Low Charge Ammonia System Controlled by Gas Quality Sensors

It`s now possible to design Low Charge dry expansion systems, more energy efficient than conventional flooded and pump circulation systems.

The potential for the application of this new technology is very large, and used with the latest energy efficient evaporators and compressors it can achieve energy savings of up to 30%. Description of new possibilities using gas quality measurement and facts from a new Cold Store in Melbourne using HB Products HBDX gas quality sensor.



Facts concerning gas quality control versus superheat and pump circulation systems:

It is now possible to control and regulate an evaporator in a refrigeration system using a sensor mounted on the evaporator outlet, as an alternative to traditional superheat measurement/control based on pressure and temperature where two sensors employ calculations to indicate superheat.

For low-temperature ammonia systems in particular, superheat control has not functioned optimally. There are many reasons for this: - Inappropriate selection of evaporator materials with poor thermal conductivity of the evaporator pipe - A lack of exposure of the internal pipe surfaces in the air cooler to the boiling refrigerant - The presence of oil, which leads to oil fouling and hence reduced heat transmission - The presence of water, which results in an increase in the refrigerant’s boiling point towards the end of the evaporation process - Inappropriate evaporator design, which does not adequately consider the fundamental thermodynamic principles that should be incorporated in all heat exchangers during the design process - Time delay and critical placement of the temperature sensor for measurement of the boiling point are often associated with a faulty indication

For example, water content of about 1% in the ammonia will lead to an increase in the boiling point towards the end of the evaporation process of 5-6°C in low-temperature evaporators. The use of superheat-based injection regulation would not be able to differentiate between superheat and a boiling point increase caused by water. This situation may result in liquid flood-back from the evaporator to the compressor. Using a new kind of regulation such as the HBDX gas quality sensor, which can differentiate between the presence of liquid/vapour and gas, would naturally improve this situation.

The refrigeration industry has thus far attempted to solve this problem using pump circulation or liquid overfeed. A pump circulation system circulates a quantity of refrigerant that is several times larger than that evaporated in the evaporator. This eliminates a large part of the effect of the boiling point increase. The disadvantages are an increase in the refrigerant inventory, increased pressure loss in the suction pipes and riser pipe, and the resulting increased energy usage.

Using an HBDX gas quality sensor, which measures refrigerant/liquid content as degree of wetness at the outlet of the evaporator, results in a measurement signal that can be used directly for controlling liquid injection into the evaporator, without all the disadvantages of a superheat calculation/measurement. Furthermore, a substantially better efficiency is achieved than with pump circulation systems and flooded systems. Among other things, this is due to the elimination of liquid content in the suction pipes and riser pipes, where the presence of liquid can increase the pressure drop by a factor of 7-10 compared with dry suction lines.

DX diagram

In general, dry-expansion evaporator design requires that users are perfectly aware of the capacity range within which the evaporator will be used. These considerations are also important for evaporators that are installed for pump circulation, but the consequences of an incorrect number of channels is easier to compensate for in a pump circulation system than in a dry expansion system with limited refrigerant charge.

In a dry-expansion system (especially using ammonia as the refrigerant) the choice of evaporator tube diameter is often of great importance. For evaporators with small capacities, the distribution of refrigerant can also result in difficulties simply because the distributor pipe diameters become so small that commercially available pipe material is not available. In these situations, other distributor types must often be used rather than the conventional choices (Colmac tank distributors) or design evaporators with one circuit.

It is not possible to generalise regarding the optimal average vapour velocity in the evaporator. Small pipes will have acceptable results with lower vapour velocities than larger pipes. Pipes with internal surface enhancements would also yield acceptable results with lower vapour velocity than smooth pipes. In general, dry expansion evaporators are designed for a substantially higher refrigerant pressure drop than pump circulation evaporators. To avoid this pressure drop resulting in a reduction of the logarithmic temperature difference, it is often necessary to use innovative evaporator circuiting methods/designs such as physical parallel flow (thermodynamic counterflow), consideration of the possible effect of gravity on the evaporator function, oil accumulation, sensor placement, etc.

Other important parameters for attaining good thermodynamic performance include that the evaporator pipes are made of a material with high thermal conductivity, such as aluminium. Today, there are many evaporator manufacturers offering industrial evaporators manufactured from aluminium. To ensure high efficiency as well as optimal boiling/flash gas formation, we recommend aluminium pipes for ammonia. 

The system must be designed with a small liquid separator designed as a subcooler, which can collect the small refrigerant excess and use the evaporation energy for subcooling. The liquid separator can be located locally by the evaporator or as a common liquid separator/intercooler located centrally at the compressors.

Experience gathered from four ammonia systems set up in Australia shows that the systems with gas quality measurement/control are more energy efficient and do not result in pressure variations of the same magnitude as DX systems based on superheat controlled refrigerant injection. Sensor set-up  in preparation for start-up is very simple and consists of a zero calibration, with the span value being set beforehand so the measurement range covers “X” 0.7 to 1.0, i.e. the 0.7 value (wet) corresponds to 20mA and 1.0 (dry) corresponds to 4mA. Using an internal or external controller, “X” is set to 0.98; to increase gas wetness a lower “X” value is chosen. With regard to the adjustments/optimisation, we recommend that this is done in line with a maximum increase/reduction of “X” of 0.02 (experience shows that X values between 0.95 and 0.98 yield slight superheat of 1 to 5 ᵒK).

Log diagram DX Australia

The graphs show the correlation between the measurement of the refrigerant gas quality (Wet/dry) and opening degree of the liquid expansion valve.

  • Setpoint of gas quality is "X" = 0.96.
  • The graph of overheating is almost non visible and indicating that the evaporator works very efficient.
  • The suction pressure is very stable, which indicating that the evaporator load is very constant.

A new cold store in Melbourne uses:                                                                                                                   

Freezing: 3 Colmac Coil DX evaporators -31°C evaporating temperature, unit refrigeration capacity approx. 60kW, refrigerant operating charge of 1.42 kg per evaporator.                                                                                                                                                                      

Medium temperature: 1 Colmac Coil DX evaporators -3°C evaporating temperature, refrigeration capacity approx. 37W, refrigerant operating charge 2.5 kg.

Ante room: 2 Colmac Coil DX evaporators, evaporating temperature - 3°C, unit refrigeration capacity  approx. 58kW, refrigerant operating charge 4.4 kg per evaporator.

The main points regarding the implementation of the system in Melbourne are:

- Use of aluminium evaporators with patented tank distributors and internal surface enhancement in low-temperature evaporators

- Use of piston compressors with very low oil carry-over (<3 ppm at the design point – less at reduced rotational speed)

- Use of piston compressors arranged for air cooling

- Frequency converters on all compressors and fan motors

- The possibility of switching between superheat based refrigerant injection control and gas quality based injection control

- Oversized evaporative condenser

- Use of internally smooth stainless steel pipes for all interconnecting refrigerant pipe lines

- High-pressure float control between the condenser and the intercooler

- Return of the condensate formed during hot gas defrost directly to the intercooler, using a high-pressure float and dedicated condensate return lines

- Two hot gas solenoid valves per cold store evaporator enabling warming of the drain pan prior to defrosting the evaporator coil

- Desiccant dryer in the freezer, which distributes the dry air to the doors using air distribution ducts

- Automatic evaporator fan speed control, optimised on the basis of the entering temperature difference between the air and ammonia for the individual evaporator

- Possibility of regulating the cold store temperature directly using compressor capacity modulation

- Evaporative condenser made entirely in stainless steel

- System refrigerant charge, 480 kg – low temperature capacity approx. 177kW, high-temperature capacity approx. 140kW, total volume facility approx. 42,600 m³, estimated yearly energy consumption 21-23kWh/m³*a


Australia 1                                                                                                            


Use of the gas quality measurement allows optimization of the most important key factors in a refrigeration system:

• The elimination of superheat measurement for evaporator control can be used for reducing the temperature difference “ETD” between the ammonia and air temperature (At the plant in Melbourne is measured a “ETD” down to 2.3K) – especially in case of partial load. This reduces the system energy consumption because the compressors operate with the highest possible suction pressure in all operational situations. It also minimize the volume of the gas, thus the compressor works less and thereby uses less energy.
• Elimination of wet suction lines in Ammonia systems and to prevent the challenges with wet return piping in industrial refrigeration systems. Wet suctions cause large pressure drop in industrial plants which mounting of the evaporators with riser pipping and long distances to the compressor room.


Rules of thumb - Suction pressure - 1°C decrease mean approx:

Temperature          Capacity                COP                    Power           
+10°C -3.6 % -5.0 % + 5.2 %
    0°C -4.0 % -4.3 % + 4.5 %
-10°C -4.4 % -3.8 % + 4.0 %
-20°C -5.1 % -3.5 % + 3.6 %
-30°C -5.5 % -3.9 % +4.1 %
-40°C -6.5 % -4.4 % +4.6 %
-50°C -7.3 % -5.0 % +5.2 %


Ex. On the increased energy consumption at -25ᵒC: 10ᵒ superheat increases energy consumption up to 10%. Optimal evaporator design can further improve the efficiency.

0.3bars Pressure drop in suction lines increases energy consumption by 24%.

The potential for the application of this new technology is very large, and used with the latest energy efficient evaporators and compressors it can achieve energy savings of up to 30%.




Covid-19 Update:

HB Products is still at full capacity and able to handle orders as usual. We have a big stock of critical parts and can supply most of our products without any delay. Please be aware that transport time can be longer due to the development in international travel restrictions. We are following the guidance from our authorities and we have taken the necessary steps to ensure the health and safety of our employees.

Sign up to our newsletter

We will keep you up to date on products, news and knowledge about sensors for refrigeration systems and heat pumps.

Invalid Input

Invalid Input

Invalid Input