DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

NEW HYBRID MODEL FOR EFFICIENCY OPTIMIZATION OF INDUCTION MOTOR DRIVES Branko D. Blanuša, Petar R. Matić, Branko L.Dokić Presented by: Branko Blanuša University of Banja Luka, Faculty of Electrical Engineering E-mail: [email protected]

Niš, Serbia, November 11th- 14th, 2010

-1-

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

Main goal: Define optimal control strategy for a given operating conditions so the drive operates with minimal energy consumption.

Niš, Serbia, November 11th- 14th, 2010

-2-

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

Content 1. Introduction 2. Power loss modelling 3. Hybrid model for efficiency optimization 4. Simulation results 5. Conclusion

Niš, Serbia, November 11th- 14th, 2010

-3-

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

Results of applied algorithms for efficiency optimization highly depends from the size of drive (Fig.1) and operating conditions, especially load torque and speed (Fig. 2)

Fig. 1 Rated motor efficiences for ABB motors (catalog data) and typical converter efficiency. Niš, Serbia, November 11th- 14th, 2010

-4-

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

Fig 2. Measured standard motor efficiences with both rated flux and efficiency optimized control at rated mechanical speed (2.2 kW rated power). Niš, Serbia, November 11th- 14th, 2010

-5-

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

Important conclusions 1. It is possible to minimize power losses by variation of magnetizing flux in the machine, so the balance between cooper and iron losses are obtained. 2. Best results in efficiency optimization of induction motor drives can be achieved for a light loads.

Niš, Serbia, November 11th- 14th, 2010

-6-

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

According to the literture, there are three strategies dealing with the problem of efficiency optimization of the induction motor drive 1. Simple State Control - SSC , 2. Loss Model Control - LMC and 3. Search Control- SC.

Niš, Serbia, November 11th- 14th, 2010

-7-

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Optimal state reference

Projekt „ISSNBS“

fe Control

Vs

Converter I.M. fr

Control state variable (measured or estimated)

Fig. 3. Control diagram for the simple state efficiency optimization strategy. The first strategy (SSC) is based on the control of one of the variables in the drive. This strategy is simple, but gives good results only for a narrow set of operation conditions. Also, it is sensitive to parameter changes in the drive due to temperature changes and magnetic circuit saturation. Niš, Serbia, November 11th- 14th, 2010

-8-

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

fe fr,ref

Efficiency optimization control

Vs

Converter I.M.

Drive loss model

fr

Fig. 4. Block diagram for the model based control strategy. For LMC methods, a power loss model is used for optimal drive control. These algorithms are fast because the optimal control is calculated directly from the loss model. But, power loss modeling and calculation of the optimal operating conditions can be very complex. This strategy is also sensitive to parameter variations in the drive. Niš, Serbia, November 11th- 14th, 2010

-9-

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

fe fr,ref

Control

Vs

Converter I.M.

ψ

r,min

Pγ

ψr

Power loss calculation Pγ = Pin- Pout

fr

Fig. 5. Block diagram of search control strategy. In the search strategy, the on-line procedure for efficiency optimization is carried out. The optimization variable, stator or rotor flux, increases or decreases step by step until the measured input power is at a minimum. This strategy has an important advantage over others: it is insensitive to parameter changes. Niš, Serbia, November 11th- 14th, 2010

- 10 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

Hybrid methods Hybrid method combines good characteristics of two optimization strategies Search Control and Loss model control. During transient process LMC is used, so fast flux changes and good dynamic performances are kept. SC is used for efficiency optimization in a steady state of drive. Hybrid method obtains fast convergence to optimal flux and negligible sensitivity to parameter changes.

Niš, Serbia, November 11th- 14th, 2010

- 11 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

Power loss modelling The overall power loss in electrical drive consists of converter loss and motor losses, while motor power loss can be divided in copper and iron loss: Ptot = Pmot + Pinv

Pmot = PCu + PFe Overall flux-dependent losses are usually given by: Pinv = Rinv ⋅ is2 = Rinv ⋅ (id2 + iq2 )

Niš, Serbia, November 11th- 14th, 2010

- 12 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

Loss model of drive is developed in d-q rotational system in such way that that rotor side variables do not depend on leakage inductances while the effect of leakage inductances is incorporated into other variables. vs = is Rs + ( p + jωe )L's is + ( p + jωe )L'm im i s = im + i f − ir = im +

' L'm ( p + jωe )im − Lm [ p + j (ω' e − ωr )] im , Rm Rr

Fig.6. Space vector model of induction motor drive.

Niš, Serbia, November 11th- 14th, 2010

- 13 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

vsd = Rs isd − ωe L's isq vsq = Rs isq + ωe L's isd − ω sl L'mimd + ωr L'mimd isd = imd L'm L'm isq = i f + irq = ωe imd − ω sl ' imd Rm Rr

Fig. 7. Steady state model of IM in a rotor flux oriented reference frame, d-axis equivalent circuit, q-axis equivalent circuit. Niš, Serbia, November 11th- 14th, 2010

- 14 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

Total power losses Ra = Rinv + Rs + (ωr L'm ) / (Rm + Rr' ) 2

2 2 Ptot = Ra isd + Rbisq ,

Rm Rr' Rb = Rinv + Rs + Rm + Rr' 3 Lm2 3 ' isd isq = PLmisd isq Tem = P 2 Lr 2 = kekvisd isq , Ptot =

2 Ra isd

2 Tem + Rb 2 2 kekvisd

kekv = 3 / 2 PL'm . Tem2 ∂Ptot = 2 Raisd − 2 Rb 2 3 = 0 kekvisd ∂isd * sdLMC

i

Niš, Serbia, November 11th- 14th, 2010

=4

Rb Tem2 2 Ra kekv - 15 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

Search controller ωr * i sd * isq

SCALING FACTOR CALCULATION

Ig

Pg Δ Pin(k)

Pin(k)

-1

z

Δ Pin(p.u) FUZZY INFERENCE

Pin (k-1)

* (p.u) Δ ids

* Δ ids

-1

* (p.u) L Δ ids

z

Fig.8. SC efficiency optimization controller

Niš, Serbia, November 11th- 14th, 2010

- 16 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

Fig. 10. Overall proposed block diagram of efficiency optimization controller in IMD. Niš, Serbia, November 11th- 14th, 2010

- 17 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

Simulation results 1 speed reference load torque 0.8

0.6

0.4

0.2

0

-0.2

-0.4

0

2

4

6

8 time (s)

10

12

14

16

Fig. 11. Load torque and speed reference. Niš, Serbia, November 11th- 14th, 2010

- 18 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

2

magnetization curremt p.u. mechanical speed p.u. torque reference p.u

1.5

1

0.5

0

-0.5

-1

0

2

4

6

8 time (s)

10

12

14

16

Fig.12. Magnetization current, mechanical speed and electromagnetic torque. Niš, Serbia, November 11th- 14th, 2010

- 19 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

1.5

magnetization curremt p.u. mechanical speed p.u. torque reference p.u

1

0.5

0

-0.5

-1 4.4

4.6

4.8

5

5.2

5.4 5.6 time (s)

5.8

6

6.2

6.4

Fig.13. Magnetization current, mechanical speed and electromagnetic torque. Niš, Serbia, November 11th- 14th, 2010

- 20 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

180 hybrid method nominal flux

160 140

Power loss (W)

120 100 80 60 40 20 0

0

2

4

6

8 time (s)

10

12

14

16

Fig.14. Graph of power loss for nominal flux and applied hybrid method. Niš, Serbia, November 11th- 14th, 2010

- 21 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

180 hybrid method LMC method

160 140

Power loss (W)

120 100 80 60 40 20 0

0

2

4

6

8 time (s)

10

12

14

16

Fig.15. Graph of power loss for LMC method and hybrid method. Niš, Serbia, November 11th- 14th, 2010

- 22 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

hybrid method LMC method

144.5

Power loss (W)

144

143.5

143

142.5

4.5

5

5.5

6 time (s)

6.5

7

Fig.16. Graph of power loss for LMC method and hybrid method. Niš, Serbia, November 11th- 14th, 2010

- 23 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

Conclusion

If load torque has a value close to nominal or higher, magnetizing flux is also nominal regardless of whether an algorithm for efficiency optimization is applied or not. For a light load hybrid method for efficiency optimization gives significiant power loss reduction (figs. 14,15 and 16). Also, it shows good dynamic performances (figs. 12 and 13) and negligible sensitivity to parameter changes.

Niš, Serbia, November 11th- 14th, 2010

- 24 -

DAAD Deutscher Akademischer Austausch Dienst German Academic Exchange Service

Projekt „ISSNBS“

Thank you

Niš, Serbia, November 11th- 14th, 2010

- 25 -