Optimizing Electromechanical and Control Circuit Parameters of Brushless Motors
Optimizing the performance of modern brushless DC motors typically requires evaluating both the electromechanical and control circuitry design factors – examining either in isolation yields only partial improvements.
Consider the following example in which we wish to improve the design of an Interior Permanent Magnet (IPM) motor by:
1. Increasing the response speed
2. Minimizing the drive’s overshoot
This dual objective design task is much more complex than it appears: we require (1) a sophisticated co-simulation of transient electromagnetic fields exchanging data with the transient solution of the drive control circuits and (2) a compatible optimizer which can search for improvements within manufacturing constraints.
Previous examples posted on our website have highlighted the capability of MagNet and PSIM (from PowerSim Inc) to address part (1) of the above, notably IPM Motor with PWM vector control in PSIM. We wish to illustrate how OptiNet can be used to further refine the results obtained by co-simulations.
Optimization of model varying the motor and control circuit parameters
)
|
| Click image to see full size version |
Design Parameters
The electromechanical parameters are the position and dimension of the magnet (indicated by L, d and h in image on right, click to enlarge). They are discrete in variation with a step of 0.1 mm.
There are two advantages in considering discrete variables: firstly we can take into account the manufacturing tolerances that are always present in industrial designs and secondly the true optimal solution is found. In fact the optimizer can search for the optimal solution in a continuous space and then the variables can be conveniently rounded, but the corresponding solution may be not as good as the continuous one.
)
|
| Click image to see full size version |
The parameters defined above, including their permitted range of values, are uploaded in OptiNet which utilizes the co-simulation link between MagNet and PSIM to search for improved IPM motor designs as defined by our objectives.
)
|
| Click image to see full size version |
Results of the Optimization
The chart on the right compares the optimized IPM design response to both the initial design and the ideal response curves.
With respect to the objectives:
- The response speed has increased from 38% to 7%
- The overshoot has decreased from 17% to 11%
)
|
| Click image to see full size version |
Comparing Parameter Values Before and After Optimization
| Configuration | Initial | Optimized |
| L [mm] | 22.7 | 21.1 |
| d [mm] | 1.2 | 1.7 |
| h [mm] | 8.4 | 8.5 |
| P | 0.1 | 0.3 |
| I | 0.01 | 0.008 |
| Beta [deg] | 50 | 48 |
Optimizing the electromechanical motor parameters alone would yield an improved response time thanks to higher torque (from 38% to 15%) but overshoot would not be improved (from 17% to 20%).
Conversely, optimizing the control parameters only would produce a smaller overshoot thanks to improved control (from 17% to 8%) but much slower response due to lack of torque (from 38% to 30%).
Only the optimization of the motor and control circuit co-simulation, by means of MagNet, OptiNet and PSIM provides faster response speed and smaller overshoot.
The results documented in this gallery example were taken from the paper:
U. Piovan1, B. Forghani2, L. De Fina1, D. N. Dyck2, P. J. Weicker2, M. Negro1
"Optimization of a Brushless DC Drive by Change of Power Electronic Control Parameters and Motor Geometry"
presented at Inductica 2006 in Berlin, German.
1 Weidmann Transformerboard Systems AG, Rapperswil, Switzerland
2 Infolytica Corporation, Montreal, Québec, Canada