International Journal of Mechanical & Mechatronics Engineering IJMM E-IJENS Vol:10 No:01 7 Temperature Profile and Thermocline Thickness Evaluation of a Stratified Thermal Energy Storage Tank Joko Waluyo1) and M Amin A Majid2) continuous profile of temperature distribution. Difficulty in determining thermocline thickness arises for the case of discrete temperature data, since the profile formed could not be used to estimate the thermocline thickness. This paper discusses a practical method for formulation of thermocline thickness of stratified thermal energy storage. Curve fitting by iterative method was adopted to identify the functions which could represent the S -curve of temperature distribution. Based on the functions, thermocline thickness was formulated using functional relationship of temperature profile. Results identified two functions which could represent S -curve of temperature distribution, namely sigmoid dose response (S DR) and four parameter sigmoid (FPS ) functions. Both functions were observed to well fit the temperature distributions having coefficient determination more than 0.99. Based on evaluations the formulations were capable to be utilized for evaluation of thermocline thickness of the stratified TES . The methods offer an advantage to obtain an exact value of thermocline thickness. Index Term-- temperature distribution profile, thermocline thickness, S tratified thermal energy storage. I. INT RODUCT ION Thermal Energy Storage (TES) systems are useful for maximizing the thermal energy efficiency for meeting the fluctuating cooling demands by shifting energy use from on peak to off-peak hours. This is achieved by charging the TES tank during the off-peak hours and discharging it later during the peak hours. Many studies have been undertaken related to stratified TES tanks. An important parameter in evaluating performance of charging and discharging of a TES tank is the thickness of the thermocline being formed. A thinner thermocline is desired since a thicker thermocline indicates larger degradation of stratification [1]. The thickness of the thermocline indicates the extent of mixing occurred due to inflow streams during the cycles . This factor influences the degradation of stratification, beside heat transfer losses from the tank [2]. Thermocline thickness is determined based on water temperature distribution inside the tank. The water temperature distribution profile formed could move either upward or downward during charging or discharging cycles [3, 4]. Many researches on thermocline thickness parameter in relation to stratified TES tank performance have been undertaken based on continuous profile of water temperature distribution. 1) Mechanical engineering department, Universiti T eknologi PET RONAS, Bandar Seri Iskandar, Tronoh, Perak, Malaysia. (phone: +60165943445), e-mail: [email protected]). 2) Mechanical engineering department, Universiti T eknologi PET RONAS, Bandar Seri Iskandar, Tronoh, Perak, Malaysia. (e-mail: [email protected]). Using continuous profile, thermocline thickness is accurately identified as asymptote regions with limit points located on the edge of profiles [5, 6, 7, 8, 9]. The difficulties of measuring thermocline thickness arise if the temperature is available as a discrete data resulted from long interval-distance of temperature sensors. The un-continuous profiles that are formed could not be used to determine the thermocline thickness due to its ambiguity in defining the limit points. A method to determine thermocline thickness from discrete data of temperature distribution is investigated in this study. The study focuses on developing an approach of establishing the profile by adopting fitting method and using the profile to formulate the thermocline thickness. Data acquired from an operating TES tank equipped with long interval-distance sensors recorded based on hourly basis, was used in this study. II. M ETHODOLOGY Normally water temperature distribution in the stratified TES tank consists of 3 regions with warm water at the top, cool water at the bottom and thermocline region in the middle. The water temperature forms S-Curve profile consisting of two asymptote curves as shown in Fig. 1. Average cool and warm water temperature is formed by the asymptote values of Tc and Th . Position of the thermocline, C, defines the boundary line of cool and warm water in the tank. It also can be interpreted as the cool water depth occupied in the tank. Thermocline thickness, WTC, is determined as the region limited by the edges of asymptote curve. Th Tank Height, H Abstract— Determination of thermocline thickness requires a WTC C Tc = Th = WTC = C = Average cool water temperature Average warm water temperature Thermocline thickness Midpoint of thermocline Tc Fig. 1. S-curve of temperature profile [7] The temperature profile formed could be represented as a function (1) with one variable of x and four parameters of Tc, Th, C and S. (1) T ( x) f (Tc , Th , C, S , x) The temperature distribution in equation (1) was used as a basis for determining thermocline thickness in this study. The analysis steps are as follows: 108601-2424 IJMME-IJENS © February 2010 IJENS IJE NS International Journal of Mechanical & Mechatronics Engineering IJMM E-IJENS Vol:10 No:01 Case I (flow rate 393 m3/hr) 16 18.00 19.00 14 20.00 Temperature ( C) 12 o S-curve profile using iterative method. Identifying the temperature profile functions. Determining the parameters in the functions. Formulation of thermocline thickness. Evaluation. Temperature data of the TES system of a district cooling plant were acquired for the study. The TES system consists of two 1,250 tons of refrigeration (RT) of steam absorption chillers (SACs) and four 325 RT electric chillers (ECs) and one 5,400 m3 storage TES tank with designed capacity of 10,000 RTh. Inlet nozzle is made from 20” NPS located at elevation 3.4 m height, while outlet nozzle is 12” NPS at elevation 12.3 m. Both nozzles are provided with diffuser on its end-connection in the storage tank. Overflow line is connected at elevation of 14.025 m. The entire tank is externally insulated. The tank is equipped with 14 temperature sensors, installed at approximately 1 m vertical interval, to measure the water temperatures. The lowest temperature sensor is located at 0.51 m height. All temperatures are hourly recorded with acquisition data system. The schematic flow diagram of the system is shown in Fig. 2. 21.00 10 22.00 23.00 8 24.00 6 01.00 02.00 4 0 2 4 6 8 10 12 14 Sensor Elevation (m ) (a) Case II (flow rate : 524 m3/hr) 16 18.00 19.00 14 20.00 12 21.00 o iv. v. vi. vii. profile. This profile move upward from hour 18.00 to the final condition at 02.00. Higher movement of the profile of case II is occurred than of that for case I. These movements, therefore, reveal a more temperature distribution span of case II compare to case I. Temperature ( C) i. Acquiring temperature data from an operating TES. ii. Plotting the temperature distribution. iii. Fitting of the functions that could represent the 8 22.00 10 23.00 8 24.00 01.00 6 02.00 4 0 2 4 6 8 10 12 14 Sensor Elevation (m) (b) Fig 3. T emperature distributions for (a) case I and (b) case II Fig. 2. Schematic flow diagram of charging cycle TES tank is charged by the ECs during off-peak hours. Normally, the charging is served by three or four of ECs. For the purpose of this study, hourly temperature records during charging period of 9th September 2008 and 15th April 2009 were used. The charging cycle was operated continuously from 18.00 hours to 02.00 hours of the following day. These two charging cycles were served by 3 and 4 units of ECs, and this is represented as case I and case II, respectively. The flow rates of the charging cycles for case I and case II were 393 m3 /hr and 524 m3 /hr, respectively. III. RESULT AND DISCUSSION Temperature distribution data The plot temperature distribution for case I and II are depicted in Fig. 2 (a) and (b), respectively. The hourly temperature distribution within charging periods are presented with respect to sensor elevation in the TES tank. Each hourly charging course form a continues S-curve Fitting of the functions Data as depicted in Fig 3 (a) and (b) were used to fit some possible functions to represent the S-curve temperature profile. The functions consist of parameters Tc, Th , C and S. Fitting was done by utilizing commercial software of Sigmaplot [10] using non linear regression by iterative method. Two functions were identified that could represent the S-curve: sigmoid dose response (SDR) and 4 parameters sigmoid (FPS) function. The first function was formed as a modification from the sigmoidal dose response (variable slope) function and the second function was obtained from the modification of 4 parameters sigmoid function. SDR function form as the following, Th Tc T Tc 1 10( C X ) S1 FPS function is expressed as, T Tc Th Tc 1 e(C X ) / S2 (2) (3) Both SDR and FPS functions relate temperature distribution to one variable of X and the four parameters. SDR function has four parameters of Tc, Th , C and S 1 . While FPS has parameters of Tc, Th , C and S 2 . Parameters Tc and Th are cool and warm temperatures (o C). X variable expresses the dimensionless elevation (x.N/H), where x is the elevation of the temperature sensors (m), H is effective 108601-2424 IJMME-IJENS © February 2010 IJENS IJE NS International Journal of Mechanical & Mechatronics Engineering IJMM E-IJENS Vol:10 No:01 height of the tank content of water (m) and N number of stratified layers. Parameter C implies of a dimensionless elevation unit. S 1 for SDR and S 2 for FPS are constant parameters related to slope parameter of the functions. The plotting of fitting profile are presented in Fig 4 (a) and (b) and Fig 5 (a) and (b) for SDR and FSP, respectively, covering of case I and II. Referring to Fig 4 (a),(b) and 5 (a),(b), it is indicated that the temperature in the mixing region fitted to the asymptotes curves, and the remains approached to the flat line. It can be seen that the adjusted temperature resulted to more clear demarcation of the temperature in the mixing region of the asymptotes curve following the fitting functions. Indicated from Fig. 4 (a) and 5 (a), both SDR and FPS functions have the same curvature profiles in the fitting of temperature data. Similar curvature profiles were also noted for 4 (b) and 5 (b) in the fitting of case II. The coefficient of determination, R2 of both function, depicted in Table I and II, are greater than 0.99, indicating that the temperature data were fitted well to the functions. Determination of parameters Parameters of Tc, Th , C and S were obtained from the fitting of SDR and FPS functions. The parameters of both functions are tabulated in Table I and II, covering both case I and II. The values of R2 , coefficient of determination, for evaluation the goodness of fitting are also provided. From Table I and II, it is noted that fitting of case I using SDR and FPS functions revealed the same values of Tc, Th , and C on each hourly charging cycle. On the other hand , it is observed that the parameters of S 1 and S 2 are different. Similar observation was found for the case II as well. It is noted that SDR and FPS have similar profiles and parameters of Tc, Th , and C in the fitting. In addition, the values of coefficient of determination, R2 , are equal for SDR and FPS functions in the fitting of cases I and II. This indicates that both SDR and FPS function are enable to fit the S-curve of temperature distribution. Either one of them can be chosen for the implementation. The different of these functions is due to usage of the basis number 10 and natural number, e. As indicated in Table I and II, the values of Th and Tc decrease for both case I and case II. The decreased in values of Tc was due to incoming supply of cooler water at the bottom of the tank from the ECs. While the decreased in values of Th , was due to diffusion of thermocline region. The values of parameter C are also shown in Table I and II. For both cases the values of C increase with charging time, with case I smaller than case II. The trends occurred due to increased cool water depth as a result of more cool water was generated with charging time. These trends were also noted by (Nelson et al, 1998) and (Karim, 2009) through their experimental investigations on stratified tanks. Formulation of thermocline thickness By identifying the function of temperature profile, thermocline thickness could be formulated based on functional relationship. The concept for determining the thermocline thickness was S-curve temperature profile as illustrated in Fig. 1. The formulation the thickness of thermocline was achieved refer to the function of temperature profile. 9 Formulation of thermocline thickness using SDR function was obtained by rearranging equation (2) into the form Th Tc 1 10(C X ) S1 T Tc (4) The left term of equation (4) is re-arranged using dimensionless cut-off temperature =(T-Tc)/(Th -Tc), Musser (1998b), describing the limit point of the thermocline thickness. 1 (5) 1 10( C X ) S Distance from C to X express the half- thickness of the thermocline, and represented as follows: 1 Log ( 1) (6) CX S1 Therefore, thermocline thickness is defined as 1 WTC 2.Log ( 1 1) (7) S1 Using the similar analysis, thermocline thickness for FPS function is determined as 1 (8) WTC 2.Ln.( 1).S 2 Equation (7) and (8) were used to determine thermocline thickness for a predetermined value of . Conceptually, the dimensionless cut-off ratio values are in the range of 0 to 0.5 covering minimum and maximum thermocline thickness. For =0 indicate that the thermocline edges profile are located at Tc and Th , therefore gives a maximum thickness. With =0.5, the limit points are at midpoint of thermocline region resulting zero values of the thickness. The evaluations were conducted by performing calculation using varies of from 0 to 0.5, with parameters S 1 and S 2 as described in Table I and II, for SDR and FPS function, respectively. Evaluated results of thermocline thickness of SDR function are tabulated in Table III and IV, for case I and II, respectively. In the calculation using FPS function, the result has similar value with maximum deviation of 1.5 % from SDR values. This is due to the usage of rounded value up to 2 digits decimal of the parameter. If it is required to reduce the deviation, us ing more decimal digits of the fitting parameters is recommended. From Table III to IV, it is noted that higher thermocline thickness occurred at the lesser values of . Thermocline has maximum thicknesses at =0.001. In addition, with equal to 0 and 0.5, revealed ∞ and 0, respectively. Therefore, range 0<<0.5 can be chosen for its implementation, depending on the appropriateness of the requirement. The smaller value takes advantage of expressing the real thermocline thickness in the cycles. It is highlighted that the above method is justified as a practical method to determine the thermocline thickness. This is achieved by converting the discrete data of temperature distribution to a continues profile content of parameters average cool and warm water temperature, midpoint of thermocline position and slope gradient. Thermocline thickness was then determined using functionally relationship of the temperature profile function. In the method, the goodness of fitting of the discrete data has significant role to the result in reflecting the thickness of 108601-2424 IJMME-IJENS © February 2010 IJENS IJE NS International Journal of Mechanical & Mechatronics Engineering IJMM E-IJENS Vol:10 No:01 thermocline. Determination of thermocline thickness based on functional relationship takes beneficial in generating exact values of the parameters rather than estimation. The above calculation was obtained based on temperature profile with has coefficient of determination more than 0.99 to the acquired data. This indicates that the obtained calculations are reliable to be applied for evaluation of stratified TES. Confirmation the magnitude values of the parameters, however, can not be conducted since the parameters might vary over the wide range depend on configuration and operating condition of the stratified TES (Zurigat and Ghajjar, 2002). Thermocline thickness evaluation of stratified TES For evaluation of thermocline position during charging cycle, results from Table I and II indicate that thermocline position moves upwards from the bottom part to upper part of TES tank during charging cycle. With regards to thermocline thickness, for case I it is lower than that for case II, as seen from Table III and IV. This indicates that higher flow rate resulted to thicker thermocline. The finding on the variations of thermocline thickness for different flow rates as noted in this study were also reported by (Karim, 2009) and (Musser and Bahnfleth, 1998b). For thermocline thickness growth evaluation, Table I,II and III,IV were used to relate parameter WTC and C. Using observation at =0.1, it indicates that thermocline thickness change with respect to time. For case I, the maximum thermocline thickness is 1.35 m, in the charging of hour 20.00 when the midpoint of thermocline position at 3.64 m elevation. Case II has maximum thermocline thickness of 1.84 m, also charging at 20.00, with midpoint of thermocline position at 4.32 m elevation. The occurrence of maximum thermocline thickness at the lower part of the storage tank are noted, this is due to its position nearby inlet diffuser where the mixing has more influence. 10 REFERENCE [1]. Zurigat YH and Ghajar AJ, 2002, Heat Transfer and Stratification in Sensible Heat Storage System, in Thermal Energy Storage System and Applications, Eds. Dincer and Rosen, Wiley, New York. [2]. Dincer I and Rosen MA, 2002, T hermal Energy Storage System and Applications, John Wiley and Sons. [3]. Nelson JEB, Balakrishnan, Srinivasa Murthy S, 1999, Experiments on Stratified Chilled-Water T anks, International Journal of Refrigeration 22, pp. 216-234. [4]. Karim MA, 2009, Performance Evaluation of a Stratified ChilledWater T hermal Storage System, World Academy of Science, Engineering and T echnology 53, pp. 326-334. [5]. Bahnfleth WP and Musser A, 1998, T hermal Performance of Full Scale Stratified Chilled Water Storage Tank, ASHRAE Transaction 104(2), pp. 377-388. [6]. Musser A and Bahnfleth WP, 1998a, Evolution of T emperature Distributions in a Full-Scale Stratified Chilled-Water Storage T ank with Radial Diffusers, ASHRAE T ransactions, Vol. 107(1). [7]. Musser A and Bahnfleth WP, 1998b, Field-Measured Performance of Four Full-Scale Cylindrical Stratified Chilled-water T hermal Storage T anks, ASHRAE T ransaction 105 (2), pp. 218-230. [8]. Homan K, Sohn C and Soo S, 1996, T hermal Prformance of Stratified Chilled Water Storage T ank, HVAC&R Research 2(2): 158-170. [9]. Yoo J, Wildin MW and T ruman CR, 1986, Initial Formation of T hermocline thickness in Stratified T hermal Storage T ank, ASHRAE T ransaction 92 (2A): 280-292. [10]. Systat Software Inc, 2008, SigmaPlot 11 User’s Guide, CA USA. IV. CONCLUSIONS Results from the study indicate that there were two functions which could be used to represent the S-curve of temperature distribution, namely sigmoid dose response (SDR) and four parameter sigmoid (FPS) function. Both functions have similar curvature in the fitting of temperature distribution. Using these functions, the average cool and warm water temperature and cool water depth in storage tank could be ascertained. The method offer an alternative solution for formulating thermocline thickness based on functional relationship of temperature profile. It takes advantage to simplify the determination of thermocline thickness and eliminate the ambiguities of limit points on the edge of thermocline profile. This approach to figure out the temperature profile and determine the thermocline thickness based for functionally relationship will assist in the evaluation of performance of TES tank. A CKNOWLEDGM ENT The authors would like to acknowledge the support of Universiti Teknologi PETRONAS for this research. 108601-2424 IJMME-IJENS © February 2010 IJENS IJE NS International Journal of Mechanical & Mechatronics Engineering IJMM E-IJENS Vol:10 No:01 11 Case I - SDR function 16 Temperature (oC) 14 12 10 18.00 19.00 20.00 21.00 22.00 23.00 24.00 01.00 02.00 8 6 4 0 2 4 6 8 10 12 14 X (dimensionless elevation) (a) Case II - SDR Function 16 12 o Temperature ( C) 14 10 18.00 19.00 20.00 21.00 22.00 23.00 24.00 01.00 02.00 8 6 4 0 2 4 6 8 10 12 14 X (dimensionless elevation) (b) Fig. 4. Fitting SDR function for (a) case I and (b) case II Case I - FPS function 16 Temperature (oC) 14 12 10 18.00 19.00 20.00 21.00 22.00 23.00 24.00 01.00 02.00 8 6 4 0 2 4 6 8 10 12 14 X (dimensionless elevation) (a) Case II - FPS Function 16 12 o Temperature ( C) 14 10 18.00 19.00 20.00 21.00 22.00 23.00 24.00 01.00 02.00 8 6 4 0 2 4 6 8 10 12 14 X (dimensionless elevation) (b) Fig. 5. Fitting FPS function for (a) case I and (b) case II 108601-2424 IJMME-IJENS © February 2010 IJENS IJE NS International Journal of Mechanical & Mechatronics Engineering IJMM E-IJENS Vol:10 No:01 12 T ABLE I. P ARAMETERS ON THE TEMP ERATURE DISTRIBUTION OF SDR FUNCTION SDR 18:00 19:00 20:00 21:00 22:00 23:00 0:00 1:00 2:00 FPS 18:00 19:00 20:00 21:00 22:00 23:00 0:00 1:00 2:00 Case I (Flow rate:393 m3/hr) Tc 6.94 6.93 6.91 6.91 6.90 6.89 6.87 6.85 Th 13.61 13.58 13.56 13.55 13.54 13.54 13.53 13.51 13.49 6.83 Tc 6.94 6.93 6.91 6.91 6.90 6.89 6.87 6.85 C 1.98 2.71 3.64 4.61 5.55 6.51 7.47 8.44 9.39 Case II (Flow rate:524 m3/hr) 2 S1 R 2.11 1.60 1.41 1.42 1.44 1.48 1.71 1.83 1.83 0.9983 0.9987 0.9994 0.9995 0.9995 0.9995 0.9996 0.9996 0.9997 Tc 7.08 7.06 7.02 7.02 6.95 6.91 6.89 6.86 6.73 Th 13.83 13.74 13.74 13.71 13.63 13.57 13.53 13.38 13.36 C 1.94 3.10 4.32 5.63 6.86 8.13 9.35 10.53 11.79 S1 R2 1.42 1.20 1.04 1.03 1.24 1.31 1.29 1.32 1.30 0.9932 0.9942 0.9965 0.9967 0.9975 0.9983 0.9987 0.9996 0.9993 T ABLE II. P ARAMETERS ON THE TEMP ERATURE DISTRIBUTION OF FPS FUNCTION Case I (Flow rate:393 m3/hr) Case II (Flow rate:524 m3/hr) 2 R S2 S2 Th C Tc Th C 6.83 0 0.001 0.1 0.2 0.3 0.4 0.5 13.61 13.58 13.56 13.55 13.54 13.54 13.53 13.51 13.49 1.98 2.71 3.64 4.61 5.55 6.51 7.47 8.44 9.39 0.21 0.27 0.31 0.31 0.30 0.29 0.25 0.24 0.24 0.9983 0.9987 0.9994 0.9995 0.9995 0.9995 0.9996 0.9996 0.9997 7.08 7.06 7.02 7.02 6.95 6.91 6.89 6.86 6.73 13.83 13.74 13.74 13.71 13.63 13.57 13.53 13.38 13.36 1.94 3.10 4.32 5.63 6.86 8.13 9.35 10.53 11.79 T ABLE III. T HERMOCLINE THICKNESS OF CASE I USING SDR FUNCTION WTC - Case I (Flow rate : 393 m3/hr) - SDR 18:00 19:00 20:00 21:00 22:00 23:00 0:00 1:00 0.31 0.36 0.42 0.42 0.35 0.33 0.34 0.33 0.33 R2 0.9932 0.9942 0.9965 0.9967 0.9975 0.9983 0.9987 0.9996 0.9993 2:00 ∞ ∞ ∞ ∞ ∞ ∞ ∞ ∞ ∞ 2.85 0.91 0.57 0.35 0.17 0 3.74 1.19 0.75 0.46 0.22 0 4.26 1.35 0.85 0.52 0.25 0 4.22 1.34 0.85 0.52 0.25 0 4.16 1.32 0.83 0.51 0.24 0 4.06 1.29 0.82 0.50 0.24 0 3.52 1.12 0.71 0.43 0.21 0 3.14 1.00 0.63 0.40 0.18 0 3.15 1.00 0.63 0.40 0.18 0 T ABLE IV. T HERMOCLINE THICKNESS OF CASE II USING SDR FUNCTION 0 0.01 0.1 0.2 0.3 0.4 0.5 18:00 WTC - Case II (Flow rate : 524 m3/hr) - SDR 19:00 20:00 21:00 22:00 23:00 0:00 1:00 2:00 ∞ ∞ ∞ ∞ ∞ ∞ ∞ ∞ ∞ 4.23 1.35 0.85 0.52 0.25 0 5.01 1.59 1.01 0.61 0.29 0 5.80 1.84 1.16 0.71 0.34 0 5.84 1.86 1.17 0.72 0.34 0 4.85 1.54 0.97 0.59 0.28 0 4.57 1.45 0.92 0.56 0.27 0 4.67 1.48 0.94 0.57 0.27 0 4.55 1.45 0.91 0.56 0.27 0 4.61 1.47 0.92 0.56 0.27 0 108601-2424 IJMME-IJENS © February 2010 IJENS IJE NS
© Copyright 2026 Paperzz