| 摘要 |
<P>PROBLEM TO BE SOLVED: To solve the problem that when model predictive control is performed, there is a risk of an increase in switching frequency of a switching state of an inverter since an operation state minimizing a deviation of a predicted current from a command current is selected at each time. <P>SOLUTION: After a voltage vector V(n) is output in a step S10, predicted currents ide(n+2) and idq(n+2) are calculated in a step S14 assuming that the number of phases to be switched of a switching state from a current voltage vector V(n) is equal to or less than "1" for a next voltage vector V(n+1). When it is determined in a step S20 that there is a predicted current having an error edq(n+2), equal to or less than a threshold eth, from a command current idr or iqr between the predicted currents ide(n+2) and iqe(n+2), a predicted current having the smallest absolute value of a change value between the predicted currents ide and iqe is selected in a step S24. <P>COPYRIGHT: (C)2013,JPO&INPIT |
