Stefano FILIPPINI ^(a), Umberto MERLO ^(a), Lorenzo NOTO LA DIEGA ^(a), Matteo Carmelo ROMANO ^(b), Ennio MACCHI ^(b).

^(a) LU-VE GROUP, Uboldo, 21040, Italy, [email protected]
^(b) Politecnico di Milano University, Milano, 20133, Italy, [email protected]

ABSTRACT

Industrial refrigeration uses ammonia unit coolers for the principle processes of conservation and freezing of foodstuffs. This paper describes the recent evolution of a unit cooler for NH3 application, aiming to lower the refrigerant charge. The main topics are the use of small tube diameter (1/2”), a new concept of refrigerant distribution, a new methodology of performance calculation, aiming to operate with reduced fluid charge. In fact, the revised construction allows operation with a very limited recirculation ratio number, from the typical 3-4 down to values of 1.8 with major containment of the ammonia charge in the unit cooler and consequently within the entire plant. The article describes the basis of the configuration, the experimental activities carried out to optimize performance and the final product arrangement.

Keywords: ammonia (refrigerant), dry expansion evaporator, evaporator, flooded evaporator, heat exchanger, NH3 (refrigerant), refrigerant charge

1. INTRODUCTION

Since the beginning of refrigeration, ammonia has been judged as one of the best working fluids for medium-large scale refrigeration plants, mostly due to its favourable thermodynamic properties (critical pressure and temperature, molecular complexity and mass): typically, a flooded ammonia system is significantly more efficient than its traditional refrigerant counterpart. Moreover, ammonia has other advantageous characteristics, including low specific cost and good heat transfer properties.

More recently, well known environmental issues have given a strong boost to the adoption of “natural” refrigerants, including ammonia, since they are the most environmentally friendly refrigerants, exhibiting both GWP (Global Warming Potential) and ODP (Ozone Depletion Potential) equal to zero.

However, all ammonia systems have to be designed with safety in mind, since the fluid is a toxic refrigerant, and it is also flammable in certain concentrations. That is why it must be handled with care: in particular, it is of primary importance to reduce the ammonia charge.

2.    THE NEW RANGE OF LU-VE INDUSTRIAL AMMONIA UNIT COOLERS

2.1. Design philosophy

The concept, introduced to the refrigeration industry by LU-VE is to adopt compact heat exchanger geometries, achieved by:

• tube diameters lower than conventional solutions (½” ⅝”)
• louvered, high efficiency fins

This enables the mass, volume and weight of the heat exchanger to be minimized and increases economic competitiveness. In the particular case of flooded ammonia unit coolers, this concept

brought about a dramatic reduction of the internal volume of the coil, and consequently of the ammonia charge. Moreover, given the high heat transfer coefficients achievable by adopting proper mass velocity and heat flux in the tubes, a large spacing (55 mm) in both row and tube is adopted. This not only gives an optimized heat balance between internal and external surfaces of the exchanger, but also offers a proper large surface for frost deposition. Another relevant feature of LU-VE ammonia unit coolers is the co-current flow arrangement, which enables the most favourable air-ammonia temperature differences across the heat exchanger and facilitates oil discharge thanks to top feeding.

2.2. The new development: low recirculation rates

Conventional evaporators are designed for a recirculation ratio (defined as the ratio between the actual inlet flow rate and the flow rate of generated vapour) of around 4-5. The reason for adopting such high values is based on the need to ensure a correct, nearly uniform, flow distribution among the various circuits and to obtain high internal heat transfer coefficients by means of relatively high mass velocities. However, there are substantial advantages in lowering the ammonia recirculation ratio in the unit cooler:

• lower ammonia mass in the evaporator
• lower flow rates (energy saving in circulation pump, smaller inlet piping and collector diameters, etc.)

The research described in this paper aims to investigate whether it is possible to preserve high unit cooler performance also at low recirculation rates, by the correct selection of circuits and orifices. The research activity was performed in two steps: (i) experimental activity on two unit coolers, with the same coil but with different circuit arrangement – one representative of current production, the other designed for low recirculation rates – and (ii) development of a computational tool, able to predict the complex fluid-dynamic and thermodynamic phenomena occurring in a real ammonia unit cooer.

3. EXPERIMENTAL CAMPAIGN

3.1. Test set up

The tests were performed at the Danish Technical Institute (DTI). The test set up is represented in Fig. 1: it is possible to vary the operating conditions (air temperature, evaporating temperature, ammonia flow rate) to investigate the evaporator capacity variation with these variables. In particular, as discussed below, the ammonia flow rate was varied (and therefore the recirculation rate) for fixed air and evaporating temperature.

Figure 1 Diagram of DTI ammonia test plant.                                     

Figure 2 Additional probes.

Three sets of measurements are used to (i) identify the thermodynamic cycle, (ii) verify the temperature uniformity of the air at evaporator inlet, (iii) calculate the evaporator capacity.

The capacity can be computed by two methods: (i) the product of the vapour mass flow rate by the enthalpy difference between saturated vapour and liquid and (ii) the energy balance of the climatic room, given by sum of the electrical power of contrast heaters and thermal power entering the climatic room through the insulated walls. The test is done in steady-state conditions and all quantities are integrated vs time for the test duration (about 30 min). The measure of the liquid flow rate is used to calculate the recirculation rate. Some additional measurements were performed, as depicted in Fig. 2: three differential pressure measurements, located on the upper, medium and top circuit and two temperature measurements, located at the exit of the bottom and top circuit.

Through measurements and calculation pressure test connections permit the evaluation of the pressure drops in the nozzles and along the circuits of the top circuit and the bottom circuit. From these it is possible to assess the mass flowrate distribution through the circuits.

The temperature measurements permit the evaluation of whether superheating occurs at the outlet of the top and bottom circuit.

3.2.                Results in term of capacity vs flow rate

The tests were performed on two unit coolers (current production and low recirculation one, see Table 1), at two different operating conditions:

• “High Temperature”: Evaporating Temperature= -8°C, Air Temperature = 0°C
• “Low Temperature”: Evaporating Temperature= -30°C, Air Temperature = – 22°C

Table 1 Main features of the tested evaporators – C = current design, N = new design.

Results are represented in Fig.3, respectively at high and low temperature, as a function of the inlet flow rate.

Figure 3 Capacity (referred to the design value at low temp.) vs ammonia flow rate for the two tested evaporators: (left) high temp. (-8°C) and (right) low temp. (-30°C). Numbers refer to the recirculation rate. For points inside the dotted circle superheating at the outlet of the upper circuit was found.

As expected, in all cases there is a value of flow rate which maximizes the capacity. At lower flow rates the capacity decreases, both because the heat transfer coefficient decreases with mass velocity and because the flow distribution experiences increased variations between top and bottom circuits. At higher flow rates the increase of pressure drops caused by large velocities inside the tubes more than counterbalance the increase of heat transfer coefficient. The design flow rate results from the best compromise between these two effects. The measured performance is well predicted for the current production evaporator: actually, the measured capacity reaches its maximum at flow rates slightly larger than the design one, at values of recirculation rate of about 4.8. The new model reaches similar or even higher capacities at flow rates much lower than in the previous case. The capacity remains high in a wide range of recirculation rates, say between 1.5 and 3, thus offering a good performance in a real plant operation, where the capacity can vary for a variety of situations.

The capacity decrease at very low rates is well detected by the temperature measurement at the exit of the top circuit (Fig. 4), which shows the occurrence of superheat in the top circuit: the temperature increases above the evaporation temperature when the pressure drop across the heat exchanger becomes close to the hydrostatic pressure of liquid column in the inlet collector. Fig. 4 depicts also the total (orifice + circuit) pressure drop across the evaporator, at high and low evaporating temperature respectively. At equal flow rate, pressure losses are larger for the low temperature case, due to lower densities of the ammonia vapour.

Figure 4 Total pressure drop (solid line) and temperature (dot line) at the end of the top circuit as a function of the inlet mass flow rate. Data refer to the “new evaporator” tested at: (left) high temperature (evaporation temp. = -8°C) and (right) low temperature (evaporation temp. = -30°C)

3.3. Defrost test

Defrost test were carried out to verify that the amount of hot gas flowing inside the heat exchanger was sufficient to perform a satisfactory defrosting also with the new arrangement, with fewer, longer circuits. Fig. 5 describes the flow pattern during defrosting: hot gas enters the tray and then flows inside the exchanger pipes. During the defrost tests the temperature measurements at the outlet of top and bottom circuits of the evaporator were used to detect real end of defrost.

Fig. 6 describes stages of the defrost test procedure. At the beginning a pump down stage is performed for 15 minutes; the liquid line and the fan are shut off while the compressor is running. Then a soft stage of 5 minutes starts, where defrost is initially performed by a small amount of hot gas. Looking at Fig. 6, it is possible to understand that during this stage the heat introduced is used to bring the evaporator near to the temperature of ice melting. The core part is the 23 minutes hard

stage which lasts longer than needed. Indeed from Fig. 6 it is possible to assess that ice melting ends after 5 minutes of hard stage, which means a total time of 10 minutes of defrost. Finally, there is a 5-minute dripping stage; the liquid line is open while the fan is still shut off. The measured defrost efficiency, defined as the ratio between the ideal/real heat required to complete defrost, was as high as 91%.

Figure 5 Flow arrangement during defrost

Figure 6 Various stages of the defrost procedure

4. CALCULATION MODEL

The fluid-dynamic and thermodynamic phenomena occurring in a real ammonia flooded evaporator are complex, given that each circuit experiences different heat fluxes, flow rates, vapour qualities at the exit, etc. In the attempt to develop a computer code capable of simulating the performance of such a complex component, the following approach is followed.

1. In order to account for differences among the various circuits, the heat exchanger is divided into five groups of parallel circuits, each group being characterized by a different hydrostatic
2. Each circuit is then discretized into 10 sections, following the evolution of fluid quality, pressure, heat transfer coefficient, etc.
3. The distribution of mass flow rate through each group of circuits is calculated by equating the overall pressure change (hydrostatic head + nozzle + heat exchanger tube) between inlet and outlet.
4. Given the mass flow rate for each circuit, the evolution of the heat transfer coefficient along the tube is calculated.
5. When the convergence is reached, it is possible to calculate the energy and mass balance of the whole evaporator.

The problem is solved through two main iteration loops: (i) an external loop, changing the overall pressure drop across the evaporator until the total ammonia mass flow rate is found and (ii) an internal loop, changing the mass flow rate through each circuit until the calculated pressure drop match the value set in the external loop of the previous point.

Empirical correlations are used to calculate heat transfer coefficient and pressure drops along the heat exchanger.

Figure 7 Schematic of the flooded evaporator           

Figure 8 Flow pattern map by model of L. Wojtan et al. assumed as basis for the present calibration.

Relation between heat transfer coefficient, mass velocity, vapour quality and heat flux are derived by L. Wojtan et al. (2005), using the flow regime map by L. Wojtan et al. (2005) (Fig. 8), with an additional coefficient calibrated to fit the experimental data.

Relation between two-phase fluid mass flow rate and pressure drop across the nozzle is derived from CFD calculations, validated by the test with reasonable accuracy. Pressure drops along the heat exchanger tube are calculated with the Lockhart-Martinelli correlation (Fig. 9).

Figure 9 Comparison between experimental data and adopted correlation for predicting pressure drop across the circuit vs average mass flow rate for circuit.

A comparison between the predicted and the measured cooling capacity is given in Fig. 10 for the 7 circuits heat exchanger operated at high and low temperature.

Figure 10 Comparison between experimental and calculated cooling capacity, (left) high temperature and (right) and low temperature cases.

Since the simulation code yields complete information about the velocity field and the vapour quality inside the evaporator, it is possible to calculate the distribution of outlet quality, flowrate and capacity for each circuit (Fig. 11) as well as the void fraction distribution in the heat exchanger, thus calculating with fair accuracy the ammonia charge in the evaporator. As is well known, the charge increases with the recirculation rate.

Figure 11 Influence of low and optimum recirculation ratio on outlet quality and capacity: at low recirculation ratio there is uneven flowrate distribution which causes superheat in the top circuits with the consequent capacity loss.

The next table shows an example of these results, giving a comparison of ammonia charges of new and current production evaporator.

Table 2 Ammonia charge in LU-VE flooded unit cooer.

If the ammonia charge/capacity ratio found in the present paper is compared with conventional flooded evaporators, as the ones described by J. Kristofersson et al. (2017) in a recent paper, we find the results described in table 3.

Table 3 Comparison between a conventional unit cooler and a LU-VE low-charge unit cooler in term of charge/capacity ratio.

We find that LU-VE current production, mainly due to the adoption of a smaller tube diameter, exhibits a reduction of over 50% of ammonia charge for equal capacity versus a conventional unit cooler. The adoption of lower recirculation rates reduces this value for a further 20%.

5.    CONCLUSIONS

The future of industrial refrigeration using NH3 as refrigerant has to put the reduction of fluid charge as a main priority. A fundamental contribution is coming from the use of low charge unit cooler design. Efficient heat transfer geometry, with small tube diameter (1/2”), combined with low recirculation rate (1.8) is the solution described in the present paper. An intensive redesign activity coupled with experimental results allowed the reduction of the ratio between liquid holdup and capacity [kg/kW] by 67% compared to conventional unit coolers.

Adopting low recirculation rates is beneficial not only in terms of reduced charge, but also in reducing cost and energy consumption of pump and piping.

The R&D activity continues, aiming to further investigate the possibility of reducing the ammonia charge in the heat exchangers.

REFERENCES

Wojtan L., Ursenbacher T., Thome J.R., 2005. Investigation of flow boiling in horizontal tubes: Part II—Development of a new heat transfer model for stratified-wavy, dryout and mist flow regimes. International Journal of Heat and Mass Transfer, 48 2955–2985.

Wojtan L., Ursenbacher T., Thome J.R, 2005. Investigation of flow boiling in horizontal tubes: Part I—A new diabatic two-phase flow pattern map. International Journal of Heat and Mass Transfer, 48 (2005) 2955–2969.

Kristófersson J., Vestergaard N., Skovrup M., Reinholdt L., 2017. Ammonia charge reduction potential in recirculating systems – calculations. doi: 10.18462/iir.nh3-co2.2017.0005.