Assessment of Energy Recovery Potential in Urban Underground Utility Tunnels: A Case Study

Abstract

Underground spaces contain abundant geothermal energy, which can be recovered for building ventilation, reducing energy consumption. However, current research lacks a comprehensive quantitative assessment of its energy recovery. This research evaluates the energy recovery potential of the Xingfu Forest Belt Urban Underground Utility Tunnels. Field experiments revealed a 7 °C temperature difference in winter and a 2.5 °C reduction during the summer-to-autumn transition. A computational fluid dynamics (CFD) model was developed to assess the impact of design and operational factors such as air exchange rates on outlet temperatures and heat exchange efficiency. The results indicate that at an air change rate of 0.5 h⁻¹, the tunnel outlet temperature dropped by 10.5 °C. A 200 m tunnel transferred 8.7 × 10¹⁰ J of heat over 30 days, and a 6 m × 6 m cross-sectional area achieved 1.1 × 10¹¹ J of total heat transfer. Increasing the air exchange rate and cross-sectional area reduces the inlet–outlet temperature difference while enhancing heat transfer capacity. However, the optimal buried depth should not exceed 8 m due to cost and safety considerations. This study demonstrates the potential of shallow geothermal energy as an eco-friendly and efficient solution for enhancing building ventilation systems.

1. Introduction

Urban expansion and population growth have increasingly strained land resources, posing significant challenges for sustainable urban development. As a result, cities are shifting from horizontal expansion to vertical development, making efficient land use an essential trend in modern urban planning [1]. A critical component of this trend is the development and utilization of underground spaces, which have become integral to contemporary urban strategies [2]. By 2022, China’s urban underground space had reached 2.962 × 10⁹ m² [3], highlighting its substantial potential for energy recovery [4]. Large-scale expansion of shallow geothermal energy within these spaces could play a crucial role in significantly reducing greenhouse gas emissions [5]. In this regard, the integration of urban underground space and geothermal energy as an alternative to traditional fossil fuels has attracted global attention [6].

Shallow geothermal energy refers to the heat stored near the Earth’s surface, typically within the top 100 m. Shallow geothermal energy, a renewable resource primarily derived primarily from solar radiation [7], is predominantly used for space heating and cooling, domestic hot water, and seasonal thermal storage [8]. The process involves harnessing ground heat to cool or warm outdoor air [9], which is then circulated indoors to aid in enhancing thermal comfort and reducing energy consumption. When harnessed efficiently, this energy provides a sustainable solution for both heating and cooling buildings.

In recent years, the development of shallow geothermal energy has accelerated, driven by its environmental benefits and vast potential [10,11]. In China alone, the annual exploitable shallow geothermal energy is estimated to be equivalent to 700 million tons of standard coal [12]. Research on shallow geothermal energy has primarily focused on its applications in various sections, including agricultural greenhouses [13], power plants [14], tunnels [15], wine cellars [16], and industrial processes [17]. Numerous experimental and modeling studies have confirmed its potential for reducing emissions and enhancing cooling and heating delivery efficiency, particularly in tunnel ventilation and heat recovery systems. For instance, Davies et al. [18] designed a cooling and heat recovery system for underground railway tunnels utilizing heat pumps to produce heat, with a direct connection to district heating networks. Based on their calculations, the technology is expected to reduce carbon emissions from the London Underground by 360,000 tons. In another study, Ahmed et al. [19] studied building-integrated horizontal Earth-to-Air Heat Exchangers (EAHEs) and found that it can save a maximum of 415.92 kWh of energy and AUS 190.31 in costs. Additionally, Lyu et al. [20] developed an integrated HVAC system that leverages shallow geothermal energy, achieving 29% energy savings compared to traditional ground source heat pumps. At the same time, the system allows a reduction of 7 kg per square meter in CO₂ emissions. Moreover, extensive theoretical investigation has also been conducted on heat exchange using shallow geothermal energy with multiple models developed and implemented. In one of these investigations, a model for EAHE was developed to assess the impact of multiple factors, including air temperature, relative humidity, inlet wind speed, and tunnel wall temperature, on system performance. The results indicated that lower inlet air temperatures lead to a slower decrease in air temperature along the EAHE’s length [21]. In their study, Le et al. [22] analyzed and evaluated a measurement model to examine the shallow geothermal potential and temperature distribution profiles at depths ranging from 0.1 m to 3.6 m. Their findings revealed that the daily temperature difference between the outdoor air levels and depths of 1.5 to 3.6 m was 10–25 °C in winter and −15–5 °C in summer. This makes the system highly effective for cooling in summer and heating in winter. Rana et al. [23] introduced multiple fins in EAHE to increase the contact area and showed that the outlet temperature of the four-fin block was reduced by 14.52% compared to the single-fin block.

To further explore the influencing factors of shallow geothermal energy ventilation, Derbel et al. [24] aimed to investigate the impact of various factors on cooling and heat recovery efficiency in underground tunnels, in addition to evaluating their potential for emission reduction and pre-cooling/heating. To attain this, they measured soil temperatures at different depths in the Southern Mediterranean climate zone. These measurements were then compared against temperatures predicted by thermal models to assess energy transfer performance. Similarly, Guo et al. [25] developed a heat transfer model for underground tunnels in Beijing, revealing an effective heat exchange length of 1545 m, with an optimal length of 628 m. In their work, Kumar et al. [26] explored the influence of geothermal gradients, surface conditions, and moisture content, and proposed a numerical model to predict the energy-saving potential of EAHE. The developed model was validated using empirical data collected from the Mathura Tunnel in India. To minimize negative effects on the surrounding rock liner structure during long-term operation, the optimal wind speed range was investigated by a laboratory test model [27]. Additionally, Ji et al. [28] analyzed the effects of inlet water temperature, flow rate, and surrounding rock types on heat transfer performance using TRNSYS, incorporating a capillary heat exchanger at the front end of a heat pump system that utilized waste heat from the Qingdao subway. While a large number of researchers have developed soil–air heat transfer models, it is worth noting that the energy savings from ventilation in underground tunnels vary significantly by region. There is a lack of empirical research on shallow geothermal applications, especially in cold climates.

To address these issues, this study examines the Xingfu Forest Belt Urban Underground Utility Tunnels in Xi’an, Shaanxi Province. The research results are validated through real-field tests, ensuring a high level of practical applicability. The case study is regarded as one of the world’s largest projects integrating underground space for landscape greening, municipal roads and infrastructure, comprehensive pipeline tunnels, and subway support. This paper aims to study and assess the factors affecting the long-term ventilation and heat exchange operation within the examined tunnels. The study investigates how various operational parameters influence outlet air temperature and heat exchange capacity by conducting on-site environmental parameter measurements and developing a computational fluid dynamics (CFD) numerical simulation model for tunnel-air heat exchange. The study provides a quantitative assessment for evaluating underground utility tunnel ventilation, filling a key research gap, and revealing important thermodynamic insights. These findings offer guidance for optimizing ventilation design and control strategies. Other cities face similar challenges in developing underground spaces, and this research can provide valuable guidance.

2. Materials and Method

2.1. Experimental Site

In this study, a ventilation system within an underground utility tunnel is utilized to efficiently harness shallow geothermal energy. During winter, heat is extracted from the ground and transferred to buildings for heating. In summer, the system reverses the process, moving heat from the outdoor air to the cooler subsurface, providing effective cooling. Since the underground temperature remains stable throughout the year, regardless of surface weather fluctuations, this energy transfer method offers a reliable and sustainable solution for optimizing thermal management.

This study focuses on a segment of the underground utility tunnel associated with the Xingfu Forest Belt Project in Xi’an, Shaanxi Province. It is a typical continental monsoon climate area, which makes the utilization of underground space for heating and cooling very representative. Construction of this extensive project commenced in 2016, and it became operational on 1 July 2021. As depicted in Figure 1, the pipeline tunnel is divided into four distinct sections: comprehensive, pressure, and gas compartments, which accommodate pipelines for water supply, reclaimed water, electricity, communication, heating, and natural gas. The tunnel spans a total length of 12,428 m, including 5755 m in the Happiness Road section, 5305 m in the Wanshou Road section, and 1368 m in the east–west connecting tunnels and power ditches. Additionally, the tunnel is buried beneath an average soil cover of 7 to 11 m, with the shallowest section covered by approximately 2 m.

2.2. Temperature Sensor Layout

To elucidate the relationship between the internal temperature of underground utility tunnels and the outdoor temperature, long-term monitoring was conducted for both tunnel temperature (T₁) and outdoor temperature (T_a). This monitoring covered both winter and the transitional seasons. Measurements were taken in the Changle Road section of the underground utility tunnels, with the winter monitoring period spanning from 16 November to 9 December 2020, and the summer–autumn transitional period from 25 September to 7 October 2021. The specific measurement locations and their respective layout are shown in Figure 2.

To ensure effective heat exchange between the air inside the tunnel and the surrounding soil before measurements, a temperature and humidity sensor was installed 250 m from the tunnel entrance and 2.1 m above the tunnel floor. The variation in the outdoor temperature was monitored by an automatic weather station located directly above the measurement site. The primary instruments employed in the experiments included a TH12R-EX temperature and humidity sensor from Inste Technology Co. Ltd, Shenzhen, China, and a TRM-ZS2 automatic weather station from Jin Zhou Sunshine Meteorological Technology Co., Ltd, Jinzhou, China, with temperature data recorded at two-minute intervals.

Table 1. Measurement instruments’ specifications.

Parameter	Testing Instrument	Model	Range	Accuracy
T1	Temperature and humidity sensor	TH12R-EX	−40 °C–85 °C	±0.2 °C
Ta	Automatic weather station	TRM-ZS2	−40 °C–80 °C	±0.2 °C

presents the specific measurement ranges and accuracy levels of the measurement instruments.

2.3. Numerical Simulation

2.3.1. Physical Model

In this study, the physical model for the numerical simulation was based on the Xingfu Forest Belt Underground Utility Tunnel, with necessary and reasonable simplifications applied to its details. The modeling and meshing were conducted using ANSYS 19.0’s ICEM software. According to the established standards, fire doors were installed every 200 m within the tunnel, establishing the usable length of a single tunnel segment as L = 200 m. The simulation area included the tunnel inlet, outlet, and all walls inside the tunnel. The reference buried depth of the tunnel was based on the shallowest cover depth, with the distance from the tunnel top to the ground surface set as H₀ = 2 m. The tunnel’s width and height are denoted as D × H = 4 m × 4 m, with the inlet and outlet vents level with the ground surface. Additionally, the vents had the same width as the tunnel, with dimensions l₀ × d₀ = 2 m × 4 m. To evaluate the temperature variations in the soil surrounding the tunnel under ventilation conditions, the tunnel was enclosed on three sides, excluding the top, by a 30 m thick soil layer (L₁, L₂ = 30 m). This setup ensured that the temperature of the outermost soil layer remained unaffected by tunnel ventilation during the simulation period, while also providing sufficient thermal mass for the tunnel. The physical structure and cross-section of the model are shown in Figure 3.

2.3.2. Mathematical Model

Computational Fluid Dynamics (CFD) was utilized to simulate the airflow and heat transfer processes in the tunnel ventilation system. FLUENT was chosen as the computational tool for its strong ability to handle complex fluid dynamics challenges. The heat exchange process between the tunnel walls and the internal air was simplified to a single sensible heat transfer process. Additionally, the airflow and heat transfer characteristics within the computational domain were determined by solving the governing equations for the conservation of mass, momentum conservation, and energy conservation. These governing equations are expressed in the following unified form [29].

$${\displaystyle \frac{\partial \left(\rho \varphi \right)}{\partial t}}+div\left(\rho U\varphi \right)=div\left({\mathrm{\Gamma }}_{\varphi }grad\varphi \right)+S_{\varphi }$$

(1)

where $\varphi$ represents a general variable that can denote velocity, temperature, concentration (relative humidity), turbulent kinetic energy, or other solution parameters [30]. ${\mathrm{\Gamma }}_{\varphi }$ is the generalized diffusion coefficient; and $S_{\varphi }$ denotes the source term. The governing equations were discretized using the finite volume method, with specific schemes applied as follows: an implicit Euler scheme for the temporal term, a second-order upwind difference scheme for the convection term, and a central difference scheme for the diffusion term. In this study, the full set of governing equations was solved using the computational fluid dynamics software FLUENT 18.0. Also, pressure–velocity coupling was managed using the SIMPLE algorithm, and the standard $\kappa -\epsilon$ turbulence model was selected.

2.3.3. Boundary Conditions

This study aims to examine the impact of long-term heat exchange between outdoor air and the soil surrounding the tunnel on the tunnel’s heat exchange capacity. On this basis, it is necessary to consider the seasonal temperature variations in the soil above and around the tunnel in the other three directions, along with the temporal changes in temperature due to heat exchange. Additionally, the seasonal fluctuations in the temperature of the introduced outdoor air must be factored in to understand and characterize the evolution of the tunnel’s heat exchange capacity throughout the year. Therefore, a transient heat transfer model is necessary to ensure accurate calculation results. To simplify the modeling and simulation process, the following assumptions are made:

• The thermal properties of the soil are assumed to be constant and unaffected by temperature variations.
• Diurnal temperature variations in the soil are neglected, and initial soil temperature is considered dependent solely on buried depth.
• Only convective heat transfer is included during the operation of the tunnel ventilation system.
• Moisture migration in the soil is ignored, and the focus is solely on heat exchange between the soil surrounding the tunnel and the air within the tunnel through the tunnel walls.

The CFD simulations are based on a typical summer day, specifically 31 July, in Xi’an, China. The boundary condition at the tunnel entrance is set as a velocity inlet, with varying inlet velocities specified, to simulate different air changes per hour. On the other hand, the outlet is configured as an outflow boundary, allowing air to exit the tunnel without specifying additional conditions, ensuring that the airflow can leave the domain naturally and the pressure at the boundary remains balanced. The tunnel walls, which are constructed from concrete, are treated as a gas–solid interface, with the wall boundary defined as a coupled boundary, which can accurately reflect the heat transfer effect. The solution to the problem is obtained through automatic iteration. Additionally, the soil boundaries on the horizontal lateral and lower layers surrounding the tunnel are assumed to have infinite thickness, so the heat transfer through the outer surfaces of these soil layers is set to zero. Due to minimal temperature fluctuations in the deep soil above the tunnel, and the thickness of the soil layer exceeding 1.5 m, diurnal temperature variations in the solid domain soil are neglected. Moreover, the initial time is set to the hottest day of the year, with July 31 chosen as the day with the highest average daily temperature in the region. The air inside the tunnel is defined as the fluid domain, with air as the material. Similarly, the surrounding soil is defined as the solid domain, with soil as the material.

Table 2. Thermophysical properties of different materials.

Material	Air	Soil	Concrete
Density (kg/m3)	1.225	2000	2500
Specific heat (J·kg−1·°C−1)	1006.4	1300	920
Thermal conductivity (W·m−1·°C−1)	0.0242	1.5	1.74

presents the main thermal properties of the air, surrounding soil, and reinforced concrete tunnel walls. As the soil boundary temperature varies with depth and time, and the inlet air temperature fluctuates over time, these variations are defined using time-dependent functions. In this regard, the detailed setup of the soil boundary condition and the inlet air temperatures is described as follows:

• Soil

To ensure that the external soil layer is unaffected by heat transfer from the tunnel, the surrounding area is modeled as an idealized infinite expanse. Consequently, in addition to the top layer of soil above the tunnel, a 30 m thick soil layer is maintained on three sides of the tunnel, in addition to the top layer of soil above it. The external surfaces of these soil layers are treated as adiabatic boundaries, preventing any heat transfer across them. For simplification, the concrete linings of the tunnel walls have been excluded from the model. It is noted here that the simulation accounts for variations in soil temperature as a function of both depth and time. On the other hand, air inlet conditions are dynamically adjusted based on the system’s operational duration.

The heat transfer in the shallow soil can be treated as a two-dimensional transient heat transfer process along the radial direction. The distribution of its temperature field can be obtained from the Fourier heat conduction differential equation if the earth is considered as a semi-infinite homogeneous object, as shown below:

$$\begin{array}{c}{\displaystyle \frac{\partial \theta }{\partial \tau }}=\alpha {\displaystyle \frac{{\partial }^2\theta }{\partial y^2}}\\ {\theta }_{(y,\tau )}=t_{(y,\tau )}-t_d\end{array}$$

(2)

where θ represents the excess temperature at any point within the strata (°C), τ denotes the time elapsed since the occurrence of the annual peak surface temperature (h), and y is the vertical depth in the soil (m). t (y, τ) is the soil temperature at depth y and time τ (°C), t_d is the average surface temperature (°C) and α is the thermal diffusivity of the soil (m²/h).

The surface soil temperature can be considered to vary synchronously with the air temperature in contact with it.

$${\theta }_{ps}=A_d\cos {\displaystyle \frac{2\pi }{T}}\tau$$

(3)

In the above equation, θ_s represents the excess temperature at the surface at any given moment (°C), θ₀ is the annual amplitude of temperature fluctuation at the surface (°C) and T is the period of the temperature wave (h).

The relationship between the excess temperature and soil depth y and time τ in a semi-infinite homogeneous body under periodic heat exchange conditions can be derived. Consequently, the relationship between ground temperature $t_{(y,\tau )}$, time $\tau$, and soil depth y can also be determined.

$${\theta }_{(y,\tau )}=A_d{e}^{-y\sqrt{{\displaystyle \frac{\pi }{\alpha T}}}}\cos \left({\displaystyle \frac{2\pi }{T}}\tau -y\sqrt{{\displaystyle \frac{\pi }{\alpha T}}}\right)$$

(4)

$$t_{(y,\tau )}=t_d+A_d{e}^{-y\sqrt{{\displaystyle \frac{\pi }{\alpha T}}}}\cos \left({\displaystyle \frac{2\pi }{T}}\tau -y\sqrt{{\displaystyle \frac{\pi }{\alpha T}}}\right)$$

(5)

In Xi’an, summer temperatures peak at 33.5 °C, while winter lows reach −6.5 °C. The average annual temperature is 13.5 °C, with a fluctuation range of 20 °C. The soil’s thermal diffusivity is α = 0.0032 m² h⁻¹, with a fluctuation period of 8760 h. These parameters guide the simulated annual outdoor temperature, omitting daily temperature variations. Consequently, Equation (6) is derived, representing the relationship between soil temperature variations with depth and time, which is described as follows:

$$t_{\left(y,t\right)}=13.5+20e^{-0.33477y}\cos \left(0.00072\tau -0.33477y\right)$$

(6)

Assuming a consistent ventilation period of three months (90 days) for both winter and summer seasons, Equation (6) is implemented into Fluent using a User-Defined Function (UDF). This UDF sets the surface soil as the y-axis origin, effectively enabling the regulation of temperature fluctuations in the soil layer above the tunnel over time. It also enables the precise control of the initial temperature settings within the soil strata and allows for adjustments of both ventilation duration and airflow velocity.

• Inlet air temperature

To verify the consistency between the simulation results obtained using the UDF and the inlet air temperatures at specific times, polynomial fitting was applied to the outdoor air temperature data from 31 July 2020, and the data came from the measured temperature of the Happiness Forest Belt. Specifically, a fourth-degree polynomial fitting curve was used, and it is represented as follows:

$$\begin{array}{l}y=29.09843+\left(-3.02503\times {10}^{-4}\right)t+\left(8.56749\times {10}^{-9}\right)t^2\\ +\left(8.01768\times {10}^{-14}\right)t^3+\left(-1.67959\times {10}^{-18}\right)t^4\end{array}$$

(7)

where y denotes the inlet temperature (°C) and t represents the time of the operation day (h) (Figure 4).

2.4. Simulation Cases

This study evaluates the impact of various parameters on the heat exchange capabilities of urban utility tunnels, including air changes per hour (ACH), tunnel length (L), buried depth (H₀), and tunnel cross-sectional area (D × H). To facilitate this analysis, the following simulation scenarios were established:

• Scenario 1: The outdoor temperature was set to 33.5 °C. The tunnel length (L) was 200 m, and the buried depth was H₀ = 2 m. The inlet wind speed was 0.56 m/s, and the tunnel cross-sectional area was D × H = 4 m × 4 m. Different rates of air change per hour (ACH) were considered, with 0.5, 1.0, and 5.0 h⁻¹, and their corresponding effects on the tunnel’s heat exchange capabilities were investigated to optimize tunnel heat exchange performance.
• Scenario 2: With an air change rate of ACH = 5.0 h⁻¹, tunnel lengths of 100 m and 200 m were evaluated to explore the impact of tunnel length on heat exchange capabilities. All other parameters in this scenario were the same as in Scenario 1.
• Scenario 3: The air change rate was set to ACH = 5.0 h⁻¹, with tunnel buried depths of 0.5 m, 2 m, and 12 m examined to assess the influence of buried depth on the tunnel’s heat exchange performance. All other parameters in this scenario were the same as in Scenario 1.
• Scenario 4: Tunnel cross-sectional areas were varied to 2 m × 2 m, 4 m × 4 m, and 6 m × 6 m, with an air change rate of ACH = 5.0 h⁻¹, to evaluate the impact of the cross-sectional area on the heat exchange capabilities. All other parameters in this scenario were the same as in Scenario 1.

2.5. Grid Independence Analysis and Model Validation

To ensure the accuracy of the simulation results, a refined structured mesh was applied near the tunnel walls, while coarser meshes were used in areas farther from the tunnel to improve computational efficiency, as shown in Figure 5a. To determine the optimal mesh size, six different computational grids were tested, with elements of 3.80 × 10⁵, 7.00 × 10⁵, 1.08 × 10⁶, 2.24 × 10⁶, 3.07 × 10⁶, and 5.05 × 10⁶ elements. The axial velocity near the tunnel exit was evaluated under identical conditions for each considered mesh scenario, as shown in Figure 5b. At the outlet of the tunnel, the velocity variation trend is the same, but the number of different grids shows slight differences. The results show that when the number of grids increases from 3.80 × 10⁵, the wind speed also increases. When the number of grids increases from 3.07 million to 5.05 million, the difference in wind speed is minor, and the average deviation of the obtained results is only 0.28%, indicating that further increasing the grid density has a negligible influence on the accuracy of the calculated results. Therefore, the grid with 3.07 million elements was selected for the numerical simulation of the flow field, ensuring computational efficiency without compromising accuracy.

The accuracy of numerical simulation results depends on various factors related to modeling and the conditions applied. In this study, the precision of the ventilation flow field simulation results was validated by comparing the results with data from existing literature. The numerical model settings presented by Zhang et al. [31] were used as a benchmark. After simulation calculation under the same conditions, the comparison results are shown in Figure 6. The simulation results of the numerical model used in this study demonstrate excellent consistency between the two sets of calculations. The temperature decreases gradually from 30 m at the tunnel inlet to the 200 m outlet. Compared to the reference data, the temperature difference at the outlet is 0.45 °C, with an error of only 0.15%. These results confirm the reliability of the numerical model and simulation conditions employed in this research.

3. Results

3.1. Characteristics of Temperature Distribution in Underground Utility Tunnels

Figure 7 presents a comparative analysis of the indoor tunnel temperature (T₁) and the external ambient temperature (T_a) throughout the test period. During winter, the air temperature inside the tunnel remained remarkably stable and significantly warmer than the outside air temperature, with an average difference of 7 °C and a peak difference reaching 21.8 °C. These results highlight the tunnel’s considerable potential as an auxiliary natural source for both heating and cooling. During the transitional season, the average external temperature was 18.9 °C, closely aligning with the tunnel’s average temperature of 18.5 °C. Notably, during working hours (9:00 to 18:00), the tunnel’s temperature was, on average, 2.5 °C cooler than the outside, with a maximum difference of 10.5 °C. Overall, the air temperature inside the tunnel shows better stability and is less affected by fluctuations in the outdoor environment. This demonstrates that even during the relatively cooler transitional seasons, underground tunnels maintain significant cooling capacity and can serve as efficient sources of ventilative cooling.

Figure 8 presents segmented statistical data on the temperature differences (ΔT) between the tunnel interior and exterior during the testing period. Additionally,

Table 3. Percentage of time intervals with temperature differences (ΔT) inside and outside the tunnel during testing.

ΔT (°C)	>0	>1	>2	>3	>5	>7	>10
Winter (%)	98.6	97.6	95.7	90.0	73.6	49.5	15.5
Transitional season (%)	60.0	43.3	32.2	30.0	22.2	11.1	-

provides the percentage of these temperature differences. It is noted that during the winter testing period, external temperatures were higher than the tunnel’s internal temperatures in only 1.4% of the recorded data. In this regard, the tunnel’s internal temperature remains at least 3 °C cooler than external temperatures for over 90% of the time. During the transitional period, although external temperatures generally exhibited a gradual decrease, they were higher than the internal tunnel temperatures 60% of the time during the 90 working hours between 9:00 and 18:00, and the temperature difference above 7 °C accounted for 11.1% of the time. The proportion is greater when the temperature difference is small. These findings demonstrate that underground tunnels offer significant potential for ventilation applications during both winter and transitional seasons, highlighting their value in the development of sustainable urban infrastructure.

3.2. The Impact of Air Change Rate (ACH)

The air change rate (ACH) has a substantial impact on airflow within the tunnel, which in turn affects thermal recovery. In this regard, increasing the air velocity can enhance the efficiency of heat exchange between the air and the tunnel walls, thereby improving heat transfer rates. However, the limited thermal capacity of the surrounding soil may result in smaller temperature fluctuations within the tunnel as the airflow increases. In this study, air change rates of ACH = 0.5, 1.0, and 5.0 h⁻¹ over 720 h were analyzed to evaluate the effect of different airflow rates on thermal recovery and average air temperature at the tunnel outlet.

Table 4. Simulated cases with different ACH.

Case	ACH (h−1)	Ta (°C)	L (m)	H0 (m)	D × H (m × m)	Duration (h)
1	0.5	33.5	200	2	4 × 4	720
2	1.0	33.5	200	2	4 × 4	720
3	5.0	33.5	200	2	4 × 4	720

summarizes the conditions of the different simulated cases.

Figure 9 and

Table 5. Heat transfer under different air change rates.

Parameter	Case 1	Case 2	Case 3
Inlet (°C)	33.5	33.5	33.5
Outlet temperature after 3 h of ventilation (°C)	20.9	23.0	25.3
Final outlet temperature (°C)	23.0	25.3	29.2
Total heat transfer (J)	1.6 × 1010	2.6 × 1010	8.7 × 1010

depict the variation in air temperature at the tunnel outlet. Before activating the ventilation system, the initial temperature at the tunnel outlet was set to match the external ambient temperature of 33.5 °C. Once ventilation began, it was observed that the outlet air temperature rapidly decreased due to the tunnel’s initial internal air temperature of 16.5 °C, eventually stabilizing at this lower level. After reaching the minimum, the temperature gradually increased and stabilized, with the rate of stabilization inversely proportional to the air change rate.

Furthermore, in cases 1–3, the stable air temperatures achieved were 23.0 °C, 25.3 °C, and 29.2 °C, respectively, corresponding to heat transfers of 1.6 × 10¹⁰ J, 2.6 × 10¹⁰ J, and 8.7 × 10¹⁰ J. As the airflow rate increased, the ability to recover heat from the soil through the supplied air improved, leading to more effective heat extraction. The total heat transfer can be expressed using the formula $Q=cm\mathsf{\Delta }T$, where c is the specific heat capacity of air, m is the mass flow rate, and ΔT is the temperature difference between the inlet and outlet. However, this also resulted in higher supply air temperatures, which negatively affects its suitability for indoor ventilation.

In cases 1 and 2, the supplied air temperatures were 23.0 °C and 25.3 °C, respectively, remaining below the indoor thermal comfort threshold of 28 °C. This suggests that the air could be used directly for building cooling. However, a lower ACH limits the amount of heat recovered, reducing the potential for reuse and decreasing its effectiveness for building cooling systems. Conversely, with an ACH of 5.0 h⁻¹, the supplied air temperature increased to 29.2 °C, exceeding the 28.0 °C comfort limit, making it unsuitable for direct use in comfort ventilation. To optimize this heat recovery, it is essential to integrate the recovered heat with other active cooling systems, especially at higher ACH levels where the total amount of recovered heat is significantly greater. The approach enhances the engineering potential of the heat recovery system and improves its overall effectiveness in managing indoor temperatures.

3.3. The Influence of Tunnel Length

Tunnel length (L) significantly influences both heat exchange efficiency and supply air temperature. Typically, as the tunnel length increases, the temperature gradient resulting from the cooling or heating of the supplied air becomes more pronounced. Additionally, the length of each tunnel segment must adhere to fire zone regulations. In this study, numerical simulations were performed for tunnel lengths of 100 m and 200 m, with an ACH = 5 h⁻¹ and an inlet air temperature of 33.5 °C. The simulations obtained characterized the ventilation and heat exchange in the tunnel over 30 days.

Table 6. Simulated cases with different lengths.

Case	L (m)	Ta (°C)	ACH (h−1)	H0 (m)	D × H (m × m)	Duration (h)
4	100	33.5	5.0	2	4 × 4	720
5	200	33.5	5.0	2	4 × 4	720

presents the specific parameters used in these simulations.

Figure 10 shows the temperature distribution inside the tunnel for the air inside the corridor and the surrounding soil along the tunnel length after 30 days of ventilation in case 4 and case 5. After this period of heat exchange, the tunnel outlet temperatures stabilized, indicating effective control of the heat transfer process. It is noted here that the elevated air temperatures within the tunnel significantly impacted the surrounding shallow soil, especially the soil above the tunnel. However, the outlet temperature of the 200 m tunnel was noticeably lower compared to the 100 m tunnel. This indicates that a longer tunnel length results in improved cooling performance under the same air change rate.

As presented in Figure 11 and

Table 7. Heat transfer in the tunnel after 30 days of ventilation.

Parameter	Case 4	Case 5
Inlet temperature (°C)	33.5	33.5
Outlet temperature after 3 h of ventilation (°C)	27.8	25.3
Final outlet temperature (°C)	30.9	29.2
Total heat transfer (J)	5.7 × 1010	8.7 × 1010

, under the same ventilation conditions, the air temperature in case 4 and case 5 rapidly drops below 20 °C and then rises. After 3 h of ventilation, the outlet temperature reaches a stable level. The outlet temperature of the 200 m tunnel in case 5 was 2.5 °C lower than that of the 100 m tunnel (Case 4) after 3 h of ventilation. This indicates that a longer tunnel not only results in lower air temperatures at the outlet but also enhances overall heat transfer. Additionally, for case 4, 3 h of ventilation reduced the outdoor air temperature from 33.5 °C to 27.8 °C, representing a decrease of 5.7 °C. However, it is shown that after 30 days of continuous ventilation, the average outlet temperature was 30.9 °C, only 2.6 °C below the inlet temperature. This highlights a significant reduction in cooling efficiency and heat transfer effectiveness. On the other hand, the heat transfer capacity reported for case 5 was around 1.5 times that of case 4. The findings of this study indicate that while longer tunnels enhance heat transfer, practical constraints such as ventilation placement and fire safety regulations limit the maximum feasible length of heat exchange tunnels.

3.4. The Effect of Tunnel Buried Depth

The initial soil temperature in various regions depends on soil depth and time, with tunnel buried depth playing a key role in determining the initial soil temperature. To explore how different buried depths affect heat recovery and the average outlet temperature, this study analyzed three specific depths, H₀ = 0.5 m, 2 m, and 12 m, corresponding to cases 6–8. These buried depths resulted in soil cover at the tunnel surface of 4.5 m, 6 m, and 16 m, respectively. The effects of these buried depths on both the temperature at the ventilation outlet of the tunnel and the overall temperature distribution were evaluated and analyzed. The simulation conditions for these cases are detailed in

Table 8. Simulated cases with different depths.

Case	H0 (m)	Ta (°C)	ACH (h−1)	L (m)	D × H (m × m)	Duration (h)
6	0.5	33.5	5	200	4 × 4	720
7	2.0	33.5	5	200	4 × 4	720
8	12.0	33.5	5	200	4 × 4	720

.

Figure 12 presents the temperature distribution within the tunnel and the surrounding soil after 30 days of system operation. It is evident that at a buried depth of 12 m, the tunnel’s internal air temperature decreased more rapidly than at shallower depths of 0.5 m and 2 m, exerting a smaller effect on the surface temperature and the outlet temperature shows a better cooling property.

In addition, Figure 13 depicts the temporal variation in tunnel outlet temperature, showing a significant cooling effect on the outdoor air as it enters the tunnel. In cases 6–8, the outlet temperatures sequentially dropped to 19.2 °C, 16.9 °C, and 14.2 °C, respectively, stabilizing within 0.5 h. Overall, for the same initial temperatures, increased buried depth resulted in lower outlet temperatures and improved heat transfer. On the other hand, shallower buried depths resulted in more pronounced fluctuations in soil temperature, affecting initial heat transfer performance significantly. As shown in

Table 9. Heat transfer under different depths.

Parameter	Case 6	Case 7	Case 8
Inlet temperature (°C)	33.5	33.5	33.5
Outlet temperature after 3 h of ventilation (°C)	27.5	25.3	23.0
Final outlet temperature (°C)	29.9	29.2	27.4
Total heat transfer (J)	6.6 × 1010	8.7 × 1010	1.1 × 1011

, after 30 days of ventilation, the outlet temperatures for case 6 and case 7 converged. This indicates that the heat storage capacity of the soil at shallow depths is limited. Moreover, the temperature trends for the tunnel buried at 2 m and 12 m remained consistent over the 30-day ventilation period, reflecting stable heat exchange between the surrounding soil and the tunnel’s ventilated airflow. This stability was associated with an outlet temperature difference of 1.8 °C and a significant increase in total heat transfer. While deeper buried depths generally improve heat exchange, the economic implications of deeper tunnels should be considered. Therefore, buried depths should not exceed 8 m, as soil temperature stabilizes beyond this point.

3.5. The Effect of Tunnel Cross-Sectional Area

The structural dimensions of the tunnel play a crucial role in its heat exchange performance. The tunnel’s cross-sectional area directly affects the capacity of convective heat exchange between the surrounding soil and the airflow within the tunnel. Additionally, the tunnel’s dimensions impact wind speed and airflow volume. To assess the impact of the tunnel cross-sectional area, three cross-sectional dimensions were considered for evaluation in this study: 2 m × 2 m, 4 m × 4 m, and 6 m × 6 m, corresponding to cases 9–11. To ensure a consistent tunnel center depth of 4 m, the buried depths of the tunnel’s top were adjusted to 3 m, 2 m, and 1 m in the three cases 9, 10, and 11, respectively. Considering these conditions, numerical simulations were conducted to evaluate the impact of these varying cross-sectional areas on the tunnel’s heat exchange capacity.

Table 10. Simulated cases with different cross-sectional dimensions.

Case	D × H (m × m)	Ta (°C)	L (m)	H0 (m)	ACH (h−1)	Duration (h)
9	2 m × 2 m	33.5	200	2	5.0	720
10	4 m × 4 m	33.5	200	2	5.0	720
11	6 m × 6 m	33.5	200	2	5.0	720

presents the detailed simulation parameters of the three investigated cases.

Figure 14 illustrates the temperature distribution within the tunnel and the surrounding soil after 30 days of operation. It is evident that larger tunnel cross-sectional areas significantly influence the temperature of the shallow soil. In case 11, with the largest cross-sectional area of 6 m × 6 m, the cooling effect at the tunnel’s air outlet was less pronounced compared to cases 9 and 10. The air above the tunnel is always at a higher temperature level because the distance from the surface is too limited. This reduced effectiveness is attributed to the fact that air in closer contact with the tunnel walls experiences more substantial heat exchange, while air nearer the center of the tunnel exhibits less temperature change as the distance from the tunnel walls increases. Consequently, smaller cross-sectional areas result in a reduced distance between the air and the tunnel walls, enhancing the efficiency of heat exchange.

In addition, Figure 15 and

Table 11. Heat transfer under different cross-sectional areas.

Parameter	Case 9	Case 10	Case 11
Inlet temperature (°C)	33.5	33.5	33.5
Outlet temperature after 3 h of ventilation (°C)	21.8	25.3	26.9
Final outlet temperature (°C)	27.4	29.2	29.9
Total heat transfer (J)	6.2 × 1010	8.7 × 1010	1.1 × 1011

present the temperature variations over time during the system’s operation. With an initial outdoor temperature of 33.5 °C, the temperature inside the tunnel for cases 9, 10, and 11 rapidly decreased to 16.9 °C upon entering the tunnel before gradually increasing. After 0.5 h of operation, it is noted that the outlet temperatures stabilized. After 3 h, the temperature drops observed were 11.7 °C, 8.2 °C, and 6.6 °C for cases 9, 10, and 11, respectively, and case 9 showed the best cooling performance. Over a continuous 30-day operation period, the soil’s pre-cooling capacity declined, leading to final outlet temperatures of 27.4 °C, 29.2 °C, and 29.9 °C, respectively. Moreover, increasing the tunnel’s cross-sectional area from case 9 to case 10 resulted in a more rapid reduction in outlet temperature, with a decrease of 1.8 °C. However, the overall system heat transfer efficiency is influenced by both the temperature drop and the ventilation volume. Although a larger cross-sectional area results in a smaller temperature drop, it allows for a higher ventilation volume, thereby enhancing overall heat transfer. In case 11, with the largest cross-sectional area, the heat transfer reached 1.1 × 10¹¹ J, a 9-time increase in cross-sectional area and a 77% increase in heat transfer compared to case 9. Therefore, for applications requiring a significant temperature differential with lower airflow requirements, a narrower tunnel system is the optimal choice. Conversely, for situations with lower temperature differential requirements but higher airflow needs, a wider tunnel system is more appropriate.

4. Discussion

Overall, the influence of ACH on outlet temperature and heat transfer performance is particularly significant. Lower ACH results in a greater decrease in tunnel outlet temperature. Therefore, the design and operation of the ventilation system must carefully balance air volume with cooling load to ensure optimal performance. Additionally, reducing tunnel buried depth can lower construction costs without significantly affecting heat transfer performance, offering greater flexibility for engineering applications. Future designs should prioritize these factors to enhance the economic efficiency and adaptability of the system.

To evaluate the heating and cooling performance of the Xingfu Forest Belt Tunnel, an economical analysis of the system was conducted based on the actual heating and cooling seasons in Xi’an in case 3. The tunnel has a significant heat exchange efficiency, as shown in Figure 16. In winter, the air temperature after passing through the tunnel averages 4.4 °C above the outdoor temperature, while in summer, it is 3.1 °C lower. For a 200 m tunnel with an airflow of 16,128 m³/h, the heat exchange capacity reaches 6.0 × 10⁷ J/h in summer, saving about 5.9 tons of standard coal during the cooling season. In winter, the capacity increases to 8.4 × 10⁷ J/h, saving approximately 8.3 tons of standard coal during the heating season. With a total tunnel length of 12,428 m, the system can save 7.17 million kWh of electricity and 880.4 tons of standard coal each year, with a reduction of 2377 tons of CO₂ emissions each year. In addition to avoiding additional piping required for standalone shallow geothermal systems, this system provides significant economic and energy savings benefits.

This study evaluated the effectiveness of using shallow geothermal energy in tunnel ventilation systems. Compared to renewable energy sources such as solar and wind, shallow geothermal energy offers several advantages, including greater stability, smaller land requirements, and continuous operation, making it a promising solution for broader implementation. However, some limitations persist. Outdoor climate conditions and soil characteristics are critical factors influencing heat exchange efficiency. Extreme weather, such as prolonged periods of high heat or cold, along with fluctuations in soil temperature, may reduce the system’s overall effectiveness. Additionally, prolonged operation may result in soil degradation or incur high operational costs, all of which must be carefully addressed during the design phase to ensure long-term feasibility and efficiency.

5. Conclusions

This study introduces an innovative method for harnessing shallow geothermal energy to pre-cooling air in underground tunnels. Aiming to validate the effectiveness of tunnel ventilation for cooling, the impacts of various parameters on heat exchange performance were evaluated and analyzed, including air change rate, tunnel length, buried depth, and cross-sectional area. The main conclusions of this study are as follows:

(1)

Long-term field experiments conducted in Xi’an demonstrated significant pre-cooling and heating potential for underground utility tunnels. In winter, the tunnels achieved a notable temperature difference of 7 °C compared to the outdoor environment, while during the summer-to-autumn transition, a temperature reduction of 2.5 °C was observed. The consistent internal temperature of the tunnels throughout the year highlights the significant potential of utilizing tunnel soil as a reliable thermal source.

(2)

The air change rate has a direct effect on tunnel outlet temperature. At an air change rate of 0.5 h⁻¹, the temperature drop at the tunnel outlet was 10.5 °C. Overall, higher air change rates significantly enhance heat recovery, leading to greater improvements in heat exchange performance.

(3)

Greater tunnel buried depth leads to lower outlet temperatures. Balancing soil temperature stability with economic considerations, the optimal tunnel buried depth for tunnels should not exceed 8 m.

(4)

Tunnel dimensions significantly affect heat exchange efficiency. A tunnel length of 200 m transferred 8.7 × 10¹⁰ J of heat over 30 days, while a cross-sectional area of 6 m × 6 m achieved a total heat transfer of 1.1 × 10¹¹ J. Although larger cross-sectional areas result in higher total heat transfer, they also produce higher outlet temperatures compared to smaller cross-sections.

Therefore, the tunnel ventilation system design should prioritize a buried depth of less than 8 m and an air exchange rate of 5 h⁻¹ or higher to maximize heat transfer efficiency. Additionally, policymakers should consider integrating underground utility tunnels into urban planning to capitalize on their energy recovery potential, reduce building energy consumption, and support urban sustainability.

The results indicate that these tunnels can effectively reduce supply air temperatures, thus reducing building energy consumption. This research presents design concepts for the sustainable development of underground spaces and can be integrated with energy recovery systems in future urban planning to improve energy efficiency.