For the estimation of flux instead of using pure integrator, i am using programmable low pass filters (PLPF) with two filters in cascaded
y1(k) = a*y1(k-1) + K*u(k-1);
y2(k) = a*y2(k-1) + K*y1(k-1);
where
K = (sqrt(G)*Ts)/tau;
Ts = Sampling time;
a = 1-(Ts/tau);
G = (1/we)*sqrt([1+(tau*we)^2]^n)
tau = (1/we)* tan(pi/(2*n));
so G = (2/we);
tau = (1/we);
we = Syn_speed in rad/sec (estimated)
n = Number of filters in cascaded.
By using above equations it has to reduce the offset and drift problems, but motor is not rotating smoothly where the file attached shows the wave forms of estimated fluxes (sid(yellow),siq(green)) ,voltage(pink) and current in (light-blue) and the applied DC voltage 80v, as the voltage increases the motor stops drawing high currents
Kindly suggest me .
Thanks in advance
there is a simple filter structure may can help you, see the attached file
The sensorless control has advantages in terms of costs and reliability. It infers the rotation speed of the rotor and the flux amplitude of the measurement of voltages and stator currents. It requires that an estimator for open loop or closed loop observer.
The main drawback of the open loop is that the maximum torque can be issued for too low a speed, typically 0.8 Hz, while for the closed loop, it can be issued to stop
I am not entirely sure of this, but I think the problem is a more fundamental one than what is suggested by Mr. Ammar and Mr. Lafifi.
You may benefit if you use a higher carrier frequency for the inverter, for the same motor inertia. Your motor is small, neither the inertia nor the inductances filter adequately, and your PWM needs a redesign.
Does that help ?
Looking at this I have a bunch of questions:
- what type of motor is this? Brush-type? Brushless?
- if brushless, is a shaft position sensor involved in the voltage output calculation?
- what is the load? The shaft is apparently turning - is it experiencing a stick-slip load?
- what do you have for a motion sensor? It would be interesting to look at the motor motion on the same graph.
- Why is the voltage so "choppy"? Is it because of:
- latency in the filter?
- latency with jitter?
- stick-slip motion of the motor?
- Why is the voltage mirroring (noisily) the green trace? Or is it mirroring the magnitude of the combined green plus yellow traces with delay?
I think the questions posed by Eric Jackson deserves a response about the globaly question and see precisely sizing the filter loan.
Hello Eric Jackson,
1. It is three phase induction motor with ratings 220/380v 2850 rpm ,2 pole
2. The technique used to control the motor is direct torque control of induction and there is no shaft position involved .
3. It is a sensor less control technique. There is no motor motion on the graph the first two sinusoidal wave forms are the flux estimated values which are obtained by (v-i*r) through the two cascaded filters. Generally the cascaded filter technique is used to reduce the offset and DC drift produced by the pure integration of (v-i*r) because of presence of offset in the measured currents.
4.The third wave form is the voltage at he motor terminals and why it is like that i am unable to understand ?
5. The other waveform is the current wave form which is phase current into the motor
- Badges
- Science topic
- Similar topics
- Engineering
- Control Theory
- Control Systems
- Control Systems Engineering