CADFEM-Symposium Energieeffizienz – Elektrische Antriebe und Wandler Evaluation of Iron Losses and Efficiency Improvements in Electric Machines EAA Gurakuq Dajaku Dieter Gerling FEAAM GmbH D-85577 Neubiberg, Germany Universität der Bundeswehr München D-85577 Neubiberg, Germany Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Overview Various Estimation Methods for Iron Losses – Fundamental Steimnetz Model – Extended Analytical Method – Iron Loss Model Considering Harmonics – Direct Flux-Density Method Consider the Major and Minor Hysteresis Loops – Iron Loss Model with Variable Iron Loss Coefficients Efficiency Improvement using Novel Windings – Concentrated Windings – New 12-teeth/10-poles Winding – New 24-teeth/10-poles Winding EAA Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Evaluation of Iron Losses Power Losses Generally, the losses of electrical machines are decomposed into: Copper losses; >> DC-Ohmic losses, skin and approximyty effect losses Mechanical losses; >> bearing friction, windage on the rotor, ventilation fan and cooling pump Iron losses; >> hysteresis and eddy current losses Eddy current losses; >> in rotor permanent magnets (PM) With the increasing demand on high efficiency and high power density in high-speed applications such as electric vehicles (EVs), it is important to estimate the iron loss of electrical machines accurately EAA Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Evaluation of Iron Losses Iron Losses Iron losses in a magnetic material occur when the material is subjected to a time varying magnetic flux density Hysteresis loss - is due to the energy expended in the redirection of the magnetic domains of the material during every flux direction reversal Eddy current - is the circulating current induced in the material by changing magnetic flux as a consequence of the electromagnetic induction EAA Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Evaluation of Iron Losses Analytical calculation of iron losses Steinmetz model for determination of iron losses: piron = phys + pedd = khys ⋅ f ⋅ Bˆ β + kedd ⋅ f 2 ⋅ Bˆ 2 The above iron-loss expression is only valid for sinusoidal flux density In most electric machines, the variation in flux density in the stator core is far from sinusoidal Minor hysteresis loops are due to the flux fluctuations (air-gap harmonics) a) Minor hysteresis loops; EAA b) Definition of flux variations. Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Evaluation of Iron Losses Analytical calculation of iron losses Improved analytical model for iron loss calculation: piron = phys + pedd = kch ⋅ khys ⋅ f ⋅ Bˆ β + kce ⋅ kedd ⋅ f 2 ⋅ Bˆ 2 c N Correction factors: kch = 1 + ∑ ∆Bi Bˆ i =1 2 N and kce = B1 ⋅ ∑ i ⋅ Bi Bˆ i =1 B1 2 Extended model: piron = phys + pedd + pexc T 1 k = kch ⋅ khys ⋅ Bˆ β ⋅ ω + ∫ edd2 T 0 2π EAA 2 T 1 kexc dB (t ) ⋅ ⋅ dt + T ∫0 2π 43 dt ( ) 3 dB(t ) 2 ⋅ ⋅ dt dt Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Evaluation of Iron Losses Harmonic iron losses using Fourier transformation This method is based on the Fourier transformation of the waveforms of the flux density at each finite element The total iron losses is assumed to be the sum of the losses that are caused by the radial and peripheral components of the flux density M Pedd = ∑∑ kedd ⋅ mi ⋅ ( nf ) ⋅ { Br2,n + Bθ2,n } i =1 2 n M Phys = ∑∑ khys ⋅ mi ⋅ ( nf ) ⋅ { Brβ,n + Bθβ,n } i =1 n This method is very useful to understand the each harmonic loss component However, the hysteresis loss cannot be decomposed into harmonic components because of its nonlinear characteristics. Also, the hysteresis losses are over-estimated using this method EAA Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Evaluation of Iron Losses Iron loss calculation using the waveform of the flux density Eddy current losses: 2 T k 1 dB (t ) = edd2 ⋅ ∫ dt 2π T 0 dt pedd ⇒ Pedd N Step M N Step 2 2 kedd 2 = 2 ⋅ N Step ⋅ f ⋅ ∑ mi ⋅ ∑ ( Br ,k +1 − Br ,k ) + ∑ ( Bθ ,k +1 − Bθ ,k ) 2π i =1 k =1 k =1 Hysteresis losses N pr β N pθ β = khys ⋅ f ⋅ ∑ mi ⋅ ∑ Bˆr ,ij + ∑ Bˆθ ,ij i =1 j =1 j =1 M Phys i i Using this method, the major and minor hysteresis losses are considered The hysteresis loss are calculated as the sum of the area of all hysteresis loops EAA (4) Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Evaluation of Iron Losses Iron loss calculation using the waveform of the flux density The current ripple in the current curve leads to minor hysteresis loops. The total hysteresis loss is the sum of the main and minor hysteresis loops. EAA Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Evaluation of Iron Losses Comparison of results B_total B_1 B_3 B_5 1,4 1,0 Eddy current Loss 0,2 3 -0,2 -0,6 -1,0 2,5 1,5311 2 0,79616 1,5 1 1,5847 1,5847 0,5 -1,4 0 4 8 12 16 20 0 Cal. FFT time [ms] EAA Hys. Loss 3,5 Iron Loss [W] B [T] 0,6 Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Cal. Direct Evaluation of Iron Losses Variable Iron loss Coefficients Conventional iron loss model with constant loss coefficients are not suitable for the machine using field-weakening control where the frequency and the flux density of the flux waveform are varied dramatically according to its driving condition Model with variable coefficiens piron = khys ( f , B) ⋅ f ⋅ Bˆ 2 + kedd ( f , B ) ⋅ f 2 ⋅ Bˆ 2 Iron loss coefficients depends on the flux density and frequency EAA Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Evaluation of Iron Losses Variable Iron loss Coefficients Division of the iron loss model by f and B² leads to the following linear equation, piron = khys ( f , B) + kedd ( f , B) ⋅ f f ⋅ Bˆ 2 Linear model with slope identifying kedd and the intersection with y-axis corresponding to khys Measured specific losses piron f ⋅ Bˆ 2 y = ahys + bedd ⋅ f EAA Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Evaluation of Iron Losses Variable Iron loss Coefficients Iron loss coefficients are derived using curve fitting method Eddy-current loss coefficient Hysteresis loss coefficient kedd ( B ) = kedd ,0 + kedd ,1 ⋅ B + kedd ,2 ⋅ B 2 + kedd ,3 ⋅ B 3 khys ( B) = khys ,0 + khys ,1 ⋅ B + khys ,2 ⋅ B 2 + khys ,3 ⋅ B 3 -5 11.5 x 10 0.04 11 Khys [W/kg/(f*T²)] Khys [W/kg/(Hz*T²)] Kedd [W/kg/(Hz*T²)] 0.035 10.5 10 9.5 0.03 0.025 9 0.02 8.5 8 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0.015 0 Induction [T] EAA 0.2 0.4 0.6 0.8 1 1.2 Induction [T] Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling 1.4 1.6 1.8 2 Efficiency Optimisation of Electric Machines Winding Topologies Winding MMF Winding MMF Distributed Winding: Concentrated Winding: - Sinusoidal MMF distribution - High nr. of slots per pole - Large end-winding length; High parasitic end-effects - High copper losses - High production costs - Compact design - Shorter and less complex end-windings - Low copper losses - Low production costs EAA - More space harmonics Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Efficiency Optimisation of Electric Machines PM machine with 12-teeth/10-poles winding topology MMF Harmonics [p.u.] MMF [p.u.] 12-Teeth/10-Poles Winding topology EAA - 5th or 7th => working harmonic - Cause undesirable effect such as losses, noise and vibration Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Efficiency Optimisation of Electric Machines New 12-teeth/10-poles winding with different turns per coil side Comparison of Rotor (magnet) Losses 35 New 12-teeth/10-poles winding with reduced subharmonics Eddy Current Loss [ W ] 30 no-load conv. winding, under load new winding, under load 25 20 15 10 5 0 500 1000 1500 2000 2500 speed [ rpm] New 12-teeth/10-poles winding Comparison Conventional 12-teeth/10-poles winding EAA Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling 3000 3500 Efficiency Optimisation of Electric Machines New 24-teeth/10-poles winding New 24-teeth/10-poles winding with reduced sub- and high harmonics PM machine with conventional distributed winding, q=2 PM machine with new 24-teeth/10-poles winding Comparison Winding distribution of one-phase (phase-A) EAA - No skewing is required for the new machine Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Efficiency Optimisation of Electric Machines New PM machine Driving Cycle Simulations Conventional PM machine Efficiency difference: Eff.new – Eff.conv New 24-teeth/10-poles winding EAA Losses reduction up to 20% with the new machine Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Conclusion Different methods for iron losses calculation are presented – Conventional analytical methods are valid only for sinusoidal flux density variations – Harmonic iron losses method is very useful to understand the each harmonic loss component. It overestimate the hysteresis losses – Iron loss calculation using the waveform of the flux density consider the major and minor hysteresis loops – Iron loss model with variable loss coefficients increase the calculation accuracy Different efficiency optimisation techniques – Tooth concentrated winding are characterised with high MMF harmonic contents – A simple technique using coils with different turns per coil-side reduce the sub-harmonics – 24-teeth/10-poles winding is another solution for simultaneously reducing the sub- and high harmonics – Loss reduction up to 20% per driving cycle are possible with the new winding EAA Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling Thank You for Attention Gurakuq Dajaku FEAAM GmbH, Neubiberg, Germany gurakuq.dajaku@unibw.de EAA Elektrische Antriebstechnik und Aktorik - Electrical Drives and Actuators Univ.-Prof. Dr.-Ing. Dieter Gerling