Abstract:The development and utilization of underground spaces can ease the shortage of urban land resources, ensure urban safety, and improve the urban ecological environment. Human overall comfort and work efficiency in underground spaces are affected by several environmental factors, such as thermal, acoustic, and lighting. Owing to the particularities of an underground space at any given location, the guarantee of a comfortable indoor environment is different from that in aboveground buildings. Based on the thermal, acoustic, and illumination characteristics of an underground space, the main differences between underground spaces and aboveground buildings, limitations of current standards/codes, and imperfections of internal environmental evaluation indicators are summarized and analyzed, and human comfort and work efficiency in terms of one-way and three-way interactions are discussed based on the literature published since 2000. The findings reveal that the current standards/codes for underground spaces mainly refer to aboveground buildings. Where the parameter index is single, there exists a large difference between the underground environment and the design standard, and there is no evaluation index for human comfort in underground spaces. In the existing research, two methods have been adopted: field surveys and climate room experiments. These have mainly focused on the effects of thermal, acoustic, and lighting environment unidirectional control on personnel comfort, while only a few studies have been conducted on work efficiency. Research on the three-directional interactions involving thermal, acoustic, and lighting environments in underground spaces is lacking. The creation of a microenvironment based on local cooling and heating, and the technologies generated under the concepts of imitating the ground environment, imitating the natural environments, and thermal–acoustic–light coordinated control are expected to play important roles in the construction of underground space environments in the future, and lay the foundation for the construction of an ecological environment system in underground spaces.

Improvement of thermal comfort for underground space: Data enhancement using variational autoencoder

Towards inclusive underground public transportation: Gender differences on thermal comfort

Parameter optimisation of the ventilation system for an underground power space using a hybrid model

A thermal sensation model for naturally ventilated indoor environments based on deep learning algorithms

Effect of Thermal, Acoustic, and Lighting Environment in Underground Space on Human Comfort and Work Efficiency: A Review

Probabilistic occupancy forecasting for risk-aware optimal ventilation through autoencoder Bayesian deep neural networks

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

A two-level optimisation approach for underground natural ventilation based on CFD and building energy simulations

Study of Thermal Comfort in Underground Construction Based on Field Measurements and Questionnaires in China

Modeling effective heat transfer and ventilation in deeply buried underground tunnels

Occupant-centric HVAC and window control: A reinforcement learning model for enhancing indoor thermal comfort and energy efficiency

Real-time indoor thermal comfort prediction in campus buildings driven by deep learning algorithms

BIM and Data-Driven Predictive Analysis of Optimum Thermal Comfort for Indoor Environment

Working in Underground Spaces: Architectural Parameters, Perceptions and Thermal Comfort Measurements

Predictive Monitoring of Built Thermal Environment Using Limited Sensor Data: A Deep Learning-Based Spatiotemporal Method

Research on Online Temperature Prediction Method for Office Building Interiors Based on Data Mining

A novel deep generative modeling-based data augmentation strategy for improving short-term building energy predictions

Modelling ventilation and convective heat transfer in deeply buried underground tunnels based on boundary layer theory

A hybrid machine learning model to optimize thermal comfort and carbon emissions of large-space public buildings

Thermal-Adaptation-Behavior-Based Thermal Sensation Evaluation Model with Surveillance Cameras

Indoor environment data time-series reconstruction using autoencoder neural networks