Summary

Being by nature very inquiring about energetic problems, I thought that something could be done about RC (Radio Controlled) helicopter models. At this time, early 1997, the state of the art in flight duration of electric helicopters on the market was about 6 to 7 minutes ; this being done in hover with 1.7Ah NiCd battery cells. My idea was then to check whether it was possible to bring a significant  improvement. This work would do for any power plant :  electric motors and piston motors. The HBE project , was then born.

Discussing about what I had in mind with a friend of mine, Pierre Hermet, engineer, he came very nicely to this idea and started the job on the go. He could therefore exercise his real talents of  theorist. Although this domain was far from our professional activities, he went right in the matter, building up a  rotary wing mathematical model on his PC. I was impressed by the realistic results of his model so we needed to check that with reality. Our work method was then born, consisting in matching theoretical model (simulator) and experimental result to select the right way in our future developments.

Not being experts in aerodynamic matters, it was necessary to start from the sources and almost discover all the problems. So we came to the idea to work out a test bench to drive a rotary wing and to measure the lift and the torque at a given spinning speed and pitch angle. The first bench, too simple, provided a bad speed control and enabled only to compares relative performances between blades (blades from the market at that time). We realized it was then required to perform an accurate speed and pitch control in order to get a significant comparison of the results with the theoretical model.

Summary

The second bench is divided in a control unit and a drive unit. The control unit is based on a computer (PC 386) and an interface cabinet (designed and home made) with the drive unit via a control cable and a power cable. On its side, the drive unit is based on a framework that includes a brushless motor with its controller, mechanical reducer, clutch, 4 blade rotor head and its collective pitch control servo (no cyclic control) ; in addition, the sensors dedicated to lift, torque, rotor speed, motor current/voltage are also included. The pitch is designed to be adjusted by steps of 1/10 of degree and memorized in tables as any other configuration parameters. The rotor speed is servo controlled within a few rpm's. For each measurements an hundred of samples are picked up and averaged. The complete bench insured a good reliability in the measurements which are made either on ground effect (IGE) or out of ground effect (OGE) using solid legs to rise up the drive unit. The software (made by my son Vincent) enables the configuration of the bench, the programming of automatic measurement sessions and to view the real time tracing of « lift to speed » « torque to speed » graphics on the screen.

Summary

Doing measurements on blades coming from the market is good, but making our own blades to our own requirements is much better. A solution that doesn’t require industrial means had to be found. Taking an aluminum sheet folded on itself was the basic idea, closed at tip and trailing edges and then injected with epoxy resin. This technique is quite efficient and provides a high technology added value. However this can be reached under the condition of implementing an adequate set of tools. This one includes essentially a multifunction folding/rolling device, a sort of plane used to machine the wooden blade templates, a mounting press and a warming chamber for resin injection and hardening. Thanks to that different blades were made with symmetrical airfoils (NACA) or cambered (Eppler) with rectangle or trapeze shapes, plane or twisted for right or left rotation. In airfoil analysis, we got some help from a bank of airfoils UIUC (http://amber.aae.uiuc.edu/~m-selig/ads/coord_database.html) and from the Web site of Martin Hepperle (http://beadec1.ea.bs.dlr.de/Airfoils/calcfoil.htm) for airfoil computation.

Summary

Computation and experimental results turned out to be very close which enabled us to undertake the rotor energetic optimization with the simulator. For a given lift and rotor diameter, airfoil choice, chord and rotor speed are the most significant variables of this optimization. The biggest difficulty stays in the drag estimate at very low Reynolds number (Refer to <the helicopter blade>). Later on we have improved the simulator including the rotor conicity and the blade distortion (flexion, twist) upon centrifugal and aerodynamic constraints.

We now have acquired a good mastery of the subject which enables us to simulate the rotor within an accuracy of a few percent. Of course some coefficients are still empirical but they are not most significant. Unfortunately we cannot simulate correctly the tip vortex as this can be done in a research center, the ONERA (Office National de la Recherche Aéronautique) for instance. It might remain a potential gain of a few percent using optimized tips but no way is clearly open in this domain.

snapshot of the test bench

rotor computation

Comparing our blades to conventional ones in very similar conditions (diameter, speed) we estimate the gain in torque and power to about 22% in favor of our blades (of course !). If you adopt such blades on a conventional helicopter the motor will be significantly relieved and adjustment will be easier (less power, easy cooling). Our faithful 30 Baron testifies of it.