With the continuous advancement of the “dual-carbon” strategy, the green and low-carbon transformation of power systems and transportation systems is imminent. Integrated Energy Systems (IES) that combine multiple energy forms, such as electricity, hydrogen, and heat, play a crucial role in optimizing resource allocation and improving energy efficiency. In particular, the application of IES in new energy vehicles, including EV cars and charging infrastructure, not only promotes the popularization and efficient use of EV cars but also accelerates the construction of a green transportation system, effectively reducing the overall carbon emissions and operating costs of the energy system. However, existing hydrogen energy utilization models often suffer from low energy efficiency and inability to flexibly supply electricity, hydrogen, and heat. To address these issues, this paper proposes a collaborative low-carbon dispatch model for integrated electricity-hydrogen-heat systems that considers new energy vehicles, such as EV cars and hydrogen fuel cell vehicles.
The comprehensive utilization of hydrogen energy is a key focus. Traditional hydrogen energy systems typically involve localized coupling, failing to achieve synergistic effects across production, storage, and use. This paper constructs an efficient hydrogen energy utilization scheme that includes electrolyzer-based combined heat and hydrogen production, fuel cell combined heat and power (CHP) with variable thermoelectric ratios, and orderly hydrogen refueling for heavy-duty trucks. This approach enables multi-link heat recovery and electrical coordination, systematically addressing bottlenecks such as waste heat from hydrogen production and inflexible thermoelectric ratios in fuel cells. Additionally, new energy vehicles, including EV cars, are innovatively introduced to participate in a joint carbon-green certificate market mechanism. Using the Minkowski sum method, a vehicle-grid collaborative dispatch model is established to fully exploit the market potential of the transportation system in carbon emission reduction. Furthermore, through carbon-green certificate joint trading constraints, the system synergistically optimizes carbon emission control and renewable energy consumption.
The integrated electricity-hydrogen-heat system architecture is illustrated in Figure 1. The system includes sources such as natural gas CHP units, distributed renewable energy, hydrogen production equipment (electrolyzers), fuel cell CHP systems, and energy storage devices. The load side comprises electricity, heat, and hydrogen energy demands. Electrical loads are met by CHP units, distributed renewable energy, and fuel cells, while thermal loads are supplied by CHP units, electrolyzers, and fuel cell CHP units in coordination.
Energy supply equipment modeling is essential for system optimization. The gas turbine, as the core of the CHP unit, has output models as follows:
$$ P_{gt}^{ele}(t) = \eta_{gt}^{ele} \cdot H_{gt}(t) $$
$$ P_{gt}^{heat}(t) = \eta_{gt}^{heat} \cdot H_{gt}(t) $$
where \( P_{gt}^{ele}(t) \) is the power generation output at time \( t \), \( P_{gt}^{heat}(t) \) is the heat output at time \( t \), \( H_{gt}(t) \) is the total heat value of fuel consumed, \( \eta_{gt}^{ele} \) is the power generation efficiency, and \( \eta_{gt}^{heat} \) is the heat production efficiency.
Distributed renewable energy devices, such as wind turbines and photovoltaic (PV) systems, have output models that depend on environmental factors. The wind turbine output power is given by:
$$ P_{wt}^{out}(t) = \begin{cases}
0 & v(t) \leq v_{ci} \text{ or } v(t) \geq v_{co} \\
\frac{1}{2} \rho A v(t)^3 C_p & v_{ci} < v(t) < v_{rated} \\
P_{max}^{wt} & v_{rated} \leq v(t) < v_{co}
\end{cases} $$
where \( \rho \) is air density, \( A \) is the swept area, \( v(t) \) is wind speed at time \( t \), \( C_p \) is the power coefficient, \( v_{ci} \), \( v_{rated} \), and \( v_{co} \) are cut-in, rated, and cut-out wind speeds, respectively, and \( P_{max}^{wt} \) is the maximum output power. The PV output power model is:
$$ P_{pv}^{out}(t) = P_{max}^{pv} \frac{G(t)}{G_{std}} [1 + \alpha (T(t) – T_{std})] $$
where \( P_{max}^{pv} \) is the maximum PV output, \( G(t) \) is solar irradiance, \( G_{std} \) is standard irradiance, \( \alpha \) is the temperature coefficient, \( T(t) \) is the actual temperature, and \( T_{std} \) is the reference temperature.
Hydrogen production equipment, such as electrolyzers, consumes electricity to produce hydrogen and heat:
$$ P_{h2}^{ele}(t) = P_{ele}^{ele}(t) \cdot \eta_{h2}^{ele} $$
$$ P_{heat}^{ele}(t) = P_{ele}^{ele}(t) \cdot \eta_{heat}^{ele} $$
where \( P_{ele}^{ele}(t) \) is the electricity consumption, \( P_{h2}^{ele}(t) \) is hydrogen production power, \( P_{heat}^{ele}(t) \) is heat production power, \( \eta_{h2}^{ele} \) is hydrogen production efficiency, and \( \eta_{heat}^{ele} \) is heat production efficiency.
Fuel cell CHP systems convert hydrogen into electricity and heat efficiently:
$$ P_{fc}^{ele}(t) = P_{h2}^{fc}(t) \cdot \eta_{fc}^{ele}(t) $$
$$ P_{fc}^{heat}(t) = P_{h2}^{fc}(t) \cdot \eta_{fc}^{heat}(t) $$
where \( P_{fc}^{ele}(t) \) is electricity output, \( P_{fc}^{heat}(t) \) is heat output, \( P_{h2}^{fc}(t) \) is hydrogen consumption, and \( \eta_{fc}^{ele}(t) \), \( \eta_{fc}^{heat}(t) \) are time-varying efficiencies for electricity and heat, respectively.
Energy storage equipment, such as hydrogen storage tanks, operates with charge-discharge losses:
$$ E_{h2}(t) = E_{h2}(t-1) + P_{ch}^{h2}(t) \eta_{ch}^{h2} – \frac{P_{dis}^{h2}(t)}{\eta_{dis}^{h2}} $$
where \( E_{h2}(t) \) is the hydrogen storage level at time \( t \), \( P_{ch}^{h2}(t) \) is charging power, \( P_{dis}^{h2}(t) \) is discharging power, and \( \eta_{ch}^{h2} \), \( \eta_{dis}^{h2} \) are charging and discharging efficiencies.
New energy vehicles, including EV cars and hydrogen fuel cell vehicles, are modeled as generalized energy storage devices. The travel characteristics of EV cars follow probability distributions for arrival/departure times and driving distances. The arrival time probability density is:
$$ f(t) = \begin{cases}
\frac{1}{\sigma_{a} \sqrt{2\pi}} \exp\left(-\frac{(t-\mu_{a})^2}{2\sigma_{a}^2}\right) & \mu_{a} \leq t \leq 12 \\
\frac{1}{\sigma_{d} \sqrt{2\pi}} \exp\left(-\frac{(t-\mu_{d})^2}{2\sigma_{d}^2}\right) & 12 < t \leq 24
\end{cases} $$
where \( \mu_{a} \), \( \mu_{d} \) are expected arrival and departure times, and \( \sigma_{a} \), \( \sigma_{d} \) are standard deviations. The driving distance probability density is log-normal:
$$ f_d(d) = \frac{1}{d \sigma_d \sqrt{2\pi}} \exp\left(-\frac{(\ln d – \mu_d)^2}{2\sigma_d^2}\right) $$
To aggregate individual EV cars into a cluster, the Minkowski sum method is used. The dispatchable domain for EV car clusters is defined as:
$$ P_{ch}^{ev}(t) = \sum_{i=1}^{N_{ev}} \delta_{ev}^i(t) P_{ch}^{ev,i}(t) $$
$$ P_{dis}^{ev}(t) = \sum_{i=1}^{N_{ev}} \delta_{ev}^i(t) P_{dis}^{ev,i}(t) $$
$$ E_{min}^{ev}(t) \leq E_{ev}(t) \leq E_{max}^{ev}(t) $$
$$ E_{ev}(t) = E_{ev}(t-1) + P_{ch}^{ev}(t) \eta_{ch}^{ev} – \frac{P_{dis}^{ev}(t)}{\eta_{dis}^{ev}} $$
where \( P_{ch}^{ev}(t) \), \( P_{dis}^{ev}(t) \) are total charging and discharging powers of the EV cluster, \( \delta_{ev}^i(t) \) is a binary variable indicating grid connection status, \( E_{ev}(t) \) is the cluster’s energy state, and \( \eta_{ch}^{ev} \), \( \eta_{dis}^{ev} \) are efficiencies. Similar models apply to hydrogen truck clusters for refueling.
The carbon-green certificate joint trading mechanism integrates carbon emission rights and renewable energy certificates. The initial carbon quota is allocated based on a baseline method, and a tiered carbon trading model is adopted. The system’s carbon emissions mainly come from gas turbines. The carbon trading model is:
$$ Q_{carbon}^{initial} = \lambda_{carbon} \sum_{t=1}^{T} [P_{gt}^{ele}(t) + P_{gt}^{heat}(t)] $$
$$ Q_{carbon}^{actual} = \lambda_{fuel} \sum_{t=1}^{T} H_{gt}(t) $$
where \( Q_{carbon}^{initial} \) is the initial carbon quota, \( \lambda_{carbon} \) is the carbon quota coefficient per unit output, \( Q_{carbon}^{actual} \) is actual carbon emissions, and \( \lambda_{fuel} \) is the carbon emission coefficient per unit fuel heat value. The green certificate trading model is:
$$ Q_{green}^{initial} = \lambda_{green} \sum_{t=1}^{T} P_{load}^{ele}(t) $$
$$ Q_{green}^{actual} = \frac{\sum_{t=1}^{T} [P_{pv}(t) + P_{wt}(t)]}{1000} $$
where \( Q_{green}^{initial} \) is the initial green certificate quota, \( \lambda_{green} \) is the quota coefficient, \( Q_{green}^{actual} \) is the actual green certificate amount, and \( P_{load}^{ele}(t) \) is electrical load. The joint trading model accounts for carbon reduction from green certificates:
$$ Q_{offset} = \sum_{t=1}^{T} [\kappa_{pv} P_{pv}(t) + \kappa_{wt} P_{wt}(t)] $$
$$ Q_{carbon}^{trade} = Q_{carbon}^{actual} – Q_{carbon}^{initial} – Q_{offset} $$
$$ Q_{green}^{trade} = Q_{green}^{actual} – Q_{green}^{initial} $$
where \( \kappa_{pv} \), \( \kappa_{wt} \) are carbon reduction conversion factors for PV and wind, respectively.
The vehicle-grid collaborative optimization dispatch model aims to minimize total system cost, including equipment maintenance, energy purchase, EV car V2G incentives, and carbon-green certificate trading costs. The objective function is:
$$ \min C_{total} = C_{gas} + C_{om} + C_{v2g} + C_{trade} $$
where:
$$ C_{gas} = \sum_{t=1}^{T} c_{gas} V_{gas}(t) $$
$$ C_{om} = \sum_{t=1}^{T} \sum_{i} c_{om}^i P_i(t) $$
$$ C_{v2g} = \sum_{t=1}^{T} c_{v2g} P_{dis}^{ev}(t) $$
$$ C_{trade} = c_{carbon} Q_{carbon}^{trade} + c_{green} Q_{green}^{trade} $$
Here, \( c_{gas} \) is natural gas price, \( V_{gas}(t) \) is gas volume, \( c_{om}^i \) is maintenance cost per unit output for device \( i \), \( P_i(t) \) is output power, \( c_{v2g} \) is V2G incentive, \( c_{carbon} \) and \( c_{green} \) are carbon and green certificate trading prices.
Constraints include energy balance, equipment operation limits, storage dynamics, and EV car requirements. Energy balance constraints for electricity, hydrogen, and heat are:
$$ P_{pv}(t) + P_{wt}(t) + P_{gt}^{ele}(t) + P_{fc}^{ele}(t) + P_{dis}^{ev}(t) = P_{load}^{ele}(t) + P_{ele}^{ele}(t) + P_{ch}^{ev}(t) $$
$$ P_{h2}^{ele}(t) + P_{dis}^{h2}(t) = P_{h2}^{fc}(t) + P_{ch}^{h2}(t) + P_{load}^{h2}(t) $$
$$ P_{gt}^{heat}(t) + P_{heat}^{ele}(t) + P_{fc}^{heat}(t) = P_{load}^{heat}(t) $$
Equipment operation constraints ensure power outputs and ramp rates are within limits:
$$ P_{min}^i \leq P_i(t) \leq P_{max}^i $$
$$ \Delta P_{min}^i \leq P_i(t) – P_i(t-1) \leq \Delta P_{max}^i $$
Storage constraints for hydrogen tanks include capacity and charge-discharge limits:
$$ E_{min}^{h2} \leq E_{h2}(t) \leq E_{max}^{h2} $$
$$ 0 \leq P_{ch}^{h2}(t) \leq \delta_{ch}^{h2}(t) P_{ch,max}^{h2} $$
$$ 0 \leq P_{dis}^{h2}(t) \leq \delta_{dis}^{h2}(t) P_{dis,max}^{h2} $$
$$ \delta_{ch}^{h2}(t) + \delta_{dis}^{h2}(t) \leq 1 $$
EV car cluster constraints ensure energy states meet user expectations upon departure:
$$ E_{ev}^i(t_{depart}) \geq \gamma_{ev} E_{max}^{ev,i} $$
where \( \gamma_{ev} \) is the minimum state-of-charge ratio required by users.
The solution process involves linearizing nonlinear models using the Big-M method and solving the mixed-integer linear programming problem with solvers like CPLEX in MATLAB. The dispatch period is one day (24 hours) with a 1-hour interval. The flowchart in Figure 2 illustrates the optimization process, which includes input data processing, model formulation, constraint handling, and solution output.
A case study based on a resource-rich area in Gansu, China, validates the model. The system includes wind turbines, PV panels, gas turbines, electrolyzers, fuel cell CHP, hydrogen storage, and EV cars. Key parameters are listed in Table 1.
| Equipment | Parameter | Value |
|---|---|---|
| Gas Turbine | Max Power Output (kW) | 1000 |
| Max Heat Output (kW) | 2400 | |
| Ramp Limit (kW/h) | 200 | |
| Efficiency (Ele/Heat) | 0.38 / 0.45 | |
| Electrolyzer | Max Power (kW) | 500 |
| Efficiency (Heat/H2) | 0.25 / 0.78 | |
| Fuel Cell CHP | Max H2 Consumption (kW) | 120 |
| Max Efficiency (Ele/Heat) | 0.55 | |
| H2 Storage | Max Capacity (kg) | 400 |
Three scenarios are compared: Scenario 1 (traditional IES with electric boilers and batteries, no hydrogen utilization, random EV car charging), Scenario 2 (preliminary hydrogen utilization with fuel cell CHP, no vehicle dispatch), and Scenario 3 (comprehensive hydrogen utilization with EV car and hydrogen truck collaborative dispatch under joint carbon-green certificate trading). Results show that Scenario 3 reduces total costs by 11.7% and 4.21% compared to Scenarios 1 and 2, respectively, while improving renewable energy utilization by 6.2% and 2.5%. Carbon emissions decrease by 14.3% and 3.5% in Scenarios 2 and 3 relative to Scenario 1.
Analysis of hydrogen energy comprehensive utilization reveals that electrolyzers and fuel cell CHP units flexibly adjust to load demands, reducing gas turbine outputs and waste heat. In Scenario 3, gas turbine electricity and heat outputs decrease by 833 kW and 403 kW, respectively, compared to Scenario 1, demonstrating the strategy’s superiority. The integration of EV cars and hydrogen trucks further optimizes energy use. EV car charging loads are shifted to periods of high renewable output, and discharging during peak loads supports the grid. Hydrogen truck refueling times are adjusted to alleviate daily hydrogen demand peaks, improving system economics and stability.
The impact of renewable energy output fluctuations is assessed by varying output coefficients. As shown in Figure 3, total costs decrease with higher renewable output, but Scenario 3 maintains faster cost reduction due to better utilization through vehicle-grid collaboration. For instance, at 1.2 times the baseline output, Scenario 3 achieves near-full renewable consumption, highlighting its robustness.
In conclusion, the proposed collaborative low-carbon dispatch model effectively integrates hydrogen energy utilization and new energy vehicles, such as EV cars, into IES. Hydrogen comprehensive use reduces operating costs by 11.7% and enhances renewable utilization by 6.2%. EV car and hydrogen truck collaborative dispatch further cut costs by 4.21% and boost renewable utilization by 2.5%. These strategies significantly lower carbon emissions and improve system flexibility, providing a reference for low-carbon transformation in coupled transportation-energy systems. Future work could explore user participation willingness under incentive pricing.
The modeling and optimization approaches presented here demonstrate the potential of integrated systems in achieving sustainability goals. By leveraging EV cars as flexible resources and advancing hydrogen technologies, we can pave the way for a cleaner energy future. The continuous development of such systems will be crucial in meeting the demands of evolving energy landscapes and environmental regulations.