Mar 21 2019
According to a team of scientists from Mitsubishi Electric Research Laboratories, little motors power everything from small luxuries, like desk fans, to larger protection systems, such as oven exhaust systems; however, they could be more accurate.
In IEEE/CAA Journal of Automatica Sinica, an international partnership from Japan and Massachusetts revealed an enhanced algorithm to trace motor performance and speed estimation.
Induction motors are power-driven by an alternating current supplied via equipment called a drive. On powering a rotor suspended through a stacked cylinder of metallic windings, a magnetic field is produced, driving the rotor to rotate. The power and variability of the drive decide the speed of the rotor.
It is extremely hard to determine the speed of the rotor without the sensors to detect the speed of the drive. Although there are a few techniques to estimate the speed, Wang stated that they are insufficient.
Rotor speed estimation for induction motors is a key problem in speed-sensorless motor drives.
Yebin Wang, Senior Principal Research Scientist, Mitsubishi Electric Research Laboratories.
Yebin Wang is also the first author of the paper.
“Existing approaches have limitations such as unnecessarily assuming rotor speed as a constant parameter” Wang noted. He also noticed that some methods swap between estimation bandwidth and measurement robustness; however, they deliver uncomplicated designs that could be expanded upon.
The rotor speed could be considered as a state variable, instead of a constant variable. State variables are supposed to be true for the entire motor system unless some external force influences them and they vary. Wang and his team took the state variables and altered their coordinates to enable the system to be stable, in relation to itself. By enabling system variables to remain synchronous but movable as a whole, the researchers were able to carry out mathematical experiments to influence the system and determine particular speed variations and changes.
Experiments demonstrate the potential effectiveness and advantages of the proposed algorithm: fast speed estimation transient and ease of tuning. This paper also reveals a number of issues.
Yebin Wang, Senior Principal Research Scientist, Mitsubishi Electric Research Laboratories.
One main problem is that in order to better estimate the speed, it is necessary to know all of the variables of the system. In real-world scenarios, it is not really possible to accurately identify every variable.
Wang and the researchers aim to further develop more systematic solutions to tackle the system stability and to generalize their projected algorithm to elucidate improbabilities within the system.