【翻译】第四十八篇 The Pros of Proactive Product Servicing

本帖最后由 小编D 于 2011-12-15 17:46 编辑

你好，我是小编H。请对以下文章有翻译兴趣的组员留下你的预计完成时间，并发短信息联系小编H，以便小编登记翻译者信息以及文章最终完成时的奖惩工作。感谢支持翻译组！

ThePros of Proactive Product Servicing
Monitoringmaintenance helps avoid unexpected shutdowns
By Necip Doganaksoy, Gerald J. Hahn and William Q. Meeker
Just as athletescan experience an injury that takes them out of a game, systems can experiencecomponent failures that require downtime and repair.

We will describe three approaches1 that canminimize the cost, inconvenience or potential danger of field failures throughstatistically based proactive servicing (or just-in-time maintenance).

A prime goal ofproactive product servicing is to avoid unscheduled shutdowns. It is usuallymuch less disruptive and costly to perform repairs during scheduled maintenanceon automobiles, aircraft engines, locomotives and medical scanners, forexample, than it is to deal with unexpected failures in the field.

The emergence oflong-term service agreements—manufacturers selling not just products, but alsoguaranteed ongoing levels of service to customers—has further encouragedproactive servicing and increased manufacturers’ desire to build highreliability into design in the first place.

Even when ashutdown cannot be prevented, proactive servicing is still useful to ensurespeedy, inexpensive repair, which can reduce the deleterious impact of suchshutdowns.

The threeapproaches for proactive product servicing are:

1. Optimum product maintenance scheduling.2. Proactive parts replacement.3. Automated monitoring for impendingfailures.

Optimum product maintenance schedulingMany systems,from automobiles to aircraft engines, are serviced periodically. Examples areautomobile oil changes, scheduled thermal barrier coating of turbine componentsand the replacement of filters in air conditioners.Routinemaintenance should be scheduled to provide an optimum trade-off between thecost and inconvenience of servicing, and the likely greater cost andinconvenience due to unscheduled failures that servicing could have averted.Consider anautomobile maintenance scheduling example; similar considerations apply forother systems.Automobilemanufacturers traditionally advised buyers to change oil and lubricate carparts every 3,000 miles. Using modern sensor technology, however, automobilemaintenance scheduling can be improved by taking into account how the car isoperated, considering factors such as driving speed, number of stops and numberof cold starts.Products usedmore harshly require more frequent servicing. This has led to the developmentof systems that determine the frequency of routine maintenance based on drivingand cost considerations.Dynamic maintenance scheduling, or so-calledcondition-based maintenance (CBM),2 extends thisconcept.For automobiles,CBM might involve monitoring oil degradation and other measures ofdeterioration to determine the timing of the next maintenance.The car operatoris then advised, perhaps upon car start-up, of the next recommended servicing.Proactive parts replacementMany systemfield failures occur because relatively inexpensive parts wear out, causingunscheduled shutdowns with costs that are far greater than the cost of thepart.To avoid this,vulnerable parts should be replaced with new ones at strategically selectedtimes during routine maintenance.Sometimes, animpending failure can be detected by inspecting the part or by embeddedinstrumentation, which we will address later.In other cases,you need to rely on the estimated statistical lifetime distribution of the partfor such assessment.If the hazardfunction for a part increases over time, the probability of failure of thatpart prior to the next (evenly spaced) scheduled maintenance also increaseswith time, and proactive replacement warrants consideration.For example, this would be the case for a partwith a Weibull distribution for its lifetime with a shape parameter exceedingone.3,4You might wantto replace such parts during scheduled maintenance if their estimatedprobability of failure prior to the next scheduled maintenance exceeds aspecified threshold, say one in 1,000, or alternatively, if this probability istwice its initial value.The specificplan needs to balance the cost of prematurely replacing a part against the costof its failure in service.The resultingstatistical evaluations can again be made more powerful by considering theoperating environment.For example, indeciding when to replace a part in a locomotive, the analysis should considerthe terrain (flat or mountainous) in which the locomotive will be operating.Automated monitoring for impending failuresConcept: A further strategy for avoiding or mitigating the impact offield failures is provided by new technology: the remote, and often continuous,monitoring of products using sophisticated instrumentation. Car owners havebecome acquainted with this approach through lights flashing on theirdashboards, which signal a possible reliability problem (for example, animpending engine malfunction), safety concern (door not closed properly),environmental issue (emissions problem) or, sometimes, a false alarm(instrument malfunction).Forewarning ofan impending system failure allows for repairs in a minimally intrusive andcost-effective manner—ideally, without users even being aware of the problem.At a minimum, it can help identify what parts need to be replaced and thetechnical service personnel best suited to quickly make the repair.Here’s alocomotive engine example: Modern locomotive engines are equipped with numeroussensors that read operating parameters such as oil pressure, oil temperatureand water (coolant) temperature. The resulting information is used duringnormal operation to make automatic system adjustments (for example, to controlfuel and oil flow), using control algorithms.In addition,such data might be used to identify and avert impending failures by shuttingdown the system before serious damage occurs. In light of the inconvenience andlost revenue associated with shutdowns, it is, however, imperative to minimizethe number of false alarms so the engine is shut down only when absolutelynecessary.Failure modes: Considerthe occurrence of an apparent large drop in engine oil pressure. This could bedue to any of the three reasons shown in Table 1, together with the associatedproblem severity and the action that would be taken if it were known that thiswas, indeed, the reason for the measured oil drop. These actions range fromactivating a back-up sensor to shutting down the engine.




http://www.asq.org/img/qp/082224_table1.gif

Thus, when anoil pressure drop is indicated, we need to find the reason so we can take theappropriate action. Sensor data on how oil pressure—and oil and watertemperature—are changing over time can help make this determination.

Data acquisition: Each of the three measured oil pressure drop modes shown inTable 1 were induced in a lab test to observe the resulting sensormeasurements. Figure 1 shows the readings on three sensors—monitoring oilpressure, oil temperature and water temperature—over a 30-second time period.The solid triangle marks the point at which the problem was induced, asevidenced by an appreciable change in one or more of the three measurements.

Analysis: Each of the three incidents resulted in a sudden, sharp dropin the reading for oil pressure. The accompanying changes in oil temperatureand water temperature readings, however, differed. Thus, data from the threesensors might help distinguish among the three possible reasons for theapparent oil pressure drop:

1. For the faulty sensor, the drop in the oilpressure reading was not accompanied by any discernible change in either oil orwater temperature.2. For the cooling system failure, a sharprise in water temperature preceded the drop in measured oil pressure, which wasalso accompanied by a sharp rise in oil temperature.3. For the oil pump failure, a sharp rise inoil temperature followed the drop in measured oil pressure. Water temperatureremained unchanged.

Toward an algorithm: The data in Figure 1 were from a single engine under fixedtest conditions in a lab. Similar data were obtained on different enginesoperating in the field under varying environments and involving various oilpressure drop incidents.

Results in exampleThe resultingpatterns were sufficiently distinct and similar to those depicted in Figure 1to develop a useful algorithm for the three sensor (and some further)measurements to identify the reason for a drop in the measured oil pressure.This led, in a great majority of cases, to appropriate remedial action,including the all-important decision of whether or not to shut down the system.

The analyses required the use of advancedmultivariate5 and time series6 methods. Variousother analytic methods are also used in automated monitoring for impendingfailures, including machine learning, neural nets and Bayesian belief networks.7, 8
Statisticalmethods play an important part in proactive product servicing. Such work, asalways, is conducted in collaboration with knowledgeable engineers and subjectmatter experts.

Necip Doganaksoyis astatistician and Six Sigma Master Black Belt at the GE Global Research Centerin Schenectady, NY. He has a doctorate in administrative and engineeringsystems from Union College in Schenectady. Doganaksoy is a fellow of ASQ andthe American Statistical Association.
Gerald J. Hahn is a retired manager of statistics at the GE Global ResearchCenter in Schenectady, NY. He has a doctorate in statistics and operationsresearch from Rensselaer Polytechnic Institute in Troy, NY. Hahn is a fellow ofASQ and the American Statistical Association.
William Q. Meeker is professor of statistics and distinguished professor ofliberal arts and sciences at Iowa State University in Ames, IA. He has adoctorate in administrative and engineering systems from Union College inSchenectady, NY. Meeker is a fellow of ASQ and the American StatisticalAssociation.