ECE 8830 - Electric Drives. Topic 17 : Wound-Field Synchronous Machine Drives Spring 2004. Introduction .
Topic 17: Wound-Field Synchronous
For high power (multi-MW) applications, the high efficiency of synchronous motors makes them more appealing than induction motors. Indeed, most of today’s electrical power generators are 3 synchronous generators.
Wound-field synchronous motors require dc current excitation in the rotor winding. This excitation is traditionally done through the use of slip rings and brushes. However, these have several disadvantages such as requiring maintenance, arcing (which means they cannot be used in hazardous environments), etc. An alternative approach is to use brushless excitation which is illustrated on the next slide.
A wound-rotor induction motor (WRIM) is mounted on the same shaft as the wound-field synchronous motor. This is acting as a rotating transformer with the rotor as the primary and the stator as the secondary. The stator of the WRIM is fed by a 60Hz supply and the rotor of the WRIM rotates at a speed set by the supply frequency. The slip voltage in the rotor winding of the WRIM is rectified to provide the current feed to the rotor windings of the synchronous motor.
Thyristor current-fed, load commutated inverters (LCI’s) are very popular for high power (multi-MW) wound-field synchronous motor drives.
We will briefly review current-fed thyristor inverters and then discuss load commutated inverters in some detail. We will then see how to apply them to wound-field synchronous motor drives.
Let us first briefly review the operation of the current-fed thyristor inverter.
Initially, ignore commutation considerations.
Induction motor load is modeled by back emf generator and leakage inductance in each phase of the winding.
The constant dc current Id is switched through the thyristors to create a 3 6-step symmetrical line current waves as shown on the next slide.
The load or line current may be expressed by a Fourier series as:
where the peak value of the fundamental component is given . Each thyristor conducts for radians. At any instant one upper thyristor and one lower thyristor conduct.
The dc link is considered harmonic-free and the commutation effect between thyristors is ignored.
At steady state the voltage output from the rectifier block = input voltage of inverter.
For a variable speed drive the inverter can be operated at variable frequency and variable dc current Id.
If thyristor firing angle > 0, inverter behavior.
If thyristor firing angle =0, rectifier behavior.
Max. power transfer occurs when =.
Two inverter operation modes are established depending on the thyristor firing angle:
1) Load-commutated inverter
Applies when /2<<.
2) Force-commutated inverter
Applies when <<3/2.
Consider =3/4. In this case vca < 0 => thyristor Q5 is turned off by the load. This requires load to operate at leading power factor => motoring mode of a synchronous machine operating in over-excitation.
Let us initially consider a single-phase inverter before discussing the 3 case. A single-phase, current-fed, parallel resonant inverter with load commutation is shown below:
A phase-controlled rectifier provides the dc input and a large capacitor C provides the load commutation of the thyristors. Assuming perfect filtering of harmonics by the capacitor and the dc link inductor, the inverter load voltage and current waves are shown below:
The thyristor pairs Q1Q2 and Q3Q4 are switched alternately for angle to produce the square wave output. The fundamental of the current wave leads the sinusoidal voltage wave by . Thus, when Q1Q2 turn on, the Q3Q4 pair has a negative voltage for duration which provides the load commutation. Since =tq, the time tq must be sufficiently long for the thyristors to turn off.
The equations for the inverter circuit are:
where Rd is the resistance of the inductor Ld.
These equations can be expressed in state-variable form and solved to model the steady state and dynamic performance of the inverter.
We will now consider an approximate steady state analysis assuming that Ld is of infinite size and is lossless. We will also assume that the load is highly inductive, i.e. L>>R.
Consider the series R-L load to be resolved into parallel R1 and L1 components in which real current IP flows through R1 and reactive current IQ flows through L1. The load impedance ZL can be written as:
If the load is highly inductive (as we had assumed) R1>>L1 and
and L L1.
The fundamental component of the current is given by:
The real and reactive components of the load current are given by:
where . Through some algebraic manipulation we get:
From the above equations we can calculate the load voltage, currents, and commutation angle .
Single-phase synchronous motor; Vd=200V, f=60Hz, R=0.2, L=1.2mH, Id=240A, C=150F. Find .
There are basically two control variables for the load commutated inverter - the dc link current and the frequency. For a variable load, a variable capacitance can be used to provide desired margin of commutation angle . However, a better way to operate is to use a PLL to control the inverter frequency to just above the load resonance frequency.
The single-phase inverter concepts can be extended to 3 LCI’s. The figure below shows a three-phase LCI with lagging power factor load. Here load commutation is achieved by using a leading VAR load connected at the load terminal.
In the case of a variable load, a fixed capacitor bank is connected at the terminals and the inverter frequency adjusted so that the effective inverter load has a leading PF so that commutation occurs at a fixed angle .
As mentioned earlier, thyristor current-fed, load commutated inverters (LCI’s) are very popular for high power (multi-MW) wound-field synchronous motor drives where it is easy to maintain the required leading PF angle by adjusting the field excitation.
The fundamental frequency phasor diagram for a salient pole synchronous machine under motoring condition is shown below:
Note: The winding resistance and the
commutation overlap effect have been neglected.
A flux linkage has been included in the phasor diagram where f= field flux linkage, a=armature reaction flux linkage and S=resultant stator flux. We can write the deand qe components of a as follows:
For a salient pole machine, LdLq the phasors a and Is are not in phase.
The motor phase voltage and current waves are shown below including the commutation overlap effect:
The load commutated inverter with an over-excited synchronous machine load depends on sufficient back emf which is not available at low speeds. The critical speed required for load commutation to work is about 5% of base speed. A forced commutation approach is required below these speeds and to start the motor. (see Bose text pp. 284-285 for details).
Having seen how a current-fed, thyristor inverter can be load commutated with a wound-field synchronous motor by operating the machine at a leading power factor, we can now consider how to design a self-controlled drive system for a wound-field synchronous motor based on a load-commutated inverter drive. As mentioned earlier, this type of drive is popular for high power (multi-MW) drives for compressors, pumps, ship propulsion, etc.
A block diagram of a self-controlled load-commutated, current-fed inverter drive for a wound-field synchronous motor is shown below:
The phasor diagram for the LCI in motoring mode driving a synchronous motor is shown below:
Note: The saliency and stator resistance have been neglected.
The field flux fis established by the field current If and depends on the rotor position. The armature flux a =IsLs is determined by the stator current and stator winding inductance. The delay angle command d* sets the position of a relative to f since a leads f by ’ given by:
where = torque angle.
Thyristors require a minimum turn-off time toff for successful commutation. This corresponds to a turn-off angle =toff. For reliable operation of a LCI drive and minimum reactive current loading to the synchronous motor, turning off the thyristors at a fixed time every cycle is a good approach. A complete speed control system for a LCI synchronous motor drive incorporating constant turn-off angle control is shown on the next slide.
This drive operates in the constant torque region in motoring mode with stator flux s maintained constant (open loop). There are four control elements:
Speed and dc link current control:
r compared to r* and error goes through P-I controller and absolute value circuit -> Id*. Id and Id* compared and controls thyristor firing angle in rectifier to control dc link current.
The generated motor torque Id (see Bose text pg. 499 for derivation).
Field flux/field current control:
The command field flux f* is given by:
where s*= constant, a*=LsIs=KaId* and
*= *+k*. To obtain we need * which can either be measured or calculated using the expression:
The command flux current If* is then generated from the command flux f* by through a function generator that corrects for saturation effects. A phase-controlled rectifier can then be used to control the flux current as shown in the system block diagram.
Generation of f*, d*, * and * command signals:
We have discussed how all of the command signals can be obtained with the exception of the * angle. This is obtained from the equation:
Delay Angle Control:
For a six-step inverter we need six discrete firing pulses at /3 intervals apart within a cycle. A block diagram showing how this can be achieved is shown on the next slide.
The corresponding alignment of reference signal S1 and the waveforms for phase a in motoring mode are shown below. These diagrams can be used to determine the inverter firing angles.
The absolute position sensor can be eliminated and the machine terminal voltage signals can be used instead to estimate the rotor position for inverter firing angle determination. Details are presented in the Bose textbook pp. 504-507.
High power, wound-field synchronous motors can be operated at unity power factor when excited by phase-controlled, line-commutated, thyristor cycloconverters. Drive control for such drives can be both scalar and vector control, similar to that of the voltage-fed inverter drive.
The next slide shows a simple scalar control method for a cycloconverter drive for a wound-field synchronous motor.
There are three control variables in this control system:
The torque generated by the motor is proportional to the in-phase stator current. The command stator current Is* is generated from the error in the speed control loop.
The angle ’ and the field current If can be determined from Is as
shown in the figure. Thus,
Is* is used to generate ’*
and If* using function
The position sensor and encoder generate the cose and sine signals and the speed signal, r. The 2-phase unit signals are converted to 3-phase unit signals using the following transformations:
Each of the 3 unit signals is then multiplied by Is* and phase shifted by angle ’* to produce the three phase current command signals as follows:
The performance of the cycloconverter drive can be enhanced if vector control is used rather than scalar control. In the constant torque region, the field current must be increased if we want to increase the developed torque at the constant rated stator flux. However, the field current response is slow and this leads to sluggish motor response. In vector control we inject a transient magnetizing current in the direction of the stator flux to obtain a much faster response than with scalar control. This current is set to zero in steady state to maintain unity PF.
A vector control implementation is shown:
A phasor diagram for the vector control approach is shown below:
Notable points from phasor diagram:
- de-qe in reference frame of rotor;
- ds-qs in reference frame of stator;
- de’-qe’ with qe’ aligned with Vs and de’ aligned
From the phasor diagram, at steady state, we can write:
This equation gives the control equation for If*. Under transient conditions, the command injected transient magnetizing current, IM*, is given by:
Under steady state conditions, IM=0 and the above steady state equation is re-established.
Control features of the vector control of a wound-field synchronous motor drive:
The complete vector control feedback signal processing is shown on the next slide.