Does 'service factor' really mean what you think it means?
Few engineers would argue the point that 'service factor' is critical in
the selection of electromechanical plant, but just how is service factor
defined and is it consistent across the manufacturing spectrum? Wayne
Patton explains how different manufacturers specify this vital parameter
The most important characteristic of mass-produced geared motor units
(GMUs) is the rated torque that can be generated in continuous operation
combined with an acceptable service life. Because of the wide scope of
application of GMUs it is inevitable that the drive has to be adapted to
the widest range of loading conditions. For instance, when utilised to
drive an aerator to oxygen-enrich water in a reservoir, the units operate
24/7, 365 days per year and in that case, service factor is characterised
by time. However, crane drives used for maintenance in a turbine hall
would encounter only very intermittent duty, in which case service factor
would be characterised by torque.
To make a valid comparison between these two different applications, we
should formulate and consider a fictitious torque level MN (see Figure
1). The acceleration torque Ma and load torques ML1-3, calculated from
the individual load collectives, should be equivalent in that they should
result in the same service life under continuous operation for the
corresponding operating variables, just as if they were loaded with the
actual torque. The service factor (SF), fB, could be defined as the
number by which the rated torque (MN) of the driven machinery must be
multiplied in order to obtain a torque rating that guarantees the same
degree of certainty against operating damage arising from an action on
the drive shaft (constant over a given length of time), as the actual
torque on the drive shaft (varying with time). The design of the GMU is
correct if its continuous load capability is equal to the rated torque.
There are multiple benefits of using SFs to both designers and users of
GMUs. The driven machinery's operational shock loads can be assessed and
subsequently taken into account in the dimensioning of the gear unit, and
transmission damage can largely be avoided. However, the formulation of
an equivalent rated torque from the load collective must take into
account the operational shock loads encountered, individual load action
times and total operating hours. This can take considerable time and
effort to estimate. A greatly simplified method for GMUs is that shock
classifications are applied according to a set of multiples ranging from
1.1 for uniform operation of the machinery, through moderate shocks
(1.35), medium shocks (1.6) to heavy shocks (1.85). In truth, even if a
computer were used to make an accurate calculation, the actual load
values are generally not available. Rule-of-thumb estimation is the only
practical solution for standard GMUs.
Most manufacturers of GMUs use a simple table of daily operating hours
and shock classification parameters to determine the SF of their
products. The basic table is frequently expanded to include additional
levels for the operating hours and/or switching frequency. Although
systems of this type are very common, there is no standardisation.
Examination of the specifications of 42 European manufacturers shows that
vastly differing operating hours are assigned to a SF of 1.0 at shock
classification 1. Some 50% of manufacturers base SF of 1.0, on no more
than 8 hours daily operation. SFs are thus neither standardised nor
uniform and indeed can by highly misleading.
While there is no arguing with the 'time' parameter, the 'shock
classification' parameter is entirely subjective. Specifications and
standards are based upon 'moderate', 'medium' and 'heavy' shock loads and
these judgmental levels must be superseded by objective limit values.
There are a number of causes of torque shocks. Three phase cage motors
develop breakaway torques MA of up to 2.5 times rated torque MN when
switched DOL on a stiff supply and this is independent of load demand.
Starting under full load, and particularly where the load inertia is
large, the torque flows through the gear unit, making load inertia an
important consideration. Regenerative or plug braking generates two or
three times the level of breakaway torque. However, these particular
shock torques need not be of concern for SF calculation if they occur
only occasionally. Their short duration means they exert only a minor
influence on the overall load collective but they do gain in significance
as the moment of inertia of the load increasingly exceeds that of the
motor rotor (Factor of Inertia) and with higher switching frequencies. At
high switching frequency, say up to 1,000 operations per hour, the GMU
can be placed under increased torque for as much as 30% of the time. More
important is the fact that changes in speed can lead to dynamic peak
torques, which are more difficult to calculate than the time under load.
The effect of switching frequency is again assessed empirically and
differently from manufacturer to manufacturer.
It may seem curious that low switching frequencies demand such a high SF
evaluation, while high switching frequencies demand so little increase,
but this is for good reason. Power transmissions with play, such as
chains and claw couplings, may generate considerable peak torques in
switching operation, approaching the limit of resistance to fracture and
sufficient to demand an increased service factor for safety reasons. With
this in place there is little more to be done in terms of the fatigue
limit at higher switching or reversing duty. Additionally, it would be
over egging the pudding to further increase the size of a unit already
increased in size for shock loading considerations. It would be more
appropriate to soften the impact of switched operation by selecting more
suitable power transmission components, such as shock-absorbing shaft
couplings.
So called 'shock neutral' couplings, such as disc couplings and belt
drives, must be free of play and allow no free travel, otherwise they
will exacerbate the peak torque loading. Other power transmission
components may even be shock amplifying due to rotational play, whereby a
considerable difference in speed can exist momentarily between the
driving and the driven shafts. This is not a problem where the load is
running continuously in one direction but in switched or revered
operation, such transmissions can be highly disruptive.
So when it comes to service factor, as Humpty Dumpty might well have said
in Through the Looking Glass, service factor means just what I choose it
to mean - neither more nor less . Clearly, equipment must be up to the
task if the system is to be reliable but, as yet, the determination of
service factor is as varied as the many load characteristics acting upon
it - often requiring the specification of a unit larger than may be
immediately apparent. One thing is certain, high quality and long life
have never come cheap. In large measure you get what you pay for and in
the long run, buying cheap can mean that ultimately you have bought dear.
* Load collective refers to the sum of all loads that absorb torque and
influence the current drawn by the machine, including friction, windage,
acceleration and the load itself
Wayne Patton is with Danfoss Bauer