Kalamazoo, Michigan
|
Inertia, or in this
case rotational
inertia is an intrinsic
property of all physical objects
which have mass and are rotated about an axis. It takes
torque to
increase (or decrease) the RPM of a rotating object. The amount
of torque
required is the product of the inertia and the rate of angular acceleration. T = I x a
where T
= torque
I = moment
of inertia
a =
angular acceleration
As soon as the RPM is
stabilized, however, it requires less torque to maintain that
RPM. Only enough torque to offset the parasitic losses is
required at that point. When an motor is bolted
to a dynamometer, the motor and dynamometer rotating elements are
connected
together and represent the sum of all the rotating masses. This
sum can be
thought of as a single mass with a specific moment of inertia which
does not change throughout the testing procedure. Because of this
rotational inertia, torque is required to accelerate the motor, even
if the dynamometer is unloaded. A typical dynamometer which
measures torque from a lever arm attached to the stator housing cannot
sense this inertial or "reactive" torque because of
the way it is mechanically connected. Our advanced "reaction
torque" measurement system can. If an motor is perfoming a sweep test where the RPM is continuously increasing, then some of the torque the motor is producing is consumed by accelerating the rotating elements (T = I x a). Because of this, a conventional dynamometer will read a torque lower than the actual motor torque while the motor is accelerating. Some commercial data acquisition systems "adjust" this reading to show what the motor torque would be if the motor were being held at a constant speed. The above equation shows that the adjustment amount is directly related to how fast the motor is accelerating and the total inertia of the rotating elements. The
following diagrams will hopefully illustrate this explanation.
Note
that the torques shown in green have been depicted as positive torques
and those in red are negative. Green torques are those the motor
supplies, and red are what we normally refer to as losses. The
purple torques are what we refer to as reactive torques because,
although they don't dissipate energy in the traditional sense (convert
it to heat where it cannot easily be recovered), they do absorb
energy during periods of acceleration but return that energy during
deceleration. This is why raw data from the torque lever arm in a
conventional dynamometer reads low during acceleration and high during
deceleration.
 In
the above diagram, the motor is held at constant RPM. The net
torque of the motor is equal to the torque read at the lever
arm. Since there is no
acceleration, there is no reactive component of torque. It is
100% brake (frictional) torque.
 In
the above
diagram, the motor is being accelerated during a sweep test. A
portion of the net motor torque is being used to accelerate all the
connected rotating elements (including those which are integral parts
of the motor) and therfore lessens the force on the
lever arm. This is depicted by the purple downward vector at the
lever arm. The total torque at the lever arm is the sum of the brake
torque (yellow vector) minus
the torque required for acceleration (purple vector). Because of
this lower reading, the value must be adjusted up by the data
acquisition system to read what the steady-state value would be at any
specific RPM.
 In
this
diagram, the motor is being decelerated during a sweep test. In
addition to the net motor torque, additional torque is
being returned by the connected rotating elements being decelerated
and
therefore increases the force on the lever arm. This is depicted
by the
purple upward vector at the lever arm. The total torque at the lever
arm is the sum of the brake torque (yellow vector) plus
the torque returned by deceleration (purple vector). Because of
this higher reading,
the value must be adjusted down by the data acquisition system to read
what the steady-state value would be at any specific RPM.
How
We Do It Better At TORQUEWORKS
we capitalize upon a generally accepted law of mechanics: For every action,
there is an equal an opposite reaction.
In the dynamometer environment, it simply means that if the
motor tries to
turn the crankshaft clockwise, there is a reaction torque which tries
to turn the motor counter-clockwise. It is importatnt to recognize that
this happens even when there is no load
connected. This is why in a conventional rear-wheel drive car
the left front fender rises when the driver stabs the throttle in
neutral. The motor develops clockwise torque to accelerate
itself, and the reaction torque tries to spin the car
counter-clockwise. At TORQUEWORKS,
we simply let the motor pivot on bearings and measure how
hard this
reaction torque attempts to rotate the motor in the reverse
direction. Of course, the motor doesn’t rotate at all. It just tries to
rotate, and this torsional effort (torque) is what we measure. Note
that the pivot point is not the centerline of the crankshaft. The
only requirement to accurately measure the reaction torque is that
these two axes be parallel.
The elegance of this
torque measurement solution lies in the fact that the reaction torque
is always exactly equal to the torque the motor is
developing.....whether it is
trying to accelerate the motor and connected elements, or whether it
is working against the brake, or any combination
of the two. And
it's totally independent of any speed or acceleration
considerations.
With our dynamometer, if
we stab the throttle with the driveshaft not even connected, we still
see a
positive
torque spike during acceleration (fuel was burned, energy was
delivered)
and a negative torque value during deceleration (energy being
dissipated via friction, windage, and pumping losses). When
we do a WOT test and modulate the motor speed up and down by applying
and removing load, we find we record a similar torque value at a
specific RPM, whether we are accelerating or decelerating. And
it's independent of the acceleration rate. Because the fixture pivot
point is offset substantially to the driver's side, there is
considerable preload on the load cell when the motor is not
running. This preload is simply zeroed out electronically before
the pull. This offset pivot concept was chosen during the design
phase to to provide improved stability and to allow the testing of
reverse
rotation marine motors without any
setup change. Dual starters
integrated with the dynamometer are software selectable to provide the
correct cranking direction. If the anticipated torque of a
reverse rotation motor is large enough to actually lift the fixture
off the load cell, additional preload is applied with an adjustable
spring. Also, a safety catch is provided to limit fixture
movement
in case of a catastrophic fault. By measuring
motor reaction torque, we’re not encumbered by determining motor
acceleration or ablolute inertia values. Consequently, our reaction torque measurement (load cell
reading) needs only to be multiplied by a simple scaling
constant, and not manipulated by complex mathematical functions which
depend on many variables. Simpler is
better.
|
