Features

How to get more value from variable-speed electric motors

LARGE compressors in modern process plants demand drive powers in the region of 10-150 MW.

Traditionally, they have been driven by gas turbines. The "all electric" concept seeks to eliminate the gas turbine and use variable-speed electric-motor drives. 

The main benefit would be the added-productivity of the plant, from an average of around 340 days due to unscheduled and scheduled outages of typical gas turbine drive systems, to a theoretical 365 days every year.

Advanced electric motors and modern compressors in the clean-gas service theoretically do not need the routine maintenance (would require only a shutdown for periods of 5 to 8 years).

Accounting for unscheduled outages of the compression units of 3 to 5 days per year, the compressor train still yields a minimum of 17 to 20 added production days, with associated huge revenue gains. 

It is very important considering modern, large process plant production value of around 0.5-1.5 million dollars per day.

Other benefits associated with electric motor driven compressors are derived from the better controllability of variable speed drives, and the unlimited number of (soft) starts.

Electric motors are around 50 to 70 percent of the initial cost compared to modern gas turbines with the same ratings. Two-pole brushless synchronous motors in power ratings up to and beyond 100 MW have not been built in the past due to lack of demand, and thus a little incentive for the manufacturers to build them. 

From the manufacturer's prospective there is no reason not to build large two-pole synchronous electric motors.

Their design, materials, manufacturing methods, and analyses (electrical, mechanical and thermal studies) are identical to those of two-pole synchronous generators which used in many power plants around the world in ratings up to 400 MW (and above).

Higher speed applications, variable speed requirements and new component designs for large electric motor applications may impose some changes. 

Sometimes, the qualification process for critical components is crucial and correspondingly extensive.

An example of variable frequency drive electrical facilities.

The detailed mechanical/electrical design reviews, the rotordynamics analysis, the torsional review, the control issues, the RAM (reliability, availability and maintainability) studies, and comprehensive network stability studies should be performed to satisfy both the operating company and the manufacturer.

Independent design reviews can provide the necessary neutral assessment of the up-scaled electric motor technology.

The design of the associated variable frequency drives (VFD) of the load commutated inverter (LCI) type, on the other hand, amounts to a downsizing exercise from existing high voltage DC (HVDC) technology and is less crucial.

Induction vs. synchronous

The usual technology for electric motor driver uses an LCI (Load Commutated Inverter) converter associated with a synchronous two-pole electric motor. However, the VSI (Voltage Source Inverter) technology and the induction motors are becoming popular option.

In this section, two technical solutions, the LCI (synchronous motor) technology and the VSI (induction motor) technology for the electric motor driver are compared.

Advantages and disadvantages of each option are described (the footprint, the torque pulsations on train shaft, harmonics, initial price, and others). In addition lessons learned from design, fabrication, installation, operation and maintenance of both solutions are presented. 

Because of the reactive power consumption of its thyristor bridge, an LCI system cannot naturally power an induction motor. If this is the case, the forced commutation circuits would have to be added, resulting in a very complex, costly and less reliable solution.

The VSI stands for "Voltage Source Inverter" using a capacitive DC link with self-commutation semiconductors. It can supply a lagging and leading power factor load.

A VSI can power both the induction motors and the synchronous motors. At present, the VSI systems are available with different topologies (2, 3, 4, 5, etc levels) and different electronic power components (GCT, IGBT, and others). 

They can feed medium and large voltage electric motors for power ratings up to around 50 MW (higher power levels are under study).

A well-known VSI arrangement is a 12-pulse VSI with 2 network serial bridges associated with a 3-level PWM (Pulse Width Modulation) motor inverter.

For comparing the Load Commutated Inverters (LCI) vs. the Voltage Source Inverters (VSI), the following factors should be considered: 

  • The LCIs have been used for more than 30 years (since the early 80s) for high-power synchronous motors. On the other hand, the VSIs are more recent and they have been used (mainly up to 30 MW, with some isolated cases above 30 MW) for around 15 years.
  • The LCI technology generates torque pulsations. A harmonic filter is generally required. Because of the speed excursion of the variable speed drive application, it is mandatory to implement a comprehensive torsional study (the torque analysis) of the complete train.
  • Because of multiple pulse rectifier configurations, The VSI network diode bridges reduce the harmonic current level emission. Depending on the network short circuit level, using a 12, 18 or 24 pulse rectifier topology can result in the elimination of harmonic filter(s) in practice.

The induction motor (using VSI technology) provides the robustness and the simplicity. There is a very low harmonic content on the motor torque (using a modern fast switching device), an appropriate PWM (Pulse Width Modulation) and a sinusoidal output filter.

There is a constant "Power Factor" (PF) on the network side >0.95 with no need for an additional power factor compensation system. 

There is a very low harmonic content, particularly when using 24-pulse arrangement on the network side. There is also a smaller footprint, mainly due to the absence of the harmonic filter.

Typically the VSI solution requires only 70 percent of the footprint compared to a LCI option with the same rating. The VSI solution also offers the lowest cost. An induction motor is cheaper compared to a synchronous motor for the same application. 

There are few references in the operation of induction motors with powers higher than 30 MW. There are few references in operation of PWM-VSI technology with power level more than 30 MW as well.

Control issues

For large electric motors, the flexible rotor concept is generally used. In other words, the rotor runs super-critical (the first critical speed lies below the operating speed range).

The rotor should be dynamically balanced.  The rotor design and construction are usually such that a subsequent field balancing would not be required (whereas it is often possible).

For large high-speed electric motors, the usual balancing methods based on rigid body balancing theories are not sufficient to create an adequate balancing condition for heavy elastic rotors (with relatively wide bearing spans). 

A low speed rotor rotates around its local geometric centre over the length of the shaft. When passing the first critical speed, the local rotational centre changes from the geometric-centre to the local mass-centre, which means, the local unbalance in an elastic rotor varies with speed. Therefore, modal sets of unbalance weights should be used to balance each mode individually. 

Otherwise, only the vibration amplitudes at the location of the shaft probes are minimised but not the vibration level at other locations over the shaft length. As a minimum, the "n+2" balancing planes (n=number of modes to balance) are necessary to balance an elastic rotor.

The theoretical rotordynamics study and the practical vibration measurement are extremely important for the reliability of large electric motors.

These combined calculations and practical vibration measurements require a continuous update of rotordynamics models with measured data to get an adequate forecast for the vibration behaviour of a large electric motor. In practice, it is useful to have more than the theoretical "n+2" balancing planes. 

Considerations should be given to:

  1. The "internal" balancing planes over the length of the rotor body for balancing of individual modes.
  2. The "trim-balancing" planes which are accessible also in the assembled condition of the electric motor to adjust possible unbalances.

Thermal unbalance

A special task for electrical motors is the handling of thermal unbalances. Because of the inevitable use of various materials with very different thermal expansion coefficients, combined with a non-uniform temperature distribution under a load condition, special care should be taken to achieve a symmetrical mechanical and thermally insensitive design.

Small asymmetries can cause unacceptable load dependent unbalance conditions. Also, high power-levels can lead to more extreme temperature gradients and can thereby cause thermal unbalance problems.

Dynamic studies and performance tests of relatively high-speed large electric motors usually show high vibrations. Particularly high vibrations at bearings are reported (whether the bearing housing or the bearing locations of the shaft).

Even for some very large electric motors, vibration velocities greater than 6 mm/s (more than 3 times of "1.8 mm/s", the allowable limit by some electric motor codes) are measured. To clarify the source of these high vibrations, an accurate dynamic modal analysis should be employed. 

Based on experience in high-speed electric motors in range of 20 MW – 50 MW, even with an accurate impact test and a mode shape evaluation, the natural mode shapes of the motor structure/frame could not be properly identified.

It is usually more difficult for above 50 MW electric motors because of their design and construction.
The impact test usually indicates a lot of peaks caused by local enclosure sheet vibrations. Often relevant modes cannot be clearly extracted. 

The accurate FE (finite element) modelling with the forced vibration is the best method to properly clarify sources of these high vibrations. Usually high vibrations are caused by some kind of resonance with one of natural frequencies of the main motor structural, frame, supports, or similar.

Often the electric motor operation excites a local mode at a component or assembly within a complex motor system such as the bearing housing/shield, the metal sheet fabricated casing, structures, supports, and others. 

These modes are usually underestimated during design calculations because of a relatively low modal mass or due to an insufficient modelling of details of electric motor components.

All details including the electric motor frame, the bearing, the bearing housing, supports, structures, and others should be modelled accurately.

Dynamic displacements

Regarding the bearing housing vibration, stiffening of the bearing shield/housing and improved supports (of the bearing housing) can usually reduce the vibration level significantly (even to less than 10 percent of initial high vibration values).

Forced dynamic displacements without a resonance usually remain below the code limit. 

For large and high-speed electric motors, if it is not possible to reduce the vibration below standard limits, the achieved vibration level should be in any case below harmful fatigue levels (usually the endurance limit should be considered due to high frequencies involved).

In this case, the vibration and noise would be relatively high but a failure will not be expected. 

The practical notes and lessons learned for recent large electric motors are:

  1. Rotordynamics models/simulations should be extended accurately to non-rotating parts of both frame and foundation structures (the rotor and bearing system on springs is not sufficient).
     
  2. Often shifting all natural frequencies outside the operating speed ranges could not be practical. The use of a special bearing design to achieve a high modal damping could be considered as an acceptable solution (particularly when shifting of frequencies from a resonance zone is not possible). 
     
  3. For some high-order excitation modes, even for the 2×-excitation forced vibration (or higher-order harmonics) calculations should be performed to identify critical mode shapes of the system. The testing may not be effective to map the high-order harmonic frequencies and modes. A modal analysis with the impact test without knowledge of the critical mode shapes gives no clear indication for critical structural modes, which may be excited by a misalignment, electromagnetic forces or other excitation mechanisms to inadmissible vibration levels.
     
  4. Some excitations such as electro-magnetic forces acting on the stator often cannot be prevented from affecting the machine and the bearings just by stiffening the electric motor frame. These dynamic deflections/ excitations sometimes should be decoupled from the machine frame to stabilise the electric motor vibration.

[Amin Almasi is a rotating machine consultant in Australia. He specialises in rotating machines including centrifugal, screw and reciprocating compressors, gas turbines, steam turbines, engines, pumps, condition monitoring and reliability.]

Send this to a friend