Since December 2019, we are developing a kite for the generation of electricity, mechanical or thermal energy.
Our current prototype (September 2020) works as follows:
The system consists of two very loosely coupled parts connected by a line, the kite and the ground station.
The purpose of the ground station is to convert mechanical energy into whatever is necessary, e.g. electrical energy and to keep the line tight. It contains a generator/motor connected to a spool (making it a winch), electronics to use the generator as a motor (an electronic speed controller), power electronics to convert the output of the generator to the desired voltage (e.g. rectifier + dc-dc-converter), sensors for determining the force on the line and the speed of the line and a processor for switching between the two modes of power generation and line tightening.
The sole task of the kite is to pull the line as tight as possible for as long as possible (precise equations follow below). The kite is a (temporarily) powered aircraft with a line attachment much like during the winch start of a glider. It contains sensors and a processor to enable it to fulfil its task autonomously. In our prototype we use a powered aircraft with one motor whose airflow passes by both rudder and elevator to enable it to hover with the nose pointing up. Our kite uses only 3 controls: one propeller for thrust generation, an elevator and a rudder. Rather than using ailerons to control the roll of the aircraft we have chosen an aerodynamically roll-stable design both with respect to the gravitational force and with respect to the force exerted by the line. This spares us from using a GPS-like sensor and dedicated wind sensors or cameras, as the windflow stabilizes the roll of the kite and pushes the kite such that the line points away from wind. From this wind induced orientation we can deduce the wind direction among other parameters.
The kite has 3 basic modes of flight:
1. Flying like a powered aircraft or like a glider
2. Flying like a kite
3. Hovering nose-up
In the first mode the force exerted by the line is relatively small in comparison to the gravitational force.
In the second mode it is the other way around.
In both the first and second mode of flight a D-controller on the elevator makes it possible to reduce the length of the tail and have the center of gravity or the line attachment further aft increasing the efficiency of the aircraft, while keeping compact and the control surfaces well in the airflow of the propeller.
In kite mode we use a six-axis inertial measurement unit (gyroscope + accelerometer) and a height sensor (barometer) to determine the relative direction of the gravitational force and the distance to the ground. On this basis alone the processor controls the kite.
The hover mode uses PD-controllers for elevator and rudder to keep the kite level (nose up). The roll is stabilised by the wind or left varying in a calm. The height is controlled using a PD-controller based on the barometer.
The kite is programmed in the following way:
While being idle during a simultaneous calm in all reachable heights the kite is loosely attached to the ground station nose pointing up such that it can escape upwards. As soon as there is enough wind in some reachable height (determined by weather forecast data or by dedicated sensory equipment) the kite enters hover mode and lifts to a height that has enough wind. It might be more energy efficient to cover some of this height by flying a spiral in powered aircraft mode.
During hover mode the neutral position of the elevator is set such that the line is held tight even in a calm, ie.e. about 10 degrees upward (in local coordinates) pitch.
Once the desired height has been reached, the existence of sufficient wind can be determined by reading the P-term (or directly the pitch angle) of the elecator. Having no I-term on the elevator and a line attachment in front of the center of pressure in hover mode pitches the kite into the wind.
Given sufficient wind the propeller can be turned off and the kite be controlled by the rudder to fly figure eights or circles in order to pull on the line.
The length of the line can be determined using the height of the kite and the orientation while flying sideways (because roll is controlled aerodynamically and by the direction of the line force).
The angle of attack of the kite can be controlled by varying the neutral position of the elevator. In strong winds it might be necessary to reduce the angle of attack. The strength of the wind can be deduced from the speed of the kite which in turn can be measured as the derivative of the height while flying in vertical direction.
Furthermore when the entire line is rolled out or the kite decides to descend, the angle of attack is reduced sufficiently to allow a descent in glide mode towards the ground station.
Pulling the line in kite mode and descending in glide mode is done periodically without landing in between the phases and with the propeller turned off all along.
Just before the weather ceases to support flying or landing the kite safely or before the battery runs out, the kite descends to the ground station, hovering the last meters.
The kite should be designed such that sensors, processor, elevator and rudder never fail, e.g. through redundancy. That way a slow and thus relatively safe landing as a glider is always possible.
P Power generated
F Force on the line
v_l speed of the line
v_k speed of the kite
v_w wind speed
c_l coefficient of lift (ca. 1.2)
ld lift to drag ratio
d density of air
A wing area
The wind speed is split into the speed of the line and the relative wind speed as seen by the kite. There is an optimum of this split for maximum power generation. The kite flies ld times faster than it "decends" relative to the wind.
v_k = ld * (v_w - v_l)
The lift force is linear in the wing area and quadratic in the kite speed
F = v_k^2 * A * d * c_l * 0.5
Power is force times speed
P = F * v_l
(c_l can be decreased by using elevator)
(v_l or F can be bound by the ground station)