Galin compressor: what it is, how it works, and its advantages and disadvantages
Let us first start by setting the context of compressor technology, leading architectures, their efficiency considerations. First, what are compressors? Compressors are devices that convert energy into the potential energy of compressed gas. Most widely available compressors convert electrical energy into the potential energy of air. The most widely used traditional compressor technologies are rotary screw and reciprocating piston compressors. (There are many more varieties, some of which are mentioned here.) A rotary screw compressor is generally more expensive than the piston, requires specialised maintenance, but can operate continuously, have high(er) flow and is relatively quiet and free from vibration. A piston compressor is comparatively cheap, reliable, but cannot operate continuously (although there is no technical reason why it can’t - consider that cars/engines run for a very long time and do okay) requires a reservoir to buffer flows, is very noise, and vibrates.
Users of piston compressors often complain about noise and overheating (inability to run the compressor for longer). Noise from piston compressors pose problems in workshops where they limit the ability for workers to communicate and hear other equipment being used around the workshop. Rotary screw compressors offer a low-noise alternative, but often at a prohibitively higher cost. (Marketing efforts are currently underway to convince people that rotary screw compressors are in fact cheaper, when comparing the overall lifetime use of the compressor.) Compressor customers cannot choose a technology that is cheap, efficient, reliable, continuous, high flow and quiet at the same time.
The piston and rotary screw compressors account have about 80% share of the compressor market. New Zealand’s Energy Efficiency and Conservation Authority (EECA) estimates that about 10% of all industrial electricity is used by air compressors. Interestingly, there are currently no regulations (except in China) in place for the energy efficiency of compressors. This is due to change with introduction of guidelines in the United States, and Europe - none yet for Australia and New Zealand. The CAGI (Compressed Air & Gas Institute) has requirements for detailed datasheets that should be provided with compressors to include their isentropic efficiency and other necessary values to asses the energy use of the given compressor.
Winge: But why are CAGI datasheets only required for rotary screw compressors? Piston compressors have an almost equal overall market share. It may be that the market of compressors used by industry is dominated by rotary screw compressors….but rotary screw compressors use more energy than piston compressors, and publishing a comparison of their efficiency could lead to a move away from rotary screw compressors? Initially, my thoughts were that rotary screw compressors are not only less noisy, but also more energy efficiency than piston compressors, but in fact, this is not the case. Rotary screw compressors are required to run almost 100% of the time, and hence, when a reservoir is filled, the compressor keeps running but does zero useful work. Here’s a document with a lot more detail.
Galin compressor: is a new compressor design, fundamentally different from both the piston and rotary screw designs (and all the others). A Galin compressor comprises a cylindrical chamber, with a shaft emerging from each circular face, each shaft driven by a precision permanent magnet electrical motor. Inside the cylinder, the shafts mate but rotate independently. Each shaft has vanes attached, which extend the full height of the chamber, dividing the chamber’s volume into four sub-chambers of variable volume. The electrical motors drive the shafts to (a) vary the volume of the sub-chambers, creating the intake and compression strokes, and (b) rotate the shaft assembly within the cylinder to align high and low pressure chambers with outlet and inlet ports respectively. A Galin compressor is mechanically simple, making it cheap to manufacture and reliable to operate. The design has few moving parts.
Compared to a piston compressor, Galin compressor has no slider-crank mechanism, valves or camshaft; it has fewer parts, better mass balance and hence less vibration and noise than a piston compressor, finally, all shaft movements contribute to doing work to compress air. Compared to an equal volume single-cylinder reciprocating piston compressor which is able to perform a single compression per shaft revolution, the Galin compressor performs 8 compressions for every shaft revolution.
Compared to a rotary screw compressor, the design of Galin compressor is simpler and hence easier and cheaper to manufacture, maintain and hence we can offer it at a lower price point. Additionally, in Galin compressor the shafts have no gearing necessary, and there is no technical requirement to run the compressor 100% of the time, thereby increasing the efficiency of the machine.
Technical details:
I spent some time agonising over how best to describe and compare the operation of Galin compressor (GC) with a reciprocating piston (RP) compressor. The best I can come up with is by imagining a double ended RP cylinder bent into a torus (and with perfectly synchronised piston motion) see figure 1 below:
Figure 1 - showing a comparison and similarities between a reciprocating piston (RP) and Galin compressor (GC). The reason we are making this comparison, is that people have developed intuitions regarding operation of RP compressors, but none regarding GC. A comparison like this, potentially bootstraps intuition for GC. Note: the synchronisation of the crankshafts in the RP is not shown, and in practice difficult to achieve, whereas in GC it is a given (via position sensors on the shafts).
GC compressor takes the above torus arrangement full circle, and separates the circular chamber into four volumes. Figure 2 shows a schematic of the GC, with the rotary pistons, with one pair attached to a shaft coming out of the page, and a pair attached to a shaft going into the page. The intake and exhaust ports are labelled. Because the rotary pistons create 4 chambers between themselves, during a single rotation of the shaft 8 compression strokes are achieved. This circular arrangement allows the vanes to rotate in a single direction only, which has the advantages that there is no stop/start forces necessary and hence smoother operation is achieved.
The differences of GC and a RP are the following:
CFM:
GC: 8 compression strokes per shaft revolution
RP: 1 compression stroke per shaft revolution.
Therefore, for the same CFM, a GC can either have a smaller chamber volume, leading to an overall smaller sized pump, or decrease the speed of rotation.
Volumetric Efficiency:
GC: the swept volume is entirely software controlled, and hence volumetric efficiency is a controlled variable - the rotary pistons can be brought as close together as required.
RP: volumetric efficiency is fixed by the crankshaft mechanism and cylinder geometry.
Direct Drive:
GC: the rotary pistons are directly driven by electrical machines and require no gearing or belts.
RP: electrical machine is connected to the crankshaft via a belt.
In GC we eliminate friction losses associated with this mechanism (friction losses within the crankshaft mechanism account for 15% of all losses).
Noise/Vibration:
GC: we have no reciprocating (stop/go) motion of the rotary pistons, and the system is always balanced. Hence, we can expect lower vibration and lower acoustic noise.
Valves:
Unlike the RP there are no valves in GC. The chamber has fixed locations of intake and exhaust ports.
Maintenance/Reliability:
GC: no crankshaft mechanism and no valves. Fewer parts to break and maintain.
Figure 2 - schematic of the Galin compressor. Two intake and two exhaust ports mean that a single revolution of the shafts permits 8 compression cycles.
Simulation and performance estimation:
My GitHub repo here contains the simulation of a GC and RP compressors. The goal of this code is to compare the theoretical performance of a reciprocating piston compression to that of the Galin engine compression stroke. We are doing this comparison to figure out how the Galin architecture stacks-up relative to the reciprocating piston based compression.
Compression parameters:
polytropic index: 1.3
suction pressure: 100 kPa
suction temperature: 295 K
discharge pressure: 689.6 kPa (100 psi)
intake volume: 900 mL
Common material properties:
steel density: 7800 kg/m^2
Reciprocating piston geometry:
bore: 94 mm
stroke: 101 mm
rod: 202 mm
piston area: 6893 mm^2
piston mass: 9.7 kg
Galin compressor geometry:
vane radius: 125 mm
shaft radius: 42 mm
vane area: 6893 mm^2
vane mass: 9.7 kg
Running the simulations with the above parameters give us the following outputs:
Common output parameters
compression ratio []: 4.4
work done on the gases during compression [J]: 168.4
pressure difference, start to end [kPa]: 100.000000 -> 689.5 [14.5 -> 100.0 (psi)]
volume difference, start to end [mL]: 900.0 -> 203.8
temperature difference, start to end [K]: 295.0 -> 460.6
Reciprocating piston compression stroke performance:
stroke time [ms]: 56.4
torque requested from electrical machine as 0.8 of max [Nm]: 69.4
based on the torque load on EM machine * 0.8 we calculate the power of the EM [kW]: 3.9
CFM: 16.9
CFM/HP: 4.4
Galin compression stroke performance:
stroke time [ms]: 71.3
torque requested from electrical machines [Nm]: 138.6
electrical machine power req. (each) [kW]: 3.1
CFM: 53.5
CFM/HP: 8.8
Notes:
In the case of the reciprocating piston simulation, we calculate the force required to compress the gases as: work required to compress gases / stroke, we perform a sanity check to make sure that in both simulations the total work done by the piston and vanes on the gases is equal to the required work to compress the gases, we provide a plot versus time under pics/torque_vs_time.png (see figure 3). This plot shows that the requested torque (while higher in magnitude) from the electrical machines is constant in the case of GC, whereas in the RP case it varies wildly as the lever arm changes.
Figure 3 - comparison between the torque requirements from electrical machines in the RP (black line) and GC (blue line). The torque requirement from EMs is much greater in the GC case than in the RP case. In the case of RP, the required torque is mediated through the crankshaft mechanism, and we get a mechanical advantage because instead of having to do work "W" over a short theta, we apply torque over a theta of 180 degrees (half a rotation of the crank). So, just going slowly on this....
GC) W = Force x stroke = Torque x theta (where the applied Torque can be constant because the lever arm is constant. As the gases get compressed, they will push back harder against the vanes, and this will increase the load torque on the electrical machines, but we can in fact get away with applying a constant torque (due to the constant lever am) which is just the arithmetic average of the load torque over the period of compression.
RP) In calculating the torque requirement from the electrical machine we need to do this (note that theta varies from 0 to pi, a much larger value compared to GE):
$$W = \int_{0}^{\pi}T(\theta)\cdot\theta d\theta\\ = \int_{0}^{\pi}F\cdot l(\theta)\cdot\theta d\theta$$
where \(l(\theta)\) is the lever arm: the distance between the crank axis and the piston pin axis. The effect of this is to lower the torque requirement from the electrical machine.
Conclusions:
The simulated CFM/HP for the reciprocating piston engine is close to the often quoted figure of 3.5 (at 100psi) in industry, so our simulation isn't too unrealistic. The compression stroke is 25% quicker in the case of the reciprocating piston geometry - the advantages of this are that the system is closer to adiabatic conditions with less heat lost externally. however, the CFM is three times greater in the case of GC which is a huge improvement. CFM/HP is double even when taking into consideration that in Galin architecture we require two electrical machines.
It may be that the above calculated torque requirements of the electrical machines are severe when compared with the RP case, however, consider the reverse case when GC is run as an expander - and the phenomenal output power deliverable!