In model reference adaptive scheme (MRAS) based speed estimation, the extended Luenberger observer is often used for deterministic state reconstruction. Often, It makes use of a closed loop observer gain multiplied with the stator current error for the purpose of correction. Can anybody provide me more insight as to how can I incorporate a sliding mode or switching surface into the Extended Luenberger Observer structure. Thank you.
Hint: you merely modify the observer gain to make it adaptive by using the sliding surface "residual=0". I suggest you to read this http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=371368&isnumber=8509&tag=1 -- it's extremenly simple.
The sliding mode observer is presented as an efficient estimator because of the following advantages: no complex calculation, easy implementation, simplicity, robustness, less restrictive design and so on [1].Error current which is the difference between the actual value and the estimated current, is used to define a sliding surface hyper plane. It is forced to zero by the switching action of the regulator. The control law is designed such that the states move towards the surface in a finite time. Then the sliding mode observer obliges states to remain within the structure of control and drag it to the desired position [1, 2]. However, the standard sliding mode control suffers from the chattering phenomenon. To remedy this problem, many methods have been proposed to reduce this phenomenon as the saturation function, low pass filter, etc...
[1] Lascu C, BoldeaI, Blaabjerg F. Very-low-speed variable-structure control of sensorless induction
machine drives without signal injection. IEEE Transactions on Industry Applications 2005;41:591–8.
[2] Hajian M, Arab Markadeh GR, Soltani J, Hoseinnia S. Energy optimized sliding mode control of
sensorless induction motor drives. Energy Conversion and Management 2009;50:2296–306.
- Badges
- Science topic
- Similar topics
- Engineering
- Control Theory