Fans are the unruly teenagers of the cement plant. They often do what they shouldn’t and don’t do what they should. If we can understand why they behave as they do, they can be helped to fit better into their surroundings and maximise their output.
There are two types of fans: Axial and centrifugal. They behave differently due to their main difference: Axial fans are best used for high volumes of gas at low pressures, for example in warehouse ventilation systems. Centrifugal fans are best suited to lower volumes with higher pressures. They have a much better ability to ‘suck,’ which is important in a wide range of industrial applications, such as commonly found in dust filtration systems.
Centrifugal fan basics
Centrifugal fans are fairly simple (See Figure 1). There is a wheel with fitted blades that spins inside a housing. Air is drawn tangentially through the inlet and is expelled through the exhaust. However simple this may be, it seems that even in the cement sector, this principle is poorly understood or even ignored.
Speaking from experience, there are many fans that are poorly configured. In an extreme case, I actually took a three hour flight to a client to troubleshoot a fan system, only to discover that the fan was working in reverse. This happens more than you would think, as many mistakenly liken centrifugal fans to axial ones. If you run a centrifugal fan backwards, it will work, but it will do so at a really low efficiency. Such cases are embarassing for everyone involved. Users should not need an engineer to visit to find out that the fan wheel is rotating backwards.
Types of centrifugal fans
We can split centrifugal fans down into three categories (See Figure 2). Radial fans have straight blades that radiate from the centre of the fan wheel. They are the most common type of centrifugal fan and are the most reliable in dusty airflows. They are slightly less efficient than forward curved fans, which can be used for higher volumes, and backward curved fans, which can be used for higher pressures. Each has advantages and disadvantages that make them more or less suitable for a particular application. While the client is rarely in the best position to specify the type of fan, it is nevertheless very helpful for them to be aware that each type has its strengths and weaknesses.
System curves
System curves and fan curves help to define the necessary operating conditions in which a fan will be efficient. Figure 3 shows a typical ventilation system, as used in a dedusting system. Air is drawn from the vent point through some duct work, to the filter bags, into the clean side and out of the exhaust via the fan.
If we imagine the system with no airflow, there is no resistance to that airflow. This is the origin in the system curve shown in Figure 4. If we start to speed up the airflow through the system, the pressure increases. The faster the gas moves, the more resistance there is to the flow. The hood may contribute 2.5mbar of resistance at 10,000m3/hr, the ducting may have a resistance of 10mbar. The dust collector will have a resistance of around 15mbar, more if it is under a heavy dust load. On the clean side there will be more resistance from the outlet and stack, perhaps 3mbar. In this example the total system resistance is 30mbar. It could be more, if the system is clogged with a lot of dust, or less if it has just been serviced and has new bags. Changing the conditions changes the gradient of the system curve. More resistance, higher gradient: Lower resistance, lower gradient.
Fan curves
Fan curves are plots of flow volume against static pressure, for a certain (designed) fan speed in rpm. They are provided along with the fan and other specifications from the supplier. They become very useful when we overlap them on the system curve (See Figure 5). The operating point is where the two curves intersect. If the system is configured well, the fan will be at its most efficient at this point. If the curves intersect elsewhere, the fan will be inefficient. The target is to run the fan near this operating point.
In the region circled in Figure 5, we see that, if we restrict flow even slightly, the fan will become unstable. The gradient of the curve is very small in this region, meaning that the fan is liable to change from higher flow to lower flow with very little input. You can hear fans ‘revving’ in complaint when they operate in this range. Obviously variable flow like this is not good for steady plant operation.
Interestingly, if we were to overlap the power curve, we can see, contrary to expectation, the power does not skyrocket when flow is restricted in a centrifugal fan system. The motor actually relaxes when there is no flow in a centrifugal fan. If you restrict it the air that is already inside simply spins inside it. This is totally unlike a pump, where restricting flow will overload the motor.
Maximising fan performance
There are a number of ways to ensure that systems are optimised to best handle the required airflow and the operating requirements of the fan. The first, and one of the most intuitive, is the ductwork (See Figure 6). We can compare this to a fast-moving highway and simply need to apply some common sense. A highway will never have a bend as shown in Figure 6a. It will have a bend as shown in Figure 6b. The same principles can be extended to find optimum solutions for merging gas flows and separating them.
Sometimes we see ridiculous layouts. In Figure 7 the central duct leaving the top of the image is the duct to the bag filter. How on earth is the fan supposed to work with that? The answer is, it can’t. If the fan was specified to move 10,000m3/hr at 30mbar it might be capable of 5000m3/hr with the configuration shown in Figure 7.
Looking at the outlet velocity profile can be very instructive. It is (should be) obvious that the faster gas flow will be on the outside of the fan wheel and the slower gas will be on the inside (See Figure 8). Why does it matter? Well, let’s imagine that, instead of a nice straight lab set-up that the fan manufacturer used to take its measurements, there is an obstruction in the way of the exhaust and we have to turn the exhaust flow through 90°. Looking at the outlet velocity profile, we can see easily that the direction we choose makes a huge difference. Figure 9 shows the options. I would strongly recommend 9b, rather than 9a because the faster moving gasflow encounters a curve. The difference in resistance terms could be quite large. Figure 10 provides a further example of a fan that has little chance of performing to its specification.
Figure 11 shows a great example of fan engineering. Some might say ‘What about rain?’ The common solution is to put a ‘Chinese hat’ outlet on top. However, that represents a massive restriction in terms of the gas flow. Once again, the fan curve goes up into the unstable range and the performance drops off. So what solution has been used in Figure 11? It’s really simple. There is a 20 - 30mm gap between each of the sections of exhaust duct. As rain never comes down completely vertically, it hits the inside of the duct and then falls harmlessly onto the outside of the next section of duct. It’s a simple solution with zero restriction to the fan. However, like so many of the published guidelines surrounding fans, this is often overlooked.
Summary
This article has explained some of the characteristics of centrifugal fans and highlighted how best to optimise them in light of well-grounded physical principles. None of the information presented here is a ‘trade secret’ but it is still applied very inconsistently across the global cement sector. It is hoped that, even with this simple grounding, non-specialists should be able to easily spot poorly-configured systems and identify easy wins towards increased fan performance and process efficiency.