Design and testing of the fuel cell twin-screw air compressor

The oil-free dry air compressor plays a crucial role in the fuel cell system, particularly in commercial buses where twin-screw air compressors have proven effective. A novel rotor profile of the twin-screw dry air compressors has been developed, which simplifies the profile curve and yields excellent performance results with a volume flow of 11.99 m³/min and corrected specific power of 3.54 kW/(m³/min).


Introduction
With the advancement of dry twin-screw air compressors [1][2][3], rotor profile design has become increasingly demanding in recent years.The twin-screw compressor's profile design has undergone three generations of development, with the third-generation asymmetric profile being widely recognized for its superior efficiency, performance, and volume capabilities.Due to lubricating oil's ability to seal, lubricate, cool, and noise in twin-screw compressors [4], oil injection screw compressors have become mainstream in the market; new profiles are typically developed for such compressors [5][6][7].However, certain special applicationssuch as fuel cell air compressors that require completely oil-free air to prevent pollution from oil particles on the fuel cell systemnecessitate dedicated rotors or specialized profiles for dry screw compressors.Despite this progress, technology development for dry screw air compressor profiles remains slow: many dry compressors still use second or even first-generation profiles that exhibit low efficiency and generate significant aerodynamic noise due to outdated profile designs.
The fuel cell air compressor plays a crucial role in the cathode system of vehicle fuel cells [8].By pressurizing the air, it enhances the power density while reducing the overall volume of the system.However, it is important to note that the power consumption generated by compressed air is the most important part, which consumes about 80% of the energy the auxiliary system consumes.Therefore, optimizing its performance is of great significance to improve the efficiency, water balance, and compactness of the fuel cell system [9].The types of compressors determine their different applications in fuel cell systems [10].The twin-screw compressor has been proven to be effective in the fuel cell systems of commercial buses, where pressure ratios typically do not exceed 3, and there are substantial requirements for gas volume.This study presents a novel design and experimental evaluation of a dry twin-screw air compressor profile demonstrating promising performance.

Profile design
Figure 1 shows the mathematical relationship of coordinate systems: XY for the static coordinate system, xy for the rotating coordinate system, subscript 1 for the male rotor, and subscript 2 for the female rotor.
Relationship of O1x1y1 and O1X1Y1 is, cos sin sin cos Relationship of O1X1Y1 and O2X2Y2 is, Relationship of O1x1y1 and O2x2y2 is, cos( 1) sin( 1) cos sin( 1) cos( 1) sin To satisfy the meshing relationship between rotor profiles, existing profiles are often composed of multiple curves that require complex calculations.A novel design of a 3-5 tooth profile is proposed with an area utilization coefficient of 0.517, resulting in greater volume flow than existing oil injection screw compressor profiles (with area utilization coefficients between 0.4 and 0.5). Figure 2 illustrates the rotor profile, while Table 1 shows the composition curves of the female rotor and the corresponding curves of the male rotor.The "e-cos" and polynomial curves are used to avoid splicing AB curves.As depicted in Figure 3 and Figure 4, in polar coordinates, the slopes of points A and B are both zero (indicating tangency between the tooth tip and root arc), enabling the identification of curves that satisfy these conditions.Three such curves meeting this criterion are provided.Curve 1 is an, bn and cn (n = 1-3) can be calculated by Formula (4).where  represents the distance between any point on the coordinate system and its origin, r2 denotes the radius of the female rotor's root circle, and R2 signifies the radius of its top circle.
It is important to note that at point B (the tooth root) of the female rotor, excessive curvature should be avoided to prevent interference with the conjugate curve on the male rotor.Based on design experience, when θ is small, solving for conjugate curves becomes easy to interfere at point B. Consequently, while curve 1 satisfies mathematical tangency requirements, it is unsuitable as a tooth curve for the female rotor.Instead, curve 2 and curve 3 facilitate obtaining a greater curvature radius at point B.
In traditional profiles, engagement line 2-2' formed by BB' arc curve completely engages with its conjugate envelope (also an arc) at one moment.Still, it fully disengages in subsequent moments, leading to pumping effects in dry air compressors where gas film within narrow gaps between rotors gets disrupted, resulting in increased aerodynamic noise levels.Similarly, the pumping effect can cause oil film rupture for oil injection air compressors, thereby reducing sealing efficiency.Conversely, in an optimized profile, the AB section constitutes an uninterrupted curved segment wherein meshing points transition continuously from point 1 to point 2, ensuring a smooth operation process and effectively minimizing aerodynamic noise during compressor operation.

Experimental theory
The experimental setup and the test bench are illustrated in Figure 5.The variable frequency motor speed is controlled through the console, which drives the meshing rotors via a gearbox.The console provides a real-time display of motor speed, input power, and other parameters.The compressor's outlet pressure is regulated by adjusting the valve opening.Sensors, including pressure gauges, temperature sensors, and pressure differential meters, are installed on the outlet pipe to measure compressor volume flow accurately.An atmospheric pressure gauge and environmental thermometer are also positioned at the site for reference measurements.The noise of the entire system under operational conditions is measured using a noise meter placed at a specified distance from the compressor unit.
Upon the motor activation, the console displays its speed and power while synchronous gears drive male and female rotors to operate at a specific speed ratio.As air passes through the inlet filter, it enters the compressor inlet, where compression occurs due to the action of the rotors before exiting through an outlet pipe where the volume flow rate is measured.The collected data from this experiment include inlet pressure (Pin), inlet temperature (Tin), rotational speed (n), input power (P), temperature and pressure in front of the nozzle (Tout and Pout), pressure differential meter reading (H), as well as noise level measurements.Relevant parameters can be calculated using the formulas provided below.

Compressor performance at design point
As shown in Table 2, at design condition, the outlet pressure is 3 bar, and male rotor speed is maintained at 8000 rpm; volume flow rate, specific power, and noise level were determined by conducting standard nozzle flow tests with adjustments made for local atmospheric conditions such as ambient air temperature and atmospheric pressure variations.
Results indicate that the air compressor achieves promising performance parameters compared to dry air compressors of similar types: volume flow rate is 11.99 m³ /min; corrected specific power is 3.54 kW/(m³ /min); and noise level is 94 dB.

Compressor performance under different speeds
Figure 6 shows compressor performance at different speeds, which displays the changes in compressor volume flow and specific power under various male rotor speeds (Pout maintained at 3 bar).Figure 6 shows a linear relationship between volume flow and speed when the male rotor speed exceeds 3000 rpm.Similarly, when the rotor speed surpasses 5000 rpm, specific power, and speed exhibit an approximately linear trend.
As the compressor's speed increases, its suction end can inhale more gas.At the same time, higher linear velocity reduces leakage gas proportion in the overall flow rate, leading to a linear correlation between flow rate and speed.Regarding specific power, since the compressor itself consumes some energy, low gas volume results in a reduced consumption power ratio of compressor gas, causing a rapid rise in specific power.

Compressor performance under different outlet pressures
Figure 7 illustrates the variations in compressor volume flow and specific power under different outlet pressures while maintaining a constant male rotor speed of 8000 rpm.As the outlet pressure increases, the volume flow and specific power exhibit linear changes, indicating stable operation of the compressor at its rated speed with stable leakage and power loss.Furthermore, Figures 6 and 7 demonstrate that when operating under off-design conditions, the influence of compressor speed outweighs that of outlet pressure.

Conclusions
The design of the fuel cell air compressor profile is proposed by the third-generation method, simplifying the formation curve.Experimental results reveal a volume flow of 11.99 m³ /min and a corrected specific power of 3.54 kW/(m³ /min), showcasing superior performance compared to dry screw air compressors of similar types.These findings hold significant implications for enhancing fuel cell system efficiency and reducing vehicle device weight.
Based on spatial meshing principles in rotor profile design and operational characteristics of vehicle fuel cell air compressors, this study provides a foundation for developing green and efficient independent intellectual property rights in rotor profile design.

Figure 3 .
Figure 3. Curve AB in polar coordinate.Figure 4. Curve AB at rectangular coordinate.

Figure 4 .
Figure 3. Curve AB in polar coordinate.Figure 4. Curve AB at rectangular coordinate.

Figure 5 .
Figure 5. Experimental setups and the test bench.

where 1 
Ps represents driving power, represents electric efficiency (0.93), 2  represents transmission efficiency (0.98), Qv represents volume flow, a represents local inlet pressure correction coefficient (1.042), C represents the coefficient of the nozzle (0.993), D represents nozzle diameter (50.19 mm), Pb1 represents specific input power, and Pb represents corrected specific power.

Table 2 .
Parameters at the design point.