Experimental Validation of a Novel CO₂ Refrigeration System for Cold Storage: Achieving Energy Efficiency and Carbon Emission Reductions

Abstract

To address the high energy consumption and carbon emissions associated with cold storage operations, a novel refrigeration system is proposed, which utilizes the natural refrigerant CO₂ and integrates an innovative control strategy. Experimental validations were conducted in Changsha (a subtropical monsoon climate) and Changchun (a continental monsoon climate), which are two regions representing typical climatic zones in China, to assess the system’s energy-saving potential, temperature stability, and environmental impacts with the total equivalent warming impact and life cycle carbon performance methods. For Changchun, the total equivalent warming impact reached 78.3 kg CO_2e/kg, reflecting reductions of 99.5% in direct emissions and 58.6% in indirect emissions compared with R410A systems, as mentioned in the reference. The life cycle carbon performance was reduced by 85.1% and 72.2% compared with the two experiment cases, with indirect emissions from energy consumption comprising the largest share. The system maintained exceptional temperature stability, with vertical-layer variations remaining under 1 °C. These findings demonstrate this system’s adaptability to achieve energy and emission reductions across diverse climates, providing a sustainable framework for future cold storage design aligned with global carbon neutrality goals.

1. Introduction

As global food demands continue to rise and trade becomes more globalized, food cold chains and cold storage industries are gaining increased importance [1]. Confronted with the dual challenges of low-energy efficiency and notable environmental impacts, particularly in cold-chain transportation and storage, the industry is urgently seeking more efficient and environmentally friendly refrigeration technologies [2,3]. Using CO₂ as a refrigerant in cold storage systems presents substantial advantages and aligns with emerging trends related to carbon neutrality objectives. CO₂, a natural refrigerant with a low global warming potential (GWP), significantly mitigates environmental impacts compared with synthetic refrigerants, contributing to the reduction in greenhouse gas emissions and supporting initiatives for carbon neutrality [4]. Furthermore, to maximize the environmental benefits of CO₂-based refrigeration, its integration with carbon capture technologies is being explored as a viable strategy to further reduce emissions and align with the EU Green Deal [5,6].

In recent years, research on carbon dioxide refrigeration has significantly increased [7]. Scholars have conducted research on various aspects, including equipment improvement [8], system construction [9], and thermodynamic analysis [10]. CO₂ refrigeration systems not only exhibit excellent overall performance in diverse application scenarios but can also be coupled with traditional refrigerants in cascade systems to enhance the Coefficient of Performance (COP) [11]. However, as an emerging refrigerant, carbon dioxide requires additional experimental validation and case analyses to advance from current theoretical research toward widespread practical applications [12].

Cold storage energy consumption is a crucial metric for evaluating performance, and numerous studies have compared the actual operational energy consumption of cold storage facilities [3,13,14,15]. In 2002, the IIR estimated that cold stores used between 30 and 50 kWh/(m³/year) [16]. Tachajapong et al. [13] conducted a survey of 161 cold storage facilities in Thailand, comprising 48 chilled cold storage and 113 frozen cold storage units. The authors recorded and analyzed electricity consumption, providing empirical formulas relating cold storage volume to power consumption. In China, the measurement unit is based on the weight of stored goods, with the energy consumption of cold storage facilities estimated to be around 0.6 kWh/day per ton of food [17]. Moreover, some studies in the literature have analyzed the composition of the cold storage load. Evans et al. [14] collected and analyzed operational data from 38 cold storage facilities in Europe, discovering that optimizing the operation of refrigeration systems can result in energy savings. Low-cost improvement measures such as enhancing door insulation and optimizing defrosting also yielded significant energy efficiency benefits. Niu et al. [18] found that door leakage and personnel operating the cooling load together accounted for 27% of the total cooling demand in frozen warehouses. This significant portion offers considerable potential for optimization and improvement. Therefore, incorporating automated equipment can mitigate the operational leaks caused by personnel and reduce door dimensions, consequently reducing the overall system’s cooling load. Existing studies have analyzed cold storage energy consumption and the composition of cold loads. In addition to the performance of the refrigeration system, auxiliary equipment and human operations also clearly contribute to the overall energy consumption of cold storage.

To highlight the advantages of CO₂-based refrigeration systems for cold storage applications, it is essential to evaluate the environmental impacts of existing refrigeration options. The current research primarily focuses on selecting and improving relevant evaluation methodologies [19]. Both the pollution caused by refrigerant leakages [20,21] and the indirect carbon dioxide emissions resulting from the electricity consumption of the equipment should be considered during the operational phase of the equipment. The annual average leakage of refrigerants is between 10% and 15% [22] and is considered a significant contributor to greenhouse gas emissions. Similarly, power consumption during the operation of refrigeration systems is equivalently converted into CO₂ emissions. This aspect of emissions can be improved by reducing the system’s energy consumption [23] and is assisted by introducing new energy systems [24]. Meanwhile, scholars have evaluated low GWP refrigeration systems [25]. Gao et al. [17] analyzed refrigeration systems in cold storage using three low GWP refrigerants and assessed their performance under various parameters. Comparisons explored power consumption, lifecycle costs, the GWP coefficient, and climate performance. The results indicated that R407F is a more favorable substitute refrigerant than R404A and R407A. Beshr et al. [26] presented an open-source LCCP framework and used that framework to compare the environmental impacts of four supermarket refrigeration systems. Li Gong et al. [27] proposed a novel comprehensive evaluation method for low-GWP refrigeration units, demonstrating the overall advantages of using R448A as a substitute for R404A.

Despite the recognized environmental benefits of CO₂-based cold storage systems, optimizing system performance remains a significant challenge, particularly in large-scale applications. At the same time, existing studies have primarily focused on the theoretical analyses of small- to medium-scale systems, with a noticeable lack of experimental validation. This study proposes an innovative refrigeration system utilizing CO₂ as a refrigerant integrated with an advanced control strategy. To evaluate the system’s performance, experiments were conducted in two distinct climatic regions of China, representative of typical northern and southern conditions. Based on the experimental results, the system’s energy-saving potential, temperature stability, and environmental performance were thoroughly analyzed, demonstrating this method’s reliability and effectiveness. This work, thus, provides a practical solution for designing new cold storage facilities that prioritize both sustainability and operational performance.

2. Methods

2.1. System Description

The CO₂ refrigeration system includes a compressor, oil separator, flash evaporator, internal heat exchanger, high-pressure reservoir, expansion valve, low-pressure circulating drum, and fin tube heat exchanger, operating in a transcritical cycle, as depicted in Figure 1b on the pressure–enthalpy (p-h) diagram. Carbon dioxide is compressed by the compressor to reach a high temperature and pressure and then directed to the flash evaporator, which operates within a partial vacuum. In this low-pressure environment, fine mist nozzles spray water droplets that rapidly evaporate, enhancing cooling efficiency, reducing water consumption, and achieving effective subcooling of the CO₂ refrigerant through heat exchange with the water vapor. Subsequently, the refrigerant flows into the internal heat exchanger for further thermal exchange. Through the high-pressure reservoir and expansion valve, carbon dioxide transitions to a state of low temperature and pressure, thereby entering the low-pressure circulating drum. The medium within the low-pressure cycle barrel flows into the finned heat exchanger in cold storage, where it facilitates the heat exchange process and lowers the temperature. To mitigate the risk of freezing on the cold storage floor due to lower temperatures, a heat exchanger is strategically installed and periodically employed for de-icing. The oil separator acts as a pre-condenser for oil–gas separation, ensuring the stable operation of the system. A flowchart of this process is detailed in Figure 1a.

The cold storage facility employs a comprehensive control strategy designed to optimize the operation of fully automated systems, ensuring precise thermal management, energy efficiency, and streamlined logistics. As shown in Figure 2, this control strategy integrates automation technologies with intelligent algorithms to maintain optimal conditions throughout the storage and retrieval processes, highlighting a paradigmatic shift in cold storage facility design.

At the inlet, the control strategy begins with vehicle-arrival monitoring, where sensors detect the proper alignment of cold chain vehicles within the docking bay. The automated control system only triggers the opening of dock doors upon proper alignment, significantly reducing heat exchange and maintaining the internal thermal environment. This controlled entry minimizes refrigeration energy losses during the unloading of goods and preserves the temperature-sensitive nature of stored products. In the sorting area, the control strategy integrates automated sorting and registration with the precooling of the goods to ensure efficient and thermally stable operations. The automated system employs barcode scanners to accurately register and categorize goods based on their storage requirements, streamlining the subsequent storage process. Simultaneously, the precooling process stabilizes the temperature of the goods before they enter the cold storage environment. By maintaining a precooling room temperature of 5 °C, this system effectively prevents thermal shocks to the goods and reduces the refrigeration load on the main cold storage facility. To optimize performance, sensors continuously monitor the efficiency of the precooling process, allowing the system to dynamically adjust the precooling intensity in real time based on current conditions. This process ensures consistent precooling quality while minimizing energy consumption, thereby contributing to the overall energy efficiency of the facility. For the transport and storage phases, the strategy employs optimal path planning to calculate the most efficient routes for transporting goods within the facility. By dynamically factoring in real-time storage availability, equipment traffic, and energy consumption, the control system minimizes unnecessary mechanical movement and reduces energy use. Once the goods reach their destination, the winch operation ensures accurate positioning within labeled storage zones, optimizing space usage and maintaining thermal uniformity by avoiding unnecessary disturbances. Within the cold storage environment, the control strategy incorporates zonal temperature stabilization to maintain precise thermal conditions. Sensors continuously monitor temperature and humidity across multiple zones. This zonal control approach prevents overcooling in unoccupied areas while prioritizing active zones, striking a balance between energy savings and thermal consistency. The main equipment involved in the aforementioned processes is illustrated in Figure 3.

Unlike traditional cold storage setups, which rely heavily on manual operations, this control strategy not only automates processes but also actively monitors and optimizes each stage to reduce energy consumption and improve efficiency. The integration of intelligent control elements demonstrates a strategic focus on energy efficiency and operational accuracy. Moreover, the strategy enables a reduction in door dimensions in scenarios where only goods are moved, further limiting refrigeration energy dissipation.

2.2. Experiment

Building upon the aforementioned system’s design and control strategy, we developed a novel CO₂-based cold storage facility in two distinct climatic zones. The structural design features a three-layered wall composition, with a steel frame serving as the central support material. This frame is flanked on both sides by layers of polyurethane, which act as highly effective insulation materials.

Traditional cold storage facilities typically have a height of 10 m or less, as they are designed to accommodate manual operations and ensure convenient access to goods. In contrast, the proposed innovative facilities leverage automated mechanical processes for both loading and unloading. This technological advancement allows for a more efficient utilization of vertical space, enabling the storage of a greater number of pallets. As a result, the height of the proposed cold storage facilities exceeds 30 m, representing a significant departure from conventional designs. Both facilities are classified as large-scale cold storage facilities, with similar volumes and a storage capacity of 15,840 pallets. Daily storage includes a diverse range of miscellaneous items, with the annual load fluctuating between 40% and 90% depending on demand. For this study, an annual average load of 60% was assumed to represent typical operating conditions.

The experimental evaluation of the CO₂ refrigeration system was conducted in two industrial cold storage facilities (Case 1: Changchun; Case 2: Changsha) to assess thermal performance under contrasting climatic conditions. Case 1 represents a continental monsoon climate in northern China, while Case 2 typifies a subtropical monsoon climate in central China. Both facilities operate as freezing chambers, maintaining a target temperature of −18 ± 1 °C. Key parameters, including dimensions and testing periods during peak summer months (August for Changchun, July for Changsha), are detailed in

Table 1. Data for experimental cold storage.

Name	Location	Latitude and Longitude	Climatic Zones	Typical Month	x (m)	y (m)	z (m)
Case 1	Changchun, central of China	43°53′49″ N 125°19′34″ E	Continental monsoon climate	August (Summer)	49.3	49.3	32.0
Case 2	Changsha, northeast China	28°13′41″ N 112°56′20″ E	Subtropical monsoon climate	July (Summer)	27.0	84.0	34.0

.

As shown in Figure 4, a three-dimensional temperature monitoring network was deployed to evaluate spatial thermal gradients within each facility. Eighteen PT100 (TianKang Group, Tianchang, China) platinum-resistance temperature detectors were distributed across three vertical layers (z-axis) and symmetrically arranged at vertex and midpoint locations along both horizontal axes (x and y), ensuring comprehensive coverage of the entire storage volume to resolve spatial temperature stratification. This configuration resulted in nine measurement points per vertical layer (18 sensors total per facility, as shown in Figure 4), which were strategically placed to resolve thermal gradients across critical geometric nodes. All sensors were calibrated to an accuracy of ±0.2 °C, with temperature data recorded at 10 min intervals via a centralized acquisition system. Power consumption was monitored using a power meter (Chint Electrics, Shanghai, China) (±0.1% accuracy) to track real-time energy usage during steady-state operation (see

Table 2. Measurement equipment parameters.

Sensor	Quantity	Measurement Device	Calibration Accuracy
Temperature (°C)	18	PT100	±0.2 °C
Power (kW)	1	Power meter	±0.1%

for the measurement equipment parameters).

The experimental protocol comprised precooling stabilization and steady-state operation. During precooling, chambers were cooled from ambient conditions to −18 °C over the course of 12 h, with temperature uniformity validated to a standard deviation ${\sigma }_T$ < 1.0 °C across all layers. The steady-state operation involved continuous monitoring under thermal equilibrium, with periodic perturbations introduced via automated product loading/unloading events simulating randomized volume changes capped at 10% per event. Environmental controls included the strict prohibition of personnel access to minimize convective disturbances.

Figure 5 illustrates the temperature and humidity variations during a typical month in Changsha and Changchun based on outdoor measurements. These two regions, representing the subtropical monsoon and continental monsoon climate zones of southern and northern China, respectively, were selected for the case study to provide insights into the distinct challenges faced by cold storage systems in different climates. Changsha experiences average temperatures of around 30 °C and humidity levels exceeding 80%, characteristic of the region’s humid subtropical climate. These conditions significantly increase the cooling load on cold storage systems, potentially reducing their operational efficiency. In contrast, Changchun, with average temperatures generally below 25 °C and winter humidity levels often below 50%, offers a more favorable environment for efficient cold storage operation. The contrasting climatic conditions of these two regions highlight the importance of assessing the systems’ ability to maintain thermal stability and energy efficiency under diverse environmental conditions.

2.3. Total Equivalent Warming Impact (TWEI)

China’s commitment to achieving carbon neutrality by 2060 underscores the urgent need to address energy consumption and greenhouse gas emissions across various sectors, including the cold chain industry—a significant contributor to both [28]. To evaluate the environmental performance of cold storage systems, this study adopts greenhouse gas emission reductions as a core metric, reflecting the critical role of this industry in China’s broader climate goals.

A widely recognized tool for such assessments is the TEWI, which quantifies the overall global warming potential of refrigeration systems. TEWI accounts for both direct and indirect contributions to global warming: the direct impact from refrigerant leakage and the indirect impact from energy consumption, primarily tied to fossil fuel-based electricity generation. By incorporating these factors, TEWI provides a comprehensive framework for evaluating the environmental consequences of cold storage operations. In this study, TEWI serves as the key metric for quantifying the greenhouse gas emissions associated with cold storage systems, enabling a thorough understanding of their environmental impacts and offering insights into strategies for emission reductions.

TEWI is calculated with Equation (1) [17,29] as follows:

$$\mathrm{T}\mathrm{E}\mathrm{W}\mathrm{I}=C\times l\times N\times GWP+C\times \left(1-\alpha \right)\times GWP+E\times N\times \beta$$

(1)

where C is the refrigerant charge in kg; l is the leakage rate per annum; and N is the service life in years. GWP represents the global warming potential; $\alpha$ is the refrigerant recovery share; E is the power consumption in kWh; and $\beta$ is the CO₂ emission intensity for electricity generation in kg/kWh.

2.4. Life Cycle Carbon Performance (LCCP)

While TEWI provides an essential metric for evaluating operational emissions, it does not account for the broader lifecycle impacts of refrigeration systems. To address this limitation, the present analysis was expanded to include LCCP, offering a comprehensive assessment of greenhouse gas emissions across the entire lifecycle of the system. LCCP extends beyond TEWI by incorporating emissions from equipment manufacturing, refrigerant production, recycling, and end-of-life refrigerant loss, in addition to operational emissions. This holistic perspective enables the identification of improvement opportunities that might be overlooked when focusing solely on operational impacts.

The LCCP is calculated based on the guidelines provided by the International Institute of Refrigeration (IIR) [30] and defined using Equation (2) as follows:

$$\mathrm{L}\mathrm{C}\mathrm{C}\mathrm{P}=\mathrm{D}\mathrm{i}\mathrm{r}\mathrm{e}\mathrm{c}\mathrm{t} \mathrm{e}\mathrm{m}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}+\mathrm{I}\mathrm{n}\mathrm{d}\mathrm{i}\mathrm{r}\mathrm{e}\mathrm{c}\mathrm{t} \mathrm{e}\mathrm{m}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}.$$

(2)

The direct emissions, which account for refrigerant leakage and end-of-life losses, are calculated using Equation (3) as follows:

$$\mathrm{D}\mathrm{i}\mathrm{r}\mathrm{e}\mathrm{c}\mathrm{t} \mathrm{e}\mathrm{m}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}=C·\left(N·l+EOL\right)·\left(GWP+adp·GWP\right)$$

(3)

where EOL is the end-of-life refrigerant leakage as a percentage, and $adp·GWP$ is the GWP of the atmospheric degradation product of the refrigerant in kg CO_2e/kg. The values are taken based on the guidelines for the life cycle climate performance.

The indirect emissions arising from energy consumption, material production, recycling, and refrigerant disposal are calculated using Equation (4):

$$\mathrm{I}\mathrm{n}\mathrm{d}\mathrm{i}\mathrm{r}\mathrm{e}\mathrm{c}\mathrm{t} \mathrm{e}\mathrm{m}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}=N·E·\beta +\sum \left(m·MM\right)+\sum mr·RM+C·\left(1+N·l\right)·RFM+C·\left(1-EOL\right)·RFD$$

(4)

where MM is the CO_2e produced/kg of material in kg CO_2e/kg; m is the mass of unit material in kg; RM is the CO_2e produced/kg of recycled material in kg CO_2e/kg; mr is the mass of recycled material in kg; RFM represents the refrigerant manufacturing emissions in kg CO_2e/kg; and RFD is the refrigerant disposal emissions in kg CO_2e/kg. These values are taken from the guidelines for life cycle climate performance.

2.5. Uncertainty Analysis

The TEWI and LCCP calculations assume that critical parameters—including the GWP, refrigerant charge, annual leakage rate, and system service life—are adopted directly from sources in the literature and treated as constants with negligible uncertainty. The total propagated uncertainty in both models is governed by regionally variable emission intensity for electricity generation, measurement errors in energy consumption, and, for LCCP, additional uncertainties in manufacturing emissions. Specifically, the emission intensity of electricity generation accounts for spatial and temporal variability (±5%), while the uncertainty in power consumption arises from instrument precision (±0.1%) and variability in material production and the supply chain (±15%). Uncertainties are propagated following the methodology of Kline et al. [31]. The composite uncertainty $\delta R$, representing the total propagated error in the dependent variable R, is expressed as follows:

$$\delta R=\sqrt{{\sum _{i=1}^n\left({\displaystyle \frac{\partial R}{\partial x_i}}·\delta x_i\right)}^2}$$

(5)

where $x_i$ denotes the i-th independent variable influencing R, and $\partial x_i$ corresponds to its associated measurement uncertainty. Partial derivatives ${\displaystyle \frac{\partial R}{\partial x_i}}$ quantify the sensitivity of R to variations in each input parameter.

Based on Equation (5), the uncertainties of the main equations are listed in

Table 3. Uncertainty of main parameters.

Parameters	TEWI	LCCP
Uncertainty	5.12%	5.46%

.

3. Results and Discussions

3.1. Power Consumption

Power consumption and the temperature at various points inside the warehouse are crucial data for evaluating the performance of cold storage.

Figure 6 shows a heatmap for the Changchun case, which demonstrates a significantly lower average hourly power consumption of 80.92 kWh, equal to 0.025 kWh/(m³∙d). Here, the daily variations in power consumption are moderate, with peak consumption typically occurring during the midday hours (10:00–14:00), coinciding with a slight increase in external temperatures. During nighttime and early morning hours (0:00–6:00), power consumption is consistently low, reflecting the system’s ability to leverage cooler ambient conditions to minimize energy usage. This pattern underscores the adaptability of the refrigeration system in responding to diurnal temperature changes.

In contrast, the Changsha case (Figure 7) exhibits higher energy demands, with an average hourly power consumption of 109.0 kWh, equal to 0.034 kWh/(m³∙d). This higher energy usage reflects the challenges posed by Changsha’s subtropical monsoon climate, which is characterized by elevated temperatures and high humidity levels. These conditions impose a significant cooling load on the facility, as additional energy is required not only to lower the air temperature but also for dehumidification. The heatmap reveals distinct daily patterns in power consumption, with peak usage generally occurring during the afternoon hours (12:00–16:00), corresponding to the highest external temperatures of the day. The system’s ability to handle these increased demands without significant spikes in power consumption demonstrates the efficiency of its refrigeration and control mechanisms. Conversely, during nighttime hours (0:00–6:00), power consumption is lower, reflecting the reduced cooling requirements. This adaptive operation ensures that the system remains energy-efficient while meeting storage demands.

Both facilities exemplify the adaptability of modern refrigeration systems to regional climatic conditions. In cooler climates such as Changchun, the system takes advantage of natural conditions to minimize energy use. In warmer and more humid climates such as Changsha, advanced technologies such as dehumidification systems and dynamic control strategies are essential to optimize energy efficiency.

3.2. Temperature Distribution

The temperature distribution in the Changchun case, as recorded by 18 sensors and illustrated in Figure 8, highlights this facility’s exceptional thermal stability. After the cooling system achieves steady-state conditions, temperatures consistently stabilize around the operational setpoint of −18 °C, with a maximum observed deviation of less than 1 °C across all sensors. This remarkable uniformity underscores the system’s capability to deliver consistent cooling throughout the facility, which is a critical factor for preserving the quality and integrity of stored goods. Even minor temperature variations can lead to uneven cooling and localized spoilage, emphasizing the importance of such precise thermal regulation.

The analysis in Figure 9 further complements this assessment by providing detailed insights into inter-sensor variability and performance. The majority of sensors exhibit median temperatures close to −18 °C, with interquartile ranges confined to ±0.2 °C, indicating minimal fluctuations. Such consistency reinforces the reliability and efficiency of the cooling system. Sensor 18 in the middle layer exhibits slightly higher median temperatures and a broader range of temperature variations compared with those of other sensors due to the relatively larger volume of stored goods in this area, which directly affects the localized thermal environment. Despite these localized fluctuations, the overall temperature distribution remains compact, underscoring the effectiveness of the refrigeration and air circulation systems in maintaining thermal uniformity across the facility.

An analysis of zonal temperature stabilization provides a deeper understanding of the spatial distribution of temperatures within the facility. The sensors were categorized into three vertical layers: upper (Sensors 1–6), middle (Sensors 7, 14–18), and lower (Sensors 8–13) layers. The upper layer recorded slightly higher median temperatures, ranging from −17.8 °C to −18.0 °C, which is attributable to the natural upward movement of warmer air. The middle layer, comprising Sensors 7 and 14–18, demonstrated the most stable temperature distribution, with medians tightly clustered around −18.0 °C and minimal variance. This layer represents the ideal zone for storing highly temperature-sensitive goods due to its superior stabilization. In contrast, the lower layer, including Sensors 8–13, exhibited slightly cooler temperatures, ranging from −18.1 °C to −18.3 °C, reflecting the natural stratification of denser, colder air at lower elevations. This layer provides consistent cooling, making it suitable for less-sensitive goods.

As shown in Figure 10, the temperature profiles of the Changsha and Changchun cold storage facilities reveal distinct thermal characteristics influenced by their respective climates. The Changchun facility presents exceptional thermal stability, with maximum inter-sensor temperature differences consistently below 1.0 °C. This uniformity reflects the effectiveness of this facility’s cooling and air circulation systems, operating under the advantageous conditions of a cooler and drier continental monsoon climate. In contrast, the Changsha facility exhibits greater variability across its sensors, particularly in the upper and middle zones, where interquartile ranges often exceed ±0.5 °C. This heightened variability is attributed to the challenges posed by the subtropical monsoon climate, including higher ambient temperatures and humidity, which impose a greater cooling load on the system.

The temperature distribution boxplot analysis for the Changsha case in Figure 11 reveals notable variations across the vertical layers of the cold storage facility, providing critical insights into zonal temperature stabilization. The upper layer (Sensors 1–6) consistently records the highest temperatures, with median values ranging from −18.5 °C to −18.0 °C. In contrast, the middle layer (Sensors 7 and 14–18) shows improved thermal stability, with median temperatures predominantly clustered around −18 ± 0.3 °C. The lower layer (Sensors 8–13) exhibits the lowest and most stable temperatures, with median values ranging from −18.0 °C to −18.2 °C. Compared with the Changchun cold storage facility, the Changsha facility shows reduced zonal stratification, as both the middle and lower layers exhibit overlapping temperature ranges. In Changchun, the lower layer consistently recorded cooler temperatures due to the more pronounced stratification of cold air. This difference can be attributed to Changsha’s subtropical climate, which imposes higher cooling demands and reduces the efficiency of natural stratification. Moreover, Changsha’s higher humidity levels and external heat load create additional challenges for maintaining uniform thermal conditions across the facility.

The sensor data feedback provides an opportunity to optimize storage strategies by enabling the system to adjust the placement of goods dynamically. Priority can be given to storing products in areas with the most stable temperatures, ensuring both product preservation and operational efficiency. Integrating real-time sensor feedback into adaptive control algorithms further allows for dynamic adjustments to airflow and cooling parameters, thereby enhancing temperature consistency.

3.3. Comprehensive Evaluation

3.3.1. TEWI

The parameters for the four cases are summarized in

Table 4. Data for the TEWI analysis.

Parameters	Case 1	Case 2	Case 3 [13]	Case 4 [15]
Refrigerant	R744	R744	R410A	R410A
Refrigerant charge amount (kg/m3)	0.051	0.052	2 kg per kW cooling load [32]
Leakage rate per year (%) [22,29]	15	15	10	10
Service life (year)	15	15	15	15
GWP [16]	1	1	1924	1924
Power consumption ((kWh$·$year)/m3) [13,15]	12.4	9.1	58.5	30.0
Carbon dioxide emission factor (kg/kWh) [17]	0.571
TEWI (kg)	78.3	106.4	523.8	279.2

. Cases 1 and 2 refer to the Changchun and Changsha cold storage facilities, while Case 3 corresponds to the cold storage facility mentioned in [13], with the power consumption calculated using Equation (6). Case 4 assumes the lowest power consumption reported in [15]. Both cold storage facilities in this comparison are assumed to use the conventional R410A refrigerant. The TEWI for each case was calculated using Equation (1), which accounts for both direct and indirect emissions as follows:

$$E=1560\times {\left(store volume\right)}^{-0.2917}$$

(6)

where E is the annual power consumption per cubic meter in (kWh$·$year)/m³.

Figure 12 presents a comparison of power consumption for the four cases, with the following ranking: Case 3 > Case 4 > Case 2 > Case 1. Taking the Changchun case as an example, the power consumption is reduced by 78.88% and 58.7%, respectively, compared with Case 3 and Case 4. This reduction in power consumption is primarily attributed to structural improvements and the introduction of an innovative control strategy, which significantly minimizes heat loss from the system and mitigates the additional cooling load caused by manual operation.

Figure 12 illustrates the TEWI analysis of four cold storage cases, revealing significant differences driven by refrigerant type and energy consumption. Cases 1 and 2, which utilize CO₂ as a refrigerant, present substantial reductions in TEWI compared with Cases 3 and 4, which employ R410A. The direct emissions in Cases 1 and 2 are minimal, reflecting a 99.5% reduction compared with Cases 3 and 4 due to the low GWP of CO₂. Indirect emissions, primarily arising from energy consumption, dominate the TEWI results in all cases, with Cases 1 and 2 achieving 78.8% and 58.6% lower indirect emissions than Case 3 and Case 4, respectively, due to their superior energy efficiency. These reductions result in TEWI values of 78.3 kg CO_2e and 106.4 kg CO_2e for Cases 1 and 2, which are significantly lower than the values of 523.8 kg CO_2e and 279.2 kg CO_2e observed in Cases 3 and 4. This analysis highlights the critical role of energy efficiency in reducing indirect emissions, which constitute the majority of TEWI in all cases, and underscores the environmental benefits of adopting low-GWP refrigerants and optimizing energy consumption.

Following the methodology of Kline et al. [31], variability in power consumption (±5%) contributes significantly to the uncertainty in TEWI calculations. In Cases 1 and 2, the uncertainty from the emission intensity of electricity generation variability accounts for approximately 68% of the total TEWI uncertainty, whereas in Cases 3 and 4, this contribution is higher, reaching 92%. The higher energy consumption in R410A systems exacerbates the dependency of these systems on grid carbon intensity, amplifying the uncertainty in their environmental performance.

3.3.2. LCCP

The LCCP was calculated using the methodology described in Section 2.4. Direct emissions were determined using Equation (3), which accounts for refrigerant leakage both during operations and at the end of life. Indirect emissions, including those from energy consumption, refrigerant manufacturing, and material production and disposal, were calculated using Equation (4). The total LCCP was obtained by summing both direct and indirect emissions, as specified in Equation (2), thereby providing a comprehensive assessment of environmental impacts over the system’s lifecycle.

Table 5. LCCP analysis of the four cases.

$\mathbf{Emissions} \left(\mathbf{k}\mathbf{g}{\mathbf{C}\mathbf{O}}_{2\mathbf{e}}\right)$	Case 1	Case 2	Case 3	Case 4
LCCP	79.25	107.73	532.38	284.86
Total direct emissions	0.12	0.12	24.49	24.49
Annual refrigerant leakage	0.12	0.12	22.26	22.26
EOL refrigerant leakage	0.01	0.01	2.23	2.23
Total indirect emissions	79.13	107.61	507.89	260.37
Energy consumption	78.16	106.29	501.51	256.95
Equipment manufacturing	0.93	1.26	5.96	3.05
Equipment EOL	0.01	0.02	0.10	0.05
Refrigerant manufacturing	0.03	0.03	0.32	0.32

presents the LCCP analysis for the four cases, detailing the distribution of direct and indirect emissions. The LCCP values vary significantly among the cases, with Cases 1 and 2 yielding the lowest values and the other cases exhibiting higher emissions. The graphical representation of these results is shown in Figure 13, providing a visual comparison of the emissions across the different cases.

The direct emissions in Cases 1 and 2 are minimal, reflecting the fact that the adoption of CO₂ refrigerants significantly reduces the environmental impacts of refrigerant leakage. In contrast, the other cases show considerably higher direct emissions, primarily attributed to the use of high-GWP refrigerants.

Indirect emissions, which primarily stem from energy consumption, are the dominant contributor to the LCCP in all four cases. Cases 1 and 2, utilizing CO₂ refrigerants, show significantly lower energy-related emissions than the R410A-based systems. This is due to the superior energy efficiency of the CO₂-based systems, which reduces power consumption and, consequently, the carbon footprint associated with energy use. For example, as power consumption is a key driver of indirect emissions, the systems in Cases 1 and 2 have significantly optimized performance, resulting in energy savings that directly translate to emission reductions.

Additional indirect emissions from equipment manufacturing, end-of-life leakage, and refrigerant manufacturing represent a smaller share of the total LCCP. These emissions are influenced by factors such as system size, design, and refrigerant type, but their contributions remain secondary compared with direct and energy-related emissions. Nonetheless, the use of low-GWP refrigerants in Cases 1 and 2 further reduces the impact of refrigerant production emissions, contributing to their superior environmental performance. The LCCP analysis highlights the influence of low-GWP refrigerants on indirect emissions and, consequently, the need to carefully consider system design.

4. Conclusions

This study proposed an innovative refrigeration system utilizing CO₂ as a refrigerant integrated with an advanced control strategy. Experimental evaluations were conducted in Changsha and Changchun, representing subtropical monsoon and continental monsoon climates, respectively, with operational data collected during typical months. The analysis focused on the system’s energy efficiency, temperature stability, and environmental performance. The findings from this analysis revealed the following:

• Energy Efficiency: a 58.7% reduction in energy consumption was achieved compared to conventional systems, with an average power consumption of 0.034 kWh/(m³·d) in Changsha and 0.025 kWh/(m³·d) in Changchun.
• Temperature Stability: temperature variations were maintained within 1 °C, with the middle layers providing the most stable conditions, ideal for temperature-sensitive goods.
• Environmental Impact: TEWI reduced by 99.5% in direct emissions and 58.6% in indirect emissions in Changchun. LCCP was reduced by 85.1% compared to Case 3 and 72.2% compared to Case 4 due to improved energy efficiency and the use of low-GWP refrigerants.

Carbon dioxide refrigeration systems have garnered considerable attention in the cold chain industry due to their excellent environmental performance. However, such systems are uncommon in real cases, mainly due to their technological immaturity. By presenting the system in this study as a typical case, we aim to contribute to the development of carbon dioxide refrigeration systems for cold storage.

Limitations: Despite the promising results, the experimental data and findings presented in this paper are applicable only to cold storage systems with the same system design and in similar climate regions, such as the subtropical monsoon and continental monsoon climates of Changsha and Changchun. Further validation is required for cold storage systems in different climatic conditions to confirm the generalizability of the results.

Future Research Directions: Future work could involve further experimental analysis of system components, especially focusing on power consumption at the individual component level. This will contribute to optimizing the system and achieving even greater energy efficiency. Additionally, validating the performance of the proposed CO₂-based cold storage system in different climate regions will be essential to ensure its adaptability and robustness across various environmental conditions.