Click here to skip to content

Reliability Warriors. Downtime Destroyers.


New Twists to Diesel Vibration


Each year hundreds of thousands of non-automotive industrial-type diesel engines go into equipment systems and machinery in which the engine drives a hydraulic pump, air compressor or generator. Many of these applications are trending toward lighter weight engines as new design strategies seek better fuel efficiency through lighter components. More-efficient designs also have allowed in-line three-, four- and six-cylinder diesels to do what used to require V-8 or bigger powerplants.

While downsizing may be a blessing in terms of operating economy, it can easily aggravate the relationship between engine and driven equipment in terms of torsional vibration and the resonance that can amplify this vibration to destructive levels. As a result, the dollars saved in fuel economy -- and in other ways -- might well be lost in repair expenses and downtime costs.


Torsional vibration remains a relatively unfamiliar science. Most diesel equipment users -- as well as many engine designers and OEM service people -- do not feel this problem is significant. It may be that most users have simply become conditioned to the replacement of broken crankshafts, spline shafts, couplings, bearings, gears and seals as the normal consequence of engine operation, and therefore continue repairing the breakage without pressing their suppliers for a more permanent, preventive solution.


The purpose of this article is to examine the sources of torsional vibration problems, and to review the important role that coupling technology can play in minimizing these problems as smaller, lighter-weight engines become more popular. People concerned with drive train mechanics generally understand the coupling as a power transmitting, shock absorbing and misalignment compensating device. However, many do not recognize its role in damping vibration and tuning the engine-driven system to shift destructive resonance speeds away from the engine operating RPM range ... or the reason why such tuning is needed in the first place.

Let's begin with a look at why tuning is needed.

The inescapable fact of diesel operation is that it creates torque pulses that momentarily accelerate or decelerate normal crankshaft operation. This is caused by the firing order, the firing angle and the number of cylinders. Since the diesel engine does not use spark ignition, and must instead achieve high compression to fire the fuel (up to three times greater than spark-ignited gasoline engines) the pulses are larger than on other engines. Some degree of misfiring also can cause "pressure spikes", because ignition doesn't always occur precisely when it should and the fuel doesn't burn completely on each ignition stroke.

These torque pulses introduce harmonic excitation forces that are not part of the smooth torque output of the engine and do not add to usable horsepower. However, they can range as high as ten times the engine's normal operating torque, so they can add substantially to the total amount of torque transmitted through the system and to the rate of wear and damage that result from it.


Torque pulses ripple through the system as torsional vibrations that cannot be seen or felt like the familiar linear (up-and-down, side-to-side) type of vibration because the forces of action/reaction are distributed across different planes tangential to the drive shaft rather than confined to the same plane. Because torsional vibration is "invisible", the damage that it causes is often mistakenly attributed to some other cause ... such as shaft-to-shaft misalignment, improperly specified components, or faulty parts.


--- Systems Have Natural Frequencies ---

The stop-start forces applied by torque pulses exert a twisting pressure on the drive train. Some users who are aware of these pulses believe that a long drive shaft will absorb them. It will help tune the system, but it does not absorb torsional pulses. The drive shaft and other drive train elements simply react as all springs do; rather than absorb the twist energy, they give it back by untwisting. Assuming that the magnitude of the torque pulses will remain fairly constant (which is not always the case), the time required for twisting and untwisting will be determined primarily by the inertial relationships between engine and driven equipment, and the spring rate (stiffness) of various elements in the drive train. This twist-untwist time increment, expressed as a ratio to one minute, identifies the natural frequency of the system.

In simple two-mass systems (engine and single load, connected by a single shaft), natural frequency can be closely approximated by the following formula, in which CTDYN is the dynamic torsional stiffness (spring rate) of whichever connecting device is torsionally softest -- usually the coupling -- JA is the engine's inertia and JL is the load's inertia. (Stiffness and inertia data are readily available from all equipment and component OEMs.)

 

Precise determination is possible with a more detailed calculation known as a Holzer Analysis, in which known frequencies are induced into the system on a trial and error basis, and their resultant energy balances are calculated, until a frequency is found for which the summation of energy gained or lost through all torsional deflections balances to zero. Note that in some industrial applications, engines can have two or more power takeoff points feeding multiple loads through multi-part drive trains; in such systems, frequency computation becomes very time consuming and complex, and is best left to experienced consultants.

--- Effects Of Smaller Engines ---

Generally, in a system where engine inertia is higher than load inertia and the drive train is relatively stiff, natural frequency will be high. In a system where engine inertia is lower than load inertia and the drive train relatively soft (springy), the natural frequency will be low.

Consequently, when smaller engines replace larger ones (assuming that other components remain essentially the same) the reduced mass and inertia at the driving end tends to lower the natural frequency of the system.


Now consider that a diesel engine "excites" its driven system at specific frequencies that vary in direct proportion to engine RPM. At those RPM levels where the engine's excitation frequency and the driven-system's natural frequency coincide, a condition of resonance emerges. Each torque pulse applies its twist energy at about the same time when the shaft has untwisted to release the energy from the previous pulse, which greatly amplifies the twist/untwist motion. Those RPM levels at which resonance occurs are termed critical speeds, for obvious reasons.

For the standard 4-cycle in-line diesel engine system, the speeds at which resonance occurs -- critical speeds -- are calculated as the natural frequency divided by N/2 for the primary resonance speed and by N for the secondary resonance speed, where N represents the number of cylinders. Critical speeds for systems using smaller engines move toward the operating RPM range because the engine's excitation frequency drops with the move from multi-cylinder or two- cycle engines to four-cylinder, four-cycle engines.


Engine driven generators, compressors and centrifugal pumps generally operate from 1,000 to 2,000 RPM, and hydraulic pumps from 1,800 to 3,000 RPM. The trend is to operate at the highest possible speed so as to have smaller, more compact driven equipment. This results in a situation where faster operating speeds, lower excitation-frequency slopes and higher natural-frequency curves combine to put critical RPM values near the required operating speed ranges. The prior situation of large-inertia, multi-cylinder engines combined with big, low-speed equipment would have high natural frequencies with relatively low operating RPMs. Resonance, or critical RPM -- occurring where excitation frequency meets system natural frequency -- would not happen in the operating range.

The relationship among RPM, natural frequency and excitation frequency explains why designers (and users) will have a problem with new systems that allow the two frequencies to meet at the operating RPM level. In that design situation, careful selection of torsionally soft or stiff couplings is essential to move the natural frequency away from destructive resonance ranges.


As engine RPM moves up and down between idle and operating speeds, it can pass through the critical areas with no problems if it is done quickly. The trouble starts when engine RPM remains in the critical zone for longer periods of operation or idling.

--- Critical Windows For RPM ---

Torsional vibration resonance typically occurs across an RPM range from 0.7 to 1.4 times the critical RPM, establishing a "window" in which engine excitation forces can cause damage to system parts. For example, if resonance or critical RPM occurs at 1500 RPM, the system will see torsional vibration problems anywhere between 1050 RPM (0.7) and 2100 RPM (1.4).

To prevent such problems, new systems designed with light-weight three-, four- or six-cylinder engines must have a complete torsional vibration review. That review should include either a two-mass analysis or, if necessary due to a complex drive system, a complete Holzer analysis to identify the system's natural frequencies and resultant critical RPM ranges. If this determines that critical RPM ranges exist near idle or operating speeds, it becomes necessary to conduct a complete multi-mass torsional vibration calculation, which typically requires hiring a specialized consultant.


Even more of a problem is torsional vibration, as engines become smaller and lighter in relation to their driven equipment, the magnitude of their torsional vibration pulses tends to increase and do more damage to drive train components.


Obviously, engine RPM standards, the number of cylinders and the inertia values of both the engine and its driven equipment are not subject to change (unless these elements can be replaced with something significantly different, which is either unlikely or prohibitively expensive). Fortunately, however, the resonance of the system can be changed in order to prevent it from occurring within the operating range.


This brings us to the concept of using the coupling as a tuning device.

--- Tuning The System ---

Being one of the spring elements in the drive train, each coupling has its own spring rate, or dynamic torsional stiffness (CTDYN). When this is factored into the system's CTDYN value, the system's value must somehow be changed ... and per the formula, the system's natural frequency changes with it.


The coupling's dynamic torsional stiffness must be picked to either match the remainder of the system or tune the system's resonance out of the operating range by being the softest element. Proper coupling selection assures that if the system is operated at the wrong speed, or if a fault occurs elsewhere in the system, the coupling will serve as a fuse and break before something more expensive does.


Couplings designed for engine-driven systems basically comprise a driving component and a driven component, connected by a variety of torque-transmitting elements that offer a range of torsional stiffness values, from highly elastic rubber to an almost rigid plastic. The stiffer elements, usually used for relatively low inertia loads such as hydraulic pumps, will move the natural frequency of the system upward. Softer elements, usually appropriate for higher inertia loads such as larger compressors or multiple pumps driven through splitter boxes, will move the natural frequency downward as well as absorb or damp vibration energy.

Some coupling elements offer a progressive characteristic torsional stiffness, in which the material is much stiffer when under full load than when it is lightly loaded. This gives system designers more flexibility in certain applications where the system idles at no load with a very low natural frequency, but operates at full load with a high natural frequency, as is often the case with engine-generator sets.


This wide range of stiffness alternatives gives designers and users alike an opportunity to change, or tune, their system's natural frequency. To take advantage of that, it is necessary to look beyond the usual mechanical and torque-load specs to determine the excitation and natural frequencies within your system. With these values known, you then can decide whether it is best to tune the natural frequencies upward or downward in order to move the resonance speed away from both idling and operating speeds.

This explains why experienced and knowledgeable coupling suppliers want to know the inertia and stiffness values of all system components, and will want to prepare a torsional analysis to identify all of the system's natural frequencies and make sure that the recommended coupling will accommodate them.


Coupling selection based on torsional analysis will keep resonance speeds outside the critical 0.7 - 1.4 RPM "windows" established by engine torsional vibration, to maximize the service life of driven equipment and to greatly reduce the downtime and repair costs associated with premature fatigue and failure in key parts.


When couplings are properly sized and specified on this basis, any premature coupling failures can be taken more reliably as a signal that something else is troubling the system ... such as cylinder misfiring problems, poor fuel or filters, ambient temperature or airborne contaminants, unrealized torque overloads, misalignment, or improperly specified parts.


Remember that couplings already in place in your system may not have been selected within the disciplines described above. As one of the least expensive elements in any engine/equipment system, couplings too often are added as an afterthought rather than factored into the design up front. If your system seems to be chronically suffering the kinds of wear and breakage symptoms that could result from torsional vibration and critical frequency problems, it's very possible that replacing the original equipment coupling with one selected via torsional analysis can put those problems behind you forever.