.1 METHOD

Develop a model that will control the error to achieve stability using DTC and fuzzy logic with duty ratio.

Figure 3.1 Simulink model for direct torque control of induction motor.

A Simulink model above was developed to study the performance of the conventional DTC and fuzzy controller for 4 poles induction motor to reduce the high ripple torque in the motor. After the field work experiment, the error of the torque, flux linkage and position of stator flux linkage were used in the simulation and the data generated are in table 3.1 below.

To determine the error in the torque of the induction motor that causes vibration which lead to backlash that result in the production of less standard products.

The errors in the magnetic torque of the motor were determined using the torque ripple test apparatus.

Because we want to know the actual error in the induction motor that causes the high ripple torque in the motor.

Figure3.2Torque ripple test apparatus

A motor with torque ripple of 0.9N-m was connected to the shaft of the motor and with a load torque sensor that can measure the vibration or ripple of the shaft and will equally give the vibrational result of the motor then a DC voltage was supplied to the moto rand observed a peak to peak torque equal to 0.9N-m. The formula for torque ripple calculation was used.

Tr = Torque ripple

Peak to peak value of the ripple = 0.9Nm, 0.15Nm

Average output of the ripple = 0.15Nm

In table 3.1 below, actual torque equal 0.15N-m, measured torque equal to 0.9N-m, error in torque is equal to 0.75N-m.

Tr = Peak – to -peak x 100

Average output

Tr = (0.9 – 0.15) x 100 = 5 ÷ 0.15 = 13.33%

0.15

To determine the stator flux linkage error in the induction motor that also causes vibration.

The errors in the stator flux linkage of the motor were determined.

To help us to know the actual flux linkage error that contributed to the high ripple torque in the motor.

Figure 3.3Stator motor

The motor was dismantled and the flux meter was used to determine the coils in the slots of the stator of the motor, when the flux meter probe that have indicator at the end where it will indicate the amount of flux linkage at any instant were placed on top of the coil in the slot, it will indicate the amount of flux linkages.

At the end of the whole slot, we got approximately 170wb while the standard value is 150wb, as stated in table 3.1 below.

To determine the position of the stator flux linkage space vector in the poles of the induction motor.

The positions of the stator flux linkage space vector were determined.

Because we want to know the position of the flux linkage in the different poles of the induction motor.

In figure 2 above, the flux meter was used to measure the flux linkages in the different poles of the electric motor, in order to know the position of the flux linkage space vector of the motor. With the measurement, we observed that the flux linkage is varies per poles in the table 3.1 below.

Table 3.1 Result obtained after the analysis

Actual value Measured value Error

Torque 0.15Nm 0.9 Nm 0.75 Nm

Flux linkage 150wb 170wb 10wb

Position of the flux linkage 0.5? 5? 4.95?

Figure 3.4 Simulink model for fuzzy logic with duty ratio of induction motor.

The Simulink model were simulated and the result are in the table 4.1, 4.2, and 4.3, below.

Direct torque control (DTC) and fuzzy logic with duty ratio model were designed.

Because we want to control the induction motor drives in order to reduce the high ripple torque of the motor.

In the principles of direct torque control of induction motor, the ripples in the motor can be reduced if the errors of the torque and the flux linkage and the angular region of the flux linkage are sub-divided into several smaller sub-section then the errors should be pick and compared in other to select voltage vector with less ripples, in doing so, a more accurate voltage vector is being selected in the switching of the system hence the torque and flux linkage errors were reduced.

In the conventional DTC a voltage vector is applied for the entire switching period, and this causes the stator current and electromagnetic torque to increase over the whole switching period. Thus for small errors, the electromagnetic torque exceeds its reference value early during the switching period, and continues to increase, causing a high torque ripple. This is then followed by switching cycles in which the zero switching vectors are applied in order to reduce the electromagnetic torque to its reference value.

The ripple in the torque and flux can be minimize by applying the selected inverter vector for a complete switching period, as in the conventional DTCinduction motor drive, but only for a part of the switching period. The time for which a non-zero voltage vector has to be applied is selected just to increase the electromagnetic torque to its reference value and the zero voltage vector was applied for the rest of the switching period.

During the application of the zero voltage vector, no power was consumed by the machine, and thus the electromagnetic flux is almost constant, it was only decreases slightly.Theaverage input DC voltage to the motor during the application of each switching vector was?Vdc. By adjusting the duty ratio between zero and one, it is possible to apply voltage to themotor with an average value between 0 and Vdc during each switching period. Thus, the

Torque ripple will be low compared to when the full DC link voltage was applying for the completeswitching period. This increases the demand of the voltage vector, without an increase in thenumber of semiconductor switches in the inverter.

The duty ratio of each switching period is a non-linear function of the

Electromagnetic torque error, stator flux-linkage error, and the position of the stator fluxlinkagespace vector. Therefore, by using a fuzzy-logic-based DTC system, it is possible to perform fuzzy-logic-based duty-ratio control, where the duty ratio is determined during each switching cycle. In such a fuzzylogic system, there are three inputs, the electromagnetic torque error,the stator flux-linkage space vector position (??) within each sector assigned with the voltage vectors and the flux error where the output ofthe fuzzy-logic controller is equal to the value of duty ratio.

There are various types of fuzzy logic controller for this particular application. A Mamdani-type fuzzy logic controller, which contains a rule base, a fuzzifier,and a defuzzifier, is selected. Fuzzification is performed using membership function. The inputs and the output of the fuzzy controller are assigned Gausian membership functions. The universe of discourse for the torque error and the duty ratio is variedusing simulations to get acceptable torque ripple reduction.

The attention in the fuzzy rule is to reduce the torque ripple. Generallythe duty ratio is proportional to the torque error, since the torque rate of change isproportional to the angle between the stator flux and the applied voltage vector, the duty ratio depends on the position of the flux within each sector. The use of two fuzzy sets isthe fact that when the stator flux is greater than its reference value a voltage vector that advance the stator flux vector by two sectors is applied which result in a higher rate of change for the torque compared to the application of a voltage vector that advance the stator flux vector byone sector when the stator flux linkage is lower than its reference value.

The duty ratio is selected proportional to the magnitude of the torque error so that if the torque error is Small, Medium or Large THEN the duty ratio is Small, Medium or Large respectively. The fuzzy rules are then adjustedto reflect the effects of theflux error, torque error and position of the space vector error. If the torque error ismedium and the stator flux lies in sector with magnitude greater than its reference value then the voltage vector Vk+2 isselected. If the flux position is small, that means there is a large angle between the flux andthe selected voltage vector that makes the selected vector more effective in increasing thetorque so that the duty ratio is set as small rather than medium, the fuzzy rule is stated as IF (torque error is medium) AND (flux position is small) THEN (duty ratio is small)IF (torque error is large) AND (flux position is small) THEN (duty ratio is medium).

Using the above reasoning and simulation to find the fuzzy rules, the two sets of fuzzy rulesare summarized in Table 3.2 below.

Table 3.2 Rules for the duty ratio fuzzy controller

Flux Torque error dT=k1 Small Medium Large

Negative

d?=0 Small Small Small Medium

Large Small Medium Large

Positive

?d=1 Small Small Medium Large

Medium Small Medium Large

Large Medium Large Large

Fuzzy logic toolbox was used in the implementation of theduty ratio fuzzy controller. The Graphic User Interface included in the toolbox wasused to edit the membership functions for the inputs (the torque error and the flux position),the output (the duty ratio). The membership functions and the fuzzy rules were adjusted using thesimulation until an acceptable torque ripple reduction was achieved.

Simulate the model above in the Simulink environment and validate the result.

The model that will reduced the high ripple torque in the induction motor were developed.

To enable us study the performance of the conventional direct torque control and fuzzy logic with duty ratio controller for four (4) pole induction motor torque control and also to simulate for the same and verified for the purpose of reducing the high torque ripple in the induction motor drive.

The motor parameters

Definition of terms

Pa = Active power per phase

Qa = Phase reactive power

Ia = Phase current

Va = phase voltage

Rs = Stator winding resistance

Rr = Rotor winding resistance

Lm = Magnetizing inductance per phase

Xis = Stator leakage reactance

Lis = Stator inductance per phase

Xir = Rotor leakage reactance

Lir = Rotor leakage inductance per phase

Dc = Direct current

Rdc = Resistance in direct current

X = Reactance

Xm = Magnetizing reactance

Xn = Total reactance

DETERMINATION OF INDUCTION MOTOR PARAMETERS

The motor is a three phase 158-W, 240-V induction motor (Model 295 Bodine Electric Co.)

The motor is Y-connected with no access to the neutral point.

DC Resistance Test:

To determine R1;

Connect any two stator leads to a variable voltage DC power supply.

Adjust the power supply to provide rated stator current.

Determine the resistance from the voltmeter and ammeter readings.

As shown in figure 3.7, a DC voltage VDC is applied so that the current IDC is close to the motor rating.

Because the machine is Y-connected: RS = Rdc/2 = (VDC/IDC)/2.

From measurement, VDC = 30.6V, IDC = 1.05A.

Hence,

RS = RDC = (31.5/1.04) = 15.14?/phase.

2 2

Figure 3.7 Circuit for DC resistance test.

BLOCKED – ROTOR TEST

To determine X1 and X2

Determine R2 when combined with data from the DC test.

Block the rotor so that it will not turn.

Connect to a variable voltage supply and adjust until the blocked – rotor current is equal to the rated current.

NO LOAD TEST

To determine the magnetizing reactance, Xm and combined core, friction, and wind age losses.

Connect as in block rotor test below.

The rotor is unblocked and allowed to run unloaded at rated voltage and frequency.

The set up for no load test and blocked rotor test is shown in the figure below:

Figure 3.8 Circuit for no load and locked rotor test.

With the motor running at no load, measure V, I and P to find the machine reactance Xn =Xis+Xm

Table 4.3 Measured value

Frequency (Hz) 50

Voltage (V) 230

Current (A) 1.32

Real power (W) 158

At no load the per-unit slip is approximately zero, hence the equivalent circuit is as shown in figure 3.9 below.

Figure 3.9 Equivalent circuit of three phase induction motor under no load test.

The real power P represents,

Hysteresis and Eddy current losses (core losses)

Friction and wind age losses (rotational losses)

Copper losses in stator and rotor (usually small as no load)

Phase voltage:

Va =V = 220 = 132V

?3?3

Phase current:

la = 1.32A

Phase real power:

Pa = Pa/3 = 138.2 ÷3 = 46.1W

Phase reactive power:

Q_a = ??(VaIa)2-P2a= ?(((137 x 1.32)2)-(46.1)2)=174.86VAr?_

Xn = Qa=174.86 =100.36?

I2a 1.322

Since S ~ 0,

Xn~ Xls +Xm

3. Locked rotor test

With the rotor locked, the rotor speed is zero and per- unit slip is equal to unity. The equivalent circuit is as shown in Figure 3.10 or Figure 3.11.

Figure 3.10 Equivalent circuit of three phase induction motor under locked rotor test.

Figure 3.11 Simplified equivalent circuit of three phase induction motor under locked rotor rest.

Table 4.4 the tested value

Frequency (Hz) 50

Voltage (V) 68.52

Current (A) 1.3

Real power (W) 105.33

Phase voltage:

Va =V = 68.52 = 39.56V

?3 ?3

Phase current:

la = 1.3A

Active power per phase

Pa = P = 105.33=35.1W

3 3

Reactive power phase

Q_a = ??(VaIa)2-P^2 a= ?(((35.56 x1.3)2)-(35.1)2)=30.08VAr?_

For a class C motor.

Xls = 0.3 x Qa= 0.3 x 30.08 = 5.34?

I2a 1.32

Xlr = 0.7 xQa = 0.7 x 30.08 = 12.46?

12a 1.32

From the no – load test, Xn = 100.36?, so

Xm = Xn – Xls = 100.36 – 5.34 = 95.02?

R = Pa = 35.1 = 20.77?

12a 1.32

From figures 3.11,

R2 = R – Ris= 20.77 – 5.34 = 1 5.43?

Comparing figures 3.10 and 3.11,

R2 + jX2 = (Rr + jXir) x jXm

(Rr + jXir) + jXm

R2 =Rr X2m

Rr + (Xlr + Xm)2

Rr = R2 x (Xir + Xm)2 = 15.43 x (12.46 + 95.02)2 = 19.74?

Xm95.02

Summarizing,

Stator winding resistance Rs = 15.14?/phase

Rotor winding resistance Rr = 19.74?/phase

Magnetizing reactance Xm = 95.02?/phase

The magnetizing inductance per phase is

Lm = Xm = 95.02 = 0.3024H

2?f 2? x 50

Stator leakage reactance Xls= 5.34?/phase

The stator inductance per phase is

Lls = Xls= 5.34 = 0.0169H.

2?f 2nx50

Rotor leakage reactance Xlr = 12.46?/phase,

The rotor leakage inductance per phase is

Llr = Xlr =12.46 = 0.0396H.

2?f 2?x50

Table 4.5: Motor parameters

Rated voltage 240V

Maximum torque 1.5N-m

Poles 4

Rated speed 1440rpm

Stator resistance 15.14?

Rotor resistance 19.74?

Stator leakage inductance 0.0169H

Rotor leakage inductance 0.0396H

Mutual inductance 0.3024H

3.3 IMPLEMENTATION

MATLAB fuzzy logic tool box was used in the implementation of the duty ratio fuzzy controller. The graphic user interface included in the tool box was used to edit the membership functions for the inputs (the torque error and the flux position), the output (the duty ratio). A Mamdani type fuzzy inference engine was used in the simulation. The membership functions and the fuzzy rules were adjusted using the simulation until a particular torque ripple reduction was achieved.

To know the performance of the duty ratio controller, the simulation was run at switching frequency of 5KHz. The difference between the conventional DTC and DTC with duty ratio fuzzy control was clearly realized by monitoring the switching behavior of the stator voltage and the electric torque. The selected voltage vector is applied for the complete sampling period and the torque keeps increasing for the complete period, then a zero voltage is applied and the torque keeps decreasing for the complete sampling period and these results in high torque ripple.

The selected voltage vector is applied for part of the sampling period and removed for the rest of the period. As a result, the electric torque increases for part of the sampling period and then starts to decrease. By adjustment of the duty ratio, the desired average torque may be continuously maintained. The duty ratio controller smoothly adjusts the average stator voltage.