CHAPTER 6 RESULTS AND DISCUSSION

Transcription

1 72 CHAPTER 6 RESULTS AND DISCUSSION 6.1 INTRODUCTION The earlier chapters discuss the problems that occur due to improper winding in three phase induction motors, the methodological basis for the approaches for the solution to the problems and the non-destructive approaches including the proposed algorithms. This chapter deals with the theoretical and experimental validity of the proposed methods. 6.2 THEORETICAL VALIDITY OF THE PROPOSED ALGORITHMS The proposed Forward and Reverse algorithms discussed in this thesis are concerned with three numbers of winding variables / winding data viz., Number of Parallel circuits in each phase (NPCP), Number of Turns per coil (NTPC) and Conductor cross-sectional Area (CA). The actual values of NPCP, NTPC and CA are represented by symbols a, T c and a z respectively. In the forward algorithm to determine the standard stator, the value of the three unknown winding variables present in the stator should be verified to check whether they are as per the designer s winding specification. In the Reverse algorithm to determine the unknown winding data, the three unknown winding data in a motor has to be ascertained.

2 73 Both the algorithms make use of the EMF relation and resistance relation respectively as given below in the following equation (3.7) and equations (3.19) or (3.22). EMF induced per phase of the winding is Eph ( m / a)*(2* Ep * ke)* Tc * km (3.7) winding is Depending upon the winding type, the resistance per phase of the or R L T m a a (3.19) 2 mt c / ( z ) R L T cp a a g ' mti * c * s / ( z 2 ) (3.22) i 1 Further, visual inspection test is carried out. The conductor area obtainable through visual inspection can take any one of the three possible values viz., the CA corresponding to the actual SWG present in the stator, or corresponding to the actual SWG minus one SWG or corresponding to the actual SWG plus one SWG of the conductor used. This gives the possible values of CA that will be determined through visual inspection and therefore provides a constraint as in the following equation (6.1). The SWG can take only one of the discrete values within this constraint. a z (actual SWG 1SWG) a z a z (actual SWG + 1SWG) (6.1) In order to ascertain the validity of the Forward algorithm for determination of standard stator or Reverse algorithm for determination of the three unknown winding variables, it should be established that there exists a unique solution for the set of variables that satisfies all the three relations /

3 74 constraints. For this, let it be assumed that there exists a set of the three winding data (T c1, a z1, a 1 ) that forms a solution, distinct from the set of the three winding data (T c, a z, a) actually present in the winding. Let this solution, which is different from the actual winding present have NTPC, CA and NPCP, which are x, y and z times the actual NTPC, CA and NPCP (i.e. T xt a 1 ya, a1 za ). c1 c, z z Let the EMF per phase and resistance per phase with the winding data present in the stator be E ph and R respectively. If there exists another solution other than the winding data present in the stator that satisfies the relations/constraints, the set of three winding data, should provide the same value of E ph as provided by the actual winding data. Then, Eph ( m / za)*(2* Ep * ke)* xtc * km (6.2) Then, from equation (6.2), the relation between x and z should be x z (6.3) In that case, this set of winding data should provide the same value of R as provided by the actual winding data present in the stator. The equation (6.4) or equation (6.5) should be satisfied depending upon the whether the winding is lap connected or concentric. or R L xt m ya z a (6.4) 2 2 mt c / ( z ) R L xt cp ya z a g ' mti * c * s / ( z 2 2 ) (6.5) i 1

4 75 Then, from equation (6.4) or equation (6.5), the relation between x, y and z should be x 2 yz (6.6) Substituting the relation in equation (6.3) into equation (6.6), to make a single equation out of the two equations, it can be deduced that z 1/ y (6.7) The value of z in equation (6.7) should also satisfy the relation with y, which arises due to visual inspection test. The BIS standard, IS (2003) corresponds to enamelled copper conductor in India. According to this Standard, the enamelled copper conductors have discrete conductor sizes. As specified earlier in the constraint equation (6.1), the conductor size observed by Visual inspection test may be either the exact conductor size as actual conductor size present in the winding or will be one SWG lower or one SWG higher than the actual SWG present. Conductors have cross-sectional area in sq. mm., which is a fractional number (as can be found from SWG tables) times the CA of the neighbouring conductor in the SWG table. Stated in other terms, y can assume either 1, in the case when SWG observed by visual inspection is equal to the actual SWG present in the stator or it will be one of the fractional values times (The inverse of y will not be an integer) greater or lesser than the actual SWG if it was observed differently, as given in Table A1.1 of Appendix 1, which provides the conductor SWG table and details of ratio (y) of possible value of SWG observed by Visual inspection test to the actual value of SWG present. Table A1.2 of Appendix 1 provides details of ratio (z) of a value of NPCP other than the actual value of NPCP, to the actual value of NPCP present. The values that z can assume for all the common number of poles is one among the following: z = {0.125, 0.167, 0.25, 0.333, 0.5, 0.667, 1, 1.5, 2, 3, 4, 6, 8}

5 76 It can be found that equation (6.7) is satisfied by the CA value, which is as per actual value (in other words, only when y is equal to 1) for which the z value is one. For other values of y, the equation (6.7) is not satisfied. Hence, except for the value of z to be 1, the other values are impossible. Then, from equation (6.5), it follows that x y z 1 (6.8) Therefore, essentially the distinct solution, if any that can exist should follow the relation in equation (6.8) or it can be concluded that the solution to the problem is unique. Hence, a unique solution exists, for which all the three relations are satisfied (i.e. x y z 1). Thus, basically there are two independent relations in the form of EMF relation and resistance relation and a constraint provided by visual inspection. The NPCP can assume only possible integer values depending upon the number of poles and number of layers of the winding. The NTPC can assume only integer values. The CA also has discrete values. The winding conductor area is specified by the Standard, which is in vogue in the region of manufacture / rewind of the motor. Hence, the solution for the Number of parallel circuits in each phase, the Number of turns per coil and Conductor cross-sectional area obtained from the three relations is unique. Hence, both the proposed Forward algorithm and the Reverse algorithm, based on these relations / constraints, are theoretically valid. 6.3 EXPERIMENTAL RESULTS Experiments were carried out on a number of specially wound three phase stators. These stators were produced by specifying the special winding requirement to the manufacturer, in order to develop the non-destructive

6 77 methods. However, keeping an eye on the comprehensiveness of the thesis, it is not possible to give the details of the tests performed on all of them. Further, this will also not be helpful in driving home the point of achieving energy efficiency in three phase motors by the proposed Quality assurance approach. Hence, Table A2.1 of Appendix 2 lists only the details of the special stators utilised for development of the proposed methods. However, the tests performed on five motors according to the present approach are given in detail. Table A2.2 of Appendix 2 provides the details of the test exciters used in the proposed methods. In order to determine the accuracy of the methods developed and to ascertain the unwarranted efficiency reduction that occur due to variation in the winding variables / winding data from the original design specifications, three numbers of specially wound three phase stators were manufactured. However, the NTPC, NPCP and CA of the winding in the individual stators of the motors were unknown to the research scholar, prior to conducting the proposed method. The stators were made of identical cores used for 400 V, 50 Hz, 4 pole, 3.7 kw / 5 hp, foot mounted three phase squirrel cage induction motors. As per design data for the 3.7 kw motor provided by the designer to the winder, the stators should have Double layer lap winding with a coil span of 7 slots, one parallel circuit in each phase, 30 turns per coil and 19 SWG (BIS Standard) conductor. Since the motors are specially made for validating the proposed methods, the winder was instructed to purposefully introduce winding deviation in the form of either a reduction in Number of turns per coil (NTPC) or a reduction in Conductor cross-sectional area (CA) from the designer specified data in each of the motors. The third motor was made to stick to the winding design for 3.7 kw rating. The three stators with stator winding configuration A, B and C of those motors were subjected to the methods in the proposed Quality Assurance and Analysis approaches. In addition to these motors, in order to determine the experimental validity

7 78 thoroughly, three phase cage motors of 1.1 kw and 7.5 kw, in operational condition were taken from the field, whose winding data were totally unknown to the research scholar prior to the test, were subjected to the proposed method (to determine the unknown winding data) of the analysis approach. The stator winding configurations in them were named as configurations D and E respectively. Table 6.1 gives the details of test exciters used for the EMF test. Table 6.2 provides the details of observations of the necessary tests for the proposed approaches. Table 6.3 provides the details of the determination of standard stator for the Quality Assurance approach, according to the Forward algorithm. Table 6.1 Details of test exciters used for the EMF test A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A Lap 4 SL 3 Y A, B, C and E Lap 4 SL 3 Y D A1 - Number of Exciter Slots, A2 - Outer Diameter (mm), A3 - Core Length (mm), A4 - Type of Winding, A5 - Number of Poles, A6 - Number of layers, A7 - Coil span, A8 - Type of three phase connection, A9 - Number of Turns per Coil in Exciter, A10 - Conductor Area (mm 2 ), A11 - The configuration of stators for which the test exciter was employed, Lap - Lap winding, Con - Concentric Winding, SL - Single Layer, Y - Star connection.

8 79 Table 6.2 Details of observations of the necessary tests for the proposed approaches Motor configuration Name Resistance Measurement Test T w ( C) R Lm (Ω) R m (Ω) E L12m (V) E L23m (V) EMF Test E L31m (V) Mean E Lm (V) E pm (mv) Visual inspection Test Conductor SWG observed A B C D E T w - The winding temperature during resistance measurement( C), R Lm - Resistance between line terminals of the three phase connected stator winding obtained through measurement (Ω), R m - Resistance per phase of the winding arrived at through measurement (Ω), E Labm - Measured value of Line EMF between terminals a and b, E pm - Measured value of EMF in the search conductor.

9 Table 6.3 Details of determination of standard stator for the Quality Assurance approach Motor configuration Name Based on Observation T w ( C) R m (Ω) Resistance Measurement Test CA= mm 2 (19.5 SWG) Theoretical calculation Nom. R (Ω) CA= mm 2 (19 SWG) CA= mm 2 (18.5 SWG) A B C EMF Test Based on observations Theoretical calculation Mean E phm (V) E pm (mv) E ph (V) values for variation of T c B1 Visual Inspection Test C1 C Qualifies Does not qualify Does not qualify T w - The winding temperature during resistance measurement ( C), R m - Resistance per phase of the winding arrived at through measurement (Ω), Nom. R - Nominal value of resistance per phase obtained theoretically according to the values of NTPC, NPCP and nominal CA (for the SWG specified according to Standard for enameled copper conductor) as per winding specifications, CA - Conductor cross-sectional area (in mm 2 ), E phm - Value of Phase EMF obtained from measurement, E pm - Measured value of EMF in the search conductor, E ph - Theoretical value of Phase EMF, B1 - Assumed NTPC for which minimum difference occurs between calculated and measured values of phase EMFs, C1 - Conductor SWG as per visual inspection, C2 - Result of the Quality assurance test to determine the standard stator. 80

10 81 Table 6.4 provides the details of determination of Number of parallel circuits in each phase of the stators. The NPCP values are found out through the proposed Reverse algorithm to determine the unknown winding data. This table gives the details of determination of Number of parallel circuits in each phase, of the winding in the individual configurations. This value will be used in the subsequent steps of the Reverse algorithms. Tables 6.5, 6.6, 6.7, 6.8 and 6.9 provide the details of determination of NTPC and CA in stators of Configurations A, B, C, D and E respectively. The NTPC and CA values are determined by the proposed Reverse algorithm. These tabulations provide the results for the three unknown data determined by the approach viz., NPCP, NTPC and CA in the individual stator winding configurations. In order to understand the significance of improper winding on the performance, including efficiency, stators with winding configurations A, B and C mentioned above were manufactured. Motors were made by assembling a common squirrel cage rotor into one each of the stator configurations at a time and load tests were performed on each of the assembled motors. On each of the three motors, the same load of about the full load of a 3.7 kw motor (i.e % of full load) was applied, and observations were noted. Table 6.10 provides the details of load and predetermination tests. The winding data in all the five stators were physically determined by destructive examination, following good practices for removal of winding. Table 6.11 provides the details of stator winding data obtained through the proposed Reverse algorithm and Physical determination, in stators of configuration A through E.

11 82 Table 6.4 Details of determination of Number of parallel circuits in each phase of the stators Based on EMF Test Based on Resistance Measurement Stator configuration D1 Mean E phm E pm Cal. T c R m CA D2 D3 D4 (V) (mv) (Ω) (mm 2 ) A B C D E D1 - Assumed NPCP, E phm - Value of Phase EMF obtained from EMF Test measurement, E pm - Measured value of EMF in the search conductor (mv), Cal. T c - Calculated value of NTPC, R m - Resistance (in Ω) per phase of the winding arrived at through measurement, CA - Cross-sectional area of winding conductor (sq. mm), D2 - Calculated NPCP, D3 - Difference between Assumed NPCP and Calculated NPCP, D4 - Result for NPCP.

12 83 Table 6.5 Details of determination of NTPC and CA in the stator of configuration A (with number of parallel circuits per phase equal to 1 as observed in Table 6.4) Configuration A R ( ) R m ( ) Ass. T c CA = mm 2 (19.5 SWG) CA = mm 2 (19 SWG) CA =0.9812mm 2 (18.5 SWG) R m - Resistance per phase of the winding arrived at through measurement, Ass. T c - Number of turns per coil assumed for calculations, CA - Nominal value of crosssectional area of winding conductor (sq. mm.) as per IS 13730, R - Theoretically calculated value of resistance per phase of the winding.

13 84 Table 6.6 Details of determination of NTPC and CA in the stator of configuration B (with number of parallel circuits per phase equal to 1 as observed in Table 6.4) Configuration B R ( ) R m ( ) Ass. T c CA = mm 2 (20.5 SWG) CA = mm 2 (20 SWG) CA = mm 2 (19.5 SWG) CA = mm 2 (19 SWG) CA = mm 2 (18.5 SWG) CA = mm 2 (18 SWG) R m - Resistance per phase of the winding arrived at through measurement, Ass. T c - Number of turns per coil assumed for calculations, CA - Nominal value of cross-sectional area of winding conductor (sq. mm.) as per IS 13730, R - Theoretically calculated value of resistance per phase of the winding.

14 85 Table 6.7 Details of determination of NTPC and CA in the stator of configuration C (with number of parallel circuits per phase equal to 1 as observed in Table 6.4) Configuration C R m ( ) R ( ) Ass. T c CA = mm 2 CA = mm 2 CA = mm 2 (19.5 SWG) (19 SWG) (18.5 SWG) R m - Resistance per phase of the winding arrived at through measurement, Ass. T c - Number of turns per coil assumed for calculations, CA - Nominal value of cross-sectional area of winding conductor (sq. mm.) as per IS 13730, R - Theoretically calculated value of resistance per phase of the winding.

15 86 Table 6.8 Determination of NTPC and CA in the stator of configuration D (with number of parallel circuits per phase equal to 1 as observed in Table 6.4) Configuration D R ( ) R m ( ) Ass. T c CA = CA = mm mm 2 (23 SWG) (22.5 SWG) CA = mm 2 (22 SWG) CA = CA = mm mm 2 (21.5SWG) (21 SWG) R m - Resistance per phase of the winding arrived at through measurement, Ass. T c - Number of turns per coil assumed for calculations, CA - Nominal value of cross-sectional area of winding conductor (sq. mm.) as per IS 13730, R - Theoretically calculated value of resistance per phase of the winding.

16 87 Table 6.9 Details of determination of NTPC and CA in the stator of configuration E (with number of parallel circuits per phase equal to 2 as observed in Table 6.4) Configuration E R ( ) R m ( ) Ass. T c CA = mm 2 (20.5 SWG) CA = mm 2 (20 SWG) CA = mm 2 (19.5 SWG) R m - Resistance per phase of the winding arrived at through measurement, Ass. T c - Number of turns per coil assumed for calculations, CA - Nominal value of cross-sectional area of winding conductor (sq. mm.) as per IS 13730, R - Theoretically calculated value of resistance per phase of the winding.

17 Table 6.10 Details of Load and Pre-determination tests Motor Configuration Name Details of Load test (applying a load of 95.6% of full load of a 3.7 KW three phase induction motor) Line current (A) Input (W) Power factor Torque (N-m) Output (W) η (%) R T/A (N-m/A) Temperature rise ( C) Results of Pre-determination test Starting Torque (N-m) Motor A * Motor B Motor C η - Efficiency, R - Efficiency Reduction in percentage points compared to motor A, T/A - Torque per line Ampere (N-m/A). Note: * - Efficiency of motor A is well above the Indian Electrical Equipment Manufacturers Association (IEEMA) Standards for Energy Efficient Motors (EFF1). 88

18 Table 6.11 Details of stator winding data obtained through the proposed reverse algorithm and physical determination Motor Configuration Name Configuration A Configuration B Configuration C Configuration D Configuration E Rating of the motor as in specification sheet 400 V, 50 Hz, 3.7 kw, 3 Phase, Squirrel Cage Induction Motor 400V, 50 Hz, 1.5 HP, 3 Phase, Squirrel Cage Induction Motor 400V, 50 Hz, 7.5 kw, 3 Phase, Squirrel Cage Induction Motor Type of Winding as per specification and actual Type of Three Phase connection as per specification and actual Results of the algorithm for determination of Winding data NPCP DL Lap Delta 1 DL Lap Star 1 DL Concentric Star 2 CA (mm 2 ) (19 SWG) (20 SWG) (19 SWG) (21.5 SWG) (20 SWG) NTPC Winding data present in the stator determined by physical determination NPCP CA (mm 2 ) (19 SWG) (20 SWG) (19 SWG) (21.5 SWG) (20 SWG) NTPC DL Lap - Double Layer Lap winding, NPCP - Number of Parallel Circuits in each Phase, CA - SWG - British Standard Wire Gauge, NTPC - Number of Turns per Coil. Winding Conductor cross-sectional area, 89

19 EXPERIMENTAL VALIDITY OF THE PROPOSED FORWARD ALGORITHM In the approach for Quality Assurance of custom-designed motors, the proposed non-destructive method to determine the standard stator is the central one. To ascertain its experimental validity, the three stators under test Configuration A, B and C - were subjected to this proposed method. The designer s winding specification is that the winding in each of the three stators be the one for 3.7 kw, 4P, 50 Hz, Three phase squirrel cage induction motor. As per design data for that particular motor, the stators should have Double layer lap winding with a coil span of 7 slots, 30 turns per coil and conductor of 19 SWG. The results of physical determination in Table 6.11 shows that the stator with winding configuration A was, as per the design data for the 3.7 kw motor; whereas, stator with winding configuration B had NTPC as per the design specification and the conductor area is lesser compared to the design specification (and the stator with winding configuration A) or SWG of conductor in configuration B is higher in SWG number, than the design requirement, by 1 SWG. Stator with winding configuration C had conductor size same as that of the original value for the design variable as per the design, but the number of turns per coil was one turn less than that of the original value for the design variable. The details of the results of Forward algorithm for determination of standard stator are given in Table 6.3. In the stator with configuration A the value of resistance per phase (R m ) obtained through measurement is above the theoretical value of resistance per phase calculated, for the nominal value of the SWG conductor specified by the designer, and below the theoretical value of resistance per phase calculated for the nominal value of CA corresponding to the next higher conductor SWG. This is the condition to be satisfied as per Step 2 of the algorithm. The NTPC value for which minimum difference

20 91 between the theoretically based value of phase EMF (E ph ) and the value of phase EMF obtained through measurement (E phm ) occurs is found to be 29 turns. This satisfies the condition in Step 5 of the algorithm. As per the Step 7, the check to determine whether the observed SWG is around designer specified SWG is also satisfied. Hence, the stator with configuration A qualifies to be a standard stator for surge comparison testing. In the stator with configuration B, the value of phase resistance (R m ) obtained is above the theoretical value of resistance per phase calculated, for the nominal value of the SWG conductor used and exceeds the theoretical value of resistance per phase (R) calculated for the nominal value of next higher conductor SWG in the SWG table. Hence, the condition as per Step 2 of the algorithm is not satisfied and the stator with configuration B does not qualify to be a standard stator for Surge comparison testing. In the stator with configuration C, the value of phase resistance (R m ) obtained through measurement is below the theoretical value of resistance per phase (R) calculated for the nominal value of the conductor SWG used. Hence, this stator does not satisfy Step 2 of the algorithm and does not qualify to be the standard stator. The tolerance allowed for the resistance value measured is based on the tolerance for possible variation of conductor cross-sectional area as specified by BIS (IS 13730, 2003) and the error in the equipment employed for measuring the resistance and the winding temperature of the motor at the time of measurement. Empirically, it has been found that the measured value of resistance is always above the theoretical value of resistance. One factor that could be accounted the same is the maximum and minimum resistance per metre length for each SWG given in the BIS Standard, are for the conductors that come out of the manufacturing line, devoid of any significant stretching. However, inadvertently while winding, the wires are subjected to

21 92 tension, because of that there will be a slight change in the conductor size, though very small. Therefore, the resistance value for the particular SWG of the copper conductor will be on the higher side than the BIS specified resistance values. From the above discussions, it could be found that the algorithm can check whether the winding data under consideration are as per the designer s specification and qualifies as fit, only the stator with winding that matches the designer s specifications. 6.5 EXPERIMENTAL VALIDITY OF THE PROPOSED REVERSE ALGORITHM The proposed Reverse algorithm for determination of winding data was applied on stators under test Configuration A through E. Table 6.4 gives details of determination of the Number of Parallel Circuits in each Phase (NPCP) of the winding in the stator configurations A through E according to the algorithm. This NPCP value found is used in the subsequent steps of the algorithm. In the stator of configuration B, the NTPC was calculated to be 30 as in Table 6.6. By visual inspection observations as shown in Table 6.2, the conductor SWG present in configuration B was approximated to be 19 SWG. This is different from the actual winding SWG present in the configuration B obtained as shown in Table 6.11, which provides the details of Winding Data obtained through physical determination. Though the winding conductor SWG could not be observed correctly during visual inspection test, it did not affect the result of NPCP and its value was found to be 1. This value of NPCP is used in subsequent calculations. Table 6.6 provides the details of determination of NTPC and CA in the stator of configuration B. The

22 93 theoretical value of resistance calculated that is lesser than the measured value of resistance and also gives the minimum difference between the two values of resistance was found. The corresponding value of NTPC and conductor SWG is found to be 30 turns per coil and 20 SWG respectively. This is the result for the NTPC and CA present in configuration B, according to the algorithm. Therefore, the NPCP of 1, the NTPC of 30 turns and conductor SWG of 20 is the result of the algorithm for determination of winding data. The maximum error committed while estimating the winding conductor SWG by Visual inspection test may be the actual SWG ± 1 SWG. By means of the features in the algorithm, any mistake committed in estimating the SWG by Visual inspection will not affect the results of determination of the winding data present in the stator. The difference between the stators with winding configuration A and B, observed by physical determination, as in Table 6.11, is that Stator B has conductor of 20 SWG, whereas, Stator A has 19 SWG conductor present in it. The result of the reverse algorithm in Table 6.11 shows that the algorithm can detect Standard Wire Gauge (SWG) deviation accurately. The difference between the stator windings of configuration A and C is that configuration C has 29 turns per coil, whereas, configuration A has 30 turns per coil. As per the results of the Reverse algorithm, the NTPC in configurations A and C could be accurately determined by the method. Hence, the tests show the effectiveness of the method in determining winding data accurately. The experimental results on the operational 1.1 kw and 7.5 kw motors taken from the site, whose winding details were totally unknown, also show that the methods provide accurate results. This may be inferred by comparing the results of the proposed method and the results of the physical determination of the winding data in the motors, as detailed in Table 6.11.

23 94 The details of the winding data on all the stators of configuration A through E is provided in Table The stators of Configuration A, B and C had NPCP value of 1, whereas, the 7.5 kw motor had NPCP value of 2. The winding present in stators of configuration A, B, C, and D are lap winding, whereas, the winding in configuration E is of concentric type. The windings of configuration A, B and C were delta connected, whereas, the stators of configuration D and E are star-connected. The NTPC in the stators of configuration A, B, C, D, and E are 30, 30, 29, 28, and 35 respectively. The CA in the stators of configuration A, B, C, D, and E are of SWG (British SWG) 19, 20, 19, 21.5, and 20 respectively. Figures 6.1, 6.2 and 6.3 show the comparison of the reverse algorithmic and physical determination results for Number of parallel circuits in each phase, Number of turns per coil and Conductor cross-sectional area in SWG respectively. It can be found that the proposed reverse algorithm was able to determine all the above-said winding data in the stators of configuration A through E, accurately. Similarly, the proposed reverse algorithm was able to choose only the stator A as the standard stator, which had the winding variables as per design specification for the 3.7 kw motor. Prakash et al (2008a - List of Publications) reported an algorithm to determine the winding data in three phase induction motors. It could determine NPCP and CA accurately. However, there was a maximum error of 3.33% in the determination of NTPC. Compared with the above-said algorithm, the proposed Reverse algorithm in this thesis gives the exact values of the NPCP, NTPC and CA present. It may be noted that there is no error in Forward algorithm for determining the standard stator for the Surge Comparison test and in the Reverse algorithm for determining the unknown winding data in an operational rewound cage induction motor. This is due to the fact that the

24 2 2 Algorithmic results Physical determination results A B C D E Configuration Figure 6.1 Comparison of the reverse algorithmic and physical determination results for number of parallel circuits in each phase 95

25 Algorithmic results Physical determination results A B C D E Configuration Figure 6.2 Comparison of the reverse algorithmic and physical determination results for number of turns per coil 96

26 Algorithmic results Physical determination results A B C D E Configuration Figure 6.3 Comparison of the reverse algorithmic and physical determination results for the conductor area in SWG 97

27 98 value of the winding variables or the winding data can only take discrete values. The NTPC can take only integer values, NPCP can take only possible integer value, whereas the CA can take only discrete values according to the SWG Standard for winding conductor prevailing in the region of manufacture / rewind of the motor. Instead of the values of the above three winding variables / data being discrete, if they are continuous, the results of the approaches will not be accurate. However, due to manufacturing constraints, conductors will have only discrete cross-sectional area as per any Standard. Therefore, the results of the proposed algorithm are error free. 6.6 DISCUSSION ON THE NECESSARY TESTS EMF Test Effect of Slot Skewing In alternators, the effect of skewing is to introduce a breadth factor in the equation for induced EMF. If the EMF induced in a single conductor by the main flux of full pole pitch is E when there is no skewing, it becomes E [sin( / 2) /( / 2)] when the angle of skew in electrical radians is ψ. However, the turns per coil as in equation (3.8) or equation (3.9) is dependent upon the ratio of the EMF induced in the stator (E ph or E L ) to the induced EMF in the search conductor (E p ), both of which will have breadth factor in them if the slot is skewed. Hence, the breadth factor term gets cancelled out. Hence, effect of slot skewing must not have any effect on the accuracy of results of the test Effect of Harmonics The two harmonic effects to be considered are the EMF induced due to the pole harmonic component and the EMF induced due to the tooth ripple. The NTPC as in equation (3.8) is dependent upon the ratio of EMF

28 99 induced per phase of the stator (E ph ) to the EMF induced in the search conductor (E p ). The pole flux harmonics have the same effect on both the induced EMFs. Hence, their effects get cancelled out. Hence, the harmonic component of the traveling MMF wave does not affect the results. Langsdorf (1990) expounds that the primary cause of tooth ripples is the variable reluctance of the magnetic circuit due to the passage of the traveling MMF wave across the teeth and slots. If the extra reluctance due to the slot openings is small compared with that of the air gap (iron to iron), the tooth ripples will likewise be small enough to be inappreciable. In the experimental setup, it happens that the slot openings are small compared with the air gap. Hence, only the fundamental flux component is considered for calculation of the induced EMF. Further, the NTPC as in equation (3.8) or equation (3.9) is dependent upon the ratio of the EMF induced in the stator (E L or E ph according to the type of three phase connection or whether the individual phase terminals are available externally) to the induced EMF in the search conductor (E p ). The tooth ripples have the same effect on both the induced EMFs. Hence, the effect of tooth ripples gets cancelled out and will not affect the results of the test Search Conductor Length and Diameter In spite of search conductor length extending beyond the overhang portions and irrespective of the diameter of the search conductor chosen, the results for NTPC do not show much of deviation from the actual number of turns. This is so because, the search conductor induced EMF in the portion extending beyond the active core length will be negligible. Similarly, we intend to measure only the EMF induced in the search conductor and no

29 100 current flows through it to cause any voltage drop. In the test cases, the search conductors were chosen to have an area of mm 2 (or 20 British SWG). Further any deviation in the results for NTPC from the actual, will not affect the results of the methods, as the proposed algorithm has been developed to take care of these deviations Minimum EMF induced in the Search Conductor Table A3.1 of Appendix 3 provides the details of calibration of the Multimeter (Agilent A) employed for the tests. It shows the one year specification for measuring ac voltage. When EMF induced in the search conductor (E p ) is less than 4.6 mv, the limiting error for measurement will be greater than ±1%. Since, as per the test, the measurement error is to be kept below ±1%, the EMF induced in the search conductor is held greater than 6 mv. The key to accuracy in the EMF test is found to be in the alignment of the test exciter concentric within the stator Resistance Measurement Test The resistances of the stator windings were measured when the motors were at cold condition. The temperature at which the resistance of the stator is measured is noted. The resistivity of the copper conductor at an ambient temperature of 20 C was chosen to be Ω m. Table A3.2 of Appendix 3 provides the details of calibration of the noncontact thermometer (Raynger ST20 Pro Standard) employed, for measurement of winding temperature.

30 Visual Inspection Test The conductor area in SWG is approximately determined by visual inspection of the winding overhang. This can be done by having samples of winding wires, of known SWG, with conductor areas near about the conductor cross-sectional area present in the stator winding under test. The estimated value of SWG actually present can take only one of the discrete values within the constraint in equation (6.1). a z (actual SWG 1SWG) a z a z (actual SWG + 1SWG) (6.1) The conductor area value estimated by visual inspection can take any one of the discrete, possible values within the limits provided above. The possibilities of conductor SWG that can be determined is depicted in equation (6.1) tables. The conductor area in units of sq. mm. is obtained from the SWG Determination of Length of Mean Turn Due care is to be taken to determine the Length of Mean turn(s) as it is crucial in determining the results of the proposed test procedures. Most three phase induction motors below 500 hp have concentric windings. For the case of concentric winding, determination of length of mean turn is relatively easy. It can be obtained by measuring the sum of the lengths of the core and winding overhang as well as the arc length between the slots in which the sides of a coil lie.

31 PERFORMANCE CHANGES DUE TO VARIATION IN WINDING DATA Table 6.10 shows the results of load test on Motors A, B and C. It can be found that there is a significant reduction in efficiency of the motor due to stator winding deviation from the designer s specification. As it can be found from Table 6.10, in Motor B (that has conductor area lesser, corresponding to an SWG, than in Motor A), the efficiency is lesser by 5.1 percentage points than that of Motor A. Similarly in Motor C (that has one turn per coil lesser compared to Motor A), the efficiency is lesser by 5.35 percentage points than that of the Motor A. It may be noted that the loss in efficiency in both the cases is only due to reduction in a single winding data from the winding data specified by the design. However, due to improper winding, they had lesser efficiencies. Such reductions in efficiencies will lead to unwarranted energy consumption. Further, the Torque per ampere in Motor B (3.16 N-m/A) and Motor C (3.17 N-m/ A) are less compared to that of Motor A (3.27 N-m/A). The Power factor in Motor B (0.83) remains almost same, whereas, the power factor of Motor C (0.82) was lower compared to that of Motor A (0.83). In other words, performance reduction occurs due to winding deviation. Additionally, when subjected to the load of 95.6% of a 3.7 kw induction motor and when thermal equilibrium was reached, the temperature rise in Motor B (65 C) and Motor C (66 C) are higher compared to that of Motor A (60 C). The ten degree rule states that for each ten degree increase in winding temperature, it is estimated that the motor life would be reduced by one half. Hence, life of the motor will decrease because of increased heating in Motors B and C. This will have detrimental effect on the reliability of these motors.

32 103 The starting torque was found out from pre-determination tests on the three motors, made out of the stators (with configuration) A, B and C. In Motor B with conductor area lesser corresponding to 1 SWG than that required for the 3.7 kw motor as per design, the Starting torque (30.05 N-m) is lesser compared to that of Motor A (31.63 N-m). In Motor C with number of turns per coil lesser than that required for the 3.7 kw motor as per design, the starting torque (35.30 N-m) is higher compared to that of Motor A that has the winding as per the designer s winding data for the particular 3.7 kw design. Literature relevant to this experimental study is available and they analyse the issue of variation of performance winding variables. Hasuike (1983) has by design computations studied the effect of the change in the winding variables individually and its effect on the performance of the machine. Analysis done by Umans (1989) makes the above statement clear. According to him, If the motor in question was designed for optimal efficiency at the design stage, there is little that can be done at the repair / rewind stage to improve upon this optimization. The designer has already specified the choice of core material, the machine dimensions and the turns distribution. Changing the turns distribution can thus only move the efficiency below the optimum efficiency of the original design, thus reducing it. On the other hand, if the motor was not designed for optimal efficiency, increase in conductor area during a rewind may, in fact, improve the efficiency. This will result from well understood phenomenon. Perhaps, the motor has been rewound with the same turns distribution but with larger wire size (resulting in lower winding resistance and hence lower ohmic

33 104 losses). In this case, the result will simply be an improvement in efficiency. The graphs provided by Hasuike (1983) depict that a decrease in the stator conductor size will result in decrease in efficiency, space factor, locked rotor current and starting torque; whereas, operating temperature increases; while power factor remains almost constant. A decrease in the number of stator conductors per slot will result in a decrease in power factor, efficiency and stator conductor s space factor. It could also be deciphered from the graphs provided by him that the locked rotor current, starting torque and temperature rise increases when there is a reduction in the number of stator conductors in each slot than the original design results. Thus, the results obtained by performing actual load test (to determine efficiency, torque, power factor, temperature rise) and predetermination tests (to determine starting torque) go well with the design computational results of Hasuike (1983) and the analysis by Umans (1989). 6.8 BENEFITS OF THE PROPOSED QUALITY ASSURANCE APPROACH As per Indian Electrical Equipment Manufacturer s Association (IEEMA) Standards EFF1 category, the efficiency for 3.7 kw motor should have a minimum efficiency of 88.3%. From this, we know that the Motor A complies with the required efficiency level of an Energy-efficient design as per IEEMA. In Motors B and C, there is reduction in efficiency below this level, due to stator winding deviation from the design data. Such a reduction in performance due to improper winding may be detected only by performing end-of-line tests carried out by machine manufacturers. IEEE Standard 112 (2004) covers instructions for conducting and reporting the more generally applicable and acceptable tests of poly phase induction motors. However, in

34 105 manufacturing units, tests as per IEEE Standard 112 (2004) are not carried out on all the manufactured motors. Normally, routine tests like Resistance measurement tests, Polarity tests etc. and Surge test, to determine winding dissymmetry are carried out. Some of the motors produced in the customdesign may have deviation in the stator winding from the designer s winding specification. Bonnett (1994) elaborates that variation in efficiency and power factor from motor to motor with identical designs can exist due to manufacturing tolerances, raw material variations, and process changes. The accumulation of these tolerances can result in a significant variation. A 10-20% spread in losses is not uncommon. A difference due to winding deviation may be accounted to that of manufacturing tolerance etc. and there is a possibility that the motor may pass through the conventional quality assurance tests. Hence, there may be unwarranted performance, including efficiency, reduction. Even if the deviation in winding could be detected, it can be done so only after assembling the entire machine and on performing end-of-line tests etc. Hence, time and energy will be wasted on building a machine with improper winding in the stator and testing it. Table 6.3 is the tabulation for determination of standard stator for Quality Assurance of custom-designed motors. The stator of configuration A satisfies the EMF and resistance checks as per the proposed Forward algorithm. Hence, it qualifies to become the standard stator for Surge comparison test. If a consignment of custom-designed motors of this particular motor design would have been manufactured, this standard stator could be used for Quality Assurance of other stators. In the stator with winding configuration B, it could be found that the measured resistance value exceeds the tolerance limit for resistance. Hence, the stator does not qualify to be a standard stator for Surge

35 106 Comparison test and hence it is rejected by the proposed algorithm as having winding deviating from the design specification. In the stator with winding configuration C, the measured resistance exceeds the tolerance limit for resistance. Hence, the stator does not qualify to be a standard stator for Surge Comparison test and hence it is rejected by the proposed algorithm as having winding deviating from the design specification. The tolerance allowed for the resistance value measured is based on the tolerance for possible variation of conductor cross-sectional area of the conductor as specified in Standard (IS 13730, 2003) and the error in measuring resistance using the equipment used. The maximum and minimum resistance per metre length for each SWG is given in the BIS Standard. By empiricism, it could be found that the resistance per meter measured is above the nominal value of resistance per metre length for the particular SWG used. Thus, the proposed procedure for Quality Assurance of customdesigned three phase induction motor rejects stators with winding configurations, which deviate from the design data (here, B and C). Hence, any unwarranted performance, including efficiency, reduction is avoided. By rejecting stators that do not have winding as per designer s specification, the increased heating that happens because of improper winding is avoided. Hence, the motors produced have their normal life time and the reliability of the motors that are quality assured is higher. Further, the time and energy spent on a stator which has improper winding is also avoided.

36 UTILISATION OF THE PROPOSED ANALYSIS APPROACH FOR EFFICIENCY IMPROVEMENT National Productivity Council, India and Centre for Energy and Environmental studies, Princeton University, New Jersey (1994) provide the statistics that the electric motor is the single largest electricity using device in the world. In Indian Industry, motors account for an estimated 74 % (72 billion KWh per year) of all industrial electricity use. The other important motor-using sector is the agricultural sector, where essentially 100 % of electricity (an estimated 44 billion kwh per year) is consumed in pumping irrigation water. In a huge electricity using sector, the tendency at present is not a high favour for energy efficiency improvement by opting for Energyefficient motors. Biller (1978) provides the statistics that the ac motors ranging from hp account for over 53 percent of the total motor energy consumption. And yet, it is this same group of motors (NEMA sizes) that have been standardized and designed for low initial cost. Historically, efficiency requirements have usually been applied only to the larger selection of motors. As a result, the NEMA size motors have room for significant efficiency improvement. Since over half the installed motor horsepower is concentrated in the hp range, chances are that the average user can improve this motor efficiency. In India, Motors of rating less than 15 HP form 80 % of the motor population in India. In agriculture, the commonly used rating are 5 HP (3.7 kw) and 3 HP. And they have been standardised and designed for low initial cost. Walters (1999a) demonstrates that, in typical 1.5 kw and 15 kw motors, copper losses (and particularly the stator copper loss) dominate.

37 108 Rewinding is prevalent and most of the rewinders are not well informed. Because of the probability that winding configuration in a large population of motors may deviate from even Standard design performed according to design for lower capital cost, enormous amount of needless energy consumption occurs. This takes place predominantly in the agricultural sector and to a lesser extent in the Industrial sectors in India. Hence, due to variation of conductor size from the optimal conductor area and turns distribution from the optimal turns distribution as determined by the designer, enormous amount of power will be wasted in the form of stator copper losses. Therefore, there is a large potential for operating-efficiency improvement in such motors. Thus, needless energy consumption may be reduced and thereby enormous amount of energy could be conserved. Good rewinding practices as suggested by EASA need be followed in such cases (EASA / ANSI AR 100 (2006)). Efficiency improvement in motor is not the sufficient condition for improving energy efficiency. Though, it is a necessary condition for the improvement of efficiency Good Practices in Rewinding Darby (1986) suggests that lowest initial cost for a rewind should not be the major selection criterion. A low quality rewind will cost more in wasted electrical energy. All motor stators should be screened by a core loss test prior to stripping to identify motor with heavily damaged stator iron for possible replacement or to have the iron repaired and reinsulated if it is a critical motor that is not immediately replaceable. All the motors should have a core loss test after stripping as a proof that no damage was done in the stripping process.