Aiming to improve cooling and dehumidification performance in air conditioning systems and to meet the trend toward environmentally friendly refrigerants, this study proposes a coupled system that combines a CO₂ transcritical refrigeration cycle (CTRC) with a liquid desiccant dehumidification cycle. The system takes advantage of high-grade waste heat from the exothermic side of the CTRC to drive the regenerating process of the liquid desiccant dehumidification. A cooling evaporator is adopted to cool indoor air, while another evaporator (i.e., Evaporator II) is utilized to cool the concentrated solution, improving dehumidification capacity and enabling independent control of sensible and latent heat loads. Through thermodynamic modeling and the exergy analysis model, a mathematical model of the system is developed to examine how key parameters (such discharge pressure and the CO₂ mass flow rate ratio in Evaporator II (λ)) affect performance and to analyze exergy loss features. Results show that the system’s coefficient of performance (COP) and dehumidification coefficient of performance (COP_deh) initially rise and then fall with increasing CTRC discharge pressure, achieving an optimal pressure of around 10,500 kPa (COP up to 4.32) under a specific working condition, surpassing those of standalone CTRC systems. Properly increasing λ enhances dehumidification capacity and energy efficiency, with a low specific dehumidification energy (SDE) of 0.2033 kWh/kg, indicating high economic efficiency. Most exergy losses occur in the CO₂-solution heat exchanger and dehumidifier (over 60% of total losses). The system’s maximum exergy efficiency reaches 12.4%, leaving room for further improvements. This coupled system offers an efficient, eco-friendly way for air conditioning in high-humidity environments, combining cooling and dehumidification with the potential for energy recovery.

When marine chillers operate under complex marine conditions, they suffer from severe cross-equipment feature distribution shifts, scarce labeled fault samples in the target domain, industrial vibration noise mixed in sensor signals, and difficulties in accurately identifying subtle faults with varying severity levels. To tackle these issues, this paper improves upon the domain difference perception network (DDPN) and proposes a dual-hybrid attention feature discriminant domain-Adversarial network (DAFDAN) to realize intelligent fault diagnosis across different equipment and working conditions under few-shot scenarios. The proposed method constructs a dual-branch feature encoder consisting of a source domain compressor and a target domain extender to accommodate the distinct sensor dimensions of two heterogeneous chiller types. A hybrid attention module is formed by integrating squeeze-and-excitation efficient channel attention (SE-ECA, a module for screening channel-wise features) and spatial attention, which adaptively amplifies time-series features sensitive to faults and suppresses irrelevant noise. Residual connections (shortcut paths in deep neural networks to mitigate the vanishing gradient problem during deep-layer training) are introduced to optimize feature transmission. A dual-layer domain alignment framework is built with gradient reversal layers and maximum mean discrepancy (MMD). Combined with adversarial training (a training paradigm that learns domain-agnostic features through a game between a feature extractor and a domain discriminator), the framework achieves joint optimization of implicit feature confusion and explicit distance constraints. Meanwhile, a five-stage progressive training strategy is designed, which activates multiple loss functions, including weighted cross-entropy, mean square error (MSE), binary cross-entropy (BCE), and Kullback–Leibler (KL) divergence stage by stage. Class weighting and early stopping strategies are adopted to alleviate sample imbalance and model overfitting. In this paper, the public ASHRAE RP-1043 centrifugal chiller dataset is used as the source domain, and time-series measurement data collected from a self-developed laboratory marine screw chiller serves as the target domain. Verification experiments are carried out covering one normal steady-state operating condition and 15 gradient faults falling into five major categories with different severity degrees. Results from ablation experiments (controlled-variable comparative experiments that quantify the independent contribution of each component by comparing model performance with or without a specific module/loss), multi-algorithm comparisons, and confusion matrix visualization demonstrate that the cross-domain fault diagnosis accuracy of the proposed DAFDAN approaches is 100%, outperforming mainstream transfer learning algorithms such as support vector machine (SVM), deep neural network (DNN), MMD, correlation alignment (CORAL), and domain-adversarial neural network (DANN). Multiple ablation experiments verify that the three core components—hybrid attention, adversarial training, and semi-supervised learning—jointly boost the model’s diagnosis accuracy and operational stability. The loss curves of the complete five-stage training process converge smoothly. The confusion matrix reveals zero misjudgments and zero false alarms across all 16 refined operating states, enabling precise identification of subtle incipient faults of all severity levels. This study proves that DAFDAN can effectively address the pain points of few-shot cross-equipment fault diagnosis for marine chillers and provides a reliable algorithmic reference for the intelligent operation and maintenance of ship refrigeration equipment.

Chillers serve as the core high-energy-consuming equipment in heating, ventilation, and air conditioning (HVAC) systems. During long-term continuous operation, they tend to suffer gradual subtle degradation, with a performance deviation less than 5%. Conventional fault diagnosis methods rely on manual threshold judgment or labeled fault data, which fail to realize accurate early warning signals. In addition, existing algorithms lack multi-dimensional baseline comparisons to verify their practical engineering performance. To address these limitations, this paper proposes an unsupervised health assessment method combining an LSTM autoencoder and Euclidean distance (LSTM-AE-ED). A multi-gradient fault time-series dataset is generated via a MATLAB R2022b/Simscape mechanism model verified by both summer field measurements and refrigeration pressure-enthalpy cycles, which resolves the practical engineering challenges of scarce on-site fault samples and potential equipment damage caused by actual fault tests. The proposed model is trained solely on healthy time-series data. It extracts dynamic coupling characteristics of chillers through LSTM, constructs a dimensionless health index based on Euclidean distance in feature space, and introduces the standard deviation of health index to improve evaluation stability. Baseline comparisons with vanilla AE and single-layer LSTM are carried out. Experimental results demonstrate that the proposed method achieves an identification accuracy of 96.3% and exhibits high sensitivity to mild degradation of four typical faults, adapting to dynamic multi-working-condition scenarios. This approach requires no additional acquisition devices for derived parameters such as power consumption and COP; online assessment can be realized merely with standard temperature, pressure, and flow sensors equipped on chillers. With lightweight inference performance, it is suitable for edge monitoring terminals of chillers in data centers, providing a low-cost and practical quantitative technical scheme for predictive maintenance and hierarchical early warning signals of refrigeration equipment.