The Thermal Response of a Packed Bed Thermal Energy Storage System upon Saturated Steam Injection Using Distributed Temperature Sensing

Abstract

The effectiveness of a thermal energy storage (TES) system is typically characterized with the help of thermal stratification or temperature gradients along the direction of heat injection, which is typically the flow direction of heat transfer fluid. The steepness of temperature gradients are a direct indicator of the effectiveness or efficiency of the heat storage or dispatch process. The temperature gradient evolution along the packed bed of ceramic particles upon saturated steam injection is presented in this work. Distributed temperature sensing based on optical frequency domain reflectometry was deployed in a packed bed of ceramic particles to capture the thermal front evolution in the axial direction. The physical processes accompanying steam injection in packed beds are complex due to phase change, transitioning two-phase flow, and changes in condensate accumulation. Therefore, the variation of thermal response of the TES system for various steam injection flow rates was experimentally studied using a high-resolution distributed temperature sensing system in a chemically inert alumina particle-packed bed. Distinct zones of different heat transfer modes were observed during the steam injection experiments. A distinct conduction zone, evident from diffuse thermal fronts, was observed at low flow rates, and these thermal gradients became sharper as the flow rate increased. The diffuse thermal fronts in the heat storage media suggest a low exergy efficiency of the TES system, as energy losses started initiating before a significant fraction of the bed was saturated with steam.

1. Introduction

With the growing disparity in daily or annual energy consumption patterns, the demand for energy storage has been continuously increasing. Thermal energy storage (TES) systems have been under consideration as an alternative to batteries for storing energy and are particularly suitable for providing thermal energy for domestic or industrial applications [1]. TES systems have a diverse set of applications and compatibility challenges with various energy generation processes including nuclear power and solar power [2,3]. These TES systems have varying performance depending on the application and the associated thermal process conditions. The performance of these TES systems can be defined in terms of macroscopic parameters such as the exergy efficiency, energy efficiency, or energy density [3]. These performance metrics are dependent upon the detailed thermal behavior of the system during charging/storage or discharging/recovery processes, including wall effects on energy losses, the dispersion and distribution of temperature, or flow-pressure behavior. These detailed thermal responses are directly affected by the choice of the size and shape of the storage medium, by the choice of heat transfer fluid (HTF), and by the operating thermal-hydraulic conditions [4,5,6,7,8].

In several experimental studies, detailed spatio-temporal thermal characteristics have been modeled and experiments designed to validate these models with the objective of evaluating the feasibility of a TES system [9,10]. These studies have been performed with both solid and liquid storage media, with a sensible or latent heat type of TES. However, the significantly important HTF steam has not been experimentally studied extensively using high-resolution or high-fidelity data [11]. Several thermal power plants, process industries, and domestic heating systems use steam as the heat transfer or working fluid. Therefore, storing excess thermal energy from process steam can play an important role in making thermal energy dispatch more flexible.

The ideal design of the TES system requires the HTF to transfer thermal energy to storage media and exit the system at conditions close to ambience. In order to bring the storage medium to the top temperature while maintaining conditions close to ambience, a sharp thermal front should propagate through the storage medium as the HTF flows through the system. This sharp front propagation concept is depicted in Figure 1. High-temperature fluid (shown in red) enters the cold (shown in blue) from the left. The worst possible scenario of inefficiently storing thermal energy is presented as a well-mixed scenario shown on the left side. Under a well-mixed scenario, the entire TES system can be brought to the top (red) temperature, but the system will discharge HTF at much higher temperatures than ambient conditions at exit, thereby wasting large amounts of energy. A third alternative is depicted on the right end of Figure 1, showing that thermal diffusion or thermal dispersion causes an increase in heat losses from exit before the entire system is brought to a high temperature. The ideal sharp front is also sometimes mentioned classically as plug flow in the literature [8,12,13].

Packed bed systems are often considered realistically closest to the ideal plug flow. Due to these reasons, packed bed systems with randomly packed solid particles or rocks have been considered for TES system design. The randomly packed beds of ceramics such as alumina or solid rocks, due to their high energy density, high durability, and high local thermal diffusivity, make them leading candidates for energy storage, which have been tested with air, gases, and liquid-type HTFs [8,14,15,16]. Although the design and assembly of such systems is simple, their performance and compatibility with applications are dependent upon the temperature distribution, which is very sensitive to material properties and operating parameters. Thus, understanding the process using high quality experimental data with detailed mathematical models is essential to design an optimum system. Previously theoretical and experimental studies on the performance of randomly packed storage media of ceramic particles focused on evaluating temperature gradients along the flow direction. Most of these studies model the temperature response of the system in the axial or radial direction upon the injection of hot fluid or cold fluid into the bed during storage or recovery cycle, respectively. Thermal dispersion gives insight on how the type of fluid and the flow rate of the fluid influence buoyancy-driven thermal stratification, volumetric energy utilization, and exergy destruction due to temperature gradients. The type of material and the geometry play a crucial role in the thermal dispersion within the beds. Although there are several modelling and experimental studies on thermal storage in packed beds, the validation lacks the use of high-resolution experimental data. Typically these test systems deploy discrete thermocouples that are sparsely located in space [14,16]. Moreover, as mentioned before, these types of studies have been conducted with non-condensing HTFs, but there are very limited experimental studies with a significantly important HTF-saturated steam.

Direct steam condensation in packed beds of solids has been studied and tested with different perspectives, e.g., for separating water vapor from air [17,18,19]. Thermal dispersion in a porous media results from a combined effect of conduction or heat diffusion and convection. The large surface area provided by the numerous small particles allows for the HTF to transfer its energy to the solid particles in a radially uniform manner. This ensures a stable condensation flow rate from the packed bed. Upon steam injection, the bed can be divided into three zones: a hot vapor zone, a mushy or liquid-vapor mixture zone, and a cold liquid zone. Due to the transient nature of this problem, the relative thickness of these zones will change continuously. An inverse formulation of this problem was presented by Woods et al. [20], who studied the thermal response of injected cold liquid inside a bed of hot rocks, which leads to the formation of moving vaporization front. Conduction in the rocks, advection in liquid media, and the evaporation/phase change of the fluid have all played a dominant role in defining the thermal transport. Condensing steam flow in a packed bed involves similar multi-physics processes, such as condensate nucleation, surface capillarity effects, condensate or liquid percolation, time-scale dewetting versus heat diffusion in solids, and adverse density variations in the bed.

As the steam flows and condenses over the packed bed of rocks, the heat is transferred from the fluid (liquid or vapor) to the solid particles through convection and within the solids through conduction. The steam injection process in a much colder packed bed is associated with continuous phase change, which implies a large amount of energy injected per unit time. Unless the steam injection flow rate is very low, it will be difficult to observe the distinguishable effects of energy deposition in the packed beds; in other words, those beds can become saturated very quickly. The condensate percolation is expected to be very complex, as will the associated cross-linked thermal conduction due to these condensate channels. Understanding the effect of the HTF flow rate or the charging rates on thermal dispersion becomes important, as the higher thermal dispersion can lead to energy losses in the form of hot fluid leaving the outlet, thereby reducing the exergy efficiency of the system [8,16]. Edwards et al. [11] conducted experiments to investigate the effect of the steam flow rate on the thermal dispersion in a packed bed and showed that, as steam flow injection is decreased, thermal dispersion in the bed increases. However, these results were obtained using sparsely distributed multi-point thermocouple assembly. These few thermocouple sensor locations do not sufficiently data for an understanding of the thermal gradients or the development of future models or designs. Current work extends previous work with distributed temperature sensors that allow for temperature data at a much higher spatial resolution and thus the experimental evaluation of thermal gradients. This paper presents results from the steam charging experiments at flow injection rates in the ceramic packed bed and their impact on the evolution of axial thermal gradients.

2. Design Objectives

The design of this experimental set-up is similar to previous experiments that utilized a cylindrical vessel for a thermal energy storage system constituting a randomly packed bed of alumina particles [8,11,13,21]. Alumina particles are used as the packing media, as they have a high specific heat capacity, a high thermal conductivity, and chemical inertness, allowing for a rapid localized equilibriation of thermal energy between the fluid and solid phases. The chemical inertness allows the alumina particles to be utilized for extended amounts of time, while withstanding multiple charging–discharging cycles. With saturated steam as the heat transfer fluid, the rate of heat injection is much faster during the condensation process. However, a large heat transfer surface area due to considerably smaller particles or packing size, as compared to the overall bed dimensions, makes the thermal front propagation very steep along the flow direction. Moreover, the results of previous material behavior studies [22] showed no changes in the physical characteristics of alumina particles under steam condensation and boiling cycles. Therefore, the choice of alumina particles meets the design objective of performing reproducible thermal behavior tests.

3. Experimental Setup

The design of the experimental set-up was extended from two previous experimental studies [8,15].

A simplified schematic for the experimental set-up is shown in Figure 2. A cylindrical quartz tube randomly packed with spherical particles was chosen as the main test chamber for this experimental set-up following the design objectives described in the previous section. The size of the tube was 1 ft or 30.48 cm tall and 6.35 cm in diameter, and the design limitations on the size were only on the diameter because of the standard ceramic flanges used to seal the ends and size of the particles. The size of alumina particles was optimally selected to be 3 mm, as examined in previous studies based on a balance between thermal performance and pressure drop requirements [15]. The ratio of the tube diameter to a particle diameter greater than 20 allows for the plug-flow assumption or radially uniform dispersion of the thermal front. The thermal storage test media were spherical alumina particles procured from Norpro with the commercial name Denstone-99 particles. The resulting porosity of the bed was approximately 40%. These commercially available spherical particles were considered for this experiment because they allow for uniform isotropic heating and have been tested for their chemical inertness and robust thermo-mechanical behavior with steam [22].

An electrical steam generator was used as a steam supply at the regulated pressures by using a pressure regulator. Steam was injected into the test section from the top at a regulated pressure. The flow rate of the steam was controlled by adjusting a globe valve at the outlet of the test section. The system was filled with water (at 25 °C) before the steam injection experiments. The global valve at the bottom of the test section controlled the constant water flow rate exiting the system.

The main interest in this paper is to study axial dispersion, which is presented later, but to have more clarity on the thermal front propagation, multiple measurement techniques were deployed. X-ray imaging was used to monitor the shape of the steam condensing front, and a thermal camera was used to observe any azimuthal or radial asymmetry. X-ray images presented in Figure 3 confirm that the condensing steam front was planar. Infra-red (IR) camera images show that the thermal front propagation was uniform along the azimuthal direction, as the temperature of the outer wall was almost uniform in the azimuthal direction. FLIR A 655sc IR camera was used to obtain surface thermal maps with a spatial resolution of 10 mm at 200 frames per second. X-rays were emitted using a generator tube capable of producing an output power of 2.8 kW, a current range from 0.4 to 100 mAs, and an energy range from 40 to 120 kV. The settings used for conducting these experiments were 40 mAs and 50 kV. A high-speed X-ray imaging camera (XRD 0822 AO/AP) with a Gadox scintillator was used to monitor uncolloided X-rays. Some of the features of this X-ray imaging camera/system are as follows: 1 Megapixels, a 0.2 mm pixel pitch, 64,000 gray levels, and compatibility with X-rays from 20 KeV to 15 MeV and with up to 100 frames per second. It should be noted that both X-ray and IR cameras were used to confirm the planar steam condensing front and minimal azimuthal variation, but the objective of these experiments was to obtain axial thermal gradients with distributed temperature sensing, described in the next section.

3.1. Optical Frequency Domain Reflectometry

The temperature distribution in the experiments were measured using optical fiber-based temperature sensors. The principle of operation is based on optical frequency domain reflectometery (OFDR), where any localized strain in the uniformly doped optical fiber due to temperature change is detected by a shift in the optical frequency of the Rayleigh backscattered signal. This technique is also called swept wavelength interoferometry (SWI) and has a clear advantage in spatial resolution over more traditional Bragg grating-based OFDR sensors. The unique features of this system are that the spatial resolution can be as high as 1 mm, and the data acquisition rate can reach 250 Hz [23,24]. SWI captures the temperature gradients in the bed, which is essential to understand this problem. The high-speed distributed temperature sensing (DTS) measurement system with optical-fiber based sensors was assembled and calibrated before its use in the system.

3.2. Experimental Procedure

To begin the experiments, the packed bed test section was filled with water at 20 °C. The steam was then injected into the bed from the top by controlling the inlet flow rate and keeping the exit valve at the bottom in the throttled position.

The steam injection experiments were performed for a range of inlet steam flow rates governed by the globe valve opening. The first set of experiments consisted of the slowest injection experiments, where a regulated steam supply at atmospheric pressure was further throttled by a globe valve to nearly the lowest flow rate. The pressure regulation to atmospheric pressure was confirmed with the experimental data of a steam temperature of 100 ± 2 °C.

The steam injection experiments were performed for a range of inlet steam flow rates governed by the globe valve opening. The same procedure was repeated as described for the slow injection case. The injection flow rate of the steam for each case was determined by measuring the condensate collected per unit time. Multiple experiments were performed to ensure repeatability and consistent flow rate values for both the slow and fast measurements. Compressed air was used after the steam supply was shutoff to remove trapped condensation in the packed bed for complete measurements. In addition to the steam injection experiments, hot water (50 °C) injection experiments were also conducted in the test section filled with only tap water (no alumina particles) and the packed bed filled with tap water (at 20 °C).

4. Results and Discussion

The effect of the packed bed on axial thermal dispersion. In order to demonstrate the efficacy of the packed bed on reducing the temperature dispersion, the thermal response comparison of the test section upon hot water injection was studied with and without packing material. Experiments were conducted where hot fluid was injected into an initially cold test section and the thermal response was recorded. The thermal response measured by the OFDR system is shown in Figure 4. When the hot fluid was injected from the top of the test section with no packing material (Figure 4), a prominent thermal dispersion was observed. The lack of packing material caused significant fluid mixing in the test section, thereby causing the outlet temperature of the test section to increase while the rest of the bed remained highly unsaturated with the hot temperature fluid. On the other hand, thermal mixing due to larger eddies was avoided with the presence of packed media, which resulted in smaller thermal fluctuations comparable to molecular thermal diffusion.

Steam injection experiments. The propagation of the temperature front for each of the cases with varying injection rate while penetrating the randomly packed bed was experimentally studied using distributed temperature sensing based on the high-resolution optical frequency domain reflectometry. The transient thermal response of the packed bed at different flow rates of steam injection is discussed in this section. The thermal response of the packed bed was governed primarily by two thermal transport mechanisms—advective heat flux and conductive heat flux within the solid–fluid media. Near the entry port at the top, where the steam was introduced, the bed and fluid temperature became almost equal to the steam inlet temperature or the saturation temperature. With the steam supply continuously available, irrespective of the injection rate, the bed temperature at the top was always maintained at the saturation temperature. This constant bed temperature at the top will transport heat from the top to the bottom of the bed due to the non-negligible thermal conductivity of alumina particles and water holdup in the bed through the heat conduction mechanism. Simultaneously, due to steam injection in the bed, some amount of energy is carried as the HTF moves in the bed through the advection mechanism.

At a slow injection rate, the initial rise in the temperature at axially farther locations will be dominated by the conduction mechanism. As the steam or two-phase mixture front, which is at temperature near the saturation temperature, reached those regions located far away from injection point, there was a sudden change in the temperature. This effect can be seen from the change in the temperature profile or temperature gradients as shown in Figure 5.

The thermal response of the test section for four different steam injection rates is shown in Figure 5. The temperature was non-dimensionalized with the initial bed temperature, $T_0$, and the saturation temperature, $T_{sat}$. For the non-dimensional temperature, $\theta$, a value of 0 represents the initial bed temperature, and 1 represents the saturated steam temperature. The time was non-dimensionalized with respect to the velocity, and the height axial direction was non-dimensionalized with respect to the total height of the test section (see

Table 1. Definition of parameters.

Property	Symbol
Density	$\rho$
Thermal conductivity of bed	k
Mass flow rate	$\dot{m}$
Flow area	A
Mass Velocity	$G=\dot{m}/A$
Bed length	L
Max. temperature difference	$\Delta T$
Non-dimensional height	$z/H$
Non-dimensional Temperature	$T-T_0/T_s-T_0$
Non-dimensional time	$t/t_{max}$
Wall heat transfer coefficient	$h_w$
Wall heat loss flux	$q_w^{″}=h_w\Delta T$
Specific heat of water	$C_p$
Latent heat of vaporization	${\lambda }_{lv}$
Injected heat flux	$q_i^{″}=G(C_p\Delta T+{\lambda }_{lv})$
Conductive heat flux	$q_c^{″}\sim k\Delta T/L$
Fractional wall heat loss	${\beta }_w=q_w^{″}/q_i^{″}$
Modified Peclet number	$Pe=q_i^{″}/q_c^{″}$

).

At the slowest steam injection velocity (

Table 2. Parameter values for the experimental tests.

Case	Test 1	Test 2	Test 3	Test 4
$\dot{m}$ (kg/h)	0.39	1.00	1.29	1.61
$q_i^{″}$ (W/m${}^2$)	$3.5\times {10}^5$	$9.0\times {10}^5$	$1.1\times {10}^6$	$1.4\times {10}^6$
$h_w$ (W/°C-m${}^2$)	5	5	5	5
$\Delta T_w$ (°C)	80	80	80	80
$q_w^{″}$ (W/m${}^2$)	400	400	400	400
${\beta }_w$	0.01	0.01	0.01	0.01
$Pe$	56	144	185	231

) of 0.1 mm/s, the thermal response, shown in Figure 5, exhibited a region of an intermediate zone between the saturated (hot) bed and the cool bed. This region grew with the progression of time. This suggests that the thermal transport through the conduction mechanism transcends the thermal transport through the advection mechanism. Therefore, before the bed was completely saturated with the high temperature steam, the dominant conduction mechanism caused hot fluid to leave from the outlet of the test section. This fluid exiting the bed at a temperature higher than the initial cold temperature of the bed before the entire bed was brought to the top temperature and led to energy loss through the outlet, thus reducing the overall exergy efficiency of the system [8].

As the steam injection rate increased, the advection mechanism became more dominant, while the conduction mechanism remained constant, as it was dependent on the static bed properties, and was independent of the steam flow rates. The higher advection term implies that the total amount of influx enthalpy carried by the steam was much higher; thus, as the fluid stream moves through the bed, it equilibrated the bed to the saturation temperature at an almost constant rate at all spatial locations. Because of a much higher rate of enthalpy injection in the bed due to the advection term, the effects of conduction will not have much of an impact on the rate of temperature increase in the bed. The results in Figure 5 confirm this explanation, as evident by the decreasing width of the mixture or intermediate zone.

5. Conclusions

Thermal dispersion experiments inside a randomly packed bed thermal storage filled with alumina particles were conducted with the help of OFDR-based DTS systems. The observations show that steam injection at different flow rates results in distinct temperature dispersions, as compared to single phase heat transfer fluids such as water. The axial dispersion of the thermal front was governed by both conduction and advection effects. At low steam injection flow rates, the conduction effects became apparent in the latter section of the bed, as the steam front or the advection current took more time to reach those locations. At slow injection rates, the experiments showed more prominent conduction-based thermal fronts, as compared to faster injections. At these slow injection rates, an increase in the outlet temperature was found at very low bed saturation levels. This level of bed saturation increased asymptotically as the steam injection velocity increased. These differences in thermal responses indicate that slow steam injection velocities can lead to inefficiencies when charging a TES system. Above a critical steam injection velocity, the variation in the fraction of the bed saturated did not change significantly, suggesting that designers of TES systems should consider an optimum charging rate to minimize efficiency loss and minimize the frictional pressure drop.