Permanent Magnet Motor: Maintenance (part 2)

(cont. from part 1)

6. Condition monitoring

As an electromechanical drive becomes more complex, its cost of maintenance rises. The condition monitoring can predict when the electromechanical drive will break down and substantially reduce the cost of maintenance including repairing. Electric motors are the most vulnerable components of electrical drive systems (see Table 1) so their condition and early fault detecting is crucial.

Condition monitoring of electric motors should be done on-line with the aid of externally mounted sensors as, for example, current transformers, accelerometers, temperature sensors, search coils, etc. without any change of the motor construction, rearrangement of its parts or changing its rated parameters. The following methods can be employed:

• aural and visual monitoring

• measurements of operational variables

• current monitoring

• vibration monitoring

• axial flux sensing

Aural and visual monitoring require highly skilled engineering personnel. The failure is usually discovered when it is well advanced. Measurements of operational variables such as electric parameters, temperature, shaft eccentricity, etc., are simple and inexpensive but as in the case of aural and visual monitoring, there is a danger that the fault is detected too late.

The most reliable, rich in information, simple and sophisticated method is monitoring the input current, vibration or axial flux. A current transformer can be used as a current transducer. For vibration monitoring piezoelectric accelerometers or other transducers can be installed. For axial flux sensing a coil wound around the motor shaft can be installed. Because it is often inconvenient to wind coils around the shaft of a motor that is in service, a printed circuit split coil has been proposed.

Monitoring techniques are based on time domain measurements and frequency domain measurements.

Time domain signals can detect, e.g., mechanical damage to bearings and gears. The time signal may be averaged over a large number of periods, synchronous with the motor speed, to allow synchronous averaging. Background noise and periodic events not synchronous with the motor speed are filtered out.

A periodic time domain waveform, if passed through a narrow band-pass filter with a controllable center frequency, is converted into frequency components which appear as output peaks when the filter pass band matches the frequency components. This technique is adopted in high frequency spectral analyzers. An alternative approach is to sample the time domain waveform at discrete intervals and to perform a discrete Fourier transform (DFT) on the sample data values to determine the resultant frequency spectra. The time necessary to compute the DFT of the time domain signal can be reduced by using a fast Fourier transform (FFT) calculation process which rearranges and minimizes the computational requirements of the DFT process. Modern monitoring techniques commonly use FFT.

Vibration or current spectra are often unique to a particular series of motors or even particular motors. When a motor is commissioned or when it is in a healthy state a reference spectrum is monitored which can later be compared with spectra taken in successive time intervals. This allows some conclusion and progressive motor condition to be formulated.

In analysis of frequency spectra of the input current usually the amplitude of sidebands are compared to the amplitude of the line frequency. The following problems in PM brushless motors can be detected on the basis of time domain input current measurement:

(a) unbalanced magnetic pull and air gap irregularities (b) rotor mechanical unbalance (c) bent shaft (d) oval stator, rotor or bearings.

Common faults which can be detected by vibration monitoring of electric motors are:

(a) rotor mechanical unbalance or eccentricity, characterized by sinusoidal vibration at a frequency of once per revolution (number of revolutions per second) (b) defects of bearings, which result in frequencies depending on the defect, bearing geometry and speed, usually 200 to 500 Hz (c) oil whirl in bearings, characterized typically by frequencies of 0.43 to 0.48 number of revolutions per second (d) rubbing parts, characterized by vibration frequency equal to or a multiple of the number of revolutions per second (e) shaft misalignment, usually characterized by a frequency of twice per revolution (f) mechanical looseness in either motor mounts or bearing end bells, which results in directional vibration with a large number of harmonics (g) gear problems characterized by frequencies of the number of teeth per revolution, usually modulated by speed (h) resonance (natural frequencies of shaft, machine housing or attached structures are excited by speed or speed harmonics), which results in sharp drop of vibration amplitudes with small change in speed (i) thermal unbalance, which results in a slow change in vibration amplitude as the motor heats up (j) loose stator laminations, which result in vibration frequencies equal to double the line frequency with frequency sidebands approximately equal to 1000 Hz (k) unbalance line voltage, characterized by vibration frequency equal to double the line frequency.

By sensing the axial flux, many abnormal operating conditions can be identified, e.g.:

(a) unbalanced supply, characterized by an increase in certain even harmonics of flux spectrum proportional to the degree of unbalance (b) stator winding interturn short circuits, characterized by a decrease in certain higher harmonics and sub-harmonics of flux spectrum (c) rotor eccentricity characterized by an increase in frequency of flux spectrum equal to the line frequency and its second harmonic.

A unified analysis of the various parameter estimators and condition monitoring methods and diagnosis of electrical machines can be found in dedicated literature.

7. Protection

The motor protection depends principally on the motor importance which is a function of motor size and type of service. The bigger the motor the more expensive the electromechanical drive system is and all necessary measures should be taken to protect the motor against damage. The main function of protection is the detection of a fault condition, and through the opening of appropriate contactors or circuit breakers, the disconnection of the faulty item from the plant. The potential hazards considered in electric motor protection are:

(a) Phase and ground faults (short circuits between phases or phase and earth) (b)Thermal damage from:

- overload (excessive mechanical load)

- locked rotor

(c) Abnormal conditions:

- unbalanced operation

- undervoltage and overvoltage

- reversed phases

- switching on the voltage while motor is still running

- unusual environmental conditions (too high or too low temperature, pressure, humidity)

- incomplete feeding, e.g., rupturing of a fuse in one phase

(d) Loss of excitation (in PM motors this means the total demagnetization of magnets) (e) Operation out of synchronism (for synchronous motors only) (f) Synchronizing out of phase (for synchronous motors only).

Most of the undesired effects cause excessive temperatures of the motor parts, in particular windings. There is an old rule of thumb that says each 100C increase in the operating temperature of the winding results in a 50% loss of insulation life.

Protective devices applied for one hazard may operate for other, e.g., over load relay can also protect against phase faults. Protection can be built in the motor controller or installed directly on the motor. Motors rated up to 600 V are usually switched by contactors or solid state devices and protected by fuses or low-voltage circuit breakers equipped with magnetic trips. Motors rated from 600 to 4800 V are switched by power circuit breakers or contactors. Motors rated from 2400 V to 13, 000 V are switched by power circuit breakers.

The simplest one-time protective devices are fuses. As a result of excessive current, the fusible element melts, opens the circuit and disconnects the motor from the power supply.

The thermal relay of a "replica" type is the thermo-electromechanical apparatus which simulates, as closely as possible, the changing thermal conditions in the motor allowing the motor to operate up to the point beyond which damage would probably be caused. This relay consists of three single-phase units, each unit comprising a heater and an associated bimetal spiral element.

The bimetal elements respond to a rise in temperature of the heaters which in turn produce movement of the contact assembly. The single-phasing protection is usually incorporated in the three-phase thermal simulation relay.

The short circuit protection against short circuits in the motor winding or terminal leads is often built in the thermal relay as separate overcurrent or earth-faults elements or both.

The stalling relay is used in conjunction with the thermal overload and single-phasing relay. It consists of a control contactor and a thermal overload unit fitted in the same case.

FIG. 5. Block diagram of a typical electronic overload relay.

FIG. 6. Integrated circuit, which includes inverter for driving a PM brushless motor, drive circuits, power source circuit, circuits for controlling the speed and protection circuit for protecting the inverter from excess current according to U.S. Patent 5327064.

An alternative to the thermal relay is an electronic overload relay ( FIG. 5) using solid state devices instead of bimetallic elements.

For small and medium power motors with rated currents up to 25 A (or Pout = 15 kW), simpler and less expensive thermal trips and electromagnetic trips instead of thermal overload relays are technically and economically justified. In a thermal trip the motor current passing directly through a bimetal strip or the heater operating with a bimetal element is fed from current trans formers connected to the motor power circuit. Bimetal elements operate the mechanical trip to open the motor contactor under an overload condition. The electromagnetic trip consists of a series-wound coil surrounding a vertical ferromagnetic plunger and an associated time-lag, i.e., oil or silicone fluid filled dashpot or air vane. The adjustable overload current lifts the plunger which opens the motor contactor. The electromagnetic trips are relatively insensitive to small overloads.

Thermistors bonded to the enamel armature conductors during manufacture are commonly referred to as motor overheat protection. Thermistor connections are brought out to an electronic control unit and interposing re lay mounted separately for small motors and usually built into terminal boxes of motors rated above 7.5 kW. The relay is activated when the thermistors indicate the winding temperature and, indirectly, the phase currents, exceed their permissible values.

Undervoltage protection is necessary to ensure that the motor contactors or circuit breakers are tripped on a complete loss of supply, so that when the supply is restored it is not overloaded by the simultaneous starting of all the motors. Undervoltage release coils operating direct on the contactor or circuit breaker, relays or contactors with electrically held-in coils are usually used.

The state of the art in the motor protection are microprocessor protection relays. These advanced technology multifunction relays are programmed to provide the following functions:

• thermal overload protection with adjustable current/time curves

• overload prior alarm through separate output relay

• locked rotor and stall protection

• high-set overcurrent protection

• zero phase sequence or earth-fault protection

• negative phase sequence or phase unbalance protection

• undercurrent protection

• continuous self-supervision.

ICs for driving PM brushless motors have built-in protective features such as, for example, current limit circuits, thermal shutdown, under voltage lockout, etc. FIG. 6 shows an IC for driving a variable speed PM brushless motor with start current limit and excess current protection circuits according to US Patent 5327064.

8. Electromagnetic and radio frequency interference

Electromagnetic compatibility (EMC) is defined as the ability of all types of equipment that emit high frequency signals, frequencies higher than the fundamental supply frequency, to operate in a manner that is mutually compatible.

Designers of electrical machines should be aware of the EMC specifications which affect their overall design. All electrical motors can be a source of electromagnetic interference (EMI) and radio frequency interference (RFI). RFI is an electric disturbance making undesired audio or video effects in received signals caused by current interruption in electric circuits as, e.g., sparking between a brush and commutator.

The main reasons for minimizing EMI are the high frequency noise con ducted into the mains supply will be injected into other equipment, which could affect their operation, i.e., noisy signals being fed into sensitive loads such as computers and communication equipment, and the interference radiated into the atmosphere by electric and magnetic fields can interfere with various communication equipments. Three-phase electrical motors often draw currents that have frequency components that are odd integer multiples (harmonics) of the fundamental supply frequency. These harmonic currents cause increased heating and lead to a shorter lifetime of appliances.

Conducted noise emissions are suppressed using filters, usually with a passive low pass filter designed to attenuate frequencies above 10 kHz. Under nonlinear loads, such as those associated with adjustable speed drives and electronic power supplies, significant power dissipation can occur within these filters. Shield radiated emissions can be suppressed by metallic shields and minimizing openings in the enclosures.

The filtering required to reduce high frequency emissions may be rather costly and reduce the efficiency of the electromechanical drive.

EMI regulations set the limits for conducted and radiated emissions for several classes of products. One of the most important international standards setting organizations for commercial EMC is CISPR, the International Special Committee on Radio Interference in IEC (International Electrotechnical Commission). The European community has developed a common set of EMC requirements largely based on CISPR standards. The Federal Communication Commission (FCC) sets the limits of radiated and conducted emissions in the U.S.A. An example of a guide for electrical and electronic engineers containing complete coverage of EMI filter design.

FIG. 7. RFI circuits and filters of brush motors: (a) asymmetrical circuit of high frequency current, (b) symmetrical circuit of high frequency current, (c), (d), (e) RFI filters.

8.1 Brush motors

Commutator (brush) motors are a source of serious EMI and RFI. The stronger the sparking the more intense the interferences. Wrong maintenance, dirty or worn commutator, wrong selection of brushes, unbalanced commutator, unbalanced rotor, etc., are the most serious reasons for EMI and RFI.

RFI causes clicks in radio reception being heard in the whole band of radio waves. TV reception is disturbed by change in brightness of screen and time base which in turn makes a vertical movement of image lines on the screen.

EMI and RFI are emitted directly from their source and connecting leads as well from the power network from which the motor is fed via converter. They are received by radio or TV antennas. In addition, RFI can be transmitted to radio or TV sets by electric installation, or even by water and gas pipes and any metal bars.

The interference current of high frequency flows through the electric net work feeding the motor, then through capacitance Ce1 between the feeding wire and the earth and capacitance Ce2 between the source of interference and the earth ( FIG. 7a). This high-frequency current, called asymmetrical high frequency current causes particularly strong interference. A grounded motor frame intensifies the level of RFI since it closes the circuit for high frequency current. Symmetrical high frequency currents are closed by capacitances Cp between wires ( FIG. 7b).

To eliminate RFI, filters consisting of RLC elements are used ( FIG. 7c,d,e). The resistance R is needed to damp high frequency oscillations.

The choke has a high inductance L for the high frequency current, since its reactance XL =2pfL increases as the frequency f and inductance L in creases. The capacitive reactance XC =1/(2pfC) is inversely proportional to the frequency and capacitance C. For high frequency currents the capacitive reactance is low.

Symmetrical inductances connected in series with the armature winding at both its terminals can damp oscillations more effectively than one inductance only. In series motors, field coils are used as RFI filter inductances. In the filter shown in FIG. 7c the high-frequency current circuit is closed by capacitors C. The capacitor C1 is a protective capacitor against electric shock.

If the motor frame is touched by a person who is in contact with the ground, the current flowing through the human body to the earth will be limited by capacitor C1, the capacity of which is low (about 0.005 µF) as compared with C (from 1 to 2 µF for DC motors).

RFI filters are effective if the connection leads between capacitors and brushes are as short as possible (less than 0.3 m). This results in minimization of interference emitted directly by the source. It is recommended that the RFI filters be built into motors. If this is not possible, all connection leads must be shielded. Resistances can effectively damp oscillations and improve the quality of RFI filter (FIG. 7d). A simple RFI RL filter for three-coil PM commutator motors for toys and home appliances is shown in FIG. 8.

FIG. 8. Elimination of RFI in a three-coil DC PM brush motor: 1-?-connected armature winding, 2 - armature stack, 3 - cylindrical PM, 4 - commutator segment, 5 - frame, 6 - resistor of RFI filter.

8.2 Electronically commutated brushless motors

Brushless DC motors use inverter-based solid state converters. The current and voltage waveforms from these converters are either sinusoidal or square wave. In both cases the converter generates the desired waveforms using PWM, switched at 8 to 20 kHz typically. When a voltage abruptly changes amplitude with respect to time, the derivative dv/dt changes produce unwanted harmonics. The nonlinear characteristics of solid-state devices worsen the situation. This accounts for the large impulse currents through the power leads, which are associated with EMI and significant voltage waveform distortion in the power system.

The reduction of electrical noise is usually done by ensuring proper grounding, avoiding extended cables from inverter to motor, twisting the cables and filtering the input power to the inverter drive.

To prevent radiated noise, the motor ground wire is required to be twisted or tightly bundled with the three line wires. The motor power wiring is to be kept as far away as possible from the rotor position signal wiring and any other light current wiring. A shielded cable is recommended for connection of the encoder or resolver with the inverter motion control section.

Placing filters on power lines to inverters not only suppresses harmonics leaving the drive, but also protects the drive from incoming high frequency signals. Fig .9 shows a three-phase low pass filter used for EMI/RFI filtering.

FIG. 9. Typical EMI filter for three-phase input.

These line filters are bulky and increase the cost of the drive substantially.

FIG. 10. Power circuit of a three-phase modified converter with EMI suppression components.

Fig. 10 shows a low cost alternative: an inverter with EMI suppression components. These include:

(a) grounding capacitance C1 from both sides of the DC link to the heat sink close to the switching devices which provides a physically short path for RF ground currents flowing from the switching device to the motor;

(b) Line capacitance C2 across the DC link, close to the switching devices which provides a low impedance for differential mode RF current flowing from switching devices such as reverse recovery current of the diodes as well as from the cable-motor load;

(c) line capacitance C3 across the AC power input terminals close to the diode rectifiers which serves as another shunt circuit in combination with the DC line capacitance for differential mode noise compensation, particularly for the noise caused by the diodes of rectifier; (d) a common mode line inductance L1 inserted in each phase of the AC input power circuit of the rectifier provide a high impedance for the RF currents to the power mains; (e) a common mode inductance L2 inserted in each phase of the AC output power circuit of the inverter reduce the time derivative of output mode voltages imposed on the motor, but do not affect the line to line voltages.

9. Lubrication

9.1 Bearings

In PM machines usually rolling bearings and porous metal bearings are used.

In high speed PM machines magnetic, air or foil bearings are used, which will not be discussed in this section.

The stress levels in rolling bearings limit the choice of materials to those with a high yield and high creep strength. Steels have gained the widest acceptance as rolling contact materials as they represent the best compromise among the requirements and also because of economic considerations. Steels with the addition of C, Si, Mn and Cr are the most popular. To increase hardenability and operating temperature tungsten (W), vanadium (V), molybdenum (Mo) and nickel (Ni) are added. Basic methods of mounting rolling bearings for horizontal shafts are shown in FIG. 11.

FIG. 11. Basic methods of installation of rolling bearing for horizontal shaft: (a) two deep groove radial ball bearings, (b) one ball bearing with one cylindrical roller bearing.

Porous metal bearings are used in small or large electric motors. The graphited tin bronze (Cu-Sn-graphite) is a general purpose alloy and gives a good balance between strength, wear resistance, conformability and ease of manufacture. Where rusting is not a problem, less expensive and stronger iron based alloys can be used. Assemblies of self-aligning porous metal bearings with provision for additional lubrication are shown in FIG. 12. In most electric motor bearings the lubricant material is oil or grease.

FIG. 12. Installation of self-aligning porous metal bearings for small, horizontal shaft motors: 1 - bearing, 2 - oil soaked felt pad, 3 - key hole, 4 - oil hole, 5 - slot to take key, 6 - end cap (may be filled with grease).

9.2 Lubrication of rolling bearings

Grease lubrication. Grease lubrication is generally used when rolling bearings operate at normal speeds, loads and temperatures. The bearings and housings for normal applications should be filled with grease up to 30 to 50% of the free space. Too much grease will result in overheating. The consistency, rust-inhibiting property and temperature range must be carefully considered when selecting a grease. The grease re-lubrication period is the same as the service life of the grease and can be estimated from the formula:

tg = kb _14 × 10^6 n vdb - 4db _ h (eq. 50)

where kb is a factor depending on the type of bearings, n is speed in rpm and db is bearing bore diameter in mm. For spherical roller bearing and tapered roller bearings kb = 1, for cylindrical and needle roller bearings kb = 5 and for radial ball bearings kb = 10. The amount of grease required for relubrication is:

mg =0.005Dbwb g (eq. 51)

where Db is the outer bearing diameter and wb is its width, both in millimeters.

The advantages of using a grease lubricant are:

• it is convenient to apply and retain the lubricant within the bearing housing

• it stays to cover and protect the highly polished surfaces even when the bearing is at rest

• it helps to form a very effective closure between shaft and housing thus preventing entry of foreign matter

• it offers freedom from lubricant contamination of the surrounding areas

FIG. 13. Graphs for the selection of oil kinematic viscosities for rolling bearings versus bearing bore diameter db and operating temperature ? at constant speed n.

Oil lubrication. Oil lubrication is used when operating conditions such as speed or temperature preclude the use of grease. Ball and roller bearings must be lubricated with oil when the running speed is in excess of the recommended maximum grease speed and also when the operating temperature is over 93 deg. C.

A guide to suitable oil kinematic viscosity for rolling bearings is presented in the form of graphs in FIG. 13. The unit for kinematic viscosity is 1 m^2/s or centistoke, i.e., 1 cSt = 10-6 m^2/s. Oil viscosity is estimated on the basis of bearing bore, speed and operating temperature.

9.3 Lubrication of porous metal bearings

As a general recommendation, the oil in the pores should be replenished every 1000 h of use or every year, whichever is sooner. In some cases, graphs in FIG. 14 should be used to modify this general recommendation. The lower the bearing porosity the more frequent the replenishment. The oil loss increases with the shaft velocity and bearing temperature.

FIG. 14. Permissible time of shaft rotation without oil replenishment as a function of shaft velocity at constant bearing temperature.

Graphs in FIG. 15 give general guidance on the choice of oil dynamic viscosity according to load and temperature. The unit for dynamic viscosity is 1 Pa s = 1 N/m^2 s or centipoise, i.e., 1 cP = 10-3 kg/(ms) = 10-3 Pa s.

FIG. 15. Graphs for the selection of oil dynamic viscosity expressed in centipoises at 60 deg. C.

The following rules apply to the selection of lubricants:

• lubricants must have high oxidation resistance

• unless otherwise specified, most standard porous metals bearings are impregnated with a highly refined and oxidation-inhibited oil with an SAE 20/30 viscosity

• oils which are not capable of being mixed with common mineral oils should not be selected

• grease should be used only to fill a blind cavity ( FIG. 12)

• suspensions of solid lubricants should be avoided unless experience in special applications indicates otherwise

• manufacturers should be contacted for methods of re-impregnation.

Numerical examples

[coming soon]