Use of liquid nitrogen in CA storage: Theoretical analysis and experimental validation

Journal of Food Engineering 82 (2007) 77–83 www.elsevier.com/locate/jfoodeng

Use of liquid nitrogen in CA storage: Theoretical analysis and experimental validation P.V. Mahajan a,*, T.K. Goswami b b

a Department of Process and Chemical Engineering, University College, Cork, Ireland Department of Agricultural and Food Engineering, Indian Institute of Technology, Kharagpur, India

Received 23 August 2006; received in revised form 24 January 2007; accepted 25 January 2007 Available online 8 February 2007

Abstract Liquid nitrogen (LN2) is a colorless, odorless, low boiling cryogenic liquid. Due to its inertness and high expansion ratio (646 between liquid and gaseous N2 at 0 °C), it is an excellent material for rapid purging of the initial O2 gas from the space of the controlled atmosphere (CA) storage. The present study establishes a relationship for predicting the amount of LN2 required for reducing O2 concentration in the CA storage. Simulations were carried out for predicting the amount of LN2 required for flushing at different conditions such as purity level of LN2, O2 set point of the CA storage and flow rate of LN2. The simulation results were validated using the laboratory scale CA storage system having a storage capacity that could hold 10 kg of apples. It was found that higher the purity level lower was the consumption of LN2. At O2 set point of 3%, the consumption of LN2 was found to increase from 0.12 to 0.21 kg when the purity level decreased from 100% to 97.5%. Higher level of O2 set point was found to consume less amount of LN2 at a given purity level. Flushing of LN2 at flow rate of 1.51  106 m3/s reduced O2 concentration more efficiently and also avoided the freezing temperature inside the CA storage. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Controlled atmospheres; Cooling; Fruits; Liquid nitrogen; Rapid CA; Storage; Vegetables

1. Introduction Controlled atmosphere (CA) storage involves a system in which the storage atmosphere is modified, monitored and maintained throughout the storage and distribution of the perishable foods. At present many different methods and devices are being used for modifying and maintaining a given gas regime in CA storage, e.g., silicone membrane system, inert gas generators and scrubbers (Bartsch, Wolanyk, & Blanpied, 1985; Cavalieri, Chiang, & Waelti, 1989). These devices may take 10–15 days to bring down the storage O2 concentration to the desired level during which wellknown undesirable changes might occur (Watkins & David, 1997). They reported that rapidly established CA is very much important in order to retain fruit firmness *

Corresponding author. Tel.: +353 214902501; fax: +353 21 4270249. E-mail address: [email protected] (P.V. Mahajan).

0260-8774/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.jfoodeng.2007.01.026

and acidity in much better condition. But there is still not much published work on the equipment required for establishing the CA conditions rapidly. Although inert gas generators provide a means for rapid pull down of concentration of O2 through the introduction of N2 gas, they proved to be too expensive for full time operation (Bartsch et al., 1985). A cryogenic fluid, such as liquid nitrogen (LN2), is an inert gas which boils at a very low temperature at atmospheric pressure. Use of LN2 for CA storage seems most tempting because alongside creating a favorable low O2 atmosphere, it is characterized by certain advantages such as it can act as a means of cooling. Indeed, interest in the use of LN2 and in situ generation of N2 gas for use in CA storage has grown rapidly (Waelti & Cavalieri, 1990). In food preservation, the use of LN2 has been explored for freezing and cooling of foodstuffs and grinding of high valued spices. In USA, the ‘Polarstream transportation

78

P.V. Mahajan, T.K. Goswami / Journal of Food Engineering 82 (2007) 77–83

Nomenclature qLN2 ER NL (O2)N (O2)t

density of LN2 (kg/m3) expansion ratio of LN2 amount of LN2 (m3/s) concentration of O2 present in N2 gas (%) concentration of O2 present in the CA chamber at time t (%)

system’ of LN2-cooled transport vehicle for perishable foodstuff introduced by M/s. British Oxygen Co., has become very popular (BOC, 1965). This method consists of spraying of LN2 into the air space of an insulated vehicle in a controlled manner; but there is no control over the O2 concentration of the storage atmosphere. Bartsch (1986) studied the use of LN2 for creating low O2 atmosphere inside CA storage. The aim was to reduce O2 concentration and nothing was mentioned about the effect of LN2 flushing on storage temperature. They reported a graphical method for calculating the amount of LN2 required in order to reduce O2 concentration to a particular level. Cavalieri et al. (1989) performed simulations on N2 scrubbing for different operating parameters of CA storage such as CO2 set point, O2 set point, respiration rate of fruit, room leakage rate and nitrogen gas purity. But the simulations reported are for the maintenance phase, i.e., O2, CO2 and temperature are already established within the optimum window. The present paper deals with pull down phase in which O2 and temperature are reduced to the optimum levels. The objective of the present study was to evaluate the effect of O2 set point of the CA storage, purity level of LN2 and flow rate of LN2 on pull down of concentration of O2 and consumption of LN2. Experimental validation was performed in order to optimize the flow rate of LN2 in the lab scale CA chamber. 2. Theoretical analysis The amount of LN2 flush depends on purity level and flow rate of LN2; O2 concentration set point and free volume of the CA storage. A relationship was established considering all the four parameters. For this purpose, a CA chamber as shown in Fig. 1 having a free volume of Vf in m3 was considered. The following assumptions were made for the development of the relationship between amount of LN2 required and different operating parameters of the CA chamber. (a) The chamber is leak proof and perfectly well mixed. (b) The change in gas composition due to respiratory metabolism of stored produce during flushing is negligible. (c) Concentration of O2 present in LN2 is the same throughout the flushing.

(O2)a t Vf V_ N2 qO2 M O2

O2 concentration of ambient air (%) flushing time (s) free volume of the CA chamber (m3) flow rate of N2 gas (m3/s) density of O2 (kg/m3) molecular weight of O2 (kg/mol)

Liquid nitrogen flushing .

at a flow rate of V N 2 in m3/sec having purity of (O2)N in % LN2 Inlet

(N2 + O2) from LN2 mix with (N2 + O2) from air

Outlet of O2 + N2 mix Fig. 1. Schematic diagram of the CA chamber showing O2 and N2 gas balance.

A mass balance between the inflow and outflow of O2 in the CA chamber as shown below was made as follows: ðO2 ÞN  qO2  dt M O2 ðO2 Þt  qO2   dt M O2

Moles of O2 inflow ¼ V_ N2 

ð1Þ

Moles of O2 outflow ¼ V_ N2

ð2Þ

where, V_ N2 is the flow rate of N2 gas, m3/s; (O2)N is concentration of O2 present in N2 gas, %; (O2)t is concentration of O2 present in the CA chamber at time t, %; t is flushing time, s; qO2 is the density, kg/m3 and M O2 is the molecular weight of O2, kg/mol. Combining Eqs. (1) and (2) leads to Eq. (3). Total moles of O2 accumulation   q O2 ¼ V_ N2  ðO2 Þt  ðO2 ÞN  dt M O2

ð3Þ

Total moles of O2 accumulated in the CA chamber during flushing time of t is also equal to V f  dðO2 Þt  qO2 =M O2 . This can be equated with Eq. (1). The -ve sign indicates that the O2 concentration of the CA chamber decreases with flushing of N2 for time t. V f  dðO2 Þt ¼ V_ N2  fðO2 Þt  ðO2 ÞN g  dt

ð4Þ

P.V. Mahajan, T.K. Goswami / Journal of Food Engineering 82 (2007) 77–83

79

It was assumed that x = (O2)t  (O2)N where (O2)N is constant and does not change with time. Differentiating this equation leads to d(O2)t = dx. Hence, Eq. (4) can be presented as: V f  dx ¼ V_ N2  x  dt

ð5Þ

Re-arranging Eq. (5) with operational limits yield as x0 = (O2)a – (O2)N and x = (O2)t – (O2)N where (O2)a is O2 concentration of ambient air (average 20.95%) Z t _ Z x V N2 1 dx ¼ dt ð6Þ  x Vf x0 0 and integrating   _V N  t ¼ V f ln ðO2 Þa  ðO2 ÞN 2 ðO2 Þt  ðO2 ÞN

ð7Þ

Eq. (7) was converted to yield the amount of LN2 required for flushing.   N L  ER  t ðO2 Þa  ðO2 ÞN ¼ ln ð8Þ qLN2  V f ðO2 Þt  ðO2 ÞN 3

where, NL is the amount of LN2, m /s; qLN2 is density of LN2 which is 808 kg/m3 (ASHRAE Handbook, 1990) and ER is expansion ratio of LN2. ER was defined as the ratio of density of LN2 (kg/m3) and density of N2 vapour ðqTN2 Þ at any temperature T (kg/m3). qTN2 was calculated by ideal gas law as qTN2 ¼

n  M N2 P ¼ RT V

Fig. 2. Pictorial view of the lab scale CA chamber (1 – fruit tray, 2 – evaporator fan, 3 – evaporator coil, 4 – protection cover, 5 – outlet valve, 6 – LN2 inlet, 7 – temperature controller and 8 – refrigeration system).

sure developed when the chamber was flushed with LN2. A gas analyzer (Testo 350 GmbH Lenzkirch, Germany, accuracy ±0.2%) was used for measuring the O2 concentration of the CA chamber during LN2 flushing. The change in temperature inside the CA chamber and core of an apple was measured by using NiCr–Ni temperature sensors (Testo 454 GmbH Lenzkirch, Germany, accuracy ±0.4 °C).

ð9Þ

where P is the pressure inside the CA chamber during flushing. It was assumed to be atmospheric pressure as the exit valve of the CA chamber was kept open during flushing. R is the gas constant and T is the absolute temperature (ASHRAE Handbook, 1990). 3. Materials and methods 3.1. Controlled atmosphere chamber In order to create a low O2 atmosphere, an airtight CA chamber of about 0.5 m high, 0.5 m wide and 0.3 m long was developed which was equipped with a LN2 flushing arrangement (Fig. 2). The LN2 supply line having an outer diameter of 0.02 m was connected at the top of the chamber. It was attached with a manually operated valve. A distributor header, placed beneath the ceiling, was used to spray incoming LN2 uniformly in the chamber. It was perforated with 16 holes (8 on each side from the center) having an internal diameter of 1.5  103 m. A fan capable of producing a velocity of about 1.4 m/s measured with no load was placed inside the CA chamber and in front of the distributor header for maintaining a uniform temperature and for preventing the formation of regions of gases inside the CA chamber. A vent valve of 0.025 m internal diameter, which was also a manually operated ball valve, was fixed at the sidewall of the chamber to release the pres-

3.2. LN2 flushing assembly A cylindrical thermocole made of expanded polystyrene having a capacity of 4.6  103 m3 was used for LN2 flushing. It was placed at the top of the CA chamber and a provision was made to attach a tube at the bottom side of the container. The tube length was selected in a way such that it just penetrated into the supply line of the CA chamber. The entire tube was insulated with asbestos rope for preventing the loss of refrigeration arising out of heat leak during flushing. The container was then placed on an electronic balance (Essae Teraoka Ltd, Japan, accuracy ±0.1 g) and weight loss during flushing was observed. Flow rate of LN2 was calculated from the slope of the straight line between weight of thermocole container and flushing time as shown in Fig. 3. Loss of LN2 from the open end at the top of the thermocole container, which was unavoidable, was taken into consideration for obtaining the actual flow rate of LN2. An experiment was performed to measure the weight loss of the thermocole container when there was no LN2 flow through the tube and thereby the loss of LN2 from the open end was accounted for. 3.3. Flushing of LN2 at different flow rates Effect of LN2 flow rate on temperature and O2 concentration of the CA chamber was studied. Teflon tubes of internal diameter 0.002, 0.003, 0.004 and 0.006 m were used

P.V. Mahajan, T.K. Goswami / Journal of Food Engineering 82 (2007) 77–83

1100

O2 c o n c e n t ra t i o n , %

W e i g h t o f L N 2 i n t h e rm o c o l e , k g

1200

1000 900 800

21

150

18

125

15 100 12 75 9 50 6 25

3

700

0

600

0 0.0

0

50

100

150

200

250

300

350

C o o l i n g c a p a c i t y ,k J

80

0.1

400

0.2

0.3

LN2 , kg

Flushing time, sec Fig. 3. Weight reduction of LN2 filled thermocole container for different tube diameters (M 0.006 m, + 0.004 m,  0.003 m, s 0.002 m).

Fig. 4. Predicted amount of LN2 required for O2 pull down of the CA chamber and resultant cooling effect (M O2 pull down, s cooling capacity).

to achieve different flow rates of 1.51  106, 3.21  106, 7.04  106 and 12.3  106 m3/s, respectively as shown in Fig. 3. A hollow pipe of 0.008 m internal diameter having a length of 0.3 m was used to collect gas sample from the CA chamber. It was located horizontally at 3/4th depth from the top of the chamber. During LN2 flushing, the O2 concentration of the chamber was measured at 10 s intervals and the data was stored directly in the gas analyzer. The initial O2 concentration of LN2, measured by gas chromatograph, was found to be 1.74%. A NiCr–Ni thermocouple was placed near the gas sensor and data was collected at 1 s intervals using computer software Comfort (version 2.2, Testo, Lenzkirch, Germany). The drop in temperature and O2 concentration was also predicted at different flow rates using Eq. (8). Each flow rate was replicated three times and the average values were reported. The flow rate was optimized according to the desired O2 concentration and temperature which are 2% and 1 °C, respectively, for red delicious apples (Kupferman, 1997).

N2 gas at 1 °C as shown in Fig. 4. For precooling 10.85 kg of apples along with the storage accessories to 1 °C, the total heat to be removed was found to be 1324 kJ for which 3.2 kg of LN2 was needed. This indicated that the LN2 required for cooling apples to 1 °C was about 12 times more than that required for the O2 pull down of the CA chamber. Mahajan and Goswami (2002) found the rapid cooling of apples beneficial; hence LN2 required for precooling was given more importance than that for O2 pull down of the CA chamber. LN2 required for precooling was therefore more than sufficient. The chamber, packed with apples, was flushed with 3.2 kg of LN2. The flow rate optimized in Section 3.3 was used for LN2 flushing. The changes in concentration of the O2, chamber air temperature and core temperature of the monitored apple were continuously recorded during flushing. Chamber air temperature was measured in order to control temperature within the tolerance limit of ±0.5 °C. Flushing was stopped when the temperature reached the lower limit and resumed when it reached the upper limit. This intermittent flushing was used in order to avoid chilling injury to the stored product. Since the change in temperature of the apple would also decide upon the rate and duration of LN2 flushing, a NiCr–Ni thermocouple was put at the core of an apple sample kept at the center of the chamber to record the changes in temperature during flushing. LN2 was stopped when core temperature of apple reached 1 °C.

3.4. Testing of the CA chamber The CA chamber having a total volume of 7.29  102 m3 and considering 65% as free volume had an available storage space of 2.55  102 m3 for storing apples and other storage accessories. The volume occupied by the storage accessories was measured and found to be 1.2  102 m3 while in the remaining available storage space; about 10.85 kg of apple could be stored. The actual free volume of the CA chamber after storing 10.85 kg of apple was measured and was found to be 66% of the CA chamber. This free volume was in accordance with the standard for the commercial CA storage rooms, i.e., 65% of the storage volume (Waelti & Cavalieri, 1990). Eq. (8) was used to predict the amount of LN2 required for lowering the concentration of O2 to 2.0% (upper limit for apple CA storage) and was found to be about 0.26 kg. Theoretically, 0.26 kg of LN2 has a typical cooling capacity of approximately 105 kJ considering an exit temperature of

4. Results and discussion 4.1. Effect of O2 set point As from Eq. (8), the amount of LN2 consumption depends upon the O2 set point of CA storage and the purity level of LN2. The amount of LN2 required for pull down of concentration of O2 in the CA chamber was calculated at different purity levels (100–96.5%) and the set point of concentration of O2 (0–4%) by using Eq. (8). The results are

P.V. Mahajan, T.K. Goswami / Journal of Food Engineering 82 (2007) 77–83

a b c d e f g

81

the room O2 level as shown in Fig. 6. With a 100% N2 purge stream, the O2 level was reduced to about 7.5% in the first 60 s. After 120 s of flushing, the O2 level would have decreased to less than 3%. With a 97% N2 purge stream, the O2 level was reduced to 10% and 5.5 % in 60 and 120 s of flushing, respectively. The results are in accordance with those obtained by Waelti and Cavalieri (1990). 4.3. Effect of LN2 flushing at different flow rates

Fig. 5. Contour plot showing the effect of LN2 purity and O2 set point on LN2 consumption, (a) 0.10 kg, (b) 0.11 kg, (c) 0.12 kg, (d) 0.13 kg, (e) 0.15 kg, (f) 0.17 kg and (g) 0.22 kg.

shown in Fig. 5. At a constant purity level of 99.5%, the amount of LN2 required was found to decrease from 0.22 to 0.11 kg when the set point of concentration of O2 was lowered from 1% to 4%. Hence, reducing the O2 concentration of the CA chamber to the level slightly above the given set point is beneficial. The rest of the O2 concentration could be left for respiration of apples to ultimately reduce it to the given set point. 4.2. Effect of LN2 purity It was found that greater quantities of LN2 needed to be flushed to compensate for lower purity LN2. At a set point of 3% O2 level, the consumption was found to increase from 0.12 to 0.21 kg when the purity level of LN2 was decreased from 100% to 97.5%. This shows that O2 is the impurity in the nitrogen gas. Similar results were also observed by Cavalieri et al. (1989). The higher the purity level of the purge stream, the faster is the pulled down of

4.3.1. Effect on air temperature The effect of different flow rates of LN2 on air temperature of the CA chamber is shown in Fig. 7. Flushing at 12.3  106, 7.04  106 and 3.21  106 m3/s quickly dropped the chamber air temperature to 43.1, 14.4 and 9.1 °C, respectively. The rate of cooling was found to reduce with decrease in LN2 flow rate. At the lowest flow rate of 1.51  106 m3/s, cooling rate was lowest (0.08 °C/ s), dropping the inside temperature of the CA chamber to 0.5 °C. Any temperature lower than the freezing point of fruit (1.5 to 1 °C for apple) may induce cooling injury or freeze burn (Metlitskii, Sal‘kova, Volkind, Bondarev, & Yanyuk, 1983). Hence, a flow rate of 1.51  106 m3/s has been found to be suitable for flushing of LN2 in the CA chamber. This flow rate also maintained the lower temperature for the longer period (about 360 s), which helped to pre-cool the stored apples. 4.3.2. Effect on O2 pull down The effect of LN2 flushing at different flow rates on the O2 concentration of the CA chamber is shown in Fig. 8. The O2 concentration was found to reduce faster initially and then the rate was found to be diminishing requiring more LN2 at lower O2 concentrations. The rate of reduction of O2 concentration was found to vary with the flow rate of LN2, being faster at the higher flow rate. The results are in agreement with those obtained by Chapon, Blanc, and Varoquax (2004). Flushing of LN2 at a flow rate of

35 25

Ai r t e m p e ra t u re , ˚ C

15 5 -5 -15 -25 -35 -45 -55 0

50

100

150

200

250

300

350

400

Flushing time, sec Fig. 6. Effect of LN2 purity on O2 concentration of the CA chamber at the flow rate of 1.51  106 m3/s (LN2 purity levels: —— 100 %, – – – – 99%, - - - - - - 98%,  97%).

Fig. 7. Changes in air temperature of the CA chamber during flushing of LN2 at different flow rates (M 12.3  106 m3/s, + 7.04  106 m3/s,  3.21  106 m3/s, s 1.51  106 m3/s).

P.V. Mahajan, T.K. Goswami / Journal of Food Engineering 82 (2007) 77–83

O 2 co n c en t r at i o n , %

21 18 15 12 9 6 3 0 0

50

100

150

200

250

300

350

400

Flushing time, sec

Fig. 8. Effect of LN2 flow rate on O2 concentration of the CA chamber at the purity level of 98.3 % (experimental values of O2 concentration at different flow rates: M 12.3  106 m3/s, + 7.04  106 m3/s,  3.21  106 m3/s, s 1.51  106 m3/s; predicted values of O2 concentration at different flow rates: - - –_- 12.3  106 m3/s, – - –7.04  106 m3/s, 3.21  106 m3/s, — 1.51  106 m3/s).

12.3  106 m3/s for 44 s reduced the chamber O2 concentration to only 4.1%. Due to the high flow rate of LN2, the eddy currents formed during flushing might have expelled the incoming N2 atmosphere without allowing it to mix with the existing O2 rich atmosphere resulting in a lower drop in O2 concentration. For the same reason, flushing at 7.04  106 and 3.21  106 m3/s (for 76 and 167 s, respectively) reduced the O2 concentrations to 3.9% and 2.7%, respectively. For CA storage of apple and other fruits, about 2% of O2 concentration is required (Olsen, 1980) for which either flushing time would have to be increased or flow rate of LN2 would have to be decreased. Since, increasing the LN2 flushing time at higher flow rate increased the amount of LN2 consumption unnecessarily, reduction in flow rate was a better alternative. At a flow rate of 1.51  106 m3/s, the chamber O2 concentration was reduced to 6.7% in the first 120 s of flushing after which the rate of O2 reduction was slower. After completion of LN2 flushing in 360 s, the chamber O2 concentration was found to be 1.8%, which was within the desired limit for storing of apples in CA. The effect of flushing of LN2 on pull down of the CA chamber was predicted at different flow rates using Eq. (8). The predicted values of O2 concentration are shown in Fig. (8). Flow rate was found to affect the rate of reduction of concentration of O2 in the CA storage, reduction was faster at higher flow rate. The rate of pull down of concentration of O2 was higher during the initial period when the difference between the O2 level of the chamber and LN2 was largest. As this difference became smaller, the rate of pull down of concentration of O2 became slower and less efficient at the end. Fig. (8) shows that predicted values of O2 concentration were in agreement with those obtained experimentally at 1.51  106 m3/s flow rate. However, at other flow rates the observed values of O2 concentration are higher than the predicted values. It indicates that flush-

ing of LN2 at a slower rate was better in order to achieve proper mixing which in turn yielded lower O2 concentration in the laboratory model CA chamber. It was assumed that the chamber is perfectly mixed. But during the experiments it was found that shorter flushing periods, especially at higher flow rate, expelled the N2 gas quickly perhaps without mixing properly with air which led to less drop in O2 concentration. Indeed it was evident that at the lower flow rates (1.51  106 m3/s) of LN2, N2 gas seemed to mix well with the chamber air, purging off almost all the air in the chamber, leading to lesser O2 concentration in the chamber, which was almost equal to the O2 concentration in LN2. 4.4. Testing of the CA chamber The changes in temperature and O2 concentration of the CA storage chamber during flushing of 3.2 kg of LN2 are shown in Fig. 9. The LN2 loss from the open top end of thermocole container was compensated by adding LN2 @ 0.12  106 m3/s. The O2 concentration of the CA chamber was reduced to 3.8% after only 180 s of LN2 flushing. Subsequently, it remained almost constant at about 1.8% for the rest of the LN2 flushing period. This was because the LN2 used contained 1.74% of O2 as impurity. Flushing of LN2 was stopped manually when the temperature of the storage air went down to 1 °C. The storage temperature of the CA chamber was reduced to 0.3 °C due to residual cooling as shown in Fig. 9. At this moment the core temperature of apple was recorded to be 17.5 °C indicating that further cooling of the chamber was still necessary. The moment flushing of LN2 was stopped, temperature of the storage air started increasing. Therefore, flushing was again started when the chamber temperature increased to 1.5 °C. The on and off control of the LN2 flushing was continued as and when the temperature of the chamber exceeded the upper limit of 1.5 °C and reduced to the lower limit of 1 °C till the core temperature of apple reached the desired level, i.e., 1 °C. The time required for the tempera30

30

25

25

20

20

15

15

10

10

5

5

0

0

-5 0

500

1000

1500

2000

2500

O 2 c o n c e n t ra t i o n , %

24

T e m p e ra t u re , ˚ C

82

-5 3000

Flushing time, sec

Fig. 9. Effect of LN2 flushing on temperature and pull down of concentration of O2 in the CA chamber ( ——— Chamber air temperature, —— Apple core temperature, - - - - - - - - - O2 concentration).

P.V. Mahajan, T.K. Goswami / Journal of Food Engineering 82 (2007) 77–83

ture to reach from 30 to 1 °C was calculated from Heisler’s Chart (Geankoplis, 1993) and was found to be 1770 s as compared to 2100 s observed experimentally (Fig. 9). A highly damped level of temperature control was obtained because of intermittent flushing of LN2. However, there is a scope to improve it using automatic controllers at commercial installations. From the above results it can be said that LN2 must be discharged slowly into the CA chamber to prevent the chilling injury to the fruit. The refrigeration system can be kept off during LN2 flushing, since cold gaseous N2 is capable to remove the field heat of the stored produce. After the field heat is removed (simultaneously reducing O2 concentration) mechanical refrigeration system could be used to maintain the temperature throughout the storage life. 4.5. Economic feasibility In CA storage, the temperature has to be reduced to the desired level so that equipment for O2 reduction starts without delay (Kupferman, 1997). Furthermore, rapid cooling of fruits after harvest reduces the rate of respiration, which helps in maintaining its quality during longterm storage. Cooling using conventional refrigeration system is difficult to maintain whilst the room is being loaded with fruits. This type of cooling system, if used in the CA storage where fast filling is required, would increase the demand on available refrigeration capacity (Watkins & David, 1997). Once the fruit is loaded there may be considerable pull down periods in a conventional system due to the refrigeration system being sized for a much lower heat extraction rate than needed during pull down. Hence, CA storages require more refrigeration capacity than the conventional cold storages. Once the fruit is cooled, only about 10% of the installed evaporator capacity is needed to maintain long-term storage temperatures (Waelti, 1994). The rest of the installed evaporator capacity remains idle during the maintenance phase. Hence, use of LN2 for initial cooling to remove the field heat of the store and for reducing the initial O2concentration of the storage atmosphere is suggested. Once the temperature and O2 concentration reached the desired level, shifting of the LN2 refrigeration system to the mechanical refrigeration system (designed to supplement the maintenance phase without considering the field heat) can be of a suitable choice. Thus, the utilization of LN2 for CA storage seems tempting because alongside creating a favorable lower O2 atmosphere, liquefied nitrogen is characterized by an advantage of faster cooling when compared with mechanical refrigeration system alone. The methodology developed in this paper predicts the amount of LN2 required for purging O2 from the CA storage. It also investigates the effect of different factors like flow rate, purity, set point on consumption of LN2. All these factors are important in modifying and controlling the gas regime in the CA storage. Further investigation is needed to explore the use of this system for the larger scale CA storages.

83

5. Conclusions Flushing of LN2 was found to be effective not only for achieving low O2 atmosphere rapidly but also for precooling of apples during the pull down phase of the CA storage. However, LN2 required for pre-cooling was about 12 times higher than that required for O2 pull down. The amount of LN2 required for flushing can be halved if the O2 set point of the CA chamber was increased from 1.0% to 4.0%. This increase in O2 set point could be compensated by the produce respiration rate, which would reduce the O2 concentration to the optimum level. Similarly, LN2 required for flushing can also be halved if the purity level of LN2 was increased from 97.5% to 100%. Flushing of LN2 at a flow rate of 1.51  106 m3/s was found to maintain the temperature of 1 °C and reduce the O2 concentration to 1.8%. This flow rate also obviated the freezing temperature of apples.

References ASHRAE Handbook (1990). Refrigeration Systems and Applications. In American Society of Heating Refrigeration and Air Conditioning Engineers. Atlanta, GA. Bartsch, J. A. (1986). Creating a low oxygen atmosphere with liquid nitrogen. In Agricultural Engineering Facts. Ithaca, NY: Cornell University. Bartsch, J. A., Wolanyk, A. M., & Blanpied, G. D. (1985). Economic analysis of systems used to establish CA atmospheres. In Proceedings of the fourth international controlled atmospheres research conference. North Carolina. BOC (1965). Manual of Polarstream refrigeration system. British Oxygen Company. Cavalieri, R. P., Chiang, W. C., & Waelti, H. (1989). Nitrogen scrubbing of CA storage: a simulation study. Transactions of the ASAE, 32(5), 1709–1714. Chapon, J. F., Blanc, C., & Varoquax, P. (2004). A modified atmosphere system using a nitrogen generator. Postharvest Biology and Technology, 31, 21–28. Geankoplis, C. J. (1993). Principles of unsteady-state heat transfer. In Transport processes and unit operations (second ed., pp. 336). PrenticeHall, NJ. Kupferman, E. (1997). Observations on storage regimes for apples and pears. Tree Fruit Postharvest Journal, 8(3), 3–5. Mahajan, P. V., & Goswami, T. K. (2002). Effect of rate of establishment of controlled atmosphere conditions on apple quality. Agricultural and Biosystems Engineering, 3(1), 10–17. Metlitskii, L. V., Sal‘kova, E. G., Volkind, N. L., Bondarev, V. I., & Yanyuk, V. Y. A. (1983). Engineering equipment for creating and regulating controlled atmosphere. In Controlled Atmosphere Storage of Fruits (pp. 68–94). India: Amerind Publishing Ltd. Olsen, K. L. (1980). Rapid CA and low O2 storage of apples: a new approach for long storage of ‘Golden Delicious’ and more effective storage of red ‘Delicious’ apples. In Proceedings of the Washington State Horticulture Association, WA. Waelti, H. (1994). Energy conservation in CA storages-Program 101, FYBriefing Report, Cooperative Extension, Department of Biological Systems Engineering, Washington State University. Waelti, H., & Cavalieri, R. P. (1990). Matching nitrogen equipment to your needs. Tree Fruit Postharvest Journal, 1(2), 3–13. Watkins, C. & David, R. (1997). Cornell fruit handling and storage newsletter. Department of Fruit and Vegetable Science, Cornell University, Ithaca, New York.