Repair probability

In such cases it is probably best to model repair probability as a periodic function rather than as a constant (Ref. 1). T is is discussed in the following section, "Average Unavailability with Periodic Inspection and... 

Markov models can show redundancy with different levels of redimdant components. Figure D-2 shows a system with two subsystems where only one is required for successful system operation. All failures are immediately recognized and the repair probability is modeled as a constant. 

An estimate of repair probability can be made. Consider the single component Markov model of Figure G-1. Assume that failures are immediately detected when they occur. One can accurately model the repair process with a discrete time Markov model. Using a delta t of one hour, one must estimate the probability of repair for each hour after reaching state 1. 

Assume that it is estimated that the repair probability will vary with time (non-homogeneous). Table G-1 shows a set of example repair time statistics. These statistics indicate that the repair time of a set of 64 repairs varied from one to six hours. Six units were repaired within one hour. Sixteen units were repaired within two hours. Other units took longer. Based on these numbers the average repair time is approximately 3 hours and the probability of repair in a particular hour is shown in Figure G-2. 

An alternative argument (see Ref. 2) has been made for using the value of 2 Mu. This argument recognizes that only a single repair crew is available. However, the second failure occurs only from state 1 where one component is already failed and is likely under repair. Since the repair of the first component is on average half complete, that repair wiU take only half the time to complete. When that repair is complete the Markov model is back in state 1. Therefore, the repair probability from state 2 is twice as likely than the repair rate from state 1 to state 0. This is a reasonable argument and justifies the use of 2 Mu in situations where aU failures are immediately detectable. 

Figure 11-2. Non-repair probability of a component as a function of the time available.

With the assumption that failure occurrence probability and time to repair probability have exponential distribution and maintenance costs probabihty has normal distribution (Sirok Neugebauer, 2005), then equation (17) transforms to form ... 

When the test itself is the source of a failure, the operating condition just before a test can lead either to an operating condition after the test (probability 1 ), or to a state of repair (probability y). This is modeled by the test matrix. 

Breakdown and subsequent repair is clearly non-scheduled, but gives rise to nonavailability of the item. Some non-critical items may actually be maintained on a breakdown basis, as discussed in Section 11.3. Flowever, an item which is critical to keeping the production system operating will be designed and maintained to make the probability of breakdown very small, or may be backed up by a stand-by unit. 

Space needs to be provided for the auxiliaries, including the lube oil and seal systems, lube oil cooler, intercoolers, and pulsation dampeners. A control panel or console is usually provided as part of the local console. This panel contains instmments that provide the necessary information for start-up and shutdown, and should also include warning and trouble lights. Access must be provided for motor repair and ultimate replacement needs to be considered. If a steam turbine is used, a surface condenser is probably required with a vacuum system to increase the efficiency. AH these additional systems need to be considered in the layout and spacing. In addition, room for pulsation dampeners required between stages has to be included. Aftercoolers may also be required with knockout dmms. Reference 8 describes the requirements of compressor layouts and provides many useful piping hints. 

Depolarization times greater than 48 hours should be allowed when applying the 100-mV criterion. Comparison of measured potentials is only possible where conditions of temperature and moisture are similar. Figure 19-8 shows the dependence on season and temperature which can probably be ascribed to different aeration. In repair work, potential probes should be provided at different places on the object to reduce IR errors in off potential measurements arising from equalizing currents. 

DEE eombustors have pre-mix modules on the head of the eombustor to mix the fuel uniformly with air. To avoid auto-ignition, the residenee time of the fuel in the premix tube must be less than the auto-ignition delay time of the fuel. If auto-ignition does oeeur in the pre-mix module then it is probable that the resulting damage will require repair and/or replaeement of parts before the engine is run again at full load. 

Polyesters. The main application of this material is as a matrix for glass fibre reinforcement. This can take many forms and is probably most commonly known as a DIY type material used for the manufacture of small boats, chemical containers, tanks and repair kits for cars, etc. 

In Section 2.5.4, we found the availability of a repairable emergency generator (EG) by Markov methods. If a plant requires that two identical, independent EGs must both work for time T for success. What is the probability of this Assume the failure rates are A., A., and the repair rates are p, p-,. 

The probability of the redundant system failing is the product of the single system time-dependent unavailability which includes approximately linearity between test and repair. There are two ways of performing the surveillance tests on these two systems sequential and staggered testing. 

Since dependency analysis is not needed, we can go on to the BUILD program. Go to FTAPSUIT and select 5 "Run Build." It asks you for the input file name including extender. Type "pv.pch," It asks you for name and extender of the input file for IMPORTANCE. Type, for examle, "pv.ii . It next asks for the input option. Type "5" for ba.sic event failure probabilities. This means that any failure rates must be multiplied by their mission times as shown in Table 7.4-1. (FTAPlus was written only for option 5 which uses probabilities and error factors. Other options will require hand editing of the pvn.ii file. The switch 1 is for failure rate and repair time, switch 2 is failure rate, 0 repair time, switch 3 is proportional hazard rate and 0 repair time, and switch 4 is mean time to failure and repair time.)... 

Accident progression scenarios are developed and modeled as event trees for each of these accident classes. System fault trees are developed to the component level for each branch point, and the plant response to the failure is identified. Generic subtrees are linked to the system fault trees. An example is "loss of clcciric power" which is analyzed in a Markov model that considers the frequencies of lo,sing normal power, the probabilities of failure of emergency power, and the mean times to repair parts of the electric power supply. 

In 1977. the technical press reported that a major leak from a 20,000-m liquefied propane tank in Qatar had ignited and that the resulting fire and explosion had killed seven people and caused extensive damage to the rest of the plant [18]. There had also been a leak the year before, but it had not ignited, and the tank had been repaired. The propane was stored at -42°C and atmospheric pressure. No detailed report on the incident was issued, for legal reasons, but a member of the company concerned published several papers [19-21], which gave new recommendations for t,he construction of tanks for refrigerated LFG, and it is thus possible to read between the lines and surmise what probably happened. 

Appendix III contains failure rate estimates for various genetic types of mechanical and electrical equipment. Included ate listings of failure rates with range estimates for specified component failure modes, demand probabilities, and times to maintain repair. It also contains some discussion on such special topics as human errors, aircraft crash probabilities, loss of electric power, and pipe breaks. Appendix III contains a great deal of general information of use to analysts on the methodology of data assessment for PRA. 

The event or scries of events that will initiate an accident has to be identified. An event could be a failure to follow correct safety procedures, improperly repaired equipment, or a safety mechanism The probability tliat tlie accident will occur lias to be determined. For e.xample, if a chemical plant has a 10-year life, wliat is the probability that... 

In dry air the stability of zinc is remarkable. Once the protective layer of zinc oxide formed initially is complete, the attack ceases. Even under under normal urban conditions, such as those in London, zinc sheet 0 -8 mm thick has been found to have an effective life of 40 years or more when used as a roof covering and no repair has been needed except for mechanical damage. The presence of water does, of course, increase the rate of corrosion when water is present the initial corrosion product is zinc hydroxide, which is then converted by the action of carbon dioxide to a basic zinc carbonate, probably of composition similar to ZnCOj 3Zn(OH)2 . In very damp conditions unprotected zinc sometimes forms a loose and more conspicuous form of corrosion product known as wet storage stain or white rust (see p. 4.171). 

Aqueous environments Neutral chloride solutions do not corrode titanium alloys at ambient temperatures, and smooth statically loaded specimens of susceptible alloys do not exhibit failure. In order to nucleate cracking it appears probable that the protective oxide film on alloys must be destroyed and its repair must not occur. If this breakdown occurs then cracking is observed in susceptible alloys. Consequently, the type of test and the type of specimen employed in any selected test are both important considerations, particularly in alloys exhibiting low susceptibility. 

On the other hand, the Hoar Committee s estimate for the UK did not include some significant factors, and some costs that were considered have increased in real terms since the estimates were made. Larger plants and structures are more common, and even when there is no increase in size more intensive use of equipment is demanded. As a result, the real cost of downtime or unavailability, and of dislocation to users of, for example, motorway viaducts while repairs are made, have increased appreciably. Moreover, maintenance and rectification are labour intensive activities, and hence particularly susceptible to the effects of inflation. The increases probably outweigh the savings mentioned, and the current cost of corrosion in the UK is probably around 4% of GNP. As future savings depend on the improvement being maintained despite pressures to reduce first costs, a sound economic approach to corrosion is no less important than it was in 1970. 

For a complete panel replacement, the refinisher starts with a panel preprimed in the appropriate stoving primer. For spot repairs or larger repairs without replacement of metal, there will be areas which have to be rubbed through to clean metal. Any indentations then have to be filled with a stopper or spray filler, probably based on unsaturated polyester resins and styrene, with cure initiated by mixing in an organic peroxide. After sanding, remaining bare metal areas are sprayed with a two-pack etch primer. 

The relative importance of these functions also depends to a considerable extent on the solution conditions. Under favourable conditions of pH, oxidising power and aggressive anion concentration in the solution, Function 1 is probably effective in preventing film breakdown. Under unfavourable conditions for inhibition, localised breakdown will occur at weak points in the oxide film, and Functions 2 and 3 become important in repairing the oxide film. 

If sufficient data are available, much more information can be provided when different curves for various percentages of failure are plotted. Where such data are available, reasonable design criteria would be based on some probability for failure, depending on how critical the effects of failure occur. If a large, expensive repair of a complex mechanism would result from the fatigue failure of one product, then a 10 or even 1 % probability of failure would be a more likely design criterion than the 50% suggested above. 

Although mast cells and basophils probably account for >90% of stored histamine in the body, histamine is also present in platelets, enterochromaffin-like cells, endothelial cells, and neurons. Histamine can act as a neurotransmitter in the brain. Histaminergic nerves have their cell bodies within a very small area of the brain (the magnocellular nuclei of the posterior hypothalamus) but have axons in most areas of the forebrain. There is also evidence for axons projecting into the spinal (Fig. 1) cord. Finally, there is evidence that histamine synthesis can be induced in tissues undergoing rapid tissue growth and repair. In certain neonatal tissues (e.g. liver), the rate of synthesis of this unstored diffusable histamine (termed nascent histamine) is profound and may point to a role for histamine is cell proliferation. 

---