Activating and Enhancing the Energy Flexibility Provided by a Pipe-Embedded Building Envelope: A Review

Abstract

Building thermal mass offers a cost-effective solution to enhance the integration of energy supply and demand in dynamic energy systems. Thermally activated building systems (TABS), incorporating embedded heat tubes, shows strong potential for energy flexibility. However, the significant thermal inertia of TABS also imposes challenges to precise load shift and indoor climate control. This review synthesizes key research on the effective demand-side management of TABS from multiple perspectives. It examines and compares various TABS configurations, including floor, ceiling, and wall systems. Differences in heat transfer performance between heating and cooling result in distinct application preferences for each type. The integration of advanced materials, such as phase change materials (PCM), can further enhance energy flexibility. TABS flexibility is primarily activated through adjustments to indoor operative temperature, with relevant influencing factors and regulatory constraints analyzed and discussed. Key aspects of optimizing building energy flexibility, including simulation methods and control strategies for TABS, are reviewed from both theoretical and practical perspectives. The energy and economic performance of TABS under various control strategies is analyzed in detail. This review provides insights to support the optimal design and operation of TABS within dynamic energy systems and to enhance the energy flexibility of building envelopes.

1. Introduction

Over the past century, buildings have accounted for a gradual increase in global energy consumption and CO₂ emissions. According to the International Energy Agency [1], buildings and the building construction sector account for nearly 30% of the world’s total energy consumption and nearly 15% of direct CO₂ emissions. Moreover, the difficulty of balancing energy supply and demand is becoming increasingly difficult due to the expansion of renewable penetration [2]. Consequently, the actual efficiency of the energy system is greatly hindered due to the large curtailment of renewable energy [3,4]. Therefore, improving demand-side energy flexibility is increasingly vital. It accommodates renewable energy’s intermittency and variability [5] while maintaining the power system’s supply–demand balance.

Building energy flexibility has great potential to resolve the mismatch between the energy supply and demand sides by conducting effective demand response [6] with heat storage potential from the building thermal mass. According to IEA EBC Annex 67 [7], building energy flexibility is defined as “The ability to manage its demand and generation according to local climate conditions, user needs, and energy network requirements”. It is more cost-effective compared to traditional active thermal storage [8] since no additional investment in storage equipment is required.

TABS presents substantial energy flexibility due to the embedded thermal pipes inside the building structure. Along with the heat-carrying medium flowing through the embedded pipes, heating/cooling energy is released to the surrounding envelope layers, thereby activating the energy flexibility of the building thermal mass [9]. The large heat transfer area allows heat to be transferred with a smaller temperature gradient, which is suitable for high-temperature cooling [10] and low-temperature heating [11]. Therefore, low-grade energy resources, such as geothermal or low-temperature waste heat [12,13], can be applied in TABS with higher exergy efficiency. By controlling the temperature or the flowrate of the heat-carrying medium [14], it is possible to dynamically regulate the heat transfer between the indoor and outdoor areas of buildings. Therefore, TABS is also considered as the variable-thermal-resistance building envelope [15].

Previous studies have made significant progress in optimizing the design parameters of TABS (or their combinations [16,17,18]), focusing on factors such as envelope materials [19,20], embedded pipe geometry [21], and pipe location and spacing [22], etc. [23,24]. Several studies have also explored operational strategies like intermittent control and scheduled approaches to enhance energy efficiency by leveraging the system’s thermal inertia [23,25]. However, these efforts often overlook the dynamic interactions between TABS and the energy supply side, particularly in the context of fluctuating renewable energy sources. With high energy flexibility from the thermal mass, TABS holds great potential as a demand-side solution for future complex energy systems. However, a comprehensive review of the available technologies, constraints, and strategies for maximizing energy flexibility remains lacking. This gap underscores the need for a thorough evaluation of TABS, particularly in terms of their capacity to deliver energy flexibility and their integration within fluctuating energy supply networks.

To address the research limitations, the focus of this review is to providing critical insights and useful information for the design and operation of TABS to support its adaptation to an increasingly dynamic and renewable-based energy system in the future. Both exogenous and endogenous factors that affect the effective utilization of TABS energy flexibility are carefully considered. The impacts from the physical properties, such as different TABS configurations and the advanced materials, are discussed from multiple perspectives, including heat transfer coefficients, heat capacity, and suitability for specific application scenarios. Moreover, schemes for activating and harnessing the energy flexibility of TABS are explored in detail, addressing operational constraints, modeling approaches and control strategies. This provides actionable guidance for achieving an appropriate operation of TABS concerning both comfort and efficiency. Lastly, the energy and economic performances of TABS under diverse scenarios are systematically reviewed and quantitatively assessed, with the aim of establishing a clear framework for leveraging the energy flexibility provided by TABS in a range of energy applications.

The structure of this article is organized as follows: A brief description of TABS and the role of the energy-flexible buildings in a dynamic energy system is given in Section 1; Section 2 presents the methodology adopted in this review; Section 3 introduces the critical physical properties of TABS that would affect the energy flexibility potential; Section 4 describes the temperature control requirements and influencing factors for regulating energy flexibility in TABS; Section 5 presents the modeling approaches for TABS; Section 6 presents the approaches to activating and utilizing the energy flexibility of TABS; the energy and economic performance of TABS under different scenarios are presented in Section 7; and last, the conclusions, limitations and recommendations for future research are summarized in Section 8.

2. Literature Survey

Despite the extensive research conducted on pipe-embedded building envelopes and TABS, the terminology used to describe these systems remains inconsistent across different studies. For example, different researchers have used terms such as “pipe-embedded building envelope” or “thermally activated building system” as keywords in their publications, although both essentially refer to the same technological system or research domain. To address this issue, this paper adopts the Sub-keyword Synonym Search (SSS) [26,27] approach for the literature review. Given that each core term often has multiple synonyms or extendable sub-keywords, this study adopts a search strategy within the SSS framework based on the combination of “core keywords + sub-keywords”. Core keywords are used to define the research object, ensuring that the retrieved literature focuses on studies related to “pipe-embedded building envelope” and “TABS”. Specifically, the core keywords include “thermally activated building system”, “TABS”, “pipe-embedded building envelope”, “pipe-embedded wall”, and “embedded pipe system”, along with their relevant synonyms. Sub-keywords further refine the research theme, highlighting studies related to “energy flexibility”, and include terms such as “energy flexibility”, “building thermal mass”, “thermal inertia”, and “load shifting”. This method involves multiple rounds of keyword searches using various synonyms and sub-keywords, aiming to minimize the omission of the relevant literature caused by limited keyword coverage.

The studies were retrieved from the Web of Science database, focusing on studies published between 2011 and 2025 that are related to pipe-embedded building envelopes and TABS. As of June 2025, a total of 179 relevant publications were retrieved (document types limited to articles and review articles). Although all retrieved records contained either core and sub-keywords, a portion of them were found to be irrelevant or weakly related to the objective of this review. Therefore, each abstract was manually screened and evaluated for relevance, resulting in the exclusion of 27 publications due to insufficient alignment with the review objectives. Ultimately, 152 publications were selected for detailed analysis. Although the literature search was conducted within the timeframe of 2011 to 2025 to reflect recent research progress and technological trends, a few key pre-2011 studies were also cited to provide essential theoretical background and early-stage technological insights, thereby enhancing the overall comprehensiveness and historical context of the review.

As illustrated in Figure 1, VOSviewer 1.6.20.0 [28] was employed to conduct a keyword co-occurrence and clustering analysis based on the selected literature published between 2011 and 2025. In the visualization, solid circles of varying sizes and colors represent keywords, where the size indicates the frequency of occurrence in the reviewed articles. The lines connecting the nodes reflect the co-occurrence strength between keywords within the same document, with thicker lines representing stronger associations [29]. As shown by the relatively prominent nodes in Figure 1, publications on TABS are commonly accompanied by discussions of energy efficiency and thermal comfort.

3. Impact of TABS Physical Properties on Energy Flexibility

The physical parameters determine the energy flexibility potential of TABS and should be carefully considered at the design stage. This section elaborates on the configurations, the material properties, and the recent trend of TABS.

3.1. Different Topologies of TABS

According to the location of embedded pipes, TABS can be sorted into floor, ceiling, and wall types [30,31]. The floor-type TABS is one of the most widely used forms that can be utilized for both heating and cooling supply. The ceiling-type TABS is also widely used, but mainly for cooling purposes. In terms of the wall type, the application is relatively scarce in practice due to the construction difficulty and the comfort issue caused by radiation asymmetry. The three configurations of TABS show different heat transfer performances. By using the wall as the heating/cooling surface, the heat transfer coefficient is 8–9.3 W/m²K, while the figure is 9.0–13.2 W/m²K for the ceiling cooling system [32]. The floor-type TABS manifests the highest heat transfer coefficient for heating, which is 9–11 W/m²K; however, the value is 5.7–7.0 W/m²K for cooling [32]. Figure 2a compares the 24 h thermal storage performance across wall-, ceiling-, and floor-based TABS configurations. Even though ceiling radiant cooling is widely used for its better heat transfer coefficient, the interest in applying the floor-type TABS in the large space buildings, such as airports or railway stations, is increasing remarkably [33]. The reason is that the floor cooling system can immediately offset direct solar radiation in the space, thereby avoiding the overheating risk. Nowadays, hybrid systems, such as the floor–ceiling radiant system, receive more attention to overcome the shortage of the single TABS.

According to the different implementation purposes, TABS can be sorted into the radiant heating/cooling system and the thermal barrier system. The radiant heating/cooling system is applied to cover the heating/cooling load of the occupants [34]. In contrast, the thermal barrier system is used to isolate the indoor environment from disturbances from the outside [35,36] so as to buffer weather fluctuations. However, the distinguishing line between the radiant system and the thermal barrier is vague, and the two types can interconvert by regulating the control parameters [37]. As illustrated in Figure 2b, Zhao et al. (2025) [38] developed double-layer pipe-embedded walls (DPEWs), which can enhance the thermal performance of buildings and reduce heating energy demand.

Figure 2. The combined application of PCM and TABS. (a) The thermal energy stored (Estored) in the structure over 24 h of continuous operation [30]. (b) A schematic of steady-state temperature and heat transfer profiles of the studied DPEW in heating mode [38]. (c) A schematic of the ventilated hollow concrete block integrating PCM [39]. (d) The lateral surface temperature under different configurations [39]. (e) The construction process of PCM-integrated timber structures [40]. (f) Temperature variations in PCM-integrated and conventional timber wall panels [41].

Figure 2. The combined application of PCM and TABS. (a) The thermal energy stored (Estored) in the structure over 24 h of continuous operation [30]. (b) A schematic of steady-state temperature and heat transfer profiles of the studied DPEW in heating mode [38]. (c) A schematic of the ventilated hollow concrete block integrating PCM [39]. (d) The lateral surface temperature under different configurations [39]. (e) The construction process of PCM-integrated timber structures [40]. (f) Temperature variations in PCM-integrated and conventional timber wall panels [41].

3.2. Core Layer Material and Heat Carrier Medium

The thermal behavior of TABS can be characterized by several key parameters, including the surface heat transfer coefficient, thermal resistance, and heating/cooling capacity. The energy flexibility potential primarily depends on the thermal capacity of the core layer material—where the pipe system is embedded—and the properties of the heat carrier medium. Air can be used as the heat carrier for TABS, especially for the natural ventilated cooling purpose. However, for most TABS, water is utilized as the heat transfer medium because of its appropriate thermal properties and reduced pumping power.

Concrete and mortar are the most common materials for the core layer of TABS concerning their large thermal capacity. A concrete floor slab of 100 m² with a thickness of 30 cm can store 20 kWh of thermal energy along with the temperature gradient. A more straightforward indicator for evaluating energy flexibility is from the time domain [42]. Ning et al. (2017) [43] used response time τ95 to characterize the thermal inertia of a TABS. In general, a conventional heating or cooling TABS can maintain the indoor temperature within an acceptable range for several hours if a supply cut-off occurs [16].

In order to enhance the thermal flexibility of the TABS, materials with larger heat storage capacities have been considered to replace or be combined with the conventional material [44]. The incorporation of PCM in TABS has become increasingly popular in recent studies [45,46]. Ren et al. (2021) [47] adopt PCM with the radiant ceiling for both heating and cooling purposes, and the application of PCM effectively solves the discomfort caused by the rapid temperature variation in the radiant panel. With the optimized charge/discharge schedule, the combination of PCM and the floor radiation systems can provide the required room temperature by storing cheap energy during the night for 8 h and releasing the heating during the daytime [48]. Moreover, the incorporation of PCM can reduce the heating load of the radiant system by up to 40% and circulating fluid up to 25% for cooling [49]. Apart from the water-based system, the placement of PCM around the ventilation ducts can stabilize the room temperature and improve the building’s thermal flexibility [39]. Figure 2c shows a ventilated hollow concrete block with integrated PCM, and Figure 2d presents the temperature evolution on its lateral surface under different configurations.

Other materials may also be applied in TABS due to special structural design requirements or the availability of local resources. For instance, timber has been widely used in building construction in many regions because of its high durability and density. Moreover, the thermal storage capacity of timber is relatively high, making it more suitable for TABS applications than concrete in certain cases [41,50]. Yang et al. (2024) [41] proposed a timber-based double-layer pipe-embedded wall system integrated with PCM. Figure 2e illustrates the process of combining PCM with timber and constructing a pipe-embedded wall structure. Figure 2f compares the peak temperatures of PCM-impregnated and conventional timber exterior panels, showing that the former exhibits a 3.0 °C lower peak temperature. The results indicate that PCM significantly improves the thermal performance of timber wall panels. In order to lower carbon emission, more biomaterials for the building construction are being studied nowadays. Some recycled bio-building materials show great performances in respect to thermal stability and load shift potential [51]. The incorporation of TABS can be a promising alternative for a low-carbon society in the future.

4. Impact of Operation Temperature Requirements on Energy Flexibility

The adjustment of the indoor temperature within the set range is the main approach to activating building energy flexibility. By regulating the indoor temperature, the building thermal mass can be charged and discharged as passive thermal storage. The efficient activation of building thermal mass has therefore become a focus of research [52]. The set range of the indoor temperature depends on a few factors, which can vary among different types of buildings or different preferences of the occupants. However, due to the large heat capacity of TABS, 1 °C temperature difference can result in significant discrepancies in the usable heat flexibility. Therefore, a clearly defined range of the operation temperature is of great importance to improve the utilization of the flexible energy from the building mass. This section provides a detailed overview of the requirements of the TABS operation temperature from multiple considerations.

4.1. Requirements Due to Thermal Comfort

The requirements due to thermal comfort are of priority for designing a suitable regulating range of the operative temperature. ISO 7730 [53] defines the comfort range of the operative temperature for humans with a sedentary activity level of 1.2 met and a clothing thermal resistance of 1.0 and 0.5 clo as between 20 and 24 °C in winter and 23–26 °C in summer, respectively.

Both the air temperature and the radiant temperature impose significant effect on human thermal comfort. For the TABS, the radiant surface leads to different air temperature and radiant temperature values for the indoor environment [54]. The indoor operating temperature [55], which integrates the impact of the indoor air temperature and mean radiant temperature (MRT), is therefore proposed as an indicator for indoor thermal comfort [43]. Chandrashekar et al. (2024) [56] conducted simultaneous investigations in studio-type classrooms under three cooling scenarios: operating the TABS as a standalone system, integrating the TABS with an indirect–direct evaporative cooling (IDEC) system to introduce natural ventilation, and combining the TABS with a fan to enhance forced air circulation. Figure 3a,b, respectively, illustrate the 12 h variations in mean radiant temperature and operative temperature under different TABS cooling scenarios.

Indoor air velocity, heating strategies, and the types of HVAC terminals significantly influence the heat transfer performance of the radiant heating/cooling systems, thereby impacting overall indoor thermal comfort. Given that radiation serves as the primary mode of heat transfer, the indoor air temperature setpoint for TABS can differ slightly from that of all-air systems, even under a comparable thermal comfort requirement [2]. Some studies [62,63] investigate the correlation between indoor comfort and MRT; however, the current comfort standards do not clearly characterize the effect of MRT. By comparing the effect of the direct and reflected radiation on human thermal sensation, it was found that the difference in MRT between surfaces of different materials can be up to 4.5 °C [64], which can lead to a significant difference in the air temperature setting.

Radiant asymmetry is another important factor related to comfort [65,66] that constrains the indoor temperature settings of TABS. Radiant temperature asymmetry is defined as the difference between the plane radiant temperature of the opposite sides of a small plane element or of the environment on opposite sides of a person [62,67]. Based on the substantial experiments by Fanger et al. (1985) [57], the acceptance of radiant temperature asymmetry strongly correlates to the different types of radiant surfaces. For example, people are more sensitive to temperature asymmetry caused by a warm ceiling; only 4 °C is acceptable. In contrast, for radiant asymmetry caused by a warm wall, this value is up to 23 °C. To avoid discomfort due to radiant asymmetry, the temperature of the radiant surface needs to be controlled carefully. Imanari et al. (2021) [68] investigated the thermal performance of a radiant cooling system and reported that the ceiling surface temperature should be maintained within 18–22 °C for cooling and 27–30 °C during heating to ensure thermal comfort. Zhou et al. (2019) [58] investigated the impact of exposure duration on thermal comfort. Two satisfaction–radiant asymmetry curves for a floor cooling system, considering exposure durations of 2 h and 8 h, were developed to reflect the influence of exposure time on thermal comfort, as shown by the red and blue lines in Figure 3c.

Moreover, there is also a discomfort risk due to the too warm/cold radiant surface that the occupants tend to come into directly contact with. The relevant requirements mainly orient toward the floor- and wall-type radiant systems. The floor surface temperature should be no less than 19 °C. The upper boundary of the floor surface temperature is 29–35 °C depending on whether it is the occupied zone, while it is 40 °C for the wall [53]. However, the criteria can be different concerning user behavior (activity level, clothing level, etc.), floor surface materials [69] and exposure duration [58]. Therefore, the design of the optimal operation temperature range of TABS should consider the actual situation and requirements comprehensively.

4.2. Requirements Due to Condensation Issues and Indoor Air Quality

Condensation has become a major issue that jeopardizes the performance of a radiant cooling system in practice [68,70,71], especially in hot and humid regions. Based on the thermodynamic theory, the temperature of the radiant cooling surface is lower compared to the indoor air temperature. The condensation of moist air occurs on the cooling surface if the dew-point temperature is reached. Long-term moisture condensation can lead to the growth of fungi, which could damage the radiant terminals or building envelop [72]. In order to prevent condensation on the cooled surfaces, the operation temperature of a TABS is constrained according to the dew-point temperature of the conditioned space. Considering the large thermal inertia of TABS, the operation temperature is often controlled to be 1–2 °C above the dew-point temperature for safety [73,74]. Olesen et al. (2017) [69] stated that the supply water temperature should be controlled to be equal to the dew-point temperature specifically for TABS so as to provide the maximum amount of cooling without condensation risks. In order to leave more flexibility for temperature regulation, other condensation prevention methods are often combined [75,76]. For example, integrating a dehumidification system with a radiant cooling system can enhance the cooling capacity of the latter [59], as illustrated in Figure 3d. Coupling this with ventilation systems is also reported to be efficient in preventing the condensation risk associated with TABS operation, including natural ventilation systems [77], mechanical ventilation systems [78], decentralized dedicated outdoor air systems [79], dedicated outdoor air systems [80], displacement ventilation systems [81,82] and air handling units [83].

The temperature of the radiation system can also affect the indoor air quality, and the issue relates more to the heating scenario. As the heating pipes are embedded in the building structure, the building envelope and the adhesive covering are also heated. A higher temperature can potentially accelerate the emission of volatile organic compounds (VOCs) from the building/furniture materials and result in a higher concentration compared to the ventilation/radiator system [84]. For example, An et al. (2010) [60] investigated formaldehyde emission under the flooring temperatures at 20 °C, 26 °C and 32 °C. The 32 °C scenario results in a 1.6–2 times higher concentration of formaldehyde compared to the other two scenarios with lower temperatures, as illustrated in Figure 3e. Chen et al. (2016) [61] simulated formaldehyde emissions from particleboard under non-isothermal conditions using floor heating and air circulation systems. Fig. 3f presents a comparison of transient formaldehyde concentration distributions at various heights and different bottom surface temperatures. Kang et al. (2013) [85] found that the floor heating temperature at 35 °C can increase VOC emission compared to the temperature at 25 °C, and such an effect is more severe for the flooring/adhesive assembly. Therefore, the temperature of the radiant component is better controlled below 30 °C concerning chemical emission, which also affects the operation of TABS in practice.

5. Modeling Methods for Dynamic Thermal Behavior of TABS

The effective utilization of TABS energy flexibility depends on accurately characterizing its dynamic thermal behavior, which in turn requires high-fidelity simulation models. The heat transfer process involves multi-stage coupling—including the fluid within embedded pipes, the interaction between pipes and the building structure, and the subsequent heat exchange with the indoor environment. To reflect real-world conditions, both endogenous and exogenous factors (e.g., material properties, environmental disturbances) should be appropriately integrated into the model. The simulation outputs are expected to quantify heat fluxes across different stages to assess energy flexibility performance and resolve multi-dimensional temperature fields to support indoor thermal comfort and prevent condensation risks. From the operation point of view, a precise system simulation is the basis of more advanced and prediction-based control strategies [86,87]. Therefore, this section reviews existing TABS simulation methods classified by modeling mechanisms.

5.1. Physics-Based Model

The mechanism model reveals the dynamic heat transfer law of TABS at the principle level and provides a theoretical benchmark with physical interpretability for the prediction of system thermal performance and parametric design [88]. According to differences in the solution methods, the mechanistic models can be further classified into two categories, analytical models and numerical models [89], which are suitable for the idealized theoretical verification and refined simulation of complex engineering scenarios, respectively.

5.1.1. Analytical Model

The analytical model aims at describing the actual heat transfer process by mathematical equations. Therefore, the differential equations, the boundary conditions, and the assumptions need to approach to the exact situation as much as possible. The analytical model aims at describing the actual heat transfer process by mathematical equations. Therefore, the differential equations, the boundary conditions, and the assumptions need to approach to the exact situation as much as possible. Since the establishment of analytical models is rather complex and difficult, the scope of their application is limited mostly to one-dimensional or two-dimensional heat transfer problems under the steady state. However, the solving process of the analytical model is simpler and it is easier to couple with other simulation tools.

Lauoadi (2004) [90] developed a two-dimensional semi-analytical model to analyze the heat transfer between the pipe, its surrounding contact and between the floor slabs. The model resulted in improved accuracy in building load forecasting compared to the one-dimensional model. The resulting error is less than 9%. Wu et al. (2020) [87] proposed a simplified method for calculating the average surface temperature of a TABS by assuming that the temperature gradient along the flow direction is negligible compared to that along the vertical plane. As illustrated in Figure 4a, heat transfer within the TABS slab can thus be considered two-dimensional. Figure 4b presents the temperature profiles at various depths below the floor slab surface under typical conditions. The numerical solution used for model validation shows excellent agreement with the steady-state analytical solution, with a maximum relative deviation of only 0.65%.

5.1.2. Numerical Model

The numerical model computes the temperature distribution and heat flux of TABS by solving a set of algebraic equations that approximate the differential equations of the specific heat transfer process. The solution methods are normally sorted into the FEM (finite element method) [91], FDM (finite difference method) [92], and FVM (finite volume method) [93] according to how the finite elements are defined. The numerical models can provide simulation results with satisfying accuracy. However, due to the huge computation cost, it is time-consuming and difficult to be coupled with the other simulation tools with a different focus.

Qu et al. (2020) [86] developed a three-dimensional simulation model of a Thermally Activated Wall (TAW) system using the commercial software COMSOL. An experimental platform equipped with TABS was constructed in Beijing, China, to validate the model. The study conducted a comprehensive analysis of the thermodynamic performance and characteristics of the TAW in ultra-low energy buildings and proposed optimized TAW design guidelines applicable to most climatic zones in China. Jiang et al. (2023) [94] employed ANSYS Fluent to numerically solve heat transfer equations for a DPEW and investigated the effects of key design parameters on its heating performance. Figure 4c illustrates a typical application scenario of the DPEW system. Based on the identified design parameters, the heat transfer process of the DPEW was simulated, and the results are presented in Figure 4d. Through comparative analysis with other wall configurations, Jiang et al. (2023) [94] demonstrated that the DPEW exhibits superior potential in enhancing indoor thermal comfort. Larwa et al. (2021) [95] investigate the temperature distribution and heat flux of the radiant floor heating system integrated with the macro-encapsulated PCM under different hydronic conditions using COMSOL Multiphysics. The simulated mean floor surface temperature of the model is only 0.5 °C lower compared to the experiment results.

Figure 4. Modeling methods of a TABS. (a) A detailed view of the TABS and its two-dimensional heat transfer diagram [87]. (b) Temperature profiles at various depths below the floor slab surface [87]. (c) A diagram of a room with DPEWs [94]. (d) The numerical simulation results of the temperature distribution in the DPEW [94]. (e) A schematic structure diagram for a simplified dynamic model of the PE wall [96]. (f) Surface temperature curves calculated by the CFD model and 5R3C mode [96].

Figure 4. Modeling methods of a TABS. (a) A detailed view of the TABS and its two-dimensional heat transfer diagram [87]. (b) Temperature profiles at various depths below the floor slab surface [87]. (c) A diagram of a room with DPEWs [94]. (d) The numerical simulation results of the temperature distribution in the DPEW [94]. (e) A schematic structure diagram for a simplified dynamic model of the PE wall [96]. (f) Surface temperature curves calculated by the CFD model and 5R3C mode [96].

5.2. Resistance–Capacitance (RC) Model

Compared to the numerical and analytical methods, the RC model method is much simpler [96]. The RC model is based on an analogy of the Kirchhoff’s law. The basic theory is to use the resistance and capacitance in a circuit to simply represent the thermal resistance and heat capacity of the heat transfer process. Due to the simplification of the modeling concept [97], it leads to a fast computation speed and acceptable accuracy. In the building simulation field, the RC approach is widely applied to establish the gray box model for the load prediction. With the RC model, the heat transfer and surface temperature of the TABS can be fast calculated with an accuracy that meets the engineering requirements.

In order to promote the application of the RC model in special conditions, modifications of the model’s framework and input constraints are conducted. Weber et al. (2005) [98] compared a star RC model with a triangular RC model in calculating the heat flow through the building envelope in both the frequency domain and the time domain. The results show that the star RC model is simple, but the application is limited, while the triangular RC model is computationally complex but can be applied more broadly. Hongn et al. (2022) [99] further compared the performances of the triangular and umbrella RC models. The frequency domain finite difference was applied to investigate the accuracy of the models. The umbrella RC model has been found to have higher accuracy with the normalized root mean square error (nRMSE) of 3%. Fan et al. (2024) [96] proposed a 5R3C simplified dynamic model with a straightforward parameter configuration scheme and compared it with other pipe-embedded (PE) wall models, including the finite element–finite difference model, the computational fluid dynamics model (CFD) model, and the 5R2C simplified dynamic model. The schematic structure diagram for the 5R3C simplified dynamic model of the PE wall is illustrated in Figure 4e. Compared to these models, the 5R3C model offers a simpler structure, higher accuracy, and greater computational efficiency. Figure 4f indicates the surface temperature on the external wall surface, internal wall surface and internal pipe surface of the PE wall calculated by the CFD model and the 5R3C model. Combined with the effective number of heat transfer units model, the RC model is capable of simulating the heat transfer process of a TABS along the pipe direction and along the wall width direction [100,101]. As a result, the dynamic thermal behavior of a TABS can be described with better accuracy.

6. Control Strategies for Activating the Energy Flexibility of TABS

The variability of the thermal demand and the stochastic nature of the renewable energy sources lead to supply–demand mismatch, grid fluctuation, and energy distribution insufficiency [102]. The large thermal capacity of TABS is used to achieve active load control on the demand side so as to enhance the network synergy. Therefore, suitable control strategies are required to activate energy flexibility precisely [103]. However, uncertainties from the internal and external environment, as well as the large thermal inertia of TABSs [104], impose significant difficulties for control. This section elaborates on the characteristics and performances of different control strategies for building energy flexibility utilization.

6.1. Rule-Based Control Methods

Rule-based control (RBC) is a control strategy that achieves automatic system regulation by utilizing predefined logical conditions and operational instructions. It includes on/off control, temperature profile control, unknown-but-bounded (UBB) method, intermittent control [105], etc. The RBC method eliminates the need for establishing accurate mathematical models while offering straightforward implementation, positioning it as a cost-effective entry-level strategy for activating energy flexibility. The implementation of the RBC strategy in TABS can lead to specific energy savings and load shift effects. However, overshooting is inevitable due to the control mechanism. The efficiency of RBC can be evaluated by the ratio of the energy used for room heating/cooling and the overall energy supply to the TABS.

On/off control is applied according to the operation range of the key parameters (e.g., room temperature). It is able to buffer the demand side variation in TABS to some extent, thereby improving the overall efficiency if supplied by randomly fluctuating renewable energy systems. However, such a control strategy ignores the dynamic heat transfer of the building system and normally results in the overshooting of the heating/cooling supply. This issue can be more serious for TABS with large thermal inertia, which can cause energy waste and thermal discomfort problems.

Temperature profile control depends on the fixed operation temperature curve, which is the function of a few factors. Temperature curve control normally applies to the supply temperature, and it is strongly related to practical experiences. The common temperature curve control method can be divided into room temperature feedback control, supply/return water temperature control with outdoor temperature compensation, and control using the supply and return water temperature difference. Qu et al. (2019) [106] used outdoor temperature compensation control, night pre-cooling control, and water supply temperature control for a radiant cooling system, which helped to reduce the nominal capacity of the system facilities by 35%. Liu et al. (2024) [107] investigated a variable water temperature control strategy for an apartment floor heating system compatible with an air-source heat pump. Compared with the constant supply temperature strategy, the proposed method reduced the average supply water temperature by 9.2 °C and 5.7 °C in two representative periods and decreased system power consumption by 56.7% and 29.5%, respectively.

However, at the design stage, no operation data are available for defining the specific control curve. The UBB control method determines both the upper and lower boundaries of heat gain and the operating range of the supply temperature, based on the thermal characteristics of the TABS and the building system itself. As long as the heat gain of a room remains within the variation range, indoor thermal comfort can be ensured. The UBB control method can automatically switch the TABS between indoor heating and cooling modes to meet the thermal comfort criteria [108]. Heating/cooling curves were then developed based on the actual internal heat gain and solar radiation. However, the effect of such heat gains on indoor thermal comfort is not directly correlated. Therefore, instead of the actual heat gain values, only the upper and lower limits of the heat gain are used for the control strategy.

Pulse width modulation (PWM) control is another heat-gain-based control method. It can be regarded as a high-frequency form of intermittent control, enabling smoother and more precise system regulation. Combined with the UBB control method, integrated control can reduce the power consumption of the circulating pump by 80%, and the whole system saves 55% of the total energy consumption [109].

RBC is suitable for application in simple scenarios, such as the initial stages of system design, or situations where energy supply and demand are balanced, time-of-use electricity pricing is fixed, and occupant behavior is predictable (e.g., regular energy usage patterns in office buildings), in order to activate the energy flexibility of TABS. Currently, most studies tend to use RBC as the baseline control method, comparing its control effectiveness with that of other optimization control methods.

6.2. Optimal Control Methods

In contrast to the rule-based control method, optimal control can optimize the trajectory of the control input to achieve the chosen objective(s) while meeting the requirement boundaries. Optimal control enables a more accurate activation of TABS energy flexibility based on predictions concerning unknown disturbances, such as outdoor weather, indoor occupant activity, fluctuations in the energy supply side, etc. According to [110], optimal control based on a prediction can reduce energy consumption by 30–40% compared to conventional control strategies under equivalent or enhanced comfort levels. Moreover, optimal control can also effectively avoid the thermal rebound effect caused by the rule-based control strategies, thereby resulting in more stable grid operation. Optimal control can be divided into two groups according to the different methodologies used, namely a group based on control theory and a group based on machine learning. Model predictive control (MPC) and reinforcement learning (RL)-based control are representative methods that can control large and complex systems, such as buildings comprising various variables.

6.2.1. Model Predictive Control

Model predictive control comprises a single cost function or combination of multiple cost functions, a system model with a receding horizon to predict the future system state, and constraints of the parameters. MPC is an optimization control method based on state-space models. When applied to the energy flexibility management of TABS, it relies on an accurate description of the dynamic thermal behavior of the TABS. Commonly used modeling methods include three types: white-box models [111], gray-box models [112], and black-box models [113]. White-box models require the use of detailed physical information about the building to construct, enabling an accurate understanding of the system’s dynamic model. Gray-box models typically involve simulating a physical model based on the building’s thermal dynamics, integrated with operational data, to provide reliable state inputs for MPC. Black-box models adopt a purely historical data-driven approach, utilizing machine learning techniques to predict the future states of the system without requiring physical modeling. From an operational perspective, accurate model construction serves as the foundation for more advanced prediction-based control strategies.

The solution to the optimization problem is the core of MPC, which can be obtained by various approaches [114]. Numerous studies have demonstrated that MPC can significantly enhance energy efficiency while ensuring indoor thermal comfort for occupants, thereby activating buildings’ energy flexibility potential. However, the application of MPC in large-scale systems is sometimes limited by the high dimensionality or the deficiency of robust system identification.

Chen et al. (2021) [115] used MPC and compared it with conventional PID control. MPC performs with almost no overshoot in the control zone concerning the uncertainty of the ambient temperature and solar radiation. Compared to the conventional strategy, a system with MPC achieves energy savings of 17–27%. As seen in [116], MPC can save 30% of thermal energy in winter and 75% of thermal energy in the transition season compared to conventional RBC for a TABS building. With multi-step forward prediction and least squares methods, Prívara et al. (2013) [117] developed MPC for a radiant system, which can achieve 20% energy savings comparing to a tuned conventional controller.

In addition to the control of indoor temperature, some researchers attempt to apply MPC for indoor comfort maintenance. Woo et al. (2020) [118] developed an MPC strategy to prevent wall condensation by considering the hygrothermal transfer process and time lag, which also reduced energy consumption by 21–29.6%. Based on dynamic and steady-state RC models for TABS, Shmelas et al. (2016) [119] develop adaptive predictive control that enables the system to form a heating curve and adapt to unknown disturbances automatically.

6.2.2. Reinforcement Learning-Based Control

RL-based control methodologies constitute a subset of machine learning techniques. The mathematical framework of RL algorithms is based on the Markov Decision Processes, which enables the agent to make decisions by interacting with the environment [120,121]. Compared to MPC, the configuration of the objective system is unnecessary [122], while substantial history data are required to capture the system behavior accurately [121]. A more detailed comparison between MPC and RL can be found in [120,123].

The application of RL for TABS control is not as extensive as MPC, but there are a few studies related to building energy flexibility enhancement using RL. Silvestri et al. (2024) [124] employed a deep reinforcement learning controller based on the Soft Actor–Critic algorithm to optimize the energy consumption and thermal comfort of TABS in buildings and compared its control performance with RBC and MPC. Compared to RBC, the deep reinforcement learning control strategy reduced TABS energy consumption by 15% to 50% and decreased temperature violations by 25%. While the deep reinforcement learning controller achieved a comparable temperature control performance to the ideal MPC, it required 29% more energy consumption. Arroyo et al. (2022) [125] applied combined RL- MPC in a single family house with floor heating; the combined control algorithm effectively reduced the constraints violations compared to single RL and improved the adaptability to uncertain environments compared to single MPC. Energy flexibility enhancement by deep reinforcement learning control was further tested in a practical test building in [126]. The control algorithm can successfully utilize building energy flexibility to minimize the energy cost while maintaining thermal comfort. Knowing that the learning efficiency of RL is one of the major burdens hindering its practical implementation, a few researchers are working on solutions for integrating the extra models/algorithms to guide the RL training process [127,128], and the effects are significant.

A summary of the literature with respect to the control strategies for TABSs is presented in

Table 1. Summary of reviewed control methods applied for TABS demand response management.

Reference	System	Control Type	Main Contribution
Gwerder et al. (2008) [108]	Radiant floor system	Unknown but bounded	This control method can maintain good indoor thermal comfort even with large indoor heat gain.
Lehmann et al. (2011) [109]	Radiant floor system	Pulse width modulation control	Compared to continuous operation, pulse width modulation can save more than 50% of the pump’s operating consumption. In combination with a separate zone return pipe, 20–30% energy savings can be achieved.
Ma et al. (2013) [129]	Radiant floor system	Supply water temperature control	Compare the difference in indoor temperature fluctuation with different building heat capacities and different set temperatures using all-day operation and night operation modes.
Lim et al. (2014) [130]	Radiant floor system	Operational guideline control	The thermal characteristics of the building are analyzed and the building load is zoned. The load is divided into six zones for the heating season and four zones for the cooling season to facilitate the operation of the engineers.
Schmelas et al. (2015) [131]	Radiant floor system	Adaptive predictive control	It improves thermal comfort while reducing pump run time by 81%, and the control method can be easily integrated into building automation systems.
Hu et al. (2019) [132]	Radiant floor system	Model predictive control	It can improve thermal comfort in the initial phase, reduce energy consumption during peak hours, and save customers up to 18.65% on their electricity bills.
Qu et al. (2019) [106]	Radiant floor system	Optimal precooling, supply water temperature control	Optimization Controls 1 and 2 increased energy use by 10.9% and 14.6%, but Control 1 shifted loads to cheaper night-time hours, reducing peak demand and making the increase acceptable.
Arroyo et al. (2022) [125]	Radiant floor system	MPC, RL	The combined control is applied to a single-zone floor heating system. RL is used to truncate the non-linear program of MPC, while the state estimation, forecast and optimization are still conducted by MPC.
Zheng et al. (2024) [133]	Radiant floor system	RC-MPC, ANN-MPC	Both RC-MPC and ANN-MPC significantly improve comfort and reduce operational costs, achieving 30–95% reductions in discomfort and 17–34% savings in energy expenses compared to rule-based control.
Qu et al. (2024) [24]	Radiant floor system	On/off control	Compared to the continuous heating mode, the four heating control strategies proposed in the study achieved energy savings of 22.3%, 25.8%, 40.4%, and 48.4%, respectively.
Široky et al. (2011) [134]	Ceiling radiant system	Model predictive control	The energy saving potential after adopting MPC is between 15% and 28%, but cost effectiveness should also depend on other factors, such as implementation costs.
Prı’vara et al. (2013) [117]	Ceiling radiant system	Model predictive control	MPC combined with prediction error minimization and partial least squares can result in energy savings of more than 20% in a heating season.
Feng et al. (2015) [135]	Ceiling radiant system	Model predictive control	The room is in thermal comfort more than 95% of the time, while the energy consumed by cooling towers and pumps is reduced.
Woo et al. (2020) [118]	Ceiling radiant system	Model predictive control	Compared to on/off control, the MPC framework for TABS achieved 2.5–10.0% greater site cooling energy savings.
Hassan et al. (2022) [136]	Ceiling radiant system	Adaptive predictive control	APC collaboratively manages air and radiant sides of the system. APC reduces energy consumption and peak power by 2.7 and 18.8%.
Krzaczek et al. (2012) [137]	Thermal barrier	Gain scheduling control	Buildings with thermal barrier systems using GSC’s control strategy can maintain indoor thermal comfort throughout the year.
Zhao et al. (2025) [14]	Thermal barrier	On/off control, climate-compensated control strategy	By adopting the hybrid LTAE-HTAE approach, heating energy consumption was reduced by 48.92%, cooling energy consumption by 70.31%, and operational carbon emissions by 51.16%.
Romaní et al. (2018) [138]	Hybrid radiant system	Peak load shifting, solar predictive control	By applying a peak load shifting strategy combined with solar predictive control to charge the radiative walls, grid energy imports during peak periods can be minimized.

.

7. Energy and Economy Performances

TABS is considered to have higher exergy and energy efficiency when combined with low-temperature heating or high-temperature cooling [139], making it more suitable for dynamic energy systems integrated with renewable energy sources [140,141]. Moreover, further energy savings can be achieved by implementing control strategies that activate the flexibility of TABS as passive thermal storage for both short-term and long-term thermal response needs [142]. Great flexibility potential helps to improve the synergy between the supply and demand, which is useful to optimize the dimensions of the system facilities and stabilize the grid operation [143].

By coupling the geothermal heat source [144], TABS can reduce CO₂ emissions by 85% and save energy costs by 76%. Zhao et al. (2025) [145] integrated the DPEW system with a ground source heat pump and conducted residential heating case studies based on representative cities. The results indicate that the coupled DPEW–heat pump system achieves significantly lower supply temperatures compared to a conventional floor heating system. In the selected cities, the proposed system demonstrates 12–18% seasonal energy savings, along with a dynamic payback period of 1.6–4.3 years and a reduction in CO₂ emission intensity by 1.2–3.2 kg/m². Long et al. (2025) [146] embedded heat pipes into concrete walls to utilize the absorbed solar radiation for heating domestic hot water within the embedded pipes, thereby reducing the amount of heat transferred into the indoor space. Simulation results based on a building in Hong Kong indicate that the system reduces wall heat transmission to 76.1%, achieves a water gain efficiency of 16.7%, and saves 162 kWh/m² of electricity annually. Bojić et al. (2015) [147] further developed a floor–ceiling radiant system coupled with a geothermal source heat pump and photovoltaic array. The system shows better performance compared to the single radiant floor system, ceiling radiant system, or wall system. Compared to the conventional ice storage system, TABS can further save 10% energy consumption by avoiding energy loss due to the phase change process of the ice storage system [148].

Moreover, with the demand response potential of TABS, the peak load of the energy system can be effectively removed [149]. The load shifting capability of TABS can be further enhanced by coupling the materials with larger thermal capacity, such as PCMs [150,151]. It is possible to move the peak load to the off-peak period completely. Abdel-Mawla et al. (2024) [152] integrated PCM into TABS, achieving a 9.75% reduction in total energy consumption and a 41.2% decrease in peak cooling load while maintaining the same level of thermal comfort. Lu et al. (2022) [153] investigated the energy flexibility potential of a near-zero energy building in Beijing. The experimental results show that 23.53 Wh/m² more energy consumption was shifted to the off-peak hours by energy flexibility. Chen et al. (2024) [149] proposed a PCM-based pipe-embedded wall system, which achieved an additional 15.78% reduction in peak demand and a 10.18% reduction in energy consumption during demand response events compared to conventional PCM passive walls.

In terms of economic performance, the reduced peak load leads to the decreased dimension of the system equipment, which is useful for investment control. Moreover, the operation cost can also be saved by utilizing lower-price energy and load shifting with energy flexibility [154]. The optimal control strategy can further improve the economic performance of TABS by setting energy-related expenses as the optimization objective. According to Hu et al. (2019) [132], end-users can save 1.82–18.65% of their daily electricity bills by applying a more effective load shifting strategy to TABS. Pedersen et al. (2017) [155] use EMPC for building control to achieve up to 6% cost savings and a 3% reduction in CO₂ emissions. Hedegaard et al. (2017) [156] simulated an EMPC-controlled building in Denmark for four months, and the total savings was about 2.9% to 5.6%.

8. Discussion

Existing research has demonstrated the significant advantages of TABS in enhancing building energy flexibility, particularly when integrated with renewable energy sources and advanced control strategies. Various TABS configurations exhibit distinct performance characteristics in heating and cooling applications. In addition, innovations in structural topology and the integration of advanced materials—such as double-layer pipe-embedded envelopes, fin-enhanced embedded pipe systems, and the incorporation of PCMs—have been shown to further improve the thermal performance and responsiveness of the system, especially in reducing peak loads. In terms of modeling, approaches have evolved from detailed numerical simulations to widely adopted RC models, providing a solid methodological foundation for flexibility assessment and control optimization. Numerous studies have confirmed that the accurate modeling of the dynamic thermal behavior of TABS, in combination with advanced control strategies (particularly MPC), can significantly enhance load shifting capacity and reduce energy costs.

Nevertheless, to ensure a more comprehensive and balanced assessment, it is essential to compare TABS with alternative passive envelope solutions and to evaluate their long-term viability in practical applications. Conventional passive envelope systems, such as those relying on optimized thermal inertia or night-time ventilation in summer, are largely constrained by external climatic conditions and occupant behavior, often resulting in delayed and less controllable thermal responses. In contrast, TABS can deliver more flexible and precise thermal regulation by integrating advanced control strategies, thereby achieving more stable load management and enhanced thermal comfort. However, in climates with significant diurnal temperature variation, passive solutions may achieve considerable energy savings at minimal operational cost. From a construction cost perspective, TABS generally involves higher initial investment than passive systems, primarily due to the complexity introduced by embedded piping and automation control. Nevertheless, existing studies widely acknowledge that such costs can be gradually offset over the system’s lifecycle through enhanced energy flexibility, peak load reduction, and improved operational adaptability, especially under fluctuating electricity prices or in systems with high renewable energy penetration. In terms of durability and maintenance, the service life of TABS is largely dependent on the material quality of the embedded pipes and the workmanship of the installation. At present, TABS typically utilizes materials such as cross-linked polyethylene, which offer excellent corrosion and pressure resistance and can achieve a service life of approximately 50 years under normal operating conditions, with relatively low maintenance requirements. However, limited accessibility to embedded piping within the envelope may lead to invasive repair operations in the event of leakage or blockage. Compared with passive systems, TABS allows for enhanced operational flexibility and comfort control, which may better accommodate building types with more stringent energy management requirements. Currently, systematic comparisons between TABS and passive systems remain limited, with a pressing need for comprehensive techno-economic evaluations under consistent boundary conditions and long-term operational scenarios.

Although considerable progress has been made in the design optimization, modeling, and control of TABS, several key challenges remain in their practical deployment. Most existing modeling approaches have limited accuracy in capturing unsteady thermal transfer and dynamic load disturbances, making it difficult to balance computational efficiency with a comprehensive representation of system behavior. Research on the integration of novel materials has largely focused on the thermal storage and peak-shaving capabilities of PCM, while studies on the coupled heat and moisture transfer performance of other advanced materials in TABS are still insufficient. Moreover, current control strategies face limitations in multi-objective optimization, long-term robustness, and autonomous adaptability, making it difficult to respond effectively to complex energy pricing mechanisms and behavioral uncertainties.

Future research should focus on the integration of high-fidelity modeling approaches with data-driven techniques, the co-design of advanced materials and structural systems, and the enhancement of adaptive control strategies. Additionally, the deployment of intelligent control frameworks under representative scenarios and long-term performance validation should be strengthened to support the widespread adoption of TABSs in the context of high renewable energy penetration.

9. Conclusions

This article reviews the development of TABS from design, simulation, control, and optimization perspectives. Since building energy flexibility is becoming the research focus to enhance the integration between renewable energy sources and the demand side, the aim is to provide knowledge of activating and regulating the energy flexibility from the building envelope more effectively. Factors such as TABS topology, thermal comfort requirements, simulation methodologies, and control strategies have a significant impact on system performance, and the key findings from this study are summarized as follows:

• TABS can be categorized into the wall type, ceiling type and floor type, each suited to specific application scenarios. The floor-type TABS shows the highest total heat transfer coefficient for heating, while the ceiling-type TABS shows a superior heat transfer coefficient for cooling. Increasing the heat capacity of the core layer and the heat carrier medium enhances the energy flexibility potential of TABS.
• The regulating range of indoor temperature is important for the activation of TABS heat flexibility. From a thermal comfort perspective, the mean radiant temperature should be considered for the determination of the temperature boundaries. The risk of condensation and VOC emission also impose limits on the optimal indoor temperature range.
• The thermal behavior of TABS can be simulated under different modeling mechanisms with varying levels of accuracy. Among them, the resistance–capacitance model is widely used due to its computational efficiency, adaptability across different scenarios, and compatibility with advanced control algorithms.
• Compared to rule-based control methods, optimal control methods can effectively mitigate the overshooting issue of TABS by accounting for uncertain disturbances. MPC has been widely applied to optimize the operation of TABS concerning their energy flexibility potential, while RL learns actions through long-term interactions with the environment.
• TABS improves system energy performances by enhancing the synergy between the energy supply/demand side and also by enabling high-temperature cooling and low-temperature heating. With the optimal control strategy, TABS demonstrates significant cost-saving potential by effective load shifting in response to price variation.

As the penetration of renewables in energy generation keeps increasing, the meticulous design and operation of energy-flexible buildings will be increasingly required. Further investigations are therefore suggested in the following topics:

• The refinement of the heat transfer simulation of TABS concerning transient heat damping along the heat transfer direction and the impact of dynamic heat sources.
• Integration with new materials to enhance the energy flexibility potential or heat and moisture transfer performance of TABS.
• The development of effective control methods for TABS with a high degree of automation and adaptability to the uncertain disturbances.