Individual Blade Control for Vibration Reduction of a Helicopter with Dissimilar Blades

Abstract

A control method is proposed to reduce vibrations in helicopters using active trailing-edge flaps on the rotor blades. The novelty of the method is that each blade is controlled independently, taking into account possible blade dissimilarities. This is different from previous control approaches that assumed blades were identical and generated a single control input, which is applied with adequate phase shift to each blade.

The controller is developed in discrete time, with the control inputs updated every rotor revolution. The method consists of performing simultaneous system identification (using Kalman filtering technique) and closed loop control (using a deterministic control law) at each time step. For the system identification, different inputs are applied to each blade, and the relationship between the individual blade inputs and the resulting loads in the fixed frame is estimated on-line, assuming a linear-time-periodic model of the helicopter.

A comprehensive rotor analysis, including all blade degrees of freedom and a free wake model for computing the inflow across the rotor disk, was used to investigate the controller performance in detail. The rotor model is based on a modern bearingless rotor that includes detailed modeling of trailing edge flap effects. The controller performance was tested at advance ratios from 0.10 to 0.40, both for a baseline rotor with identical blades and a damaged rotor with dissimilar blades. In the case of the dissimilar rotor, comprehensive analysis predicts that allowing independent control inputs for each blade dramatically improves the vibration reduction compared to restricting the control inputs to be specific phase shifted versions of each other.

In order to test the controller experimentally, a Mach-scale rotor model was fabricated. The rotor model consists of 4 blades with piezo-ceramic actuated trailing edge flaps. A new type of hinge using flexures was designed to improve the flap articulation and incorporated in each blade.

The smart rotor model is then fitted on a bearingless model-scale hub and tested both on a hover stand and in the Glenn L. Martin wind tunnel. Both rotating-frame as well as fixed-frame vibratory loads were targeted in the closed-loop control tests. These tests demonstrate the controller's ability to account for blade dissimilarities and generate different optimal inputs for each blade.

For example, in hover, at 500 RPM, the 1/rev bending moment at the root of three of the blades was simultaneously reduced by 77% using three active blades. In forward flight, the controller could simultaneously reduce the baseline 4/rev fixed frame vibration as well other harmonics of vibration such as 1/rev and 3/rev arising from blade dissimilarities. It was also possible to minimize vibration in the fixed frame for several loads simultaneously. However, for most control tests, increases in other loads (not included in the control objective) were observed. During most closed loop tests, the maximum allowable input to the actuators was reached. It was found that the method used to account for actuator saturation and maintain actuator input within acceptable limits had an important effect on controller performance. The best controller performance was obtained when control inputs are computed by solving the constrained minimization problem. However, this procedure is very time consuming and could not be implemented in real-time with the available computer.

It can be concluded that accounting for blade dissimilarities using individual inputs for each blade results in improved vibration reduction. However, to maximize the benefits of this control scheme, an efficient, practical method to limit control inputs needs to be devised.