Experimental investigation of the transient thermo-mechanical behavior of a modular sensible heat storage system

. This paper focuses on the experimental investigation of the transient thermo-mechanical behavior of a lab-scale prototype heat storage system. The experiment involves heating of the storage unit up to a set temperature of 70°C followed by a passive cooling, and monitoring the thermally induced strains and stresses in the sensible heat storage module. The results show a significant development of induced thermal strains and stresses in the heat storage unit upon heating, particularly at the interface between the embedded heating element and the soil.


Introduction
Heat demand in many places of the world is seasonal, that is, most of the heat energy is generated during the summer while the highest demand is in winter [1]. The difference between the demand and supply of heat energy is problematic for most societies of our world, forcing their continued reliance on fossil fuels and other unsustainable sources of energy.
One way of bridging this gap is via the use of clean and renewable energy schemes, such as borehole thermal energy storage (BTES) systems, where heat or cold from solar collectors or other forms of energy is stored for long periods for use in future industrial and domestic purposes [2,3]. Detailed studies on the physical modeling of the coupled heat transfer and water flow characteristics in unsaturated soil-BTES systems were reported in [4] and [5].
Such heat storage systems are usually built at underground levels supporting structures and having load bearing capabilities. When the heat storage units or structural elements are heated and cooled, thermally induced strains and stresses are added to the already present mechanical strains and stresses due to structures built on them [6]. Therefore, careful study of the thermomechanical behavior of the heat storage systems, and soil-structure interaction studies especially at elevated temperatures are essential [7,8].
In this paper, the transient thermo-mechanical behavior of a lab-scale prototype borehole thermal energy storage unit is studied experimentally. The heat transport to the app. 0.9 m 3 cylindrical heat storage barrel is facilitated via an embedded cemented borehole heat exchanger (BHE) unit placed at the center of the barrel. The set-up is heated up to a maximum storage temperature of 70°C followed by a passive cooling, and the thermally induced strains and stresses of the soil and cemented BHE are monitored.

Experimental program 2.1 Experimental set-up
The solid-liquid modular thermal energy storage set-up consists of a cylindrical plastic barrel with a diameter of 100 cm and a height of 110 cm, a cylindrical cemented BHE unit with a diameter of 15 cm and a height of 100 cm, a 20 cm insulation (i.e. PROBAU Mineralwolledämmstoff soft insulation + thick Styrodur [9] hard insulation) for the top and bottom bases of the barrel, a flow meter and a 15 liter capacity CC-215B Huber heating pump [10] (Fig. 1). The barrel plastic has a side thickness of 10 mm and bottom thickness of 20 mm, and is placed on a heavy duty wooden pallet (Fig. 1). The BHE unit is equipped with two independent U-shaped plastic pipes for the transport of heat to and fro the soil, and rests on 10 cm height of soil placed on the base of the barrel. The soil layer below the BHE provides extra insulation for the heat losses via the bottom face of the barrel. The heat carrier fluid, i.e. water without any additives, is circulated using the heating pump inside the BHE. To minimize heat losses, all the pipes carrying heated water were properly insulated.
A total of 32 K-type thermocouple temperature sensors [11], numbered T0 to T31 are used to monitor and record temperatures in the storage unit (Fig. 2). 24 of the 32 thermocouples are embedded inside the soil, 5 are embedded in the cemented BHE and 3 are placed outside the barrel to monitor boundary conditions. The thermocouple sensors are attached to a NI data acquisition unit connected to a PC, and have a measurement error of ±0.45°C [11].
Two 1-LY41-100/120 DMS HBM linear strain gauges [12], labeled S1 and S2, one on the BHE-soil interface attached to the BHE and the other on the outside of the barrel are installed to monitor the thermally induced strains (Fig. 2). The strain gauges have an electrical resistance of 120 Ω and a transverse sensitivity of 0.1%. The strain measurements are recorded using a datalogger unit connected to the PC.
Four VW push-in pressure cell stress sensors [13], numbered P1 to P4, are used to monitor and record the thermally induced lateral stresses and pore-gas pressures in the soil (Fig. 2). The pressure sensors are attached to a DT2040 40 channel VW and thermistor datalogger unit connected to the PC, and have a measurement capacity of 350 kPa, and a minimum measurement resolution of ±0.1 kPa.
And finally, eight CS640 three-rod 7.5 cm long TDR probes [14], numbered M1 to M8, are used to monitor and record the volumetric moisture contents in the soil (Fig.  2). The moisture content measurement set-up also includes a 12 Vdc power supply, a TDR100 time domain reflectometer, a SDM8X50 8-channel solid state 50 Ω coaxial multiplexer, and a CR800 measurement and control datalogger unit connected to the PC. The CS640 TDR set-up is capable of measuring moisture content of soils with bulk electrical conductivity of up to 5 dS/m, and has a timing resolution of 12.2 ps.

Materials used
An air dried naturally occurring fine sand was used as the heat storage medium. Table 1 shows a list of the obtained geotechnical properties of the soil, and in Figs. 3 and 4, results of X-ray diffraction (XRD) mineralogical analysis and grain size distribution of the soil are shown, respectively. The fine sand particles are primarily composed of quartz and albite minerals (Fig.  3).
A commercial cement-based material, named ThermoCem PLUS [15], was used to construct the BHE. After preparing the BHE using a fresh mix of the ThermoCem PLUS powder and water with a water to solids ratio of 0.802, the BHE was cured for a minimum of 28 days to complete the cement's exothermic hydration processes and achieve its operational geotechnical properties. In Table 2, a summary of the obtained physical properties of the cemented BHE material are presented. And in Fig. 5, the set-up after the installation of the soil, BHE and the sensors is shown. Unified soil classification system (USCS) SP (2) (1) : measured using Decagon KD2 Pro device (2) : poorly graded sand Table 2. Physical properties of the BHE cemented material.

Experimental procedure
After an initial circulation of water without heating to remove air from the piping system, the heating phase theating of the experiment was performed at the Tmax 70°C for around 4 hours. The duration of heating was kept relatively short in order to avoid heat reaching the lateral boundary (plastic barrel), thus avoiding extra thermally induced strains of the plastic barrel and replicating actual field BTES conditions. After the heating phase, the flow/heating of the fluid was stopped to start the passive cooling tcooling process and the data was recorded for around 46 hours. The loading procedure followed for the experiment is schematically shown in Fig. 6. Figure 7 shows the time plot of the temperature T distributions in the heat storage soil. For the given heating duration and set heating bath temperature Tmax (70°C), a maximum temperature of around 35°C is recorded at the BHE-soil interface, sensor T9 (Fig. 7, left), which is significantly lower than the average maximum temperature measured at the center of the cemented BHE (59°C), indicating a significant heat loss in the BHE, soil and the BHE-soil interface due to the high thermal resistance or low thermal conductivity of both the BHE cement and the surrounding soil. Whereas, the heat loss in the water carrying pipes (i.e. from the heating bath to the BHE inlet) is generally around 4°C, which is comparatively much lower than the heat losses recorded inside the BHE, soil and BHE-soil interface.

Temperature distribution
The plots also show no considerable increase of temperature at the sensors placed near the outer lateral boundary of the system as assumed earlier (Fig. 7, right) (i.e. similar temperatures recorded at the sensors placed in the soil near the plastic barrel boundary, sensors T11, T14 & T17, and the sensor placed outside the barrel, sensor T30, which represents the room temperature).

Stress distribution
In Fig. 8, the thermally induced lateral stress profiles at the four sensor positions are shown along with their respective temperature profiles. As expected, an increase in the measured thermally induced stress σinduced is recorded at sensors P1, P2 and P4, following the shape of the time-temperature profile. The maximum value of σinduced is recorded at the BHEsoil interface sensor P1 with a value of around 27 kPa, and no significant σinduced is measured at the outer boundary sensor P3. The generated stresses are primarily attributed to the expansion and rearrangement of the quartz and albite dominated soil grains.
It should be noted that the generated stresses in the heat storage unit may not be isotropic, with the generation of induced stresses in the axial or vertical direction expected to be of different magnitudes to the radial or lateral stresses shown in this study. Due to the direction of installation of the stress sensors used in this research, however, it was not possible to measure these axial/vertical stresses. Nevertheless, considering the direction of heating, the heat insulation system and boundary conditions provided in this study, the magnitudes of the induced vertical stresses may be lower than the lateral/horizontal stresses reported in this study.
In contrast to the thermally induced stresses, very insignificant changes in the generated gas pressures Uinduced are recorded, with only around < 1 kPa increase measured at the BHE-soil interface sensor P1. However, unlike in this study, which was performed on a soil at or near dry moisture condition, the generation of excess gas or pore water pressures may be significant in the case of partially or fully saturated heat storage soils/systems.

Strain distribution
The highest thermally induced strain εinduced is recorded at the BHE-soil interface sensor S1 with an average maximum value of around 270 μm/m, once a peak temperature condition is reached in the BHE (Fig. 9, left). As expected no strain is generated at the strain sensor attached to the outside of the barrel (Fig. 9, right).

Moisture distribution
In Table 3, the averages of the volumetric water content θ measurements taken by the eight TDR moisture probes M1 to M8 throughout the test duration are shown. Overall, no significant changes in moisture content were observed at all sensor locations during the heating/cooling operation. Comparatively speaking, similar average volumetric moisture content θavg measurements were recorded by all the moisture sensors ranging at near dry conditions between 0.3 and 1.3% (corresponding to degrees of saturation Sr between 0.9 and 3.9%), indicating the homogeneity of the used soil.

Conclusions
An experimental lab-scale prototype heat storage system was set up to study the transient thermo-mechanical behavior of sensible heat storage systems. The findings of the experiment, which included heating of the storage unit up to a set temperature of 70°C followed by a passive cooling and monitoring the thermally induced strains and stresses of the system, indicated a significant, but tolerable, increases in the measured thermally induced strains and stresses in the heat storage unit, particularly at the borehole heat exchanger-soil interface, where maximum values of induced strain of around 270 μm/m and induced stress of around 27 kPa were recorded. The induced stress level, which is mainly attributed to the expansion and rearrangement of the quartz and albite dominated soil grains, is equivalent to that generated by an over-burden soil pressure of around 1.5 m depth in field conditions, and would not be expected to cause a significant impact on the safety and operation of most sensible heat storage systems.