By deploying data acquisition sensors on the floating platform, superstructure, and other components of the building, as well as in the surrounding environment, motion response data (displacement, vibration acceleration) of the structure during daily operation and environmental data (wind speed, wave elevation) are collected. Based on the acquired data, digital twin modeling, structural response prediction, and environmental parameter inference are achieved.

Addressing the energy constraints and sensing requirements of offshore habitation, we research and develop a series of self-powered IoT nodes, including smart floors, HMI buttons, wave energy buoys, and small vertical axis wind turbines. These energy-autonomous and functionally independent nodes facilitate energy conservation, emission reduction, and ubiquitous IoT connectivity in offshore habitation projects.

By collecting real-time data from sensors—such as cameras and environmental monitors—and equipment operational status deployed both indoors and outdoors, multimodal large models are utilized to create a universal administrator for monitoring, analyzing, and predicting indoor and outdoor conditions.

An underwater robot launch ramp, located close to the water surface, is arranged on the first deck for related experiments.

This exhibition unit features a prepared 50cm × 50cm × 5cm sample of flame-retardant, ultra-insulating polymer supercritical CO₂ foaming material for construction. Compliant with building fire safety requirements, the material achieves ultra-insulation and radiative cooling effects, and the sample is embedded within the wall structure. The display aims to allow visitors to visually observe the material’s texture, structure, and performance, and to experience its lightweight and insulating properties through touch. Accompanying informational panels provide details on its excellent flame retardancy, thermal insulation, corrosion resistance, and its potential applications in offshore habitation architecture.

The off-grid low-carbon smart microgrid is centered around a retired lithium battery energy storage system, combined with photovoltaics and small-scale wind power. It forms a stable AC bus via bidirectional inverters and is managed by a carbon-intelligent Energy Management System (EMS) for unified dispatch and low-carbon optimization. The system adopts a strategy of “renewable energy priority, storage peak shaving and valley filling, and critical load assurance.” During the day, photovoltaics and wind power supply loads directly while charging the batteries; at night or during periods of low power generation, the storage system discharges to support building energy consumption. When resources are abundant, the intelligent dispatch system dynamically adjusts charging and discharging curves to minimize carbon emissions and energy waste. The EMS also integrates prediction algorithms, carbon emission factor calculations, and multi-energy flow monitoring. This achieves real-time monitoring, visualization, and optimized decision-making for energy consumption and carbon emissions, thereby creating a safe, economical, low-carbon, and sustainable building-level smart grid for energy and carbon management.

Using a UAV platform to observe surface ocean currents provides the flexibility to cover a large area based on flight settings while achieving high spatial resolution. It is adaptable to the operational environment of floating platforms and the specific characteristics of the Pearl River Estuary waters, thereby enhancing the monitoring of nearshore flows. Currently, the ADCP is rented from the EESRF Central Laboratory.

Multi-parameter water quality detection (including dissolved oxygen, pH, temperature, salinity, turbidity, conductivity, resistivity, total dissolved solids, ammonia nitrogen, nitrates, chlorides, etc.), BOD analyzer, chlorophyll content determination, and dissolved pCO2 analyzer. Currently, the equipment is rented from the EESRF Central Laboratory.

This project plans to develop a non-contact full-rotation underwater thruster, which is primarily used to test an underwater pod thruster with a power of approximately 3 kW. The focus is on verifying its 360° full-rotation performance, thrust output, rotational control precision, watertightness, and structural reliability.

This project aims to establish an indoor lighting system based on perovskite photovoltaic modules. As a third-generation photovoltaic material, perovskite is leading the energy revolution with its high efficiency, low cost, and excellent performance under low-light conditions, making it particularly suitable for indoor applications. The photovoltaic system incorporates a variety of advanced modules: semi-transparent modules vividly demonstrate the future scenarios of Building Integrated Photovoltaics (BIPV); wide-bandgap modules showcase their potential for providing high voltage in tandem cells; and standard modules serve as performance benchmarks. By using LED lights to simulate indoor light sources, the system displays power generation data in real time and powers small LED devices, achieving a closed-loop interaction between power generation and consumption.

A photothermal seawater desalination device is designed based on a “sodium alginate-calcium chloride” sponge material. By absorbing moisture at the low-temperature end and evaporating at the high-temperature end, it achieves sustainable and passive photothermal seawater desalination. The calcium chloride salt within the material endows it with hygroscopic capacity, while the cross-linking between sodium alginate and calcium chloride immobilizes the salt ions within the material, preventing salt accumulation during the moisture absorption and evaporation process. Due to the osmotic pressure difference during evaporation, the material enables ultra-fast water transport to replenish the water lost through evaporation, thereby supporting rapid evaporation.

This project focuses on ensuring human health and comfort in extreme offshore environments, aiming to construct a “sensing-diagnosis-regulation” system based on the interaction between smart furniture and the environment. Addressing core issues such as motion sickness caused by low-frequency swaying and psychological stress arising from long-term confinement in enclosed spaces, the study utilizes non-contact biosensing technology embedded in furniture terminals to monitor occupants’ physiological indicators (heart rate, body movement) and psychological states in real time. By combining Generative AI algorithms, the system drives smart furniture (such as active stabilization beds and circadian rhythm-regulating lights) to dynamically adjust physical and environmental parameters in real time, providing theoretical support and technical verification for resilient human habitats on future deep-sea platforms and in extreme environments.

By accessing real-time operational data from photovoltaic modules, wind turbines, inverters, battery clusters, and the EMS, this system achieves continuous monitoring of the status of key equipment in power generation, energy storage, and power electronics. It tracks key indicators such as power, voltage, current, temperature, SoC, and SoH. Using time series anomaly detection models, it identifies early signs of failure—such as module degradation, BMS alarms, and battery imbalance—and constructs an equipment health model. This provides the microgrid with intelligent health management capabilities encompassing “monitoring, early warning, fault localization, and recommendation.”