Simulation and Sensitivity Analysis of Energy Consumption in Floating Structures Under Typical and Typhoon Meteorological Conditions

Abstract

Floating structures are increasingly recognized as crucial infrastructure for deep-sea energy exploitation, offshore communities, and maritime hub facilities in recent years. Understanding their energy consumption characteristics under varying meteorological conditions is essential for ensuring operational efficiency and resilience. This study investigates the influencing factors and variation patterns of energy use in floating structures under normal and typhoon environments. Three representative scenarios with different scales and functions were developed based on a bionic hexagon-shaped floating unit, and their respective energy demands were defined. A systematic sensitivity analysis was conducted using DeST with Typical Meteorological Year data and field observations from Super Typhoon Yagi (No. 2411) at Qionghai Station. Results indicate that, according to sensitivity analysis using the dynamic “intraday fluctuation + daily quantile” threshold, dry-bulb temperature and specific humidity are the dominant factors influencing floating-structure energy consumption, contributing 31.1% and 7.8% increases, respectively—significantly higher than other parameters. Under typhoon conditions, total energy consumption rose slightly relative to the TMY baseline, by 0.12%, 0.49%, and 0.95% across the three scenarios, with diurnal variations within ±5%. This study provides a quantitative basis for optimizing energy storage design and enhancing the resilience of floating structures to extreme meteorological events.

1. Introduction

In response to the challenges of global climate change, countries worldwide are accelerating the development of clean and renewable energy sources such as wind and solar power, particularly offshore wind and marine photovoltaics [1,2]. Concurrently, the concept of floating cities and communities has emerged as a potential solution to the challenges posed by rising sea levels. As the primary structural base of these floating energy and urban infrastructure systems, floating platform structures have attracted extensive research interest, leading to continuous innovation in conceptual models and structural frameworks. Several representative projects illustrate this global trend. The “Oceanix City” concept (Figure 1a), proposed by Denmark’s Bjarke Ingels Group (BIG) in collaboration with Oceanix, is the world’s first sustainable floating community designed with a modular architecture to achieve self-sufficiency in energy, water, and food [3]. The Waterbuurt project in Amsterdam (Figure 1b) exemplifies successful floating housing development, demonstrating practical adaptation to future sea level rise in low-lying countries [4]. Denmark’s “Energy Island” project (Figure 1c) aims to construct the world’s first offshore wind hub in the North Sea, with an expected capacity of 3 GW and integration of energy storage and hydrogen production technologies. The Maldives Floating City project (Figure 1d), designed by Dutch architecture firm Waterstudio, adopts a hexagonal maze-like layout that accommodates residences, schools, hospitals, and commercial functions within a self-sufficient maritime community [5]. Similarly, Japan’s “Green Float” by Shimizu Corporation (Figure 1e) [6] and “Dogen City” by N-Ark (Figure 1f) [7] represent forward-thinking vision of floating urban ecosystems. The Dogen City project aims to establish a floating city at sea by 2030, capable of accommodating 40,000 residents. Collectively, these projects provide innovative solutions for climate adaptation and sustainable marine resources.

Accurate assessment of energy demand constitutes a fundamental design requirement for floating cities and communities, as it directly influences the optimization of platform-scale energy storage and supply configurations. To investigate the impact of diverse meteorological conditions on building energy consumption, numerous studies have employed dynamic simulation tools such as EnergyPlus, DeST, and TRNSYS, as well as data-driven approaches. For instance, in order to explore the relationship between building energy consumption and climatic variability Li et al. [8] used the TRNSYS software to simulate the heating and cooling energy consumption of different types of buildings in Tianjin from 1981 to 2010, and analyzed the impact of climate on extreme energy consumption. They found that the extreme energy consumption of commercial and residential buildings was affected by multiple meteorological parameters, among which dry bulb temperature had the greatest influence, while extreme cooling energy consumption was mainly affected by wet bulb temperature. Li et al. [9] explored the impact of stochastic temperature shocks on household electricity consumption and air-conditioner penetration rates, revealing significant regional and seasonal heterogeneity in households’ energy behavior. Roka et al. [10] reviewed sensitivity analysis methods applied in building energy simulations, emphasizing the importance of temperature set-points and meteorological parameters as primary determinants of building energy variation.

Recent research has also focused on improving the representativeness and reliability of meteorological datasets, particularly TMY data. Tian and de Wilde [11] conducted a probabilistic analysis based on UKCP09 projections, demonstrating that uncertainties in climate datasets can substantially affect heating and cooling demand as well as carbon emissions, thereby underscoring the importance of dataset representativeness in long-term energy predictions. Herrera et al. [12] comprehensively reviewed the development of current and future weather datasets for building performance simulation, emphasizing that the selection and generation methods of TMY/TRY and morphed climate files critically influence the accuracy of heating and cooling load predictions. Li et al. [13] proposed a clustering and principal component analysis (PCA) framework to optimize TMY parameter selection, demonstrating improved simulation accuracy in Beijing. Kilanko et al. [14] developed TMY datasets for six locations in Nigeria using the Sandia statistical method and showed that dry-bulb temperature, humidity, and solar radiation are the dominant factors affecting regional energy performance, confirming the effectiveness of accurately generated TMY data for building energy simulations. Hong et al. [15] analyzed long-term observational data from San Francisco and found that spatial variations in urban microclimate can result in multi-fold differences in building energy consumption, highlighting the necessity of incorporating microclimatic heterogeneity into urban-scale energy simulations. Elnabawi and Hamza [16] developed an urban-specific weather dataset (USWD) generation framework coupling ENVI-met 4 microclimate simulations with IES-VE energy modeling, demonstrating that localized meteorological inputs can reduce building energy prediction errors by over 50% compared with standard TMY files. Bhandari et al. [17] evaluated measured and vendor-provided weather datasets for EnergyPlus simulations and found that differences among sources can cause up to ±7% variation in annual and ±40% variation in monthly energy loads, emphasizing the need for accurate, site-specific meteorological inputs. Kočí et al. [18] compared multiple weather datasets, including the TRY and recent observed years, and reported that ongoing warming trends lead to approximately 4% lower heating and 4% higher cooling energy demands. Their findings suggest that outdated TRY datasets may no longer reflect actual climatic conditions, underscoring the importance of continuously updating design weather years for energy performance assessments under climate change.

With the increasing frequency and intensity of extreme weather events, scholars have extended building energy studies to extreme climatic conditions. Jiang et al. [19] examined heating energy responses to a severe cold wave across northern China by comparing the heating energy consumption per floor area (HECPA) with heating degree days (HDD), revealing a strong positive correlation between heating demand and cold-spell intensity and demonstrating that HECPA better captures extreme-cold-induced heating surges. Villa et al. [20] integrated 97 Building Energy Models (BEMs) with smart-meter data from 242 campus buildings across western United States to evaluate building performance during heat waves. Their findings indicated that intensified heat waves can raise cooling energy and peak demand by 20–40%, with pronounced inter-building and climatic variability. Zou et al. [21] further coupled the Local Climate Zone (LCZ) classification and the Urban Weather Generator (UWG) model to quantify the combined effects of heat waves and urban heat islands (UHI) in a hot-humid city, showing that ignoring UHI effects leads to 14–22% underestimation of cooling energy and up to 54% at night. Jin et al. [22] examined vernacular buildings in the extremely hot and arid climate of Turpan and proposed targeted optimization strategies for energy efficiency. Tan et al. [23] investigated island buildings in the South China Sea under extremely hot–humid conditions, identifying window-to-wall ratio (WWR), solar heat gain coefficient (SHGC), and shading configuration as the dominant thermal design factors affecting cooling load.

Several recent studies have begun to incorporate typhoon conditions into building performance and energy-use assessments. Pantua et al. [24] developed an integrated fluid–structure interaction (FSI) and building energy simulation (BES) framework to evaluate the wind-induced response and photovoltaic performance of low-rise buildings subjected to typhoon-strength winds. Li et al. [25] analyzed household electricity demand and grid outage records during multiple severe typhoon events in Japan, revealing strong short-term sensitivity of electrified buildings to typhoon-driven outages. At the urban scale, recent microclimate–building coupling models have enhanced the physical fidelity of energy simulations under extreme meteorological conditions. Li et al. [26] introduced a radiative-transfer-aware microclimate framework capturing air absorption and emission effects, demonstrating that high humidity and radiative interactions—commonly intensified during typhoon passages—can markedly increase cooling demand. Additionally, newly released datasets from southern China with explicit tropical-cyclone annotations reveal substantial reshaping of residential load profiles under strong typhoon events, underscoring the meteorological sensitivity of near-coastal energy systems.

In summary, most existing studies have focused on the energy-consumption characteristics of land-based buildings under typical meteorological years or specific extreme climatic conditions—such as extremely hot–arid, extremely cold, and extremely hot–humid environments—and have proposed corresponding optimization strategies for improving building energy performance under these scenarios. Although recent work has begun to consider typhoon impacts on land-based building, these efforts remain limited in scope and are largely confined to terrestrial contexts. Consequently, the energy-consumption characteristics of marine floating structures under extreme meteorological events such as typhoons remain largely unexplored, even though such structures operate in highly dynamic marine environments and must maintain short-term self-sufficiency and operational resilience during severe weather.

To address this research gap, this study employs a bionic hexagon-shaped floating unit (HS-FU) as the basic structural module to construct three marine floating scenarios with different scales and functional purposes. The DeST (Designer’s Simulation Toolkit, DeST-C) energy consumption simulation software is used to systematically analyze the sensitivity and influence mechanisms of meteorological factors on energy consumption under different design schemes. Furthermore, simulations based on TMY data and field observations from Super Typhoon Yagi (No. 2411) are compared to evaluate the diurnal and monthly variations in energy consumption across different floating scenarios. The findings of this study are expected to provide valuable insights for energy storage optimization and enhancing the energy resilience of floating structures during extreme weather events.

2. Floating Structure Scenarios and Energy Parameter Settings

2.1. Introduction of Bionic Hexagon-Shaped Floating Unit (HS-FU)

The floating structural unit adopted in this study is a bionic hexagon-shaped floating unit (HS-FU), developed under the Shenzhen High-Level Talent Team Project “Novel Bionic and Sustainable Marine Floating Structures”. The geometry configuration of the HS-FU is shown in Figure 2, and its principal dimensional parameters are summarized in

Table 1. Dimensions of the hexagon-shaped floating unit.

Partition Design	Dimension	Value (m)
Living area	H3	4.553
	L3	12.85
Mechanical area	H2	3.554
	L2	8.744
Water tank	H1	12.682
Heave plate	L1	10.744

. The HS-FU comprises three vertical functional layers: a living area, a mechanical area, and four water tank [27,28]. The HS-FU is primarily fabricated by compacted cast concrete, which provides higher strength and durability compared with conventional concrete while reducing cement consumption. This design achieves notable low-carbon and environmental benefits [29].

2.2. Configuration of Application Scenarios

The HS-FU has good aesthetics and functional scalability, allowing flexible assembly to satisfy different functional and spatial requirements [30]. Based on the functional configuration of the living area, three application scenarios of different scales were developed for this study:

(1)

Scenario 1: Functional Floating Platform

This scenario consists of a single HS-FU module with a total floor area of 439 m², as shown in Figure 3a. It is designed to serve specific functional purposes and can accommodate up to 10 occupants. The internal space is divided into four zones: seawater desalination, waste treatment, residential, and other auxiliary areas.

(2)

Scenario 2: Floating Community

This scenario consists of nine interconnected HS-FU modules assembled as shown in Figure 3b, with a total floor area of 3951 m² and capacity for 90 occupants. This configuration represents a floating community with an expanded functional layout, encompassing five primary zones: residential area, shopping center, leisure and recreation area, healthcare and medical center, and public area. The seawater desalination equipment, energy management system room, and waste treatment system room are integrated within the mechanical area of the HS-FU, ensuring centralized utility management.

(3)

Scenario 3: Floating Hotel

This scenario consists of eleven interconnected HS-FU modules, as shown in Figure 3c, forming a floating hotel with a total floor area of 4829 m² and a capacity for 120 guests. According to the Classification and Accreditation for Star-Rated Tourist Hotels (GB/T 14308-2023) [31], the floating hotel comprises multiple functional zones, including the lobby, standard suites, view suites, laundry room, kitchen, restaurant, lounge bar, storage room, gym, indoor swimming pool, supermarket, conference room, recreation room, public areas (corridors), as well as dedicated rooms for seawater desalination equipment, energy management systems, and waste treatment systems. These zones collectively form a comprehensive and self-sufficient maritime living system, integrating accommodation, services, and energy infrastructure into a single floating platform.

2.3. Energy Equipment Distribution and Control Strategy

2.3.1. Occupants, Lighting, and Equipment Control Strategy

Energy consumption simulations in this study were conducted using the DeST software package, which incorporates an integrated thermal disturbance database. The operational schedules for occupancy, lighting, and internal equipment were defined uniformly based on this database. For functional zones already included in the database (such as residential areas, shopping centers, leisure and recreation areas, healthcare and medical centers, and public activity areas), the default thermal disturbance parameters and operation schedules were directly adopted. For the special functional zones not covered by the database (such as public area, seawater desalination, and waste treatment areas), custom operational schedules were defined based on typical working hours and process characteristics. The corresponding schedules of thermal disturbance parameters are summarized in

Table 2. Indoor thermal disturbance parameter schedule.

Sites	Public Area		Seawater Desalination System Zone		Waste Treatment Zone
Occupant thermal disturbance	8:00 22:00	9:00–10:00 13:00–14:00 17:00–18:00	8:00–20:00	21:00–7:00	8:00–20:00	21:00–7:00
	0.3	0.5
	11:00–12:00	15:00–16:00 19:00–21:00	1	0.2	1	0.2
	0.2	0.7
Lighting thermal disturbance	1:00–7:00 22:00–24:00	8:00–21:00	8:00–20:00		8:00–20:00
	0.3	1	1		1
Appliances thermal disturbance	1:00–7:00 22:00–24:00	8:00–21:00	8:00–20:00		8:00–20:00
	0.3	1	1		1

.

2.3.2. Air Conditioning Control Strategy

The air conditioning (AC) system represents the dominant energy-consuming subsystem of the floating island structure, and this has a decisive impact on total energy consumption. Considering the different usage characteristics of each functional zone, two differentiated AC control strategies were designed. The first strategy is a demand-responsive control strategy based on the operating hours of each functional zone. It applies to intermittently occupied areas such as shopping centers, leisure and recreation areas, healthcare and medical centers, supermarkets, multifunctional activity rooms, conference rooms, and gyms. In these areas, the AC equipment operates only during active service hours and is switched off during non-operational periods, thereby achieving basic demand-responsive energy savings. The second strategy is an all-day temperature-control strategy, designed for zones requiring continuous temperature and humidity stability, such as hotel guest rooms, healthcare and medical centers, and storage rooms. In these areas, the AC system operates continuously throughout the day to maintain constant indoor temperature and humidity. In the simulation model, the cooling period was defined as April 1 to October 31, with a setpoint temperature of 26 °C, while the heating period spanned November 1 to March 31 of the following year, with a setpoint temperature of 22 °C.

2.3.3. Auxiliary Equipment Control Strategy

Considering the special demand for closed-loop resource utilization in the marine environment, the floating structures were equipped with auxiliary energy systems for domestic sewage treatment, waste treatment, and seawater desalination. The operation of these auxiliary systems was modeled under the “other equipment” category within the DeST simulation framework, where the daily per capita assumptions were as follows: each person generates 150 L of domestic wastewater, 1 kg of solid waste, and consumes 250 L of freshwater per day. Based on the number of occupants in the three application scenarios, the capacities and operation schedules of the auxiliary systems were determined accordingly. The detailed equipment configurations and operational parameters are listed in

Table 3. Parameter settings of auxiliary equipment.

Scenario	Population	Equipment Type	Rated Power (kW)	Quantity	Operating Time (h)
Scenario 1	10	WCBx-10B Shipborne Wastewater Treatment Unit, Wuhan Zhongzhou Env. Protection Equipment Co., Ltd., Wuhan, China	5.6	1	12
		E200 Shipborne Garbage Processor, Shanghai DanFengMarine Equipment Co., Ltd., Shanghai, China	0.45	1	8
		RDA-FSHB3 Compact Seawater Desalination Unit, Dongguan RDA Environment Protection Technology Co., Ltd., Dongguan, China	1.5	1	12
Scenario 2	90	WCBx-15B Shipborne Wastewater Treatment Unit, Wuhan Zhongzhou Env. Protection Equipment Co., Ltd., Wuhan, China	5.6	1	12
		E200 Shipborne Garbage Processor, Shanghai DanFengMarine Equipment Co., Ltd., Shanghai, China	0.45	1	8
		RDA-FSHB20 Compact Seawater Desalination Unit, Dongguan RDA Environment Protection Technology Co., Ltd., Dongguan, China	10	2	12
Scenario 3	120	WCBx-50B Shipborne Wastewater Treatment Unit, Wuhan Zhongzhou Env. Protection Equipment Co., Ltd., Wuhan, China	6.3	2	12
		JZXFJ-15 Shipborne Garbage Processor, Shanghai DanFengMarine Equipment Co., Ltd., Shanghai, China	1.5	1	8
		RDA-FSHB20 Compact Seawater Desalination Unit, Dongguan RDA Environment Protection Technology Co., Ltd., Dongguan, China	10	2	12

.

To illustrate the temporal distribution of energy demand, scenario 2 (Floating Community) were selected as a representative case. The 24 h energy use patterns of five functional zones: residential area, shopping center, leisure and recreation area, healthcare and medical center, and public activity area, are shown in Figure 4. It should be noted that the occupant’s movement among different functional spaces was not considered in this study. To ensure conservative estimation, full occupancy was assumed for all functional zones throughout the day.

3. Construction of Energy Consumption Simulation Models Under Different Scenarios

This study employed the DeST (Designer’s Simulation Toolkit) software to establish and analyze the energy consumption models of floating building structures under various operational scenarios. DeST integrates functions for building thermal environment simulation and HVAC system optimization, enabling detailed analysis of both dynamic thermal processes and system-level energy performance [32]. The software is widely applied for building thermal performance analysis, dynamic annual load calculation, and economic evaluation of terminal HVAC systems [33].

The tropical city of Qionghai, located in Hainan Province, China, was set as the construction site for the floating structures. Using the typical meteorological year (TMY) data, the DeST software was used to simulate the energy consumption of the three floating structure scenarios. All simulated models were designed as single-story buildings with a floor-to-ceiling height of 4 m and a window-to-wall ratio of 0.3. The exterior walls and roof of each HS-FU consist of 0.4 m thick compressed cast-concrete panels, while the interior walls were modeled as 0.24 m thick panels of the same concrete material. The modeling process is shown in Figure 5, and the basic model information is listed in

Table 4. Basic parameters of the simulated buildings.

Category	Parameter
Location	Qionghai, Hainan, China
Orientation	South
Height (m)	4
Number of floors	1
Window-to-wall ratio	0.3
U-value of external windows (W/(m2·K))	3.0
U-value of external walls and roof (W/(m2·K))	2.5
U-value of internal walls (W/(m2·K))	3.85

.

Given that offshore floating structures mainly rely on renewable energy sources such as wind, solar, and wave energy, all forms of energy consumption in the simulation were converted into electricity equivalents. Accordingly, the total energy use in all three scenarios included air-conditioning electricity, lighting electricity, equipment electricity, water supply and drainage electricity, and other auxiliary electricity (e.g., marine waste processors, small seawater desalination units, and wastewater treatment devices).

4. Correlation and Sensitivity Analysis of Meteorological Parameters on Energy Consumption

4.1. Correlation Analysis Among Meteorological Parameters

To facilitate the identification and analysis of factors affecting the energy consumption of floating structures, a correlation analysis was conducted using the Typical Meteorological Year (TMY) data for the Qionghai station in Hainan province, extracted from the DeST meteorological database. The Pearson correlation coefficient method was used to calculate the correlations among nine meteorological parameters: dry-bulb temperature, humidity ratio, total horizontal radiation, diffuse horizontal radiation, ground temperature, effective sky temperature, wind speed, wind direction, and atmospheric pressure. The analysis results are shown in Figure 6. A Pearson correlation coefficient with an absolute value close to 1 indicates a strong linear relationship, whereas values close to 0 imply weak or negligible correlation. The abbreviations and definitions of these meteorological parameters used in DeST are listed in

Table 5. List of meteorological parameter abbreviations.

Abbreviation	Definition	Unit
T	Dry-bulb temperature	°C
D	Humidity ratio	g/kg.dra
RT	Total horizontal radiation	W/m2
RR	Diffuse horizontal radiation	W/m2
T_GROUND	Ground temperature	°C
T_SKY	Effective sky temperature	K
WS	Wind speed	m/s
WD	Wind direction	-
B	Atmospheric pressure	hPa

.

As shown in Figure 6, strong positive correlations are observed between dry-bulb temperature (T) and several other parameters, including humidity ratio, ground temperature, and effective sky temperature, with correlation coefficients of 0.85, 0.86 and 0.77, respectively. Conversely, atmospheric pressure (B) exhibits a strong negative correlation with dry-bulb temperature (r = −0.77). The humidity ratio also shows a pronounced negative correlation with atmospheric pressure (r = −0.85), indicating the inverse relationship between moisture content and air density under tropical maritime conditions. Among all parameter pairs, ground temperature and effective sky temperature demonstrate the highest correlation (r = 0.92), suggesting a strong coupling between terrestrial and sky radiative exchange processes. In contrast, wind speed (WS) and wind direction (WD) show relatively weak correlations with other meteorological parameters (all |r| < 0.5), implying their statistical independence in the local climatic system. These findings are consistent with the results of Li et al. [34], who also identified weak interdependencies between wind-related variables and thermal-radiative parameters in building energy consumption studies.

4.2. Sensitivity Analysis of Meteorological Variables Affecting Energy Consumption

Given that meteorological observations during typhoon periods are often incomplete or unavailable, a sensitivity analysis was conducted to ensure the robustness and scientific reliability of building energy consumption simulations under extreme typhoon conditions. Based on the Typical Meteorological Year (TMY) data, seven meteorological parameters—dry-bulb temperature (T), humidity ratio (D), total horizontal radiation (RT), diffuse horizontal radiation (RR), ground temperature (T_GROUND), effective sky temperature (T_SKY), and atmospheric pressure (B)—were selected for the analysis. Wind speed and wind direction were excluded in this stage, as they are insensitive to energy consumption.

To capture the influence of meteorological variables on energy consumption across different intensity levels, while retaining their diurnal and seasonal variability, a dynamic threshold construction method was proposed. The method is based on the 8760 h data of the TMY and combines “intraday fluctuation value + daily mean percentile value”. This method preserves both the diurnal and seasonal variation characteristics of meteorological parameters while allowing the assessment of their effects under different intensity levels. Taking an arbitrary meteorological variable $x$ as an example, the processing steps are described as follows. First, the daily mean value ${\overline{x}}_d$ of the TMY data over the whole year is calculated as:

$${\overline{x}}_d={\displaystyle \frac{1}{24}}{\displaystyle \sum _{h=1}^{24}{x}_{d,h}}$$

(1)

where $x_{d,h}$ is the hourly value on day $d \left(d=1, \dots , 365\right)$ and hour $h \left(h=1, \dots , 24\right)$, and ${\overline{x}}_d$ is the daily mean on day $d$.

The 10th, 30th, 50th, 70th and 90th percentiles of the daily mean series ${\left\{{\overline{x}}_d\right\}}_{d=1}^{365}$ are then extracted as static baselines. Let $p_k\left(k=1, \dots , 5\right)$ denote the five quantile levels. The static baseline corresponding to level $p_k$ is defined as:

$$x_k^{base}=Q_{p_k}\left({\left\{{\overline{x}}_d\right\}}_{d=1}^{365}\right)$$

(2)

where $Q_{p_k}\left(·\right)$ is the empirical quantile function.

Then, based on the annual simulation results in Section 2, July, the month with the highest total energy demand, was selected for further analysis. The hourly meteorological data of July were processed by removing the daily mean to obtain zero-mean diurnal fluctuation sequences. Denoting by $D_{Jul}$ the set of days in July, the zero-mean intraday fluctuation series can be written as:

$$\mathrm{\Delta }{x}_{d,h}=x_{d,h}-{\overline{x}}_d$$

(3)

Finally, the zero-mean sequences of July were superimposed onto the five percentile-based baselines, forming five dynamic meteorological input sequences corresponding to different intensity levels, denoted as L1–L5. The dynamic meteorological series at the $k$-th intensity level (L1–L5) are expressed as:

$$x_k\left(d,h\right)=x_k^{base}+\mathrm{\Delta }{x}_{d,h}$$

(4)

The above processing steps are applied independently to all seven meteorological variables, namely dry-bulb temperature, humidity ratio, global horizontal radiation, diffuse horizontal radiation, ground temperature, effective sky temperature, and atmospheric pressure, thereby yielding five sets of dynamic meteorological series (L1–L5). Using the Scenario 1 (Functional Floating Platform) model, the five dynamic meteorological sequences (L1–L5) were sequentially imported into the built-in meteorological database of DeST for the Qionghai region to quantify energy consumption responses under varying intensity levels.

The results of the sensitivity analysis for each parameter are shown in Figure 7. As shown in Figure 7a, dry-bulb temperature (T) and humidity ratio (D) exert the most significant influences on total energy consumption. When the dry-bulb temperature increased from level L1 to L5, the total energy consumption rose from 10,886.94 kWh to 14,272.66 kWh, representing an increase of approximately 31%. Similarly, the humidity ratio produces an increase of 7.79% in total energy use across the same range of intensity levels. In contrast, radiation-related variables exhibited moderate effects. Variations in total horizontal radiation led to total energy consumption variations ranging from 13,417 kWh to 13,485.95 kWh, while diffuse horizontal radiation induced slightly greater fluctuations but remained far less impactful than dry-bulb temperature and humidity ratio. Disturbances in ground temperature and effective sky temperature produced negligible changes, indicating limited influence on energy performance. In summary, the analysis confirms that dry-bulb temperature and humidity ratio are the dominant meteorological drivers of floating structure energy consumption under tropical maritime conditions. Consequently, for subsequent simulations of building energy consumption under extreme typhoon conditions, it is essential to incorporate measured data of dry-bulb temperature, humidity ratio, wind speed, and wind direction to capture the full range of dynamic meteorological influences.

5. Energy Consumption Simulation Under TMY and Typhoon Conditions

5.1. Energy Consumption Simulation of Different Scenarios Under TMY Data

Based on the three floating structure energy consumption models established in Section 3, the Typical Meteorological Year (TMY) data from the Qionghai meteorological station (Station No. 59855) were used as the input parameters for annual energy consumption simulations. The analysis was conducted to evaluate the energy performance of floating structures of varying functional types and scales under typical climatic conditions.

The monthly energy consumption variations in different equipment throughout the year are shown in Figure 8. All floating structures exhibit obvious seasonal variations in energy consumption. The three floating structure scenarios show similar seasonal variation patterns: higher energy consumption during summer and winter and lower consumption in spring and autumn. This trend is primarily attributed to increased air-conditioning cooling demand during hot summer months and elevated heating demand during winter. In contrast, the mild transitional seasons lead to reduced HVAC operation and lower total energy use. Across the three floating structure scenarios, July corresponds to the annual peak of electricity use, with total electricity consumption of 14.2 MWh, 117.4 MWh, and 153.1 MWh for Scenarios 1–3, respectively. In contrast, February represents the annual minimum, with corresponding consumption of 8.3 MWh, 54.6 MWh, and 72.8 MWh, respectively.

To further examine daily operational characteristics in typical seasons, Figure 9 presents the hourly energy consumption for a typical summer day (21 June) and a typical winter day (21 December). The main differences among the three scenarios arise in the HVAC electricity consumption. During summer, the high ambient temperature significantly increases the cooling load and corresponding power use, whereas in winter, cooling demand decreases due to lower outdoor temperatures. The water supply and drainage system exhibits slightly higher electricity consumption in winter, likely due to the increased demand for domestic hot water. In contrast, lighting, equipment, and other auxiliary systems show minor seasonal or diurnal variation, primarily affected by functional zoning, control strategies and occupancy schedules of the floating structures. These end uses thus remain relatively stable electricity demand throughout the year.

To evaluate the energy utilization efficiency of floating structures at different scales, two performance indicators were employed: (1) Energy Use Intensity (EUI), expressed as annual energy consumption per unit floor area (kWh/m²·year); (2) Per Capita Electricity Use, defined as annual electricity use per occupant (kWh/person year). The results are summarized in

Table 6. Comparison of floor area, annual energy consumption, and per capita energy consumption across different scenarios.

Scenario	Floor Area (m2)	Occupancy	Annual Energy Consumption (×104 kWh)	Energy Use Intensity (kWh/m2·Year)	Per Capita Electricity Use (kWh/Person·Year)
Scenario 1	439	10	13.5	307.52	13,500.00
Scenario 2	3951	90	102.66	259.84	11,406.67
Scenario 3	4829	120	133.65	276.76	11,137.50

. As the scale of the floating structure increased from Scenario 1 (439 m², 10 occupants) to Scenario 3 (4829 m², 120 occupants), the annual EUI decreased from 307.5 kWh/m²·year to 276.8 kWh/m²·year, an approximate reduction of 10%. Similarly, the annual per capita electricity use decreased from 13,500 kWh/person·year to 11,138 kWh/person·year, representing a reduction of about 17.5%. These results indicate that larger-scale floating structures exhibit improved energy utilization efficiency, attributable to the scale effect and shared resource utilization. It should be noted, however, that the three scenarios differ in functional orientation, occupancy density, and spatial layout. Scenario 1 represents a single functional floating platform, while Scenario 2 corresponds to a floating residential community, and Scenario 3 is a floating hotel integrating more complex residential, service, and public space functions. Therefore, the observed efficiency improvement may not solely result from scale effects, but also from functional integration, occupant utilization rate, and equipment and resource sharing among different spatial spaces.

5.2. Energy Consumption Simulation Under Extreme Typhoon Conditions

To investigate the characteristics of building energy consumption under extreme typhoon conditions, field-measured meteorological data from the Qionghai meteorological station were utilized. The data were recorded during Super Typhoon Yagi (No. 2411), from 00:00 on 5 September to 23:00 on 7 September 2024. Five key meteorological parameters—dry-bulb temperature, humidity ratio, wind speed, wind direction, and atmospheric pressure—were selected as inputs for the simulation. These measured variables replaced the corresponding TMY dataset in the three floating structure models established in Section 3, allowing a direct assessment of building energy response to actual typhoon conditions.

Super Typhoon Yagi (No. 2411) made landfall near Wengtian Town, Wenchang City, Hainan Province, at around 16:20 on 6 September 2024, marking one of the most severe typhoons to affect Hainan province in recent years. As the Qionghai station is located west of the typhoon landfall point, it is situated primarily in the outer part of the typhoon’s 10-grade wind circle rather than in the core region near the eyewall. Consequently, owing to its geographic position and the typhoon track, the extreme meteorological conditions observed at Qionghai are weaker than those at the typhoon center. Nevertheless, these observations still provide a representative characterization of the meteorological conditions affecting building energy consumption during the typhoon event. Consequently, the dataset captures realistic typhoon-induced environmental fluctuations relevant to offshore and nearshore building performance. Figure 10 shows the temporal variations in major meteorological parameters during the 72 h landfall period. For simulation purposes, the corresponding period in the TMY dataset—from 00:00 on 5 September (the 5928th hour of the year) to 23:00 on 7 September (the 5999th hour)—was replaced with these measured typhoon data to construct a 72 h simulation window. The modified dataset was then used to perform dynamic building energy simulations for the three floating structure scenarios.

Figure 11 shows the component-based energy consumption of the three floating structure scenarios during the typhoon period. Across all scenarios, air-conditioning electricity consumption remained the dominant contributor to total energy use, particularly during the period of typhoon landfall. The electricity consumption associated with lighting, equipment, water supply and drainage, and other auxiliary systems remained stable throughout the three-day period. This stability reflects the fact that these subsystems are primarily governed by occupancy schedules, control strategies, and operational routines, which were assumed to remain unchanged during the typhoon event. In contrast, the air-conditioning electricity consumption showed slight fluctuations over time, primarily driven by variations in outdoor temperature, humidity, and wind conditions. These fluctuations affected the cooling and heating loads of the HVAC system, requiring continuous operational adjustment to maintain indoor thermal comfort. The energy performance differences among the three floating structure types are summarized in

Table 7. Energy consumption variations across different scenarios during typhoon events.

Scenario	Energy Consumption During Typhoon Period (kWh)		Difference (kWh)	Growth Rate (%)
	Typical Meteorological Year	Super Typhoon Yagi
Scenario 1	1336.80	1338.36	1.56	0.12
Scenario 2	11,091.35	11,145.79	54.44	0.49
Scenario 3	14,804.65	14,945.86	141.21	0.95

. Compared with simulations under TMY conditions, the total energy consumption during Super Typhoon Yagi (No. 2411) increased slightly. by 0.12%, 0.49%, and 0.95% for Scenarios 1–3, respectively.

Although the overall changes are relatively small, they reveal that short-term typhoon-induced meteorological fluctuations primarily influence HVAC system loads, while the energy consumption of lighting, equipment, and auxiliary systems remains nearly invariant. The underlying causes of these variations, including the dynamic coupling between meteorological disturbances and building system responses, are further analyzed in Section 6.

6. Comparison of Daily Average Energy Consumption During Typhoon Landfall

6.1. Comparison of Meteorological Parameters During Typhoon Landfall

To enable a comparative analysis of the meteorological drivers influencing the energy consumption of floating structures under typical meteorological years and extreme typhoon conditions, a detailed examination of key atmospheric variables was conducted. Figure 10 presents the temporal evolution of selected key meteorological parameters during the landfall of Super Typhoon Capricorn (2411) and the corresponding period in the Typical Meteorological Year (TMY). Because the TMY dataset provides wind speed and wind direction values at only four time intervals per day, the comparison for these two parameters focuses exclusively on the measured data during the typhoon period, whereas the other parameters are compared directly between the TMY and measured datasets.

As shown in Figure 10a, the TMY dry-bulb temperature exhibits a distinct diurnal cycle, with values ranging between 24 °C and 34 °C, reflecting the typical daytime heating and nighttime cooling patterns of the tropical climate in Qionghai. In contrast, during the passage of Super Typhoon Yagi, the dry-bulb temperature was lower overall and showed smaller fluctuations, displaying a non-periodic variation pattern. The variation in the humidity ratio, illustrated in Figure 10b, further highlights the contrast between the two datasets. During the typhoon event, the humidity ratio remained consistently higher—mostly between 20 g/kg·dra and 23 g/kg·dra—compared with the TMY range of 18 g/kg·dra to 21 g/kg·dra. This elevated moisture content reflects the saturated marine air masses and sustained precipitation typically associated with typhoon systems. Atmospheric pressure (Figure 10c) serves as a key diagnostic indicator of typhoon intensity. The simulation data reveal a pronounced pressure drop as the typhoon center approached, reaching a minimum of 975.3 hPa at the 5969th hour, substantially lower than the TMY value of 998.8 hPa for the same period. Figure 10d shows the temporal variation in the hourly average wind speed at a height of 10 m above ground level during the typhoon. The wind speed exhibited the most distinct deviation between the two conditions, increased sharply as the typhoon center approached, reaching a maximum of 13.3 m/s near the 5967th hour. Figure 10e presents the wind speed and direction rose chart during the impact of Super Typhoon Yagi, clearly showing marked shifts in wind direction and concentration of high wind speeds, consistent with the rotational flow characteristics of a typhoon. Figure 12 shows the track of Super Typhoon Yagi (No. 2411) from 00:00 on 5 September to 23:00 on 7 September 2024. The figure also marks the location of the Qionghai meteorological station (No. 59855, indicated by a red triangle) and the landfall point of the typhoon (indicated by a yellow pentagram). As shown in Figure 12, the straight-line distance between the Qionghai station and the landfall point is approximately 54.9 km, placing it roughly at the outer edge of the typhoon’s Category-10 wind circle, where wind intensities remained moderate compared to the core of the system. Since Super Typhoon Yagi weakened rapidly after making landfall, the observed hourly mean wind speed at Qionghai station was lower than at the typhoon center, but still sufficient to reflect the transitional meteorological impact.

In summary, the meteorological conditions during the typhoon landfall were characterized by lower air temperatures, slightly higher humidity, and significantly stronger and more variable winds relative to the TMY. These distinct environmental conditions form the physical basis for the subsequent comparison and analysis of floating structure energy consumption under typical and extreme meteorological scenarios.

6.2. Comparative Analysis of Energy Consumption Simulation Results

To investigate the impact of changing meteorological conditions during the typhoon period on the energy consumption of floating buildings, this section compares the energy consumption characteristics of three application scenarios under the TMY conditions and during Super Typhoon Yagi (No. 2411). The comparison is based on the meteorological data presented in Section 6.1. The analysis is conducted along two key dimensions: (1) hourly comparison of building energy consumption (Figure 13); (2) daytime versus nighttime energy consumption during the typhoon’s landfall period (Figure 14).

Figure 13 compares hourly building energy consumption of those three scenarios under both the TMY and typhoon conditions. To ensure the comparability of simulation results across different meteorological conditions, identical energy equipment control strategies were applied in all three scenarios. As shown in Figure 13, the energy consumption patterns for all three scenarios exhibit a clear diurnal fluctuation under both TMY and typhoon conditions. This periodic behavior aligns with the control strategies outlined in Section 2.3, reflecting typical load demands in accordance with operational schedules. Overall, a comparison of the hourly energy consumption between the TMY and typhoon conditions reveals slight variations in both peak and valley values, with more noticeable differences during the daytime. These variations can be attributed to the dynamic meteorological changes associated with the typhoon. Notably, the decrease in dry-bulb temperature reduced the indoor–outdoor temperature difference and the solar heat absorbed by the building envelope, leading to a lower sensible cooling load. However, the increase in wind speed enhanced the convective heat transfer on the external surface of the envelope and increased air infiltration. As a result, the air-conditioning system slightly adjusts its air supply volume and operation rhythm to maintain the desired indoor temperature and humidity, leading to a mild rise in nighttime energy consumption. Overall, while these meteorological effects led to minor changes in the peak-valley difference in the energy consumption curves, the overall fluctuation amplitude remained relatively small.

Figure 14 compares the daytime and nighttime energy consumption during the typhoon period for the three scenarios. The daytime period is defined as 8:00–18:00, with the remaining hours constituting the nighttime period. The percentage values in the figure represent the relative change in energy consumption under typhoon conditions compared with the TMY values. On 5 September, the daytime energy consumption change rates were 2.2%, 3.3%, and 4.2%, respectively. Nighttime values showed 1.9%, 2.6% and 3.8%, respectively. On 6 September, the daytime change rates were −3.3%, −3.4% and −3.1%, while the nighttime changes were −0.7%, −0.8%, and −0.2%. On September 7, the daytime changes were −1.0%, −0.7%, and −1.0%, with nighttime changes of 2.4%, 2.0%, and 2.9%, respectively. In total, the differences in daytime and nighttime energy consumption between the TMY and typhoon scenarios remained within ±5%, indicating relatively limited fluctuation overall. These results are consistent with the hourly analysis, suggesting that short-term variations in energy consumption are mainly associated with changes in air-conditioning demand. In contrast, the energy consumption of lighting and equipment remained stable across different meteorological conditions due to the constraints imposed by occupancy schedules and control strategies.

7. Conclusions

This study employed a bionic hexagon-shaped floating unit (HS-FU) as the fundamental structural module to develop three floating structure scenarios of varying scales and functional configurations. The DeST (Designer’s Simulation Toolkit) software was used to simulate and analyze their energy consumption characteristics. Comparative evaluations were conducted under both Typical Meteorological Year (TMY) conditions and during the Super Typhoon Yagi (No. 2411) event to examine how different meteorological environments influence energy use patterns. The main conclusions are summarized as follows:

• The sensitivity analysis based on the “intraday fluctuation value + daily mean percentile value” dynamic threshold method, identified dry-bulb temperature and humidity ratio as the dominant meteorological factors influencing the energy consumption in floating structures. The energy consumption variation rates for these parameters were 31.1% for dry-bulb temperature and 7.8% for humidity ratio, both significantly higher than the variation rates for other meteorological parameters.
• The energy consumption of all three floating structure scenarios exhibited clear seasonal variations. Air-conditioning loads dominated the energy consumption fluctuations during the summer and winter months, followed by water supply and drainage loads. In contrast, the electricity consumption for lighting, equipment, and other systems remained relatively stable, primarily influenced by usage patterns and control strategies. As the scale and functional integration of the floating structures increased, energy consumption per unit area and per capita decreased, highlighting improvements in energy efficiency. This trend was attributed not only to the scale effect but also to functional positioning, occupancy utilization, and the sharing of equipment and resources among different spaces.
• Under the conditions of Super Typhoon Yagi (2411), the energy consumption of the three floating structure scenarios increased slightly compared to the same period under TMY conditions. The total energy consumption for the three scenarios increased by 0.12%, 0.49%, and 0.95%, respectively. Day-to-night energy consumption differences remained within ±5%, with the most noticeable variations occurring in the air-conditioning loads. These results demonstrate that the proposed floating structure design exhibits good climatic adaptability and energy stability. The air-conditioning system serves as the key component in floating structures’ response to extreme weather conditions, highlighting the importance of optimizing its performance in future design.

Based on the above findings, clarifying the functional purpose and usage mode is essential for optimizing the energy systems of floating structures during both design and operation. To achieve energy self-sufficiency, the integration of diversified renewable energy sources is recommended. This could involve installing wind turbines on upper decks, deploying high-efficiency photovoltaic panels on roofs, and incorporating energy storage systems to balance supply and demand. These integrated strategies would not only address peak energy demands but also store surplus energy during off-peak periods, ensuring efficient and environmentally friendly operation. This study provides a theoretical foundation and technical support for energy consumption prediction and climate-adaptive design of floating buildings under extreme weather conditions. Additionally, it also offers valuable insights for the optimization of energy systems, the development of adaptation strategies, and the exploration of sustainable pathways for future offshore floating architecture.