Enter your search terms Web www.qtime.co.za Submit search form

Curriculum Vitae for J.H.M. de Swardt (Henk) - Published Papers

Life! / Lewe!Education and training Full time work experience Published papers Part time work experience Vacation employment Leadership abilities Computer Literacy Contact me

For the first time the papers I wrote are now also available in HTML format

 

Critical Speed on an Electric Motor Explained.

  

By Henk de Swardt

 Article Published in Vector[1] of February, March and April 2007 and published in Energizer[2] of March 2007:

 

Critical Speed on an electric motor explained.

By Henk de Swardt, Engineering Director, Marthinusen & Coutts

 

Introductions

 

One of our valued customers had major problems on a set of motors. These locally manufactured motors (> 1000 kW, 2 pole), driving locally manufactured water pumps, had mean time between failures of less than twelve months. All of the repair companies have had a go at trying to solve this ongoing problem with these seven motors.

 

The situation had gotten so bad that the repair companies struggled to repair a motor quick enough before the next motor failed. This was in spite of the fact that the motors were running on a rotation basis.

 

Although the exact content of the report presented to the customer to solve the problem permanently is of a confidential nature, the exhaustive analysis and research done is published in this paper to assist the industry as a whole to better understand critical speed problems on electric motors.

 

 

Existing Motor: Mean Time between Failures and Total Cost of Ownership

 

The existing motors have been running in the mine for more than ten years. The information we received is that these motors had vibration problems from the date of installation.

 

The failure rate on these motors was such that each time a motor returns from a repair company, another motor is again due for repair.

 

 

Site investigation:

 

We visited the site and found some very disturbing factors that influence the performance of these motors.

 

There is a definite ground fault line between pump number three and motor number four. A vibration level of 4.5 mm/s RMS was measured on the ground, about 1.5 meters from the actual motor base while motor four was running.

 

This ground fault results in a severe “looseness” condition that is noticeable on all of the other motors as well, even when those motors are not running, and with only motor number four running.

 

The run-up times of the motors are long for this type of application, being about 3 seconds. This is in line with the modelled design that showed a very low starting torque.

 

The run-down times are also long, but this is typical for motors fitted with white metal sleeve bearings.

 

Vibrations are transferred between all of the motor and pumps primarily through the ground, but also through the piping.

 

A major cause of induced vibration is definitely due to the starting and stopping of the motors. During these states, the motor’s vibrations levels more than tripled while the motor is in the vicinity of its critical speed. These very high vibration levels are transferred between the motors. The motors that are actually damaged the most are the motors that were not running when another motor was stopped or started.

 

The largest culprit in the transferring of this critical speed related frequencies, is the ground fault next to motor number four.

 

 

Critical speed

 

A basic calculation of this motor’s critical speed gives a value of 2165.85 rpm. This is totally inconsistent with the values seen on no-load test.

 

In order to understand the failures of these motors, we need to first explain critical speed and some related concepts.

 

Critical speed is the resonance frequency at which a shaft will resonate if it is excited by a force acting on the shaft.

 

The forces acting on a shaft are very complex, but some of the more obvious ones are:

·        Unbalanced magnetic pull due to finite number of stator and rotor slots.

·        Centrifugal force due to rotation of shaft.

·        Force due to the torque transmitted to the load.

·        Accelerating force due to stator’s rotating flux.

·        Force due to mechanical unbalance.

 

These forces do not all act at the running frequency. For instance: the rotor and stator bar pass frequencies are much higher than the running frequency.

 

An extremely important aspect is that any unbalance on the rotor will always excite at the current running frequency. Thus, is it of the utmost importance that the rotor’s balance must be done to a very high standard and by well experienced personnel, otherwise they can by means of an incorrect balancing method introduce massive vibration problems at the critical speed of the rotor, even though at a certain speed (even at synchronous speed, which is 3000 rpm. in this case) the balance might be almost perfect.

 

The rotor can be modelled as a damped harmonic oscillation.

 

At the critical speed the rotor resonates at the natural frequency, and thereby acts as an amplifier for any external force. The amplification factor is known as the “Q”-factor. The Q-factor is an indication of the amplitude as well as the band width of the amplifications. This will be explained further hereunder.

 

In order to get a better understanding of the critical speed concept, we can model the motor shaft as an RLC circuit, i.e. Resistor, Inductor and Capacitor, supplied by a variable frequency current source. Let us first consider the case of a rigid bearing mounting arrangement, as shown in this figure 1:

 

Figure 1: Rigid bearing mounting arrangement.

 

A simplified circuit diagram model of this can be as shown in figure 2:

 

Figure 2: Equivalent circuit diagram model of rigid bearing mounting arrangement.

 

If we consider the variable frequency voltage source to be only as a result of the rotating force on the shaft, we can thus quite easily draw a frequency response graph. In this, the current flowing through all the components will be dependant on the circuit’s impedance and thus on the values of L, C, and r.

 

The inductance is a model of the rotating mass. Changing the mass of the assembly will affect this value.

 

The capacitance is a model of the reciprocal of the stiffness of the system.

 

The resistance is a model of the dampening of the system, i.e. the rigidity of the system.

 

Let us for this example pick the following characteristics:

 

r = 10 [W]

L = 300 [mH]

C = 65 [mF]

V = 1000 [V]

f = 0 to 100 [Hz] where 50 Hz = the motor’s running speed = 3000 rpm

 [rad/s]

Equation 1: Angular velocity equation.

 

And:

Total circuit impedance:

 

Equation 2: Total circuit impedance equation.

 

Inductive Reactance:

Equation 3: Inductive Reactance equation.

 

Capacitive Reactance:

Equation 4: Capacitive Reactance equation.

 

We can plot the frequency response for the current as shown in figure 3:

 

Figure 3: Frequency response graph for rigid bearing mounting arrangement.

 

We can clearly see the resonance where the current (vibration velocity) is at its maximum. At this point the inductance and capacitance components cancel each other. The resonance frequency can easily be calculated as shown in equation 5:

 

Equation 5: Calculation of resonance frequency for rigid bearing mounting arrangement.

 

This 36 Hz we can now be equated to 2165 rpm, which as planned, works out to be the motor’s theoretically calculated critical speed.

 

We have however noticed that with these motors on test, the actual critical speed is much higher, in the region of 2850 rpm.

 

Remembering that this specific model explained above, is for a rigid bearing mounting arrangement, and that this motor actually uses white metal sleeve bearings, which are not a rigid, but a flexible mounting, as shown in figure 4, we need to change our diagram as follows:

 

Figure 4: Flexible bearing mounting arrangement.

 

Empirical data shows that the stiffness of a sleeve bearing arrangement compared to the stiffness of ball & roller bearing arrangement can be lower by as much as a factor of 1000!

 

As noted above, the stiffness if the reciprocal of the capacitance in our circuit model, as shown in equation 6:

 

Equation 6: Relationship between stiffness and capacitance in the circuit model.

 

If the stiffness thus reduces, then the capacitance will increase.

 

Our circuit model should now change as shown in figure 5:

 

Figure 5: Equivalent circuit diagram model of flexible bearing mounting arrangement.

 

Let us change the values as follows:

 

r = 10 [W]

L = 300 [mH]

C = 65 [mF]

C2 = 88 [mF]

V = 1000 [V]

 

We can now re-calculate the system impedance by using equation 7:

 

Equation 7: Combined system impedance equation.

 

Total capacitance:

Equation 8: Combined system capacitance equation.

 

We can plot the frequency response graph again, as shown in figure 6:

 

Figure 6: Frequency response graph of flexible bearing mounting arrangement.

 

We can clearly see that the resonance where the current (vibration velocity) is at it’s maximum. At this point the inductance and capacitance components cancel each other. The resonance frequency can easily be calculated as shown in equation 9:

 

Equation 9: Calculation of resonance frequency for flexible bearing mounting arrangement.

 

This 47.523 Hz we can now be equated to 2851 rpm, which as planned, works out to be the critical speed that we’ve seen with the motor running during a no-load test.

 

We conducted an over-speed test on the motor, and noticed that the motor’s vibration levels do not significantly decrease with an increase of the motor’s running speed.

 

This indicates to a low Quality Factor, or Q-factor. The Q-factor gives an indication of the width of the peak of the frequency response curve graph.

 

The Q-factor is calculated using equation:

 

Equation 10: Quality factor equation.

 

This gives a Q-factor of 8.95.

 

If we change our starting values in order to reduce the Q-factor:

 

r = 10 [W]

L = 3 [mH]

CT = 3742 [mF]

V = 1000 [V]

 

We get the following frequency response graph as shown in figure 7:

 

Figure 7: Frequency response graph of flexible bearing mounting arrangement with lower Q-factor.

 

We can see that the peak vibration level did not change. If we recalculate the Q-factor, we get a new value of 0.0895.

 

It is very clear that in these examples I manipulated the input values in order to get the critical speed results similar to that of the actual motor. This was not at all an attempt to fully model these motors, but as a starting point to explain the concept of critical speed. Let us now explore these concepts further.

 

Let us first define some more concepts:

 

·        Sub-synchronous motor:

 

A sub-synchronous motor is a motor with a critical speed lower than the running speed of the motor, and thus by definition, the motor would run through it’s critical speed while running-up and while coasting down to rest.

 

 

·        Super-synchronous motor:

 

A super-synchronous motor has its critical speed higher than the motor’s running speed. The motor thus DOES NOT pass through its critical speed while running-up and while coasting down to rest at all.

 

 

·        Shaft stiffness:

 

The shaft stiffness is a very important issue regarding the critical speed. The most important items that have an influence on the shaft stiffness are:

·        Mass of complete rotor assembly.

·        Shaft Material.

·        Diameter.

·        Distance between bearing centres.

 

All of these are interlinked.

 

Let’s look at the influence of each of these items on the electrical equivalent circuit, which would then help us to see the influence on the critical speed.

 

 

a.      Mass:

 

This is modelled as the inductance (L) in the equivalent circuit. If we increase the mass of the shaft, for instance by adding a fly-wheel to it, or change the diameter or material of the shaft, the inductance in the equivalent circuit would increase. This increase in the inductance would decrease the motor’s critical speed. This will also improve the Quality factor.

 

A further consequence of the change in the mass is additional torsional stresses on the shaft during start-up, due to the increased moment on inertia. A higher moment of inertia will also increase the motor’s run-up and coast down times. If a flywheel is added, care must be taken not to adversely affect the cooling circuit, i.e. air flow inside the motor and thereby increasing the motor’s temperature rise. Additional weight also adds extra static loading to the bearings, which would increase their running temperature.

 

 

b.             Shaft material:

 

These motors use a flexible shaft arrangement. A flexible shaft is designed to be able to bend (flex) during the start-up and run-down cycles. This is done by decreasing the shaft diameter and by using a low carbon shaft material. The shaft material used in these motors is most likely to be EN 3 or EN 8. These shaft materials are used because of their low cost as well as their ability to be used as flexible shafts.

 

On high speed motors, the torque is much lower for a specific power rating, than for a lower speed motor, for instance, for this motor, the maximum synchronous torque is calculated as shown in equation 1:

 

Equation 11: Calculated synchronous torque for 1100 kW, 2 pole motor.

 

For the same power rating, but a 6 pole motor, with a synchronous speed of 1000 rpm, the maximum synchronous torque will be as calculated in equation 12:

 

Equation 12: Calculated synchronous torque for 1100 kW, 6 pole motor.

 

It should be clear that the 2 pole motor would thus require a shaft that is not nearly as strong as the 6 pole motor.

 

One would think that a higher strength shaft material could be of great benefit, because of the reduced flexibility and higher strength. A lower flexibility would decrease the critical speed and improve the Quality Factor. This is true, but unfortunately the practical benefit is not that much, since the flexibility change between rigid (ball and roller) bearings to flexible (sleeve) bearings is huge, the actual difference in flexibility (referred to as the shear modulus or lame constant) is not that much different for the higher strength shaft materials compared to the lower strength materials.

 

Table 1 lists typical Tensile and Yield strengths[3] for common shaft materials. 

 

Material		Composition						Tensile strength [N.mm -2 ]	Yield strength [N.mm -2 ]
Common	BS 970	C [%]	Si	Mn	Mo	Ni	Cr
EN 3	070M20	0.2		0.7				400	200
EN 8	080M40	0.4		0.8				550	280
EN 9	070M55	0.6	0.4	0.9				690	495
EN 19	709M40	0.44	0.44	1.1	0.35		1.2	690	495
EN 24	817M40	0.4	0.25	0.7	0.25	1.85	0.8	930	730
EN 32	080M15	0.15	0.25	0.8				450	330
EN 36	655M13	0.16	0.4	0.6		3.75	1	900	730

Table 1: Typical shaft material specifications.

 

 

c.              Shaft diameter:

 

Strange enough this specific motor’s shaft diameter under the core is 170 mm. Normally a shaft of this size and design would have a diameter of between 100 and 120 mm.

 

A larger diameter shaft has the advantage of improving the shaft strength and also increasing the shaft stiffness.

 

Changing the shaft diameter can have a large difference on the critical speed. Theoretical modelling shows the following possible changes to the critical speed for different shaft diameters as listed in table 2.

 

Table 2: Theoretical critical speed for different shaft diameters.

 

Please note that this is still the theoretical critical speed calculation, and that the measured critical speed is around 2851 rpm., with the current shaft design.

 

It appears to be a logical deduction that a large reduction in the shaft diameter would significantly reduce the rotor’s critical speed and that this would solve this motor’s critical speed problem.

 

Care must however be taken with reducing the shaft diameter, since the OEM should have designed the motor with a typical shaft life of at least 5000 starts. These motors are more than ten years old. If we assume that these motors are started on average twice per day. This would give the number of starts to date to be about 7300.

 

Furthermore, the torsional stiffness of the shaft is proportional to the cube of the apparent diameter of the shaft. A 10% reduction in the shafts diameter, could reduce the shaft’s torsional stiffness thus by as much as 46%, and thereby greatly reducing the value of “r” in the equivalent circuit, as well as the Quality Factor.

 

The end result of decreasing the shaft diameter would thus probably be a lower critical speed, but a much higher vibration level at this lower critical speed, which would most likely result in a shaft failure. It is thus clear that degreasing the shaft diameter is not necessarily a good idea.

 

An interesting possibility could be to perform a “turret drill” on the shaft. This process involves drilling a hole through the whole length of the shaft. This will increase the shaft’s strength as well as a large increase in torsional stiffness. The problem with this is that the surface finish on this internal hole is absolutely crucial. A mirror polished finish would probably be necessary (referred to as N8). We do not believe that this option would be easy to achieve, since the skill and tooling involved to achieve this quality hole is mostly confined to military applications.

 

 

d.             Distance between bearing centres:

 

The distance between the bearing centres of a motor is predominantly determined by the physical construction of the frame and the bearing arrangement. For instance: A motor with only one internal fan would have shorter frame and thus also a shorter distance between bearing centres.

 

Again changing the distance between bearing centres will change the critical speed due to more than one action working. The most obvious two are: Firstly the weight of the shaft, and secondly the stiffness of the shaft. For this aspect we also carefully need to consider both a motor repair was well as a new motor.

 

The distance between the bearing centres can be changed on a motor repair by changing the frame length – which is normally a very difficult task, or by changing the bearing arrangement. In this specific motor’s case, this can be done by changing the white metal bearings from flange mounted (i.e. the bearing sits on the outside of the casing), to centre mounted (i.e. the bearings sit in the centre of the end shield, overhanging both inside the motor as well as outside the motor) bearings.

 

The centre mounted bearing (also referred to mid-counted bearing) arrangement is a very effective method of reducing the distance between the bearing centres. This modification would require the following modifications: New bearing housings, modification to end shield and most likely a new rotor shaft. We will discuss shortly the influence on the critical speed with such a modification.

 

To increase the distance between bearing centres is a lot motor difficult on this frame construction. The end shield and/or bearing housing will literally have to be extended outwards. It would probably also to be necessary to replace the shaft.

 

Let us now consider the influence on the critical speed when the distance between bearing centres are either increased or decreased:

 

a.                  Distance between bearing centres: Increased

 

As mentioned previously, a change in the distance between the bearing centres will influence the shaft weight and stiffness.

 

                                                  i.            Influence of rotor mass

 

Clearly, an increase in the distance between the bearing centres will increase the rotor weight. Paraphrasing from the above explanation:

 

“The mass is modelled as the inductance (L) in the equivalent circuit. If we increase the mass of the shaft, by increasing the shaft length, the inductance in the equivalent circuit would increase. This increase in the inductance would decrease the motor’s critical speed. This will also improve the Quality factor.

 

A further consequence because of the increase in the mass is additional torsional stresses on the shaft during start-up, due to the increased moment on inertia. A higher moment of inertia will also increase the motor’s run-up and coast down times. Additional weight also adds extra static loading to the bearings, which would increase their running temperature.”

 

                                                ii.            Influence of rotor stiffness

 

An increase in the distance between the bearing centres will decrease the rotor stiffness. Decreasing the stiffness will reduce the “r” value in the equivalent circuit, and would thus increase the critical speed. Furthermore the quality factor will be decrease, and thus the maximum vibration level will increase.

 

 

b.                  Distance between bearing centres: Decreased

 

As mentioned previously, a change in the distance between the bearing centres will influence the shaft weight and stiffness.

 

                                                  i.            Influence of rotor mass

 

Clearly, a decrease in the distance between the bearing centres will decrease the rotor weight. Paraphrasing from the above explanation:

 

“The mass is modelled as the inductance (L) in the equivalent circuit. If we decrease the mass of the shaft, by decreasing the shaft length, the inductance in the equivalent circuit would decrease. This decrease in the inductance would decrease the motor’s critical speed. This will also reduce the Quality factor.”

 

                                                ii.            Influence of rotor stiffness

 

A decrease in the distance between the bearing centres will increase the rotor stiffness. Increasing the stiffness will increase the “r” value in the equivalent circuit, and would thus decrease the critical speed. Furthermore the quality factor will be increased, and thus decreasing the maximum vibration level.

 

 

It is not really clear which situation would be more desirable in this instance: Decreasing or increasing the distance between bearing centres. We will adapt the theoretical model and calculate the critical speed for different distances between the bearings, taking into consideration that the distance can only be decreased and increased by a limited distance. The theoretical critical speeds are listed in table 3.

 

 

Table 3: Theoretical critical speed for different distanced between bearing centres.

 

Please note that this is still the theoretical critical speed calculation, and that the measured critical speed is around 2851 rpm., with the current shaft design.

 

It should now be clear that compared to, for instance a change to the actual shaft diameter, the influence on the critical speed by means of a change in the distance between the bearings centres is significantly smaller.

 

A reduction in the stiffness could however result in a very large reduction in the “r” value in the equivalent circuit, and could thus greatly increase the vibration levels at this critical speed.

 

 

·        Bearings:

 

It is clear from the preceding discussion that a major problem is the white metal sleeve bearings. Due to the inherent lack of stiffness, the use of these bearings increases the motor’s critical speed tremendously, and also reduces the system’s stiffness, and thereby reducing the “r” value in the equivalent circuit, which amplifies the maximum vibration level. Naturally one would then ask why these motors are not fitted with ball and roller bearings instead of the white metal sleeve bearings.

 

With these motors being sub-synchronous motors, and the associated high vibration levels when the rotor passes through the critical frequency, ball and roller bearings would unfortunately fail very quickly. The additional forces on these rigid bearings is directly proportional to the Quality factor, and thus could be a factor of ten or higher!

 

Ball and roller bearings would just not be able to carry these huge overloads, and will fail quickly!

 

Furthermore, on such high shaft rotational speeds, correctly sized ball and roller bearings will have reduced re-greasing lubrication intervals. The Anglo specification (538/11 rev 3) requires a lubrication interval of at least 4000 hours. This problem can however be overcome by means of either oil-lubricating the bearings or by using special long life grease and/or by providing automatic grease lubrication devices.

 

 

·        Frame stiffness:

 

Although we have not touched on this until now, the stiffness of the frame (and thus also the end shields) most likely also have a major effect on the vibration levels and critical speed of these motors.

 

On these motors the end shields are very flexible, and this would thus greatly increase the vibration at the critical speed.

 

All of these explanations aim to build a platform to discuss possible long term solutions to these problem motors.

 

 

Root cause

 

Before we examine specific possibilities to solve these motors’ problems, let us first summarise the main root causes of failures:

 

·        Critical speed:

·        The motor’s critical speed is too close to the running speed.

·        The motor’s critical speed is below the running speed.

 

·        Stiffness and strength of components:

·        The frame and end shield stiffness is poor.

·        The bearing stiffness is poor.

 

·        Run-up and coast down times:

·        The motor take much too long to run up to full speed.

·        The coast down time is much too long.

 

·        Vibration transmission between motors:

·        High vibrations are transmitted between the motors because of the ground fault.

 

 

Solutions

 

Let us now evaluate possible solutions for these problems.

 

Firstly we must certainly consider the repair option.

 

 

a.     Repair motors

 

As mentioned above, these motors have been giving problems for many years, and thus have been repaired many times.

 

From our experience on the repair of several of these motors, it is clear that they are neither identical nor standard. Over the years, various repair companies probably tried a lot of different solutions to these problems, without ever really understanding or solving the root causes of the failures. Nobody has found a long term solution.

 

We believe that through some finite element design analysis, mechanical design modelling and prototyping, through our own design and manufacturing experience of motors, there is a possibility to repair these motors to achieve long term reliability.

 

This would most likely involve the following:

 

a.      Replacement of shafts:

 

As mentioned previously, the shafts are probably nearing their designed life and could start failing soon.

 

The shaft would probably be redesigned by using a higher grade of steel (EN 19 or EN 24), which could include tempering (condition T) as well as turret drilled. This would include extensive finite element analysis to ensure a correct shaft design.

 

 

b.     Changes to end shields:

 

The inherent flexibility of the end shields can be corrected by manufacturing new, high strength end shields. This can be achieved by using boiler plate (300WA) in stead of normal mild steel, and also adding many gussets, increasing number of fixing bolts, etc.

 

 

c.      Changing bearing arrangement:

 

By using a smaller diameter sleeve bearing (90 mm instead of the current 100 mm.), the flexibility of the bearing can be decreased. This will however require a good cooling system and forced lubrication system for the bearing oil, because a smaller bearing would probably run at a higher temperature.

 

 

d.     Reducing motor run-up time:

 

This can be achieved in a number of ways:

 

·        Rewinding the stator winding in order to achieve a higher starting torque. This will result in higher losses, higher starting current, lower power factor, lower efficiency and higher temperature rise on the stator winding.

 

·        Change the rotor bar design to a double cage design. This might also require a stator rewind. The rotor laminations would have to be replaced. This is a very elegant and good solution. The result would have little effect on the losses of the motor, will reduce the starting current and have little effect on the power factor, efficiency and running temperature of the stator. Apart from ourselves  there are few repair companies in South Africa who would be able to design such a double cage rotor.

 

 

e.      Reducing motor’s run-down time:

 

This is a common solution between all the options, and will be discussed at the end of this main section.

 

 

f.       Reducing the transmission of vibrations between the motors:

 

This is a common solution between all the options, and will be discussed at the end of this main section.

 

 

Although the solution of repairing these motors, which would involve extensive design changes, would solve the problems and can provide a long term solution, it would be time consuming and probably very expensive. We do not believe that this is a viable option.

 

 

b.     New Motors: Super-Synchronous Motors

 

The best solution would be to replace all of these motors with new, super-synchronous motors.

 

The influence of the run-up and run-down times and critical speeds would thus be irrelevant.

 

Unfortunately it would be extremely difficult to persuade a motor manufacturer to build such a motor, since this is not the standard method of manufacturing these types of motors.

 

A prototype would have to be designed and built, but this is entirely achievable and would provide a “permanent” solution to this problem.

 

Also see the additional specification requirements listed hereunder for new motors.

 

 

c.      New Motors: Sub-Synchronous Motors

 

Extreme care must be taken to ensure that the new motors do not have these same inherent design mechanisms which would result in a similar unreliable situation a few years down the line.

 

Also see the additional specification requirements listed hereunder for new motors.

 

 

In order to ensure a long term solution, the following additional requirements would have to be stipulated to the manufacturer of the new motors:

 

 

Additional New Motor Specifications:

 

a.     Critical Speed:

 

Logically it is very important to define very clearly the requirements for the critical speed of the new motor.

 

The critical speed of the motor must be at least 30 % below the full load running speed of the motor. This is in line with the current Anglo Gold New Motor Specification (AAC 538/11 revision 3).

 

 

b.     Starting and Pull Out Torque:

 

The current motor’s run-up time is VERY long (relatively speaking). The run-up time can be reduced by specifying a starting torque of at least 85% of full load torque and a pull out torque of at least 210 % of full load torque, at a 20% drop in supply voltage. This is in line with the current Anglo Gold New Motor Specification (AAC 538/11 revision 3).

 

During the design of a motor, two theoretical torque values are obtained for both the starting and pull out torque. The model of the original motor gave a theoretical value for the starting torque of 35 % and 44 % as well as pull out torques of 221 % and 246 %. These values are referred to as the un-saturated and saturated values. The supplier is free to quote any of these two values, since very few customer specifications (including SANS 34, IEC 60034, SP46-11 and AAC 538/11) address this fact.

 

 

c.      Starting Current:

 

One way of increasing the motor’s starting torque is by increasing the starting current. It is very important to specify that the new motor must have a starting current not higher than 600 % of the full load current. The current Anglo Gold new motor specification (AAC 538/11 revision 3) does specify this 600 % value, BUT does allow a 20 % tolerance. The supplier can thus design the motor with a starting current of 720 % and still be within the specification!

 

The result of a higher starting current would be a larger voltage drop when the motor is started, which would reduce the starting torque by the square of the voltage reduction! Clearly this will defeat the whole object of this specification for a higher starting torque. The supplier MUST guarantee his quoted starting current value and prove these values on test prior to the motor leaving the factory.

 

 

d.     Vibration Level:

 

The supplier must ensure that the motor supplied has a very low vibration level. This is to further minimise the influence between the motors. A good value is 1.2 mm/s RMS vibrations in any direction on both bearings. (This is in line with the Eskom, Goldfields and Sasol specification.)

 

 

e.     Bearings:

 

In order ensure that smaller bearing sizes are used, it must be compulsory for the motors to be supplied with forced lubrication bearings. It is preferable that the forced cooling system should also be supplied by the motor manufacturer, to ensure a turn-key solution.

 

It must be made clear that the bearing temperatures must not exceed the AAC 538/11 specification, to ensure long term reliability.

 

 

f.       Reducing motor’s run-down time:

 

This is a common solution between all the options, and will be discussed at the end of this main section.

 

 

g.     Reducing the transmission of vibrations between the motors:

 

This is a common solution between all the options, and will be discussed at the end of this main section.

 

 

h.     Interchangeably:

 

It is important to thoroughly specify to the manufacturer that a motor is required that must be fully mechanically and electrically interchangeable.

 

The interchangeability should include at the very least the following:

·        Foot fixing

·        Shaft centre heights

·        Terminal box position

·        Terminal cables position and fixing arrangement

·        Auxiliary box position

·        Auxiliary terminals position and fixing arrangement

·        Overall dimensional limitations

·        Shaft extension size, length and key arrangement.

 

 

As promised, we will address the last two remaining failure mechanisms which must be solved for all three repair or replace options.

 

 

i.       Reducing motor’s run-down time:

 

This would involve some instrumentation. The run-down time of the motor can be reduced by closing the water value only once the motor speed has dropped below a specific value (say 200 rpm). The back pressure of the water would thus act as a brake to reduce the motor’s speed.

 

A fail safe must be installed to ensure that the valve does close, even if the speed sensor on the motor is faulty. This can be done by means of a timer.

 

 

j.       Reducing the transmission of vibrations between the motors:

 

Practically we know that the ground fault cannot be economically solved.

 

Step one would be not to run motor number four unless it is entirely necessary! This will greatly reduce the vibrations transmitted to the other motors.

 

Unfortunately all other additional solutions would involve a huge amount of money.

 

 

Conclusion

 

When taking into consideration the cost of all the repairs, production losses and time to remove and install these motors it is clear that replacing these problem motors with new motors does make economic as well as technical sense, as long as a stringent technical specification is applied in the procurement process in order to provide a long term solution.

 

 

We are proud to engineer quality solutions for our valued customers.