back to main Samples page

Light at the end of the tunnel

J P Miller, Air Products

It is more than thirty years since the invention of the cryogenic tunnel freezer. In that quarter-century, what have manufacturers learned about the process of cryogenic freezing? Not much.

We have made them easier to clean, and we have added electronics so that we can control the temperature inside the freezer — though not the temperature of the food — more precisely. What we have not done is reduce their energy consumption or make them smaller. The temperature of food leaving the freezer still varies across the width of the belt. We still have problems with excess liquid nitrogen collecting at the base of the freezer.

The food industry knows a lot about producing frozen food, but it knows much less about heat transfer and the effects of freezing on food. As a result, today’s cryogenic freezers are bulky and wasteful. In the most basic process terms they are badly designed.

A research project run by Air Products over the last three years has shown that it is possible to do things differently. There is light at the end of the tunnel.

We have not only measured what goes on inside a cryogenic freezer, we have thought about it as well. As a result, we have the first clear understanding of how to design a freezing tunnel that makes process engineering sense.

We have already supplied prototype freezers designed according to our new system to two major suppliers of frozen food in the UK. They have been in service for nearly a year now, and the customers are delighted.

The new freezers have overall heat transfer coefficients of 120 W/m2K — around twice the value found in a conventional cryogenic freezer. As a result, not only do they use less liquid nitrogen for the same freezing duty, but they are physically smaller: the new design typically takes up only half the floor space of a traditional cryogenic freezer. Freezing is more consistent, and dehydration losses are reduced.

Air Products is planning to launch the new freezer design commercially within the next year. Meanwhile, this presentation explains what we have learned from three years of research.

Problems with existing designs

Existing cryogenic freezers suffer from two fundamental problems. The first is lack of controllability, especially under changing load conditions. This shows itself as excessive consumption of liquid nitrogen.

Operators often complain that their freezers are using too much liquid nitrogen. The freezer manufacturer’s usual response is to send in the experts, who adjust the setpoint temperature, gas flowrates and belt loading to suit the current operating conditions. The problem goes away — until the next time conditions change, when the liquid nitrogen consumption shoots up again.

Lack of controllability also makes itself felt during breaks in production. Food passing through the freezer as it is being shut down tends to be over-frozen, and liquid nitrogen consumption during the shutdown period is high. When the freezer is brought back on line again the food at the front of the conveyor can be under-frozen.

The second main problem is variability in freezing performance across the width of the belt.

Freezer operators know that when food in the middle of the belt is processed correctly, food at the edges is likely to be over- or under-frozen. Before we developed our new design, one of the frozen food manufacturers with whom we have been working had to employ someone full-time to correct this problem by moving product between different cold storage areas.

Our solution to the lack of controllability involves dividing the freezer up into zones. This allows us to control the temperature profile along the tunnel much more accurately than is possible with existing designs. In effect, we have decoupled the control of heat transfer from that of temperature.

As well as improving controllability, the zones increase heat transfer rates. Other changes we have made — adjusting conditions at the spray nozzles, moving the fans and changing the aerodynamic design of the tunnel — also improve heat transfer. The result is lower consumption of liquid nitrogen and a smaller freezer.

Just as importantly, the redesign of the tunnel aerodynamics has made the gas flows within the freezer much more uniform. This in turn has dramatically reduced the variability in freezing performance across the width of the belt.

In the following sections I will explain the problems of existing freezers in more detail and explain how we arrived at the new design.

Lack of control

The fact that existing freezers have trouble adjusting to changes in load shows that they are inadequately controlled.

This may sound surprising to anyone operating a freezer equipped with the latest in self-tuning controllers. However, even the best controller cannot stabilise a system that is fundamentally uncontrollable.

The trouble is that all existing freezer designs work by controlling the temperature at a single point somewhere along the tunnel. The position of this point varies from one manufacturer to another, but in every case it is a single point. It is not difficult to show why this is not a good way to control a freezer.

We can see what is going on along the length of the freezer by drawing temperature profiles for the food product and the tunnel atmosphere.

[Figure 1: Temperature profiles along conventional freezer]

For practical purposes, in any particular application the endpoints of both these temperature profiles are fixed:

  • the inlet and outlet temperatures of the food are fixed by the requirements of the upstream and downstream processes;
  • liquid nitrogen is added at a single point (the spray zone) and at a fixed temperature. In practice this inlet temperature is around –150°C;
  • to minimise liquid nitrogen consumption the exhaust temperature of the gas should be as high as possible. To make the exhaust temperature the same as the food inlet temperature would require an infinitely long freezer, so we usually settle for a gas exhaust temperature 30–40°C below the food inlet temperature.

What is the effect of controlling the temperature at a single point part-way along the freezer? Let’s say the throughput increases:

[Figure 2: New temperature profile superimposed on original]

The liquid nitrogen inlet temperature and the temperature of the control setpoint are the only points that remains constant. The rest of the gas temperature profile changes to accommodate the new conditions by pivoting around the setpoint. As the flowrate of liquid nitrogen increases the profile becomes flatter and the nitrogen exhaust temperature falls.

In fact, the freezer responds to an increase in load by throwing away more heat in the gas leaving the tunnel. Clearly this is not a good way to operate, either for economy or for good control over the freezing conditions.

A better system would be to divide the length of the freezer into several zones — four or five is enough in practice — each fitted with its own liquid nitrogen spray. We could then control the temperature of each zone by adjusting its liquid nitrogen flowrate.

[Figure 3: Zones and temperature profiles inside the Vortech freezer, not showing fans]

This system would ensure that the freezer works as economically as possible, because it would allow us to set the exhaust temperature of the gas close to the temperature of the incoming food. The resulting temperature profile would be stable when the load on the freezer changes.

In the conventional model of a cryogenic freezer, liquid nitrogen supplied to the spray zone evaporates immediately. According to this model, therefore, it is a bad idea to add liquid nitrogen anywhere else in the tunnel because this would lower the nitrogen exhaust temperature.

But the conventional model is seriously wrong. In the next section I will show why this is so, and how the zone model makes sense for reasons of heat transfer as well as controllability.

Too much liquid nitrogen

High consumption of liquid nitrogen is partly a control problem, as explained above. There is, however, another side to the story.

According to the conventional model of a cryogenic freezer, liquid nitrogen sprayed onto the food evaporates completely in the spray zone. It is in the spray zone that most of the heat transfer takes place, and the rest of the tunnel’s length serves merely to pre-cool the food as it passes through the cold atmosphere of nitrogen.

Yet not all the liquid nitrogen does evaporate in the spray zone; some of it drips through the belt and accumulates at the base of the freezer. Manufacturers know this, and fit trays beneath the belt to catch the surplus. In doing this they are clearly admitting that the conventional model is wrong.

What are the effects of this? Spills of liquid nitrogen can damage insulation and steelwork, as many freezer operators have found to their cost. But there is a more fundamental problem.

We have proved by experiment — and in practice every freezer manufacturer realises this — that it is simply not possible to evaporate all the liquid nitrogen in the spray zone. In fact, only about a quarter of the total flow of liquid nitrogen evaporates on contact with the food or the surrounding gas. The amount that evaporates is limited by the available surface area, the short contact time available and the time taken to conduct heat through the thickness of the food product.

The remaining liquid nitrogen collects in the trays, where it is slowly evaporated by heat transferred from the returning belt trace. Now this liquid nitrogen is not surplus to requirements. Without it the heat balance over the freezer simply would not add up, and the food would come out under-frozen. Yet evaporation from beneath the belt, rather than from the surface of the food, dramatically lowers the effectiveness of the heat transfer process.

Heat transfer coefficients — the rate of heat transfer per unit area, divided by the temperature difference — are highest when droplets of liquid nitrogen evaporate directly from the surface of the food. This is why we spray the liquid nitrogen onto the food in the first place.

Once the liquid nitrogen has dripped through the belt and collected in the trays beneath, it cannot remove heat directly from the food. Instead it must evaporate from the trays, cooling the gas in the tunnel, and rely on the gas to cool the food. Gas-solid heat transfer coefficients are poor, so we end up with a freezer that is much longer than it would need to be if we could transfer all the heat directly to the liquid phase.

Note also that by cooling the belt, the liquid nitrogen in the trays reduces the temperature of the gas leaving the freezer. This is the opposite of what we want: for lowest running costs the temperature of the exhaust gas should be as high as possible.

The solution is obvious once we have identified the problem. With a single spray zone we cannot evaporate all the liquid nitrogen, yet we cannot throw away the “surplus” because this would upset the heat balance. On the other hand we would prefer to evaporate the liquid nitrogen in a spray zone (because this gives higher heat transfer coefficients) than from beneath the belt.

Clearly we need multiple spray zones. The zoned approach that we suggested for reasons of controllability turns out to fit very well with our improved model of heat transfer inside the freezer.

By adding liquid nitrogen in controlled quantities all the way down the length of the tunnel, we can use the same total amount of liquid nitrogen but allow it to evaporate directly from the food surface. Heat transfer coefficients are higher and the freezer can be made smaller.

We have also found that we can improve heat transfer coefficients in the spray zone by adjusting conditions at the spray nozzles.

[Figure 4: Nozzle pressures and heat transfer coefficients]

Higher discharge pressures give the right combination of droplet velocity and size distribution, and the liquid-solid heat transfer coefficient increases — often quite markedly. In a conventional freezer, high discharge pressures tend to aggravate the problem of excess liquid nitrogen. With the zoned design this is less of a problem.

Variability across the belt

So far we have concentrated on solid-liquid heat transfer. Gas-solid heat transfer, however, can also be very significant — if we take the trouble to get it right. It can easily become the weak point in the heat transfer chain, and we should not neglect it.

But variations in freezing performance across the width of the belt show that something is badly wrong with the gas flow in existing freezers.

[Figure 5: Heat transfer coefficient / length for Cryoquick 4 m tunnel]

Figure 5 shows how the heat transfer coefficient varies along the length of a typical 4 m-long cryogenic freezer. Notice that the heat transfer coefficients underneath the fans are quite high — 80–100 W/m2K — but very localised.

Gas velocity is one of the most important factors controlling solid-gas heat transfer coefficients. It looks as if by providing high, uniform gas velocities we could both increase overall heat transfer coefficients and eliminate performance variations across the width of the belt.

The maximum gas velocity tends to fixed by the need not to blow the food around. We therefore looked at how to distribute the gas flow evenly and achieve high, uniform gas velocities.

There are two basic solutions to a problem like this: increase the distance between the fans and the food, or use flow-straightening devices such as vanes or slotted plates. Flow straighteners are not ideal because they tend to ice up, which destroys their effectiveness. For similar reasons we decided against using fans with complex airfoil-section blades.

Instead we decided to increase the distance between the fans and the food. Why not put the fans under the belt?

[Figure 6: Aerodynamic layout of the Vortech freezer]

Our new design has base-mounted fans and a specially-shaped tunnel top that generates a pair of vortices above the belt. The vortices are important: they mix the gas, helping to make conditions uniform across the tunnel, and they ensure that gas impinges on the food product at an angle close to 90°. This gives better heat transfer than if the gas were simply blowing across the top of the food.

[Figure 7: Vortices video]

Figure 7 is a still from a video of the vortices in action. It shows how a combination of high gas velocities and good mixing generates high, uniform heat transfer between the gas and the solid.

The base-mounted fans have the extra advantage of stirring up any excess liquid nitrogen droplets that do fall through the belt.

Another significant point is that by adding liquid nitrogen in zones along the length of the tunnel we eliminate the gas transfer fans needed to move the nitrogen through a conventional freezer. This saves on fan power and simplifies the flow patterns within the tunnel.

The fan speeds are set manually according to the type and throughput of product. It is also possible to control fan speed by measuring the exit temperature of the product, but there are practical difficulties in doing this accurately. We think manual control is more reliable.

Note that most of the power absorbed by the fans ends up in the tunnel’s heat balance and so must be removed by the liquid nitrogen. This is another reason why it is essential to be able to control fan speed independently of liquid nitrogen flowrate. During breaks in production, for instance, we can slow the fans down until there is just enough gas movement to prevent stratification. The freezer remains cold but liquid nitrogen consumption is very much less than for a conventional freezer running unloaded.

How it stacks up

Table 1 shows the most important figures for the new freezer compared with the traditional design.

[Table 1: Freezer league table]

The figures speak for themselves — but do they match the feelings of our customers who have tested the new freezer?

For nearly a year now we have had two full-size test units running with two different customers. One freezer handles meat, the other bakery products and confectionery. One has a throughput in the range 300–600 kg/h, while the other processes up to 2000 kg/h.

Both these customers are delighted with the new design. As well as the saving in floor space, they have seen liquid nitrogen consumption fall by around 14%. Dehydration losses are lower, because the food is frozen more quickly. There has been a marked improvement in the consistency of product temperatures, especially during startup and shutdown.

Air Products plans to be selling freezers designed according to the new system within a year.

For the first time, and alone among the suppliers of freezing equipment, we really understand the heat transfer processes involved in cryogenic freezing. There really is light at the end of the tunnel.

back to main Samples page