Abstract-- Spatial variation of temperature and relative humidity were estimated with Computational Fluid Dynamics (CFD) at top, middle and base levels for peach storage at +1oC and 90% relative humidity and verified with measured data in a cold store with evaporative cooling system. Ansys Fluent Software was used for CFD modelling. CFD models were validated with sensors measurements. Results were evaluated by using descriptive statistics, relative error and variance analyses. Relative error of the CFD model was calculated as 9.77 for the ambient temperature and 1.29 for relative humidity for peach storage. The developed CFD models estimated the ambient factors with an acceptable error in the evaporative cold store for peach storage.
Keywords-- peach, cold storage, computational fluid dynamics (CFD), modelling, validation
I. INTRODUCTION
Losses of fruits and vegetables by stages in Turkey were determined as 4-12% during harvest, 2-8% during transportation, 5-15% in preparation for the market, 3-10 % in storage, 1-5 % in the consumer phase and 15-20 % for total (Özelkök and Kaynaş, 1991). Post-harvest crop losses occur between 5-25% in developed countries; this ratio varies between 20-50% in developing countries (Mitchell, 1992). It was known hundreds of tons of products decay before reaching the consumer and Turkish agricultural economy suffers because of great loss (Özelkök et al., 1992). According to estimates, about 10-30% of orchards products were spoiled and discarded from producer to consumer (Özdemir et al., 2009). In Turkey, product losses are so high. The total amount of loss in the agricultural production phase of the food supply chain in Turkey has been reported as approximately 13.7 million tons. This amount corresponds to approximately 12% of the total production amount. It is stated that approximately 9.48 million tons of fruit and vegetable production is lost during the agricultural production phase (Salihoğlu et al., 2018). Peach is a very sensitive fruit for stored conditions. Eris et al. (1992) found that only the weight loss for peaches reached up to 36% from harvest to consumer table. Effect of two different cooling systems on quality of cold stored peaches were investigated by Akdemir and Bal (2016a). Since peach fruit is sensitive, great care should be taken during harvest.
The cooling and quality of agricultural products in a cold store is highly dependent on the temperature distribution highly associated with the air flow areas. Xie et al. (2006) designed a two-dimensional mathematical model in a cold room and developed a computer program. Simulation results reflected the characteristics of air velocity and temperature distribution. After that, the parameters (corners, product batch) affecting the flow area were analysed. The results showed in particular that the batch of product was greatly affected by the flow and temperature. It was demonstrated that CFD was a powerful tool for designing and optimizing the flow field in cold store. A modelling method was developed to simulate heat and mass transfer in cold storage of horticultural products. This method was used to estimate the mass transfer of water vapor from packaged horticultural crops. Heat transfer paths, which are considered to be the most important factor for most products and for packaging systems, have been modelled. Experimental tests were arranged for 10 product packaging systems. Although the model supports multi-zone packaging systems, single-zone packaging was found to be sufficient in all tests (Tanner et al., 2002). Latent temperature, audible temperature, moisture loss and temperature distribution in bulk storage of fruits and vegetables were predicted by a developed model The porous media approach has been applied to model respiration, perspiration, convective heat and mass transfer (Becker and Fricke, 2001). Air flow in a cold room was investigated using computational fluid dynamics (CFD). The airflow model was made according to the conditions considered to be permanent and incompressible. Turbulence was taken into account using the k-ε model. The forced air circulation of the cooling unit has been modelled in accordance with the characteristics of the evaporator air duct and the fan, approximately an as-sociated body strength and resistance. The validity of the model was made by comparing the calculated time-average velocity values with the sensor values measured in each direction. The relative error of air velocity was observed at 26% (Hoang et al., 2000).
In a research on the evaluation of air flow in a food cleaning room; the flow distribution of the room in a pilot size, was investigated by Rouaud and Havet (2002) using a code of computational fluid dynamics based on finite volume formula. Two versions of the k-ε turbulence model were tested. These were Standard and RNG. Analyses on velocity magnitudes could not determine the difference between these two versions. As a matter of fact, the main characteristics of the flow in both models were determined by the harmony of the numerical results and the experimental results that the results obtained could quite well predict. Further experiments have shown that the RNG k-ε turbulence model can predict more eddy and complex flow states. As a result, this model was found to be more suitable for the estimation of air flow in clean rooms compared to the standard model. A three-dimensional CFD model was developed to calculate the distribution of velocity, temperature and humidity in a full and empty cold room. The dynamic behaviour of the fan and cooler was modelled. The validity of the model was tested considering the air velocity and product temperature (Nahor et al., 2005).
Spatial variability of the air velocity for a small experimental cold store were investigated by using mapping software. Using mapping software for measured data in the cold store was useful to determine weak airflow region to consider for possible solutions (Akdemir and Arin, 2005). Management Zone Analyze (MZA) was used to determine spatial variability of ambient factors in a cold store (Akdemir and Tagarakis, 2014). Modelling of the ambient factors and quality changes of the cold stored peaches were investigated in a cold store with chiller cooling system. Mean differences of the CFD model and measurements for temperature and relative humidity were calculated as 1.02ºC and 11.66%, respectively. Fruit firmness, total soluble solid, titratable acidity, pH, respiration rate, external appearance and flavour analyses were determined (Akdemir and Bal, 2016b).
One of the most important problems is high storage losses greater than allowed. Therefore, air flow, heat transfer and moisture losses were investigated in a commercial scale potato storage using fluid dynamics techniques for steady conditions. The CFD model developed was a two-dimensional simplified version of the cold storage. The model was capable of predicting air velocity with an accuracy of 19.5% and temperature with an accuracy of 0.5% (Chourasia and Goswami, 2007).
Thermal and aerodynamic study was performed on a cold store room and filled with about 11 tons of dates to be cooled from harvest conditions to about 5°C. The main object of the study was to define the suitable precooling conditions leading to homogeneous storage temperature inside the room. Volume of cold room was of about 67 m3 and equipped with an evaporator giving the required refrigeration capacity. Three-dimensional CFD model was established taking into consideration the local environmental parameters. The model SST k–ω is used to analyze the air turbulences. For normal design of the cold store room and during a precooling period of 40 h, the air velocity and temperature were determined in the different locations inside the room and vary between 0.25–7 m/s and 3–6 °C. While the product temperature remains higher than the required storage conditions of about 12°C. A new cold store room design was proposed using specific aerodynamic air deflector profiles. That permits to improve the heat transfer between the cold air and product. The precooling period is reduced of about 10 h and the average product temperature reaches 6°C (Ghiloufi and Khir, 2019).
A CFD model was presented to predict airflow patterns
and temperature profiles in ventilated packaging
Figure 1: Cold store (A), evaporator (B), condenser (C).
systems, during the forced-air cooling of peaches stored in a cold chamber. Transient CFD simulations were performed for the chamber containing four ventilated boxes and the evaluation of the results show that the temperature removal near the vent holes and the hand holes was relatively high when compared to other regions of the packaging box. The study reveals the airflow behaviour develop an uneven temperature distribution within the box. To overcome the flaws, future work was suggested to focus on modifying the vent hole design to improve the airflow phenomenon to maintain the temperature homogeneity throughout the box (Ilangovan et al., 2020).
The term of heat source was coded as detailed procedures and included into a computational fluid dynamics (CFD) model to improve the accuracy of simulation results in peaches cooling. It is found that a reasonable decrease in variations of cooling performances was obtained with sustained increase in air-inflow velocities. A maximum discrepancy in peaches volume-weighted average temperature (∆T vwa-max) is mainly concentrated in 0.1-0.3°C when the air-inflow velocity not exceeds 1.7 m/s, and its corresponded 7/8ths cooling time (SECT) is also prolonged by 1-6 min. This means that, below 1.7 m/s, these heat sources should be added as a term into the heat transfer equations for modifying the mathematical model inside peaches computational domain. Furthermore, the feasibility of this modelling method was confirmed by a great agreement with experiments, and its modified model has a higher accuracy with the decreased RMSE and MAPE values of 6.90%-11.26% and 7.28%-12.95%, respectively (Chen et al., 2020).
The objective of this study was to determine the spatial distribution of the temperature and relative humidity by using computational fluid dynamics and then validate with sensor measurements for peach storage in a cold store with evaporative cooling system. Computational fluid dynamics (CFD) models were created by using Ansys Fluent Software. Then, ambient temperature and relative humidity were measured for 3 levels (top, medium and base levels). CFD Model estimations and measured values were compared. Furthermore, the changes of temperature and relative humidity were modelled using computational fluid dynamics (CFD).
II. METHODS
A. Materials
A cold store which has evaporative cooling system was used for peaches (Figure 1).
Dimensions of the cold store were 4.60 × 4.35 × 3.41 m
(in length, width and height). Its volume was 68.23 m3. Type of
compressor was hermetic and its power was 5.25
Figure 2: Arrangement of peach cases.
Figure 3: Glohaven Variety Peach.
Figure 4: Geometry and mesh of the cold store.
kW. Condenser had axial fan and its capacity was 15 kW. It was constructed in Tekirdag Vineyard Research Institute (VRI).
Testo 177 H1 data loggers were used to measure temperature and relative humidity from 36 points in the cold store. Glohaven type peaches were stored. Sizes of plastic boxes were 0.49 m × 0.29 m × 0.1 m. Peaches were loaded in cold store within 256 cases (considering a storage of 4 ×4 × 16 cases) (Fig. 2).
Glohaven variety of peaches (Prunus persica L., Batsch; cv. Glohaven) were stored (Fig. 3).
Testo 177H1 sensors were used for measurement of ambient temperature and relative humidity in the cold store. Measurement limits of the temperature sensors ranged between -20oC and +70oC. Accuracy of the temperature sensors was ± 1%. Measurement limits of the relative humidity sensors ranged between 0 % and 100%. Accuracy of the relative humidity sensors was ± 5 %.
Spatial distribution of temperature and relative humidity was modelled by computational Fluid Dynamic (CFD) by using Ansys Fluent 14.0 software. A tetrahedral mesh was created and then refined until a converged solution was obtained (Fig. 4).
There were 239918 nodes for mesh analysis.
Storage temperature and relative humidity were assumed as 1oC and 90%, respectively. Boundary conditions were given in Table 1.
Measurements points of the sensors for different levels in the experimental cold store were given in Fig. 5.
The ambient temperatures and relative humidity were measured from 36 different points at 3 levels for 12 points.
Table 1. CFD model boundary conditions
|
Inlet |
surface of fluid inlet |
|
Outlet |
surface of fluid outlet |
|
Walls |
solid, proof against flow of fluid |
|
Convective heat transfer coefficient |
0.24 W/m K (for plastic boxes) |
|
Conductive heat transfer coefficient |
0.025 W/m K |
|
Inside and Outside temperature |
28oC and 1oC |
|
H2O mass fraction value |
0.0041432 H2O for 90% RH |
|
Solution type |
Pressure-based- double precision |
|
Flow type |
steady state |
|
Turbulence model |
k-e |
|
- for continuity, x velocity, y velocity, z velocity and k, e - for energy |
1e-03
1e-06 |
|
Density of peach |
930 kg/m3 |
|
Pressure-velocity relation |
Simple |
|
Specific heat for peach |
3.887 kJ/kgoC |
Figure 5: Measurements points in the cold store.
Figure 6: Planes and its codes for Y axis (A) and Z axis (B).
The Y planes were called as 1, 2, 3, and 4 (Fig. 6A) and Z plane were called as top, middle and base in the variance analysis and evaluations (Fig. 6B).
Differences between models and measurements data (∆t, ∆RH), descriptive statistics and variance analyses were calculated. In addition, Relative Error of the CFD model were calculated. CFD Model validated with measurements by using equation given below (Hoang et al., 2000; Nahor et al., 2005; Chourasia and Gosvami, 2007; Akdemir and Bartzanas, 2015).
(1)
(2)
(3)
(4)
Figure 7: Temperature contours at Y axes.
Figure 8: Temperature contours Z axes
(5)
(6)
where:
= Predicted ambient temperature(oC)
= Measured ambient temperature (oC)
= Difference between CFD model
prediction and measured ambient temperature (oC)
= Differences between CFD model prediction and measured
ambient temperature (%)
= Predicted relative humidity (%)
= Measured relative humidity (%)
= Differences
between CFD model prediction and measured relative humidity
= Differences between CFD model
prediction and measured or relative humidity (%)
= Relative error of CFD model estimation for
temperature and relative humidity
= Number of total measurements
= Measurements (1th, 2nd, 3rd, …).
Estimated values for the ambient temperature and relative humidity were determined from CFD models given in from Fig. 7 to Fig. 10 by using Ansys Fluent software. Variance analyses were also used by using SPSS to compare CFD model estimations and measurements, levels and Y axis and interactions of these variables
III. RESULTS
A. CFD Analyses for Ambient Temperature and Relative Humidity
CFD Model distributions of temperature for Y planes and Z axes (top, medium and base levels) were given in Fig. 7 and Fig. 8, respectively. Spatial distributions of relative humidity (RH %) in the cold store were given in Fig. 9 for Y axis and Fig. 10 for levels.
Figure 9: Relative humidity (%) contours at Y axes.
Figure 10: Relative humidity (%) contours at levels.
Figure 11: Predicted by CFD and measured ambient temperatures and differences for top, medium and base level
B. CFD Model Validation
Table 2. CFD model and measurement for ambient temperature
Differences between predicted and measured temperature were calculated between 0.2oC and 0.8oC.
Mean difference between the CFD model and sensor measurements was calculated as 0.51oC for ambient temperature. Model estimations were generally less than measured data. Relative error of the model was 9.77. Differences between the CFD model and sensors measurements for the ambient temperature was found statistically significant (F=879,759, a=0.001).
Predicted values, measured values and differences between measurements and model predictions of relative humidity for top level, middle level and base level in the
Table 3. Descriptive Statistics of CFD model and measurement for relative humidity
|
|
|
|
|
|
Mean |
97.90 |
94.43 |
3.47 |
|
Minimum |
97.90 |
93.20 |
2.10 |
|
Maximum |
97.90 |
95.80 |
4.70 |
|
Standard Deviation |
0.00 |
0.67 |
0.67 |
|
CV(%) |
0.00 |
0.71 |
|
|
Relative error (%) |
1.29 |
||
full loaded cold store with peaches were given in Figure 12. The measured relative humidity in the cold store varied between 93.2% and 95.8%. Differences between predicted and measured relative humidity were calculated between 2.1% and 4.7%.
Descriptive statistics such as mean, minimum (Min), maximum (Max.), standard deviation (SD), Coefficient of Variation (CV), differences between model estimation and the measured values and absolute error of the model were given in Table 3 for relative humidity.
Mean difference between the model and measurements for relative humidity was calculated as 3.47. Model estimations were higher than measured relative humidity values. Relative error of the CFD model for the cold store was calculated as 1.29. According to the analysis of variance for the relative humidity; differences between CFD models and sensors measurement was evaluated as statistically significant (F=879,759, a=0.001).
Chen et al. (2020) was determined relative error as 4.40%–11.54% for temperature in their modified CFD model for during forced-air cooling process of peaches. Nahor et al (2005) worked on pears to estimate temperature distribution by using air velocity magnitudes. They developed CFD models with an 20% error. Akdemir and Bal (2016b) was developed CFD models to estimate relative humidity and ambient temperature in a cold store with chiller unit for peach storage. CFD model error was 0.47 for ambient temperature and 13.91 for relative humidity.
In this research; the relative errors were determined as 9.77% for the ambient temperature and 1.29% for the relative humidity. Developed CFD models predicted the ambient temperature and relative humidity with an acceptable error when compare the literature given above. This research indicated that CFD model can be used for estimation of storage conditions for peach. Homogeneous distribution of ambient conditions in a cold storage is important for agricultural products. Therefore, it will be useful to use CFD Model estimates in cold storage design.
IV. CONCLUSIONS
In order to maximize the efficiency of food cooling and freezing processes, it is necessary to optimize the design of the appropriate refrigeration equipment required for refrigeration and freezing. The cooling time must be determined depending on the cooling load in the design of the cooling equipment. Accurate estimation of these factors depends on precise determination of surface heat transfer coefficients for cooling and freezing operations (Becker and Fricke, 2004). In this research study, CFD was used to estimate the ambient temperature and relative humidity in a cold store with evaporative cooling system for peach storage. The CFD model estimation for the temperature and relative humidity values were compared with sensors measurement to validate CFD model. The cold store was arranged for 1oC and 90% RH for peach storage. Mean difference between the model and measurements was calculated as 0.51oC for ambient temperature and as 3.47 for relative humidity. Relative error of the CFD model was calculated as 9.77 for the ambient temperature of the cold store and 1.29 for relative humidity. The developed CFD models were successful for the estimation of the ambient temperature and relative humidity.
ACKNOWLEDGEMENTS
The authors would like to express their gratitude to Scientific and Technological Research Council of Turkey (TUBITAK) for supporting the project (Project No 110O047). Data used in this article were taken from the Final report of the project.
REFERENCES
Akdemir, S. and Arin, S. (2005) Effect of air velocity on temperature in experimental cold store. J Appl. Sci. 5, 70-74.
Akdemir, S. and Bal, E. (2016a) Effect of two different cooling systems on quality of cold stored peaches. J. Environ. Prot. Ecol. 19, 1237-1248.
Akdemir, S. and Bal, E. (2016b) Modelling of ambient factors and quality changes for peach cold storage. J. Environ. Prot. Ecol. 17, 1122–1131.
Akdemir, S. and Bartzanas, T. (2015) Numerical Modelling and experimental validation of a cold store ambient factors. Tarim Bilimleri Dergisi-Journal Of Agricultural Sciences. 21, 606-619.
Akdemir, S. and Tagarakis, A. (2014) Investigation of spatial variability of air temperature, humidity and velocity in cold stores by using management zone analysis. Tarim Bilimleri Dergisi-Journal Of Agricultural Sciences. 20, 175-186.
Becker, B.R. and Fricke, B.A. (2001) A numerical model of commodity moisture loss and temperature distribution during refrigerated storage. Acta Hortic. 566, 431-436.
Becker, B.R. and Fricke, B.A. (2004) Heat transfer coefficients for forced-air cooling and freezing of selected foods. Int. J. Refrig. 27, 540–551.
Chen Y.-M., Song, H.-Y., Chen, Z.-S., Zhao, R., Su, Q., Jin, P.-Y., Yi-Shu, Sun, Y.-S. and Hao, W. (2020) Sensitivity analysis of heat and mass transfer characteristics during forced-air cooling process of peaches on different air-inflow velocities, Food Sci Nutr. 8, 6592-6602.
Chourasia, M.K. and Goswami, T.K. (2007) Steady state CFD modeling of airflow, heat transfer and moisture loss in a commercial potato cold store. Int. J. Refrig. 30, 672-679.
Eris, A., Turk, R., Turkben, C. and Ozer, M.H. (1992) A study on the losses in some important fruit types from the period of harvest till accession to consumer’s table. Proc.1st National Horticulture Congress of Turkey. Izmir, Turkey, 343-348.
Ghiloufi, Z. and Khir, T. (2019) CFD modeling and optimization of pre-cooling conditions in a cold room located in the South of Tunisia and filled with dates. Journal of Food Science and Technology. 56, 3668-3676.
Hoang, M.L., Verboven, P., De Baerdemaeker, J. and Nicolai B.M. (2000) Analysis of airflow in a cool store by means of computational fluid dynamics. Int. J. Refrig. 23, 127-140.
Ilangovan, A, Silva, P.D. and Gaspar, P.D. (2020) Airflow and thermal behavior within peaches packaging box using computational fluid dynamics - A preliminary study. International Congress on Engineering — Engineering for Evolution, KnE Engineering, 222–231.
Mitchell, F.G. (1992) Cooling horticultural commodities. in: postharvest technology of horticultural crops (Ed. A. A. Kader). Publication 3311 Univ. of California, Division of Agriculture and Natural Resources, California, USA, 53-58.
Nahor, H.B., Hoang, M.L., Verbovena, P., Baelmansb, M. and Nicolai, B.M. (2005) CFD model of the airflow, heat and mass transfer in cool stores. Int. J. Refrig. 28, 368–380.
Özdemir, A.E., Çandır, E., Dündar Ö. and Dilbaz, R. (2009) Losses in apples from the producer to the consumer and prevention methods, J. Agr. Res. Sci. 2, 165-168
Özelkök, S. and Kaynaş, K. (1991) Harvest and post harvest losses in fresh fruits and vegetables and possible measures. Journal of Agriculture, Forestry and Rural Affairs Ministry. 59, 1-9.
Özelkök, S., Ertan, Ü., Kaynaş, K. and Öz, F. (1992) Comparative studies on normal and controlled atmosphere storage of some ımportant apple varieties. Horticulture (in Turkish: Bahçe). 21, 77-90.
Rouaud, O. and Havet, M. (2002) Computation of the airflow in a pilot scale clean room using k- turbulence models. Int. J. Refrig. 25, 351-361.
Salihoğlu, G., Salihoğlu, N.K., Uçaroğlu, S. and Banar, M. (2018) Food loss and waste management in Turkey. Bioresource Technology. 248, 88-99.
Tanner, D.J., Cleland, A.C. and Robertson, T.R. (2002) A generalised mathematical modelling methodology for design of horticultural food packages exposed to refrigerated conditions: part 3, mass transfer modelling and testing. Int. J. Refrig. 25, 54–65.
Xie, J., Qu, X.-H., Shi, J.-Y. and Sun, D.-W. (2006) Effects of design parameters on flow and temperature fields of a cold store by CFD simulation. J. Food Eng. 77, 355-363.
Received: January 11, 2021
Sent to Subject Editor: January 15, 2021
Accepted: September 29, 2021
Recommended by Subject Editor Ardson Vianna Jr.