Professional Documents
Culture Documents
Oreda 2009 Topside Equipmentpdf Compress
Oreda 2009 Topside Equipmentpdf Compress
re oo Tb ee ee re i Figure 5- julti-sample problem ‘To mesge ail the samples, and estimate the “average” fatlure rate as the total number of failures divided by the aggregated time in service will not always give an adequate result, ‘The ‘eonfidence’ interval will especially be unrealistically short, as illustrated iat Figure 5. We therefore need a more advanced estimation procedure to take carc of the multi-sample problem. Below, the so-called OREDA-cstimator of the “average” failure rate in a multi-sample situation is presented together with a 90% uncertainty interval. Spjotvoll (1985) gives rationale for the estimation procedure, ‘The OREDA-estimator is based on the following assumptions; © We have k different samples. A sample may e.g. correspond to an installation (facility), and we may have data ftom similar items used on k different installations, © In sample no. # we have observed n, failures during a total time in service 1, for PDs k Satnple no. j has a constant failure rate 2, f0F 1=1,2e 00k: Due te different operational and environmental conditions, the failure rate 2, may vary between the samples “The variation of the failure rate between samples may be modelled by assuming thatthe failure rate is a random variable with some distribution given by a probability density funetion 2). The mean, of “average” failure rate is then: @= [4-a() da and the varianee is: o* = [(a-By' (A) da 2 OREDAVolume 1 - Topside Equipment 28 OREDA-2009 ‘To calculate the multi-sample OREDA-cstimator, the following procedure is used: 1. Calculate an initial estimate @ of the mean (“average”) failure rate @ by pooling the data: Tonal noo files _ 2" ‘Total time in service “GS goat hd, r XS; when greater than 0, else 4, Calculate the final estimate # of the mean (“average”) failure rate @by: 5, Let SD= 6 Jn the data tables in Part IL of the handbook @ corresponds to the mean (column 4), and SD corresponds to the standard deviation (column 6). ©. OREDAOREDA-2009 28 Volume 4 ~ Topside Equipment “The Jower and upper “uncertainty” values are given by: oer J 704) d4=90% Since the distribution a4) is not known in advance, the following pragmatic approach is used: 6. (A) is assumed to be the probability density function of a Gamma distribution with parameters crand 7. The parameters cand are estimated by: B a@ 8 where cosjy aNd coos, denote the upper 95% and 5% percentiles, respectively, of the 7f-istribution with wdegrees of freedom, see Table 6, page 33. In situations where vis ‘not an integer, an interpolation in the z*-distribution is performed, Wore 1 More detailed analysis of the OREDA data (see Vatn 1993) has indicated that there may be 8 large variation between installations. The multi-sample OREDA estimator should therefore as a rule be used instead of the m/z estimator which assumes a homogeneous sample. The variation between the samples (installations) is measured by the standard deviation SD. Both the standard deviation atid the lower and upper uncertainty values describe the distribution over the failure rate, i.e. a2), The uncertainty interval should not be mixed up with a confidence interval. A confidence interval decreases with more data, ‘which is not the case for the uncertainty intervals given in this procedure, oREDAVolume 1 - Topside Equipment 0 OREDA-2009 Note 2, In the case of &~ 1, the procedure cannot be used. In this case the n/restimate is given for the mean, and the Jower and upper valves should be interprcied as a traditional 90% confidence interval Nowe 3: ‘The subdivision of the equipment in detailed taxonomy classes as shown in the data tables Will in several cases result in a very low population and/or no. of installations. In these cases it is obvious that the confidence in the data will be very low, The cimalyst must therefore examine the trade-off between data relevance and population size. We therefore recommend in these cases to use the given data with caution and find additional means to verify the confidence of the data, Some ways this can be done are by using/comparing data from a higher taxonomy level, checking other reliability data sources andor using expert judgement, Note 4: If no failures are observed for an item, the following approach is used to obtain lower, ‘mean and upper values for “All failure modes”: 1 Let 4, denote the failure rate estimate (“mean”) one level up in the wxonomy: hierarchy 2. Lot rdenote the total time in service (operational or calendar) for the item of interest 3, Let 4. 5. The standard deviation is given by 6. A 90% uncertainty interval is given by © OREDA3a Volum Altematively, the term “All failure modes" can be replaced with the failure mode of interest. ESTIMATION OF DEMAND PROBABILITIES If information about “number of demands” is given (see Section “Data Table, Reliability Data”, page 35) itis possible to estimate the demand probability, The demand probability is always related to one specific failure mode, for example critical fail to start. The maximum likelihood estimator of the demand rate 0 p= where » is the number of failures with the appropriate failure mode, and dis the number of demands, Note that in the data table presentations the demand probabilities may apparently look different, The teason for this is that in some cases there are registered “demand failures”, but the number of demands is not recorded for one or more inventories. For these inventories, the demand failures are not included in the total number of demand failures for rat data table. a ESTIMATION OF REPAIR TIMES Ih OREDA, data on three maintenance related ‘times’ are collected, ie. dows tinte, active repair time and restoration man-hours. The difference between the first two measures is illustrated in Figure 6. Figure 6 = Maintenance times (ISO 14224) Down rime includes the calendar time trom the equipment is being stopped until itis tested after a repair and ready for being reconnected to its intended service. This parameter is not included in the handbook, only in the OREDA database. @OREDAVolume 4 = Topsic ment OREDA-2009 Active maintemance’rematr time is the calendar time during which maintenance work is actually performed on the item, irrespective of the number of persons that may work in parallel. Data on this is scarce in the OREDA database and should therefore be used with, same caution in reliability analysis, The active repair time is estimated as the mean (average) af the available active repair times, Together with the mean, also the maximum value is presented, giving an indication of an upper limit of the active repair time, It should ‘be noted that this value becomes more robust as the sample size increases. In situations With few observation (times), @.g. less than 10, the maximum value should be applied with carefulness. Restoration man-howrs are the total man-hours used to restore an item to operational status after failure. These data do normally have a high fill-in degree in OREDA and estimates are given by mean (average) and maximum values. Note that an alternative presentation of repair data is by parameters in the lognormal distribution, which is commonly used as a distribution for repair time, ref. e.g. Rausand and. Hoyland (2004) ‘WeicuTinc OREDA-2009 Dara wiTH OTHER DATA Sources Ip many RAMS analyses, data may also be available from other sources than this hhandbook. For offshore RAMS analyses, the most obvious data source in addition to this book, is the previous handbooks. A method for weighting data from e.g. OREDA-2002 and OREDA-2009 is given below, The method is based on an approach suggested in the OREDA Dats Analysis Guidelines (Vain 1993), The calculations are repeated for all failure modes of interest. Let Ay denote the mean failure rate for Phase VI and VII data (column 4 of the data table) in the OREDA-2009 handbook. Further, let SDy denote the standard deviation ($D) in column 6 of the data table. do is the corresponding mean in the OREDA-2002 handbook. A weighted failure rate ‘estimate is given by: _asa(g se Ale, 4orag(¥f, seals, . SD, ) where [s| dehotes the absolute value of x. BOREDAOREDA-2008 33 Volume 4 = Topside Equipment PERCENTAGE POINTS OF THE CHI-SQUARE DISTRIBUTION ‘Table 6 —Percentage Points of the Chi-square (f') Distribution FIZ > Fay) Via 0.995 0.990 0.975 0. 005 025 1 0.00 0.00 000 3S. 2 O01 0.02 00S O10 599 738 10.60 3 007 O11 022 035 781 935 4 021 030 048 O71 949 1114 14.86 5 O41 055 O83 LIS ILO? 1238 16.75 6 O68 087 124 164 1259 1445 18.55 7 099 124 169 217 1407 16.01 20.28 8 134 Les 218 1351 17.583 21.96 9 173 200 2.70 1692 19,02 23.59 1 216 256 325 394 1831 20.48 235.19 1 260 3.05 457 19.68 21.92 26.76 12 «307 357 523 2103 23.34 28.30 30437 41 589 2236 24,74 29.82 4 407 4.66 657 23,68 26.12 3132 IS 4.60 5.23 726 25.00 27.49 32.80 6 514 5.81 796 26.30 28.85 34.27 17 570 641 867 2759 30.19 35,72 18 626 7.01 823 930 2887 31.53 37.16 19 684 7.63 891 1012 3014 32.85 38.58 20 743 (826 959 1085 SL4l 34.17 40.00 21-803 8.90 1028 1159 3267 35.48 41.40 22 8G 954 1098 1234 33.92 36.78 42.80 23° 926 10.20 1169 13.09 3517 3808 4164 44.18 24 989 10.86 1240 1385 3642 39.36 42.98 45.56 25 1052 11.82 1312 1461 37.65 40.65 4431 46.93 26 11.16 12.20 1384 1538 3889 41.92 45.64 48.20 27 1181 1288 1457 1615 40.11 43.19 46.96 49.64 28 1246 13.56 1531 16.93 41.34 44.46 48.28 50.99 29 1312 14.26 1605 17.71 42:56 45.72 49.59 52.34 30 13.79 14.95 16.79 1849 43.77 46.98 50.89 53.67 40 20.71 2216 2443 2651 55.76 59.34 63.69 66.77 $0 27.99 29.71 3236 34.76 67.50 71.42 76.15 79.49 60 35.53 3748 4048 4319 79.08 83.30 8838 91.95 10 43.28 45.44 48.76 S174 90.53 95.02 10042 104.22 80 S117 S354 STIS 60.39 101.88 106.63 112.33 11632 90 $9.20 61.75 65.65 69.13 113.14 T1814 12412 12830 1006733 70,06 74.22 7793 124.34 129.56 _135.81_140.17 ooREDAVolume 4 - Topside Equipment 34 ‘OREDA-2009 (This page ineetionaty ben bari ORENOREDA-2008 35. ‘Volume 4 - Topside Equipment TOPSIDE DATA TABLE FORMATS: Dara TaBLe, RELIABILITY DATA Each data table contains an identification of the item and the estimated reliability Parameters. The figures provided should be intespreted on the basis of the assumptions specified in the boundary definition for cach equipment category and the estimation method applied. The format of the data tables is shown in Figure 7. Taaanamy ee Te Popaston | Wailnions “Ragas a nar PA Wao domanee ‘alongs ree Tale med Toot Tal ab ee 1 Rou Teiverep as | Wana tauree|Cowar [ Wan [Upper [80 [ahr | Moan [ Mx [Ween [ee Figure 7 - Fos of the reliability data tables ‘The various entries of the data table are explained in the following: Taxonomy nuinber and Hei Numerical identification of the equipment unit. The taxonomy number has. been chosen to comply with previous handbooks. Population Tota] number of equipment units forming the basis for the reliability estimates. Installations ‘Total number of installations (facilities) covered by the data surveillance for the equipment unit. OREDAVolume 1 - Topside Equipment 6 OREDA-2008 Axgregated tine In service ‘Two types of time scales are presented as the basis for the failure rate estimates; Calendar ime anc Operational time. The aggregated time in service for the total population is given for both time scales. Note that while the calendar time is given with high certainty, the operational time has in many cases to be based on estimates (by the data collector), Tt will also be noted that for some equipment there is great differences between calendar time and operational time. This is typical for equipmient being oceasionally started and operated for some time only (e.g. fre pumps) Equipment being tainly used for on/off functions (e.g. valves) are considered to be in an operational state as long as they are in a status ready to be used on demand. Number of demands The accumulated number of demands/cycles for the total population is given when available. In several cases these numbers. are based on estimates and not accurate measurements. Failure mode / Severity class ‘This column contains a brief description of the manner in which the failure occurred, when such information is available. The failure modes are given for each Severity class, i.¢, Critical, Degraded, Tacipient and Unknown. Number of failures Total number of recorded failures for each faiture mode. The accumulated number of failures is presented as “All modes”, Failure rate Estimated failure rate is given both under the “multi-sample” assumption, and the assumption of homogeneous data sets. In the “multi-sample” situation the failure rate is assumed to vary between installations, and cach installation represents one sample. Assuming homogenous data sets, the failure rites of various installations are identical. The following entries are incinded: Mean ‘The failure rate obtained by the OREDA estimator, i.¢. the mean failure rate among the installations for which data have been collected, dower, Upper) A 90% uncertainty interval for the failure rate, ie. an interval covering ‘90% of the variation between the multiple samples. SD A standard deviation indicating the variation between the multiple sampies. we ‘The total number of failures divided by the total time in serviee, ie, the failure rate estimate we would use for a homogeneous sample. All the entries are measured per 10° hours and refer to calendar time. e-OREDAOREDA-2008 a ‘Volume 1 = Topside Equipment Note: Failire rate values greater than or equal to 0,01 (per 10° hour) are given with two decimals places and failure rate values less then 0.01 (per 10° hour) and equal to or greater than 0.0001 (per 10° hour) are given on “E-* format, ¢.g. 1E-4 means 1.010" = 0.0001, For very. small failure rates, there is little meaning in presenting exact estimates. Therefore failure rates less than 1E-1 (per 10° hour), ie. failure rates with magnitude 10" or less are marked with Aesive repatr time (hours) ‘The active repair time columns present the mean and maximum calendar time (hours) required to repair and retum the item to a state where it is ready to resume its functions, Active repair time is the time when actual repair work is being done, It does not include time to shut down the unit, issue work order, wait for spare pars, start-up after repair etc. The active repair time is therefore normally shorter than the dawatime Whee some of the activities indicated bove may be included. Repair values that are estimated from 10 or less observations are marked with “+ Note. During the data collection exercises it has been very difficult to obtain data regarding active repair times, In the OREDA database there is @ good coverage of “restoration man-hours” data, whereas the data for “active repair time” is rather sparsc. It should also be noted that active repair hours are highly influenced by how maintenance is organised an the facility. ‘The figures for active repair times should therefore only be used as an indication of what the actual active repair times would be. It is highly recommended to use some kind of ‘expert judgement in addition to the values given in the handbook. ‘Manhours (repair) ‘The repair columns present the mean (average) value and the maximum value of the total repair times (manhours). The repair manhours represent the total sum of manhours spent by the maintenance personnel for the repair work. Repair manhours that are estimated from 10 or less observations are marked with "+", Note: Repair values are given with one decimal (for values < 10) or as integers (for values 2 10). ‘This has been found to be the most readable format the space limitation of the columns being considered. Comments When data have been available, the on-demand failure probability is given in the Comment field, DATA TABLE, MAINTAINABLE ITEM VERSUS FAILURE MODE The reliability data presented in the data table in Figure 7 p. 35, does not give information on which part of the equipment that bas failed, In the Motntainable item versus Failure ‘mode listing the relative contribution from each maintainable item to the total failure rate may be obtsined, The figures in the table represent percentages of occurrence for cach ©OREDAVolume 4 = Topside Equipment 38 OREDA-2009 combination of failure mode and maintainable item. The row sum represents the total percentage of failures that are related to the actual maintainable item. Note that several maintainable items might be assigned to each failure record, In such situations, the “score” for the actual maintainable itemfailure mode combination is set to U/n, where m is the number of maintainable items listed for that failure record. The column sum represents the Contribution for each failure mode in percentages. ‘This information is valuable input to an FMEA‘FMECA analysis. The FMEA/FMECA analysis is further a major part ofa reliability centred maintenance (RCM) analysis. As the RCM methodology focuses on failure causes, it is also important to have information regarding failure causes as discussed in the next section. Note that several maintainable items might be assigned to each failure reeard. Note 1 ‘The data tables ‘only those failure modes where failures ate registered Note 2: For some of the equipment classes the tables Mainiainable items versus Failure mode are split on two or more tables depending on the number of different failure modes. The column Maintainable item is then repeated after each split. Note 3: ‘The values in the tables are given with two decimals, This is done in order to present data in a. common format. It should be kept in mind that such accursey in the figures result from the calculated percentage distribution in the collected data, and do not reflect the real accuray in the collected data, DATA TABLE, FAILURE MECHANISM VERSUS FAILURE MODE In the Failure Mechanism versus Failure Mode listing, the relative contribution from each Failure Mechanism (cause) to the total failure rate may be obtained, The figures in the table represent percentages of occurrence for each combination of Failure Mechanism and. Failure Mode, The row sum represents the total percentage of fallures that are related to the actual Failure Mechanism. The column sum represents the contribution for each Failure Mode in percent. ‘As mentioned above, the information about failure causes is essential in an RCM analysis. For example scheduled replacement of units is only applicable if one or more failure causes may be related to ageing, wear, corrasion etc. Note “The data tables list only those failure modes where failures are registered. Note 2: For some of the equipment classes the tubles Failure mechanism versus Failure mode ate split on two or more tables depending on the number of different failure modes. The ‘column Failure mechanism is then repeated after each split, BOREDAOREDA-2009 39 Volume 1 - Topside Equipment Note 3: ‘The values in the tables are given with two decimals. This is done in order to present data in 4 comirion forntat, It should be kept in mind that such accuracy in the figures result from the calemlared percentage distribution in the collected data, and do not reflect the real accuracy inthe collected data. ‘BoREDAVolume 1 - Topside Equipment 40. OREDA-2009 (This page interionaly let ian). BOREDAOREDA-2009 4 ‘Volume 1 - Topside Equipment DEFINITIONS® The main terminology used in the OREDA-2009 handbook is defined in this section. The specific definitions of the terminology and parameters used in the statistical estimation procedures are included in the section "ESTIMATION PROCEDURES" p, 23. Terms marked with (C) are categorised in pre-defined codes. Active repair time ‘That part of the maintenance time during which a maintenance action is performed on an item, either automatically or manually, excluding logistic delays. This excludes the time to detect the failure, time to isolate the equipment from the process before repair, delay and waiting for spare parts or tools, and any time after the repair has been completed ifthe item isnot put into service immediately. Tinie for testing is included when such testing is an integrated part of the repair activity. For a more detailed description and interpretation of maintenance times, see Figure 6,p.31 Boundary ‘The interface between an item and its surroundings. Catendar time ‘The interval of time between the start and end of data collection for a particular item. Equipment unit ‘The highest indenture level including subunits and smaller entities belonging to that equipment nit, Fallure The termination or the degradation of the ability of an item to perform its re function(s) (see below). It inetudes: © Complete failure ofthe item. © Failure of part of the item that causes unavailability of the item for corrective action. © Failure discovered during inspection, testing, or preventive maintenance that requires repair, = Failure on safety devices or control/monitoring devices that necessitates shutdown, or reduction of the items capability below specified limits, ‘The following outages are not considered as faites: © Unavailability due to preventive.or planned maintenance, ‘© Shutdown of the item due to extemal conditions, or where no physical failure condition of the item is revealed. A shutdown is not to be considered a failure unless there is some recorded maintenance activity. ‘Defiitons related to this Handbook: wil also be found in the standards ISO 14224 and ISO .20815, ©OREDAVolume 1 - Topside Equipment 42 OREDA-2009 ‘A required function is defined as any fimction necessary to maintain the item's capability of providing its output at specified capacity and quality. Note that a failure could be either Complete loss of function or function degradation below an azceptable limit. A filure will normally require a work order and involvement by maintenance person. Failure mechanisnt (C) ‘An attribute of the failure event that can be easily deduced technically. The failure mechanism is the apparent, immediate cauve of the failure and is related to the fowest level in the hierarchy where it can be identified. Failure mode () ‘The effect by which a failure is observed on the failed unit. The failure modes describe the loss of required system function(s) that result from failures, or an undesired change in state or condition. The failure made is related to the equipment unit level. The failure mede is a description of the various abnormal stotes/conditions of an equipment unit, and the possible transition from correct to incorrect state. Note that the failure mode can be grouped in three main categories (ref. ISO 14224): 4) desited function is nor obrained (e.g. fail to stan): bb) specified function fost or outside accepted operational limits (@.. spurious stop, high output); ) 1 failwre indication is observed, but there is no immediate and critical impact on equipment unit function. These are typical non-eritical failures related to some degradation or ineipient fault condition (e.g, initial wear). Hem A common term used to denote any level of hardware assembly; Le. equipment unit, subunit, maintainable items and parts. Maintainabte trem (C) -An item that constitutes an assembly of parts that are normally the lowest indenture level during maintenance, Number of demands ‘The total number of times an item is required to perform its specified function(s) during the calendar time. The number of demands shall include beth test demands and demands from safety or process activation ‘Operational time The period of time during which a particular item performs its required function(s) between, the start and end of deta surveillance, Equipment being mainly used for ov/off functions (e.g. valves) are considered to be in’ an ‘operational state as long as they are in a status ready to be used on demand. @OREDAOREDA-2008 43 Volume 1 — Topside Equipment Population The total number of items of one particular type im service during the period of the event data surveillance. ‘Sample ‘The group of items of one particular type in service - described by its taxonomy code - on one instaltation during the period of the event data surveillance. ‘Severity Class Fypes (C) Each failure is categorised within one out of the four following severity classes: CRITICAL FAILURE: & failure which emuses immediate and complete loss of an equipment unit's capability of providing its output. DEGRADED FAILURE: A failure which is not critical, bur it prevents an equipment unit from providing its output within specifications, Such a failure would usually, but not necessarily, be ‘gradual or partial, and may’ develop into a critical faihure in time, INCIPIENT FAILURE: A failure which does not immediately cause loss of a unit's capability of providing its output, but which, if not attended to, could result in a critical or degraded failure in the near future. UNKNOWN; Failure severity was not recorded or could not be deduced. ‘The severity class is used to describe effect ion operational status and the severity of loss of output froin the unit, Each failure has been associated with only one severity class, critical, degraded, incipient or unknown, independently of the failure mode and failure cause. The severity classification is confined to the lacation and use of the equipment wait that has failed. Subunis (C) An assembly of items that provides a specific function that is required for the equipment Lunt to achieve its intended performance. Corresponds frequently with sub-tag number(s). Taxonomy (C) A systematic classification of items into generic groups based on factors possibly common to several of the items. e.g. functional type or medium handled, OREDA