|
|
|
Customer Success Story Index |
US Petrochemical Complex Experiences Two Shutdowns on a Critical Reciprocating Compressor
|
|
A large US petrochemical complex uses a 3500 Machinery Protection System to monitor a critical reciprocating compressor with XY
proximity probes observing the plungers. They also have a Data Manager® 2000 condition monitoring system connected to the
machine. When establishing the shutdown logic in the 3500 System, the customer considered the various failure mechanisms that could occur and made certain their alarming strategy could address them. In particular, one failure mechanism was that the plunger could shatter. To detect failures of this nature, the customer activated their gap voltage alarms and programmed the 3500 System’s logic to generate a shutdown signal if both X and Y probes were in a gap alarm condition. They also enabled conventional DANGER alarms that would shut the machine down when excessive vibration levels were present. Just such a condition occurred recently, and the gap voltage alarms from both probes tripped the unit when a plunger shattered. While the 3500 System did its job in shutting the machine down automatically, it also provided earlier indications that were, unfortunately, unheeded. Eight hours prior, the 3500 System had shut the same compressor down due to high vibration alarms. Operators, believing that there was an instrumentation problem rather than a machinery problem, restarted the unit and ran it until the 3500 System again shut the unit down on gap alarm conditions from both probes as noted above. Editor’s Note: While we wish that every machine shutdown represented a “Synopsis of Savings,” it sometimes represents a “Synopsis of Savings Lost ” because a failure progressed much farther than it needed to. Was the machine in this story ‘saved?’ Technically, yes. However,the customer missed multiple opportunities to be proactive. First,operators could, and should, have taken action when the initial alarm occurred. The machine was telling them it was in distress, but these indications were ignored under the assumption it was an “instrument ” problem, rather than a bona-fide machinery problem. Bently Nevada advocates that customers take all alarms seriously and thoroughly understand why the alarm occurred before they restart or continue to run a machine. Assuming an alarm is an instrument problem, rather than a machinery problem, is generally a very bad assumption. A recent ORBIT article dealing with the issue of relying on a single measurement and assuming alarms are machinery, rather than instrument, problems is worth your attention. (“Relying on a Single Measurement – Guilty Until Proven Innocent,” ORBIT, Vol.22, No.1, 2001, p.55.) Second, and equally important, the customer had at their disposal an online software system to help manage the machine. At minimum, they should have relied on their Data Manager® 2000 software following the first alarm to understand better what was happening in the machine, possibly providing sufficient information to prevent them from starting the machine again. Ideally, they would have been using their Data Manager® 2000 system to spot changes in the machine’s condition well before the 3500 alarms were activated. The customer undoubtedly saved some money when the machine tripped automatically, rather than allowing it to continue running after such a catastrophic failure as a shattered plunger. Thus, we’re proud of our machinery protection systems and that they act when required, as this story illustrated. However, we’re even prouder when customers use our machinery management systems to avert such failure progression. Today’s best machine saves no longer involve automatic machine shutdowns. Instead, they involve prevention of failure altogether by properly managing the machine. |
|