shown in Equations (22) and (23).

$$CEE{P}^{’}=\sum _y{CEEP}_y=\sum _y(P_{res}-P_{El}-H_i)$$

(22)

$$RES=\sum _y{P}_{res}$$

(23)

where CEEP′ is the cumulative critical excess power, CEEP is the annual critical excess power, P_res is the annual renewable power generation, P_El is the total annual power demand, H_i is the electrolyzer power consumption, and RES is the cumulative renewable power generation.

4.2. Empowerment Ranking Model Construction

4.2.1. Entropy Weight Method

The entropy weight method is an objective assignment method that determines weights based on the degree of dispersion of the indicator data [38]. The greater the data discrepancy, the smaller the entropy value and the higher the weight, which can reduce subjective bias and can handle multi-indicator evaluation. The mathematical model and calculation steps are shown below.

i.

Establish the original matrix [X_ij]_m×n with n evaluation objects and m evaluation indicators.

ii.

To eliminate the influence of dimensions, all indicators are first standardized. For positive indicators Y_ij⁺ and negative indicators Y_ij⁻, the maximum–minimum value standardization method is typically used to map them to the [0, 1] interval.

$$Y_{ij}^{+}={\displaystyle \frac{Y_{ij}-\mathrm{m}\mathrm{i}\mathrm{n}\left(Y_j\right)}{\max \left(Y_j\right)-\mathrm{m}\mathrm{i}\mathrm{n}\left(Y_j\right)}}$$

(24)

$$Y_{ij}^{-}={\displaystyle \frac{\max \left(Y_j\right)-Y_{ij}}{\max \left(Y_j\right)-\mathrm{m}\mathrm{i}\mathrm{n}\left(Y_j\right)}}$$

(25)

where Y_ij represents the value of the ith evaluation object in the jth evaluation indicator.

iii.

We need to calculate the indicator weight D_ij and perform data translation. If the standardized data contains a value of 0 for Y_ij, this will result in a weight D_ij of 0, rendering the entropy calculation meaningless. To ensure that all logarithmic values are meaningful, we perform a slight “translation” on the data, i.e., add a very small positive number d to all data.

$$D_{ij}={\displaystyle \frac{Y_{ij}+d}{\sum _{j=1}^n{(Y}_{ij}+d)}}, d=0.001$$

(26)

iv.

Information entropy q_j is a measure of information uncertainty. According to the definition of information theory, the formula for calculating the information entropy of the jth indicator is shown in Equation (27).

$$q_j=-k\sum _{i=1}^n{D}_{ij}\ln \left(D_{ij}\right),k={\displaystyle \frac{1}{\mathrm{l}\mathrm{n}\left(n\right)}}$$

(27)

v.

Perform the weight calculation as shown in Equation (28).

$$w_j={\displaystyle \frac{1-q_i}{\sum _{j=1}^n(1-q_i)}}$$

(28)

4.2.2. Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) Decision Theory

TOPSIS is a multi-indicator decision-making method. It ranks the objects by calculating the distance of each evaluation object from the optimal solution and the worst solution, which can objectively reflect the comprehensive advantages and disadvantages of each solution [39]. The mathematical model and calculation steps are shown below.

i.

The normalized initial matrix is constructed, in which the Euclidean quantum normalization method is used to construct the weighting method matrix and vector normalization, where each column element is divided by the current parameter of the column vectors to get the normalized normalized matrix Z.

$$Z_{ij}={\displaystyle \frac{x_{ij}}{\sqrt{\sum _{i=1}^n{x}_{ij}^2}}}, i=1, 2,\dots , n;j=1, 2,\dots ,m$$

(29)

In the processing of the Pareto solution set, the entire Pareto solution set constitutes the original data matrix X. The points on the Pareto front are denoted by x_ij, where i denotes each point located on the Pareto front and j denotes the number of dimensions on the target space.

ii.

The calculation of points and the normalization process is followed by decision-making. In decision-making, there are ideal and non-ideal points. Ideal point Z⁺ consists of the maximum value of the elements of each column in matrix Z, non-ideal point Z⁻ consists of the minimum value of the elements of each column in matrix Z. In deciding how to choose the best compromise program, the distance between the evaluation object and the two ideal points is considered comprehensively as an evaluation criterion. The distance between the evaluation object and the ideal point is calculated as shown in Equation (30).

$$d_i^{\pm }=\sqrt{\sum _{j=1}^m{\omega }_j{(Z_{ij}-Z_j^{\pm })}^2}$$

(30)

where ω_j is the weight of the jth attribute, i.e., each target weight.

The final calculation of the evaluation object and the optimal program closeness Y_i, 0 < Y_i < 1. The more convergent to 1, the better the evaluation object. The comprehensive evaluation parameters are shown in Equation (31).

$$Y_i={\displaystyle \frac{d_i^{-}}{d_i^{+}+d_i^{-}}}$$

(31)

5. Example Results and Analysis

5.1. Results of Energy Demand Projections

Based on the multiple linear regression method constructed in the preceding section, we predicted the energy demand of the electricity, industrial, and transportation sectors in Jinan City. Referring to relevant government planning documents, the development experience of developed countries, and related research, we set the average annual growth rates of the independent variables for the policy scenarios over the next decade. Thus, the parameters were set as shown in

Table 3. Forecast and basis for economic and social development of Jinan in the next 10 years.

Independent Variable	Setting Basis	Annual Growth Rate Setting
GDP per capita	Jinan City “14th Five-Year Plan” economic and social development requirements to reach 7%.	7.59~6.54%
Population	The Jinan Territorial Spatial Master Plan (2021–2035) proposes that by 2035, the resident population in the city area will be strictly controlled within 12 million people.	5.23–6%
Percentage of secondary sector	Responding to the national energy-saving and emission reduction policy the proportion of secondary industry will decline.	−0.36~−0.54%
Urbanization rate	Jinan City, “14th Five-Year Plan” economic and social development requirements to reach 58%.	2.45~1.92%
Consumer price index CPI	Consumer index increase controlled at around 3%.	−4.06~−2.06%

[40,41]. The robustness of these policy targets is corroborated by independent sources, including China’s economic growth forecast [42] and macroeconomic studies on China’s carbon neutrality pathways [43]. Using Equation (5) as the initial model, variable screening was conducted in IBM SPSS Statistics 27 to exclude independent variables with insignificant effects on the dependent variable, ultimately yielding optimized prediction models for each sector. The detailed regression statistical results for each model, including regression coefficients, significance levels, and R² values, are fully listed in Appendix A.

i.

Electricity energy demand forecasting model

After regression analysis, the significant independent variables ultimately included in the electricity demand prediction model were per capita GDP, the proportion of secondary industry, urbanization rate, and consumer price index. The resulting prediction model is shown in Equation (32) below, with an adjusted R² of 0.986, indicating a high degree of model fit.

$$ln{D}_{El}=24.002+0.193lnGDP+1.378lnSec+8.783lnUrb-3.586lnEte$$

(32)

According to Equation (32) and Table 3, the total annual demand for electricity in Jinan City in the next 10 years is predicted as shown in Figure 3. For the years 2026–2035, the overall demand for electricity shows a growth trend, and the increase is relatively flat, though the increase in 2033 is larger. With the future energy structure adjustment and further promotion of energy-saving and emission reduction technologies, the proportion of electricity in terminal energy consumption is expected to gradually increase.

ii.

Industry energy demand forecasting model

The main factors influencing industrial fuel demand are also per capita GDP, the proportion of the secondary industry, urbanization rate, and the consumer price index. The prediction model obtained from the regression analysis is shown in Equation (33) below, with an adjusted R² of 0.980, which also has strong explanatory power.

$$ln{D}_{In}=-14.001+0.733lnGDP+1.093lnSec-0.340lnUrb+4.020lnEte$$

(33)

The fuel demand forecast for the industrial sector is shown in Figure 4. From the perspective of qualitative analysis, the trend of energy demand changes over the next decade will be influenced by a number of factors such as economic growth, industrial restructuring, energy conservation, and emission reduction in an upward trend, with the increase slowing down by 2029, and the share of fossil energy in the structure of industrial energy consumption is expected to decline.

iii.

Transportation Energy Demand Forecasting Model

The energy demand of the transportation sector shows the strongest correlation with per capita GDP and the urbanization rate of the permanent population. Its prediction model is shown in Equation (34) below, with an adjusted R² of 0.982.

$$ln{D}_{ET}=-23.071+1.254lnGDP+1.354lnSec-1.997lnUrb+4.763lnEte$$

(34)

The results of the transportation sector energy demand forecast for the next 10 years are shown in Figure 5. From a qualitative analysis point of view, transportation energy demand shows a growth trend from 2026 to 2035. Gasoline and diesel, as traditional transportation energy sources, will continue to occupy an important position in the short term, with a slowing trend of growth from 2034 onwards, and their demand growth is expected to slow down due to the popularity of new energy vehicles and the optimization of the public transportation system.

5.2. Results of Renewable Energy Supply Projections

The projected growth of renewable energy capacity and the declining share of coal power are illustrated in Figure 6. The 2025 targets are based on Jinan’s “14th Five-Year Plan”. While policy-driven, these targets are well-aligned with China’s national ‘Dual Carbon’ strategy [44] and are corroborated by independent analyses. These analyses confirm the region’s high potential for solar and wind generation [45] and the compelling economic case for renewables, driven by their rapidly declining levelized cost of energy [46]. Accordingly, Figure 6 depicts the installed capacity of wind power, photovoltaics, and biomass reaching 3200 MW, 1050 MW, and 450 MW by 2025, respectively. This expansion of renewables corresponds with a projected gradual decrease in the share of coal power over the study period.

Although the share of coal power declines, its installed capacity is not immediately eliminated. This reflects the fundamental requirement for a stable energy transition. Conventional power plants provide essential grid services, including firm capacity and inertia, which are crucial for balancing the intermittency of variable renewable energy sources like wind and solar. Their continued, albeit decreasing, presence in the model ensures the energy system’s reliability and maintains the energy balance, especially during periods of low renewable generation or high demand. Therefore, the model represents a gradual phase-out process, where the role of conventional power will be progressively replaced by other dispatchable, low-carbon technologies as they mature.

This study models the development of two primary long-duration electricity storage technologies: Pumped Hydro Storage and Battery Energy Storage Systems, with electricity as the stored energy carrier. The projected growth is illustrated in Figure 7. This projection is grounded in regional and national energy strategies aimed at achieving the ‘Dual Carbon’ goals, with targets derived from policy documents such as China’s “14th Five-Year Plan for New Energy Storage Development” [47]. The model assumes that the initial installed capacity in 2025 aligns with policy targets, with PHS at 100 MW/600 MWh. In the future scenario, electrochemical and pumped storage are both growing, with pumped storage increasing at a relatively flat rate. The increase in electrochemical energy storage expands significantly in 2030. In the future scenario, both electrochemical storage and pumped storage show a growth trend, with pumped storage increasing at a relatively flat rate.

5.3. Cost Projections

The variable costs of integrated energy systems are highly sensitive to fuel prices. Predicting future fossil fuel prices is inherently complex, as it is influenced by conflicting factors. In a future aligned with deep decarbonization goals, factors such as the large-scale deployment of renewable energy, significant improvements in energy efficiency, and robust climate policies aimed at reducing demand for carbon-intensive fuels are expected to suppress long-term fossil fuel prices. This study analyzes decarbonization pathways, which implicitly assume a future where strong and effective climate policies are implemented. Therefore, in the cost analysis, we adopt a policy-driven price scenario aligned with long-term climate goals, referencing predictions from authoritative institutions such as the IEA in its “Net Zero Emissions by 2050” scenario [48,49,50,51]. This scenario is a globally recognized and technically robust benchmark for modeling ambitious climate mitigation. It is not a simple extrapolation but is based on a comprehensive analysis of global energy markets under a paradigm shift toward clean energy. Its projection of declining fossil fuel demand and subsequently lower prices is a logical outcome of the modeled decarbonization efforts. The price trends adopted are shown in Figure 8.

The economic parameters of the main equipment are shown in

Table 4. Economic parameters of main equipment.

Equipment	Initial Investment Cost (MCNY/MW)	Fixed O&M Cost (MCNY/MW)	Period (Years)
Wind turbine	8.50	0.12	10
Photo votaic	3.80	0.05	10
Biomass power	3.50	0.30	10
Gas-fired power	5.50	0.18	10
Coal-fired power	12.0	0.22	10

. These parameters, including initial investment and fixed O&M costs, were compiled from a comprehensive review of recent industry reports, scholarly literature, and official policy documents to ensure they reflect China’s current and near-future economic landscape [52,53,54]. The selected investment costs for wind and photovoltaic power provide a conservative baseline for long-term modeling, accounting for potential fluctuations in supply chains and raw material prices. In contrast, thermal power technologies such as coal and gas have more complex system configurations, leading to higher O&M costs. Specifically, the higher investment cost for coal-fired power accounts for the deployment of high-efficiency, ultra-low emission technologies and carbon capture readiness, aligning with national deep decarbonization goals. These carefully chosen parameters establish a robust foundation for calculating total annual system costs and evaluating the economic viability of different decarbonization pathways.

6. Results

6.1. Analysis of Hydrogen Decarbonization Pathway Simulation Results

6.1.1. Analysis of the Evolution of the Energy Structure

The evolutionary trend of end-use energy supply for each scenario of the integrated energy system in Jinan City is shown in Figure 9. The share of hydrogen energy in end-use fuel consumption in each sector gradually increases with the increase of hydrogen energy penetration in different scenarios. However, the proportion of hydrogen energy consumption in total energy consumption is small, reflecting the current cost and infrastructure challenges of hydrogen technology, but it shows some value in improving power system flexibility, promoting renewable energy utilization, and reducing carbon emissions, foreshadowing the potential growth of hydrogen technology in the future energy system.

The energy consumption structure of the integrated energy system is divided into three major sectors, namely transportation, industry, and electricity, and the simulation results of the evolutionary trend of primary energy supply in each sector from 2026 to 2035 are shown in Figure 10. The simulation results show that the share of hydrogen fuel in the energy consumption of the transportation sector has increased significantly, and the energy consumption of the transportation sector is dominated by gasoline and diesel in the CRPS scenario, with a low penetration rate of hydrogen fuel. By contrast, the total energy consumption in the transportation sector is projected to decrease significantly in the CHAPS scenario because of the special emphasis placed on the core value of hydrogen fuel in promoting the development of sustainable transportation. This indicates that hydrogen fuel technology and related infrastructure development is crucial to achieving the energy transition and the goal of sustainable development in the transportation sector.

Figure 11 shows the amount and share of renewable energy generation under the different scenarios. With the increase of the installed capacity of wind power, PV, and biomass, the renewable energy generation in all scenarios shows an increasing trend year by year. The increase of renewable energy generation is more obvious in the QPDS scenario, and the CHAPS scenario shows the highest energy share, with 89.9% and 4.9% increases in the renewable energy generation and primary energy share in 2035, respectively. This change reflects the fact that as hydrogen penetration increases, the installed capacity of the electrolyzer increases and the system is able to consume more power.

The critical excess electricity versus electrolyzer use for each scenario is shown in Figure 12. In the BEIS scenario, as the installed capacity of renewable energy sources rises, it is difficult for the integrated energy system to fully consume their power generation, resulting in a high share of CEEP. In the hydrogen energy application scenario, the CEEP decreases significantly due to the commissioning of the electrolyzer. In the CRPS, QPDS, and CHAPS scenarios, the higher the penetration of hydrogen in the terminal energy, the lower the CEEP, indicating that the rational allocation of the electrolyzer is significant in enhancing the stability of the integrated energy system.

6.1.2. Carbon Emission Analysis

The trend of CO₂ emissions in 2026–3035 under different scenarios is shown in Figure 13. Carbon dioxide emissions increase in all scenarios, and the increase slows down until 2029. Compared with the BEIS in 2030, the carbon emission reduction levels of the CRPS scenario, the QPDS scenario, and the CHAPS scenario reach 3.48%, 6.72%, and 10.08%, respectively, which indicates that with the increase of hydrogen penetration and CCS, carbon dioxide emissions can be effectively reduced.

The cumulative carbon dioxide emission reductions and percentages of different scenarios are shown in Figure 14. With the increase of hydrogen energy penetration, the cumulative carbon dioxide emission reduction and the proportion of emissions reduction of each hydrogen energy scenario show an increasing trend, among which the cumulative carbon emissions reduction of the CRPS, QPDS, and CHAPS scenarios reaches 16.34 Mt, 37.74 Mt, and 57.93 Mt, respectively, and the proportion of cumulative carbon emission reduction is 2.30%, 5.48%, and 8.67%, respectively.

6.1.3. System Cost Analysis

Figure 15 shows a comparison of annual costs under different scenarios and quantitatively analyzes the evolution of annual costs over time. As the penetration rate of hydrogen energy in terminal energy increases, the annual costs of the CRPS, QPDS, and CHAPS scenarios gradually rise, and they are all higher than those of the BASIC scenario. The application of hydrogen production and coal-to-hydrogen with CCS technology is the main reason for the cost increase. Given that the proportion of coal, oil, and other fossil energy sources in future energy applications will gradually decline, and the cost of fuel consumption will be reduced accordingly, the annual cost of each scenario will show a decreasing trend after 2029.

To more accurately compare these cost discrepancies, we calculated the RMSE of each hydrogen scenario relative to the BEIS scenario, as shown in

Table 5. RMSE comparison of each hydrogen energy scenario relative to the BEIS scenario.

Control Group	CRPS (Billion CNY)	QPDS (Billion CNY)	CHAPS (Billion CNY)
RMSE	30.57	53.40	67.09

. The RMSE values for the CRPS scenario, QPDS scenario, and CHAPS scenario are 30.57 billion CNY, 53.40 billion CNY, and 67.09 billion CNY, respectively. These values quantify the growing cost gap as hydrogen penetration deepens.

Figure 16 shows the CO₂ abatement potential and unit abatement cost of the three hydrogen penetration scenarios compared to the BASIC scenario. The CHAPS scenario shows the lowest unit abatement cost and the highest carbon abatement potential, with reductions of 29.39% and 7.98%, respectively, which demonstrates the excellent abatement performance of its integrated energy system.

6.2. Evaluation of Hydrogen Decarbonization Pathway Simulation Results

The entropy weight method and TOPSIS methods were used to comprehensively evaluate the three different hydrogen energy scenarios of the integrated energy system, using cumulative cost, cumulative carbon emissions, cumulative CEEP, and cumulative renewable power generation as the evaluation indexes. The data of the simulation results of the hydrogen decarbonization pathway are shown in

Table 6. Hydrogen decarbonization simulation results data.

Hydrogen Scenario	Cumulative Cost (Billion CNY)	Cumulative CO2 Emissions (Mt)	Cumulative CEEP (TWh)	Cumulative RES Generation (TWh)
CRPS	8744.43	709.67	5.19	134.5
QPDS	8983.30	688.27	1.15	179.3
CHAPS	9124.12	668.08	0.49	216

.

The weight distribution of the indicators is shown in

Table 7. Assignment of weights to objective assessment indicators.

Evaluation Indicators	Cumulative Cost (Billion CNY)	Cumulative CO2 Emissions (Mt)	Cumulative CEEP (TWh)	Cumulative RES Generation (TWh)
Weighting factor	0.2814	0.2502	0.2233	0.2451

. The weight distribution of each evaluation index is relatively balanced. Among them, the weight of cumulative cost is the highest, reaching 0.2814, which reflects the relative fairness of the decision-making process and the moderate focus on economic cost factors in the planning and operation of the integrated energy system, taking into account the multiple objectives of the environment, energy efficiency, and the optimization of the energy structure, which helps to achieve the sustainable development of the system on the basis of balancing the needs of all parties.

The simulation results of hydrogen decarbonization pathways under different scenarios are shown in

Table 8. Evaluation results of hydrogen energy pathway simulation.

Evaluation Indicators	CRPS	QPDS	CHAPS
Relative posting progress	0.4039	0.3154	0.5961

, in which the CHAPS scenario ranks first with a score of 0.5961, which indicates that this scenario is the closest to the ideal state after considering all the indicators. The QPDS scenario has the lowest score of 0.3154, which indicates that the penetration rate of hydrogen in end-use energy is not as high as possible, but rather, it is necessary to select appropriate hydrogen penetration scenarios in accordance with the actual situation of different energy systems so as to realize the optimal economic, environmental, and sustainabilitybenefits.

6.3. Evaluation of Electricity Supply and Demand Simulation Results

To provide a granular view of the system’s operational dynamics, this section analyzes the electricity supply and demand balance for a representative week in January 2035, comparing the BEIS and CHAPS scenarios. Figure 17 and Figure 18 highlight how hydrogen integration fundamentally alters grid operations.

Figure 17 presents the electricity demand composition. In the BEIS scenario (Figure 17a), the system struggles to absorb all the renewable energy generated. When generation from wind and solar exceeds the final electricity demand and export capacity, the surplus energy is curtailed, resulting in significant CEEP. This represents a substantial waste of clean energy potential. In contrast, the CHAPS scenario (Figure 17b) demonstrates a highly effective solution to this problem. The previously curtailed excess power is now channeled to power electrolyzers for green hydrogen production. This new, flexible demand from the electrolyzer almost entirely eliminates CEEP, effectively converting what would have been wasted energy into a valuable energy carrier, thereby significantly improving the overall system efficiency and renewable energy utilization rate.

Figure 18 presents the electricity supply composition. In the BEIS scenario (Figure 18a), the supply structure is heavily reliant on variable wind and solar, with coal and natural gas plants providing residual balancing and flexibility. This leads to periods of fossil fuel consumption to meet demand when renewable generation is low. The supply structure under the CHAPS scenario (Figure 18b) is markedly different and fully decarbonized. To complement the high penetration of variable renewables, dispatchable biomass power plants are introduced to ensure grid stability and meet residual load, particularly during periods of low wind and solar output. As defined in Section 5.2, biomass is a planned component of the future renewable energy mix. Its role here is crucial for providing the necessary firm capacity to balance the grid in a system without fossil fuels. This, combined with an expanded deployment of electricity storage, creates a resilient and flexible power system that can reliably match supply with the newly expanded demand, ensuring a dynamic balance.

7. Discussion

The study quantitatively analyzes and comprehensively evaluates the decarbonization pathways of urban integrated energy systems under different hydrogen penetration rates by constructing a coupled multisectoral EnergyPLAN model and combining the entropy weight and TOPSIS methods. The results of the study reveal the non-linear relationship between hydrogen penetration rate and comprehensive system benefits, especially the finding that the “Hydrogen Alliance Potential Scenario” is better than the “Policy-Driven Scenario”, which provides an important quantitative reference for cities to formulate hydrogen energy development strategies, challenging the assertion that the higher the hydrogen energy share is, the more the city’s energy system will be able to decarbonize. It challenges the intuitive perception that “the higher the proportion of hydrogen energy, the better”. However, in order to ensure the rigor and completeness of the study, it is necessary to discuss the limitations of this study in depth and to look forward to future research directions.

7.1. Research Limitations

i.

The use of a priori settings for cross-sectoral hydrogen allocation, with fixed hydrogen substitution efficiencies for fossil fuels, reduces hydrogen utilization in the industrial sector to coal substitution and injection into natural gas pipeline networks. Whereas hydrogen substitution efficiencies are closely related to specific application technologies (such as gas turbine hydrogen doping ratios) and are dynamic parameters, the use of fixed values makes it difficult to adequately capture the impact of technical details on system energy efficiency. The industrial sector has been neglected in terms of the use of hydrogen as a chemical feedstock or high-temperature process, which feature significant differences in economics and abatement potential, which in turn may underestimate the role of hydrogen in industrial decarbonization. Ideally, hydrogen allocation should be dynamically optimized based on the marginal abatement cost of each sector, and preset values make it difficult to achieve the global optimum of the system.

ii.

The model’s key technology costs and fossil fuel prices are projected using linear trends, and the selected future data and efficiency parameters are more optimistic. Whereas the cost reductions of hydrogen-related technologies show a non-linear “learning curve” characteristic, static or linear projections may overestimate the long-term costs of hydrogen penetration scenarios. At the same time, international fuel prices and future carbon pricing mechanisms are subject to large uncertainties, and a fixed forecast curve is difficult to make cover the risk of future fluctuations, all of which can affect the accuracy of the model’s economic assessment.

iii.

The core of this study is an EnergyPLAN-based energy flow balance simulation, which focuses on “hydrogen use side analysis”, but the hydrogen energy supply chain and infrastructure modeling is insufficient, and the costing and layout modeling of “hydrogen supply side” infrastructure, i.e., the whole chain of “production–storage–delivery–use”, has not been carried out in detail, and the model lacks a dynamic optimization mechanism for selecting among green, blue, and gray hydrogen sources. There is no detailed costing and layout modeling of the “hydrogen supply side” infrastructure, i.e., the whole chain of “production–storage–delivery–use”, and the existing models only consider hydrogen as a usable fuel input, which may underestimate the overall social costs required to realize each scenario.

iv.

The model analyzes Jinan as a single energy node without considering the spatial heterogeneity within the city, and there are limitations in the spatial resolution of the model, thus ignoring issues such as regional grid transmission bottlenecks, natural gas pipeline network hydrogen doping limitations, and mismatches between the geographic distribution of renewable energy sources, such as photovoltaics, and load centers, etc. In fact, the siting of large-scale electrolytic water-to-hydrogen plants is subject to stringent constraints in terms of grid access capacity and water resource conditions. Such spatial factors can adversely affect the overall layout and operating costs of the system.

7.2. Future Research Perspectives

i.

Future research will construct more refined sub-sector models and develop more refined sub-models for the characteristics of hydrogen use in different sectors. In the transportation sector, dynamic models will be constructed based on the ownership, penetration, and behavioral patterns of different vehicle types (fuel cell vehicles and hydrogen internal combustion engine vehicles). In the industrial sector, different uses of hydrogen as a fuel and raw material will be distinguished and modeled in conjunction with specific industrial processes to more accurately assess hydrogen’s emissions reduction potential and economic value.

ii.

In the future, we will embed dynamic learning curves for key technologies in our model, dynamically depicting how costs change over time and scale of deployment. At the same time, using stochastic planning or multi-scenario analysis to cope with the uncertainties in the external environment, such as fuel prices and carbon prices, to assess the robustness of different decarbonization paths, and ultimately to propose more adaptive strategic recommendations.

iii.

In the future, the existing energy system model will be deeply integrated with the geographic information system to carry out spatial and temporal optimization studies for the whole chain of hydrogen energy “production–storage–transmission–use”, and optimize the siting and scale of hydrogen production facilities, storage, and refueling stations, taking into account the endowment of renewable energy resources, water resources distribution, constraints of power grids and pipeline networks, land use, and safety distances. Thus, we will form a blueprint for the development of hydrogen energy infrastructure that is spatially feasible and economically reasonable.

iv.

Extend the research scope from technical and economic analysis to socio-economic areas, explore the effectiveness of subsidy policies, carbon trading mechanisms, and a hydrogen certificate system on the development of hydrogen energy, assess the pulling effect of the hydrogen industry on local employment and economic growth, and analyze the public’s acceptance of hydrogen energy infrastructure through questionnaire surveys, in order to provide more comprehensive support to policy formulation, a more accurate portrayal of the role of hydrogen energy in the urban energy transition, and guidance for the scientific operation of the path to achieve the “dual carbon” goal.

8. Conclusions

Based on the EnergyPLAN model to construct an integrated energy system coupled with electric power, transportation, and industrial sectors, four hydrogen energy penetration scenarios were set up with Jinan City as the research object. The entropy weight method and TOPSIS method were combined to evaluate the dimensions of cost, carbon emission, and sustainability and to explore the impact of hydrogen energy penetration rate changes on the decarbonization path of the urban integrated energy system. The main conclusions are as follows. Our work’s novelty stems from overcoming the limitations of common single-sector analyses by innovatively quantifying the effects of graduated hydrogen penetration, as opposed to a simple with/without comparison. This granular method reveals complex, non-linear trade-offs and synergies across end-use sectors, leading to a more realistic and actionable assessment for urban energy planning.

i.

The Hydrogen Alliance Potential Scenario has the best performance in decarbonization, with CO₂ emission reduction potential and unit emission reduction cost improve by 7.98% and 29.39%, respectively, and renewable energy power generation and primary energy share increased by 89.9% and 4.9% in 2035 compared with the baseline energy inertia scenario, showing the strongest carbon emission reduction capability and renewable energy consumption potential.

ii.

The comprehensive evaluation results show that the Hydrogen Alliance Potential Scenario ranks first with a relative closeness of 0.5961, which is closest to the ideal state in terms of the balance of cost, environment, and sustainability indicators. The Policy-Driven Scenario has the lowest score (0.3154), which indicates that hydrogen penetration is not “the higher the better” and that the selection of an appropriate solution needs to be combined with the actual needs of the system.

iii.

Hydrogen penetration has a positive effect on the optimization of the energy structure. As the share of hydrogen in end-fuel consumption increases in all scenarios, the expansion of the installed capacity of electrolysis tanks effectively reduces the critical excess power and improves the stability of the system. However, the current proportion of hydrogen in the total energy consumption is still low, reflecting that the cost of the technology and the construction of the infrastructure are still the main constraints.

iv.

In terms of system costs, the annual costs of the hydrogen-containing scenarios are all higher than those of the baseline energy inertia scenario, mainly due to the inputs of hydrogen production and coal-to-hydrogen with CCS technology. However, the costs of the scenarios take on a downward trend after 2029, which predicts that with the decrease in the proportion of fossil energy sources, the economics of the hydrogen energy system will gradually improve.