An approach to screening faulty plant machinery.
A significant portion of fires and explosions result from machinery failures, which can often be traced to deficient operational practices. Unfortunately, lost operating experience resulting from recent staff reductions in the refining and chemical industries have made it difficult for process plants to address these deficiencies. In addition, the assessment shows that some plants are incurring additional operating risks by improperly applying predictive maintenance concepts that delay inspections of critical machines.
A comprehensive reliability evaluation, which includes assessing operational, maintenance and organizational practices, is normally required to fully assess operational risks. However, a screening test can determine fire and explosion risks in process plants. Applying this test to plants that have undergone reliability evaluations indicates that those with passing grades have had significantly lower rates of machinery failures and fires than those that failed.
The screening approach consists of five questions that deal with operational procedures. Process plant operators should be able to provide specific written details on how the various functions are performed. Although the following screening questions and their underlying issues have been primarily applied in large refining and chemical facilities, they can also be used to evaluate operating risks in power plants and other process industries.
What are the testing procedures for alarms and shutdowns and how often are they performed?
Every plant has monitors that alert operators when its process exceeds allowable levels. The monitors also activate an automatic shutdown if these excessive levels endanger the plant. These vital alarm and shutdown systems are almost always tested when the plant undergoes general maintenance, but plant operators are usually reluctant to perform these tests while the plant is operating. Experience has shown that these alarm and shutdown functions should be tested at least every four to six months, which would require that these tests be performed while the plant is operating.
Unfortunately, plant operators often do not perform these tests during operation for fear that they will inadvertently cause a plant shutdown. This should not be a concern as long as the testing procedures are properly engineered and implemented. With the trend toward longer periods of uninterrupted operation and the existence of lower available levels of operating experience, systematic testing of these safety systems is even more imperative.
What surveillance procedures are utilized for evaluating the operating condition of machinery?
Surveillance procedures are an active function and so differ from conventional monitoring functions. Surveillance requires operators and technicians to test and evaluate machinery support and safety systems to assure they operate properly. Conventional monitoring, performed by most operators, is basically a passive function since it emphasizes transcribing instrument readings onto a log sheet. In recent years computerized process control systems have improved manual monitoring methods, but they still do not replace a comprehensive surveillance program.
Surveillance involves performing a physical inspection of possible leaks, loose components, unusual noises and the like; a log sheet analysis comparing values with defined safe ranges and identifying deterioration trends; tests of standby and emergency lubrication and sealing systems; and machinery operation analysis, which includes performance tests and vibration analysis. Operators who consistently perform these functions are more likely to identify and avoid problems before they endanger their plants.
What provisions exist to ensure that operators are using the proper procedures for starting and shutting down equipment?
A key part of an operator's training involves using proper start-up and shutdown procedures. While operating manuals are often available in the plant's control room, operators are not always able to review them when starting or shutting down machinery. Too often, operators depend on their memory to start up or shut down equipment. Interestingly, plants with the most stable operations are most prone to operator errors during start-ups and shutdowns because such procedures are performed infrequently. On the other hand, plants that experience many operating problems generally have fewer operator errors while starting or shutting down machinery because they do it so often.
To minimize risk of machinery damage and fires, which are a primary danger during the start-up or shutdown phase, the best plants provide operator aids, including personal procedure cards designed to fit in a shirt pocket or printed procedures adjacent to the equipment. These relatively simple provisions have helped minimize machinery failures, toxic gas and fluid releases and associated fires.
Are the vibration monitoring systems on critical machines set to automatically shut down the machinery?
Critical machines such as compressors and turbines are almost always required to have permanently mounted vibration sensors. They normally have provisions to automatically shut down a machine when vibration levels exceed preset danger limits. To minimize the risk of fires and explosions resulting from sudden machinery failures, these automatic shutdown provisions should be fully commissioned. However, many plants do not use this capability, depending instead on operators to manually shut down the equipment. This practice, which has often proved inadequate when dealing with rapidly developing problems, stems from early industry experience when the reliability of these monitors were suspect. These concerns have been successfully addressed over the years by improving the quality of components and application methods. Today, a plant can expect to significantly lower operating risks by using automatic shutdowns activated when dangerous vibration levels occur.
Are there documented procedures that help determine the contributing causes of a machinery failure? When a machinery failure occurs, most process plants attempt to determine the immediate cause. However, this is far less than needed to avoid a repetition of the failure. In addition to the immediate cause, the fundamental cause of the failure must be established. It may be due to such basic factors as faulty procedures, improper materials or inherent design flaws.
To properly determine the fundamental cause of a failure, a plant must institute several procedures: accurate, detailed maintenance records must be available; operating records of the period immediately prior to the failure must be evaluated; and plant management must hold an individual or group responsible for reviewing all failures. When these procedures have been aggressively applied, plant operators have been able to prevent repeated failures and identify deficiencies to avoid them.
To appreciate the limitations in applying predictive maintenance to critical machinery, one must understand the various maintenance approaches. Until the 1960s, the common approach was to allow machinery to operate until it performed poorly or failed. This breakdown maintenance approach led to so many catastrophic failures that it was, for the most part, replaced by inspection of critical equipment on a scheduled basis. This preventive maintenance method, which is used by most process plants, has effectively minimized catastrophic failures. Its major limitation is that the fixed maintenance schedule can sometimes result in unnecessary, costly inspections.
Responding to the limitations of preventive maintenance, the predictive maintenance approach has recently gained favor as a way to reduce maintenance costs. It is based on the use of diagnostic methods to identify signs of deterioration while the machinery is operating, thereby justifying inspection and repair postponement. When properly applied, predictive maintenance can allow certain classes of process machinery, such as small pumps and motors, to operate safely for longer periods between maintenance. However, there are limitations to applying this method to critical machinery such as compressors and steam turbines.
Identifying signs of deterioration in critical machinery involves three basic diagnostic methods: vibration analysis, performance analysis and analysis of wear debris in lubricating oil. Vibration analysis is the most commonly used method because it provides a rapid, accurate indication of machinery operation. The other two methods are difficult to employ and interpret and, consequently, are less frequently used.
Vibration analysis of critical machines is similar to an electro-cardiogram analysis of the heart. Both methods use vibratory response to determine the "health" of the respective "machines." However, both methods are limited in their ability to predict if and when the machine will fail. In the case of critical machines, serious internal faults such as cracks and corrosive effects do not consistently provide readily discernible vibratory responses. In spite of this, some process plants have been improperly using indications of an apparently normal vibratory response to justify extending maintenance periods. This practice has resulted in sudden, catastrophic failures.
The following two cases illustrate some of the limitations of using vibration analysis to extend maintenance periods of critical machines in process and power plants. In one case, a power plant's technical staff evaluated the performance and vibratory response of a turbine scheduled for maintenance. Finding no signs of a problem, they recommended that the internal inspection be postponed until the next maintenance period. Operating management, however, overruled the recommendation and decided to maintain the scheduled inspection. When the turbine was inspected, failed blades, which could have eventually contributed to a catastrophic failure, were unexpectedly discovered.
Another scenario concerns one of the worst instances of improper application of vibration analysis as a predictive tool. It occurred when a turbine-driven compressor in a chemical plant was not inspected because an analysis of the vibration response and performance were apparently acceptable. Based on this assessment, the plant decided that the turbine and compressor could go another few years without an inspection. Unfortunately, the vibration analysis did not detect the existence of a shaft crack. Three months later, there was a sudden shaft failure that resulted in the turbine exceeding its maximum operating speed. The failure caused the release of lubricating oil, which resulted in a fire that destroyed the turbine, its compressor and the adjoining process unit.
Although these cases illustrate the improper application of vibration analysis, one should not assume the technology is flawed. On the contrary, when properly applied, it is an excellent tool for determining the condition of a machine. The problem is not in the technology, but in its application.
These experiences show that applying predictive maintenance to critical machinery can introduce a higher risk of machinery failure. Until diagnostic instrumentation is developed that reliably identifies structural faults in critical machines, preventive maintenance practices should be used to minimize operating risks.
This does not mean that the process and power industries must endure the higher maintenance costs often associated with preventive maintenance. On the contrary, periods between scheduled maintenance can be safely increased, and the extent of maintenance reduced, based on an evaluation of operating history, design and materials upgrades and comprehensive equipment operation surveillance.
PHOTO : The blue sky could turn black with fiery smoke resulting from insufficient maintenance.
Gerard Muller is president of Serry-Tech Inc., a Morristown, NJ-based provider of reliability evaluations, training and process technology services to refineries, chemical plants and corporate engineering organizations.
|Printer friendly Cite/link Email Feedback|
|Date:||Oct 1, 1991|
|Previous Article:||Risks in developing nations pose an uphill battle.|
|Next Article:||Preventing a crisis from getting out of hand.|