In the 200 ms it takes for one revolution of a crankshaft, everything can change. In that brief span of time, the structural integrity of a single concealed bolt can reach the breaking point, beginning a cascading series of events that causes the the catastrophic failure of an entire machine. Consequences may include not only hundreds of thousands of dollars in equipment damage, maintenance expenses and lost productivity, but also possible emissions releases and serious threats to the safety of nearby personnel.
Sound unlikely? It’s precisely the scenario encountered by an off-gas processing plant in USA . However, the outcome was not failure, but triumph for the plant’s foresight and protective measures. Impending disaster was averted by an advanced condition monitoring system installed on a critical compressor and by the effective response of the maintenance team. This incident illustrates how effective advanced condition monitoring systems can be, especially in terms of their sensitivity to minute changes in normal operating behavior, their accurate diagnosis of what those changes mean, and their ability to activate a proper automatic response – all within a few revolutions of a crankshaft.
Plant, Compression System and Monitoring Technology
The plant processed off-gas from six refineries. A by-product once used as boiler and heater fuel, off-gas contains ethylene, ethane, hydrogen and other components that are now more valuable when recovered for use as basic petrochemicals and feedstocks. By collecting off-gas from multiple refineries, the comany attained sufficient volumes to make processing and separating these components economically viable. With all six refineries online, the company could process roughly 138 million scf of off-gas per day. From this supply, the plant generated more than 28,000 bbl per day of liquid hydrocarbons, including ethylene, ethane, propylene, propane, butane and natural gasoline, as well as 30 million scf of per day of hydrogen.
The State of Machine Monitoring
The company benefitted from: (1) the ability of their monitoring system to detect a relatively small fault at an early stage, and (2) the speed with which the system was able to take proper action (i.e., shut down the machine). The shearing of the bolt was literally a single-cycle failure. There was no gradual increase or trend to indicate an impending problem. Still, the PROGNOST®-NT system was able to shut off power to the electric drive within two seconds of the failure [Figure 4].
In order to avoid false trips from brief transient events, the PROGNOST®-NT system requires high vibration amplitude in a number of segments (typically, a minimum of five segments). It must be a high value across a wide span of crank angle degrees to distinguish a significant impact. Additionally, vibration values must exceed established safety limits for a preset number of consecutive revolutions in order to generate a trip alarm.
In this case, however, the bolt failure resulted in such a strong impact that the system recognized a violation of safety limits on both the throw 5 crosshead sensor, as well as on the crosshead sensor on the opposite throw. For these conditions, shutdown is activated after only three revolutions. This additional level of sensitivity shortened the length of time the machine operated with a single bolt keeping that side of the big end con rod together.
Conclusion
Since being commissioned in 2005 on these four critical compressors, the PROGNOST®-NT system has provided significant return on investment by detecting multiple failures of drive train components at an early stage and successfully interlocking the machine before consequential damages could occur.
