A total of 2312 hydrocarbon release incident reports have been received, checked and input during the 9½ year period since the start-up of the Hydrocarbon Releases (HCR) Database on 1 October 1992, up to 31 March 2002 inclusive.
The following subsections comprise detailed discussion of the statistics, which are illustrated in the tables and figures in sections 5.0 and 6.0 respectively.
Caution should be exercised when interpreting these statistics, in view of the limited accuracy of population data, the voluntary nature of the information supplied on form OIR/12, and the small number of releases involved in some cases.
Figure 1, in section 5.0, shows the breakdown of reported releases by hydrocarbon type. It can be seen from this pie-chart that gas releases constituted, by far, the largest proportion of releases reported, i.e. 1267 out of 2312 (54.8%). The other types, ranked in decreasing order, were oil 403 (17.4%), non-process 269 (11.6%),
2-phase 205 (8.9%) and condensate 168 (7.3%).
The higher proportion of gas releases may have arisen from the fact that almost all offshore installations handle gas, but only 37% of installations handle hydrocarbon production liquids (see also section 3.6 - 'Installations'). Other contributing factors to the higher incidence of gas releases could be that gas is usually subjected to higher pressures and temperatures than other forms of hydrocarbon, hence there are greater challenges associated with gas containment.
Figure 2, in section 6.0, shows the monthly reporting frequencies since October 1992 up to the end of March 2002. Major milestones which occurred during the reporting period are also listed.
Reporting reached a peak of 325 in 1994/1995, probably attributable to more detailed reporting following the issue of the OIR/12 form in August, 1992, and to comprehensive guidance in August, 1993. There then followed a substantial drop down to 212 in 1995/96, ostensibly due to the safety case regime and improved safety management systems taking effect. Thereafter, the number of incidents seemed to have reached a plateau with the annual totals for the following years being 226 (1996/1997), 218 (1997/1998), 234 (1998/1999) and 234 (1999/2000). The monthly reporting rate fluctuated between 18 to 20 in the same period.
However, 2000/2001 saw a rise in the overall number of releases reported to 270 with a rise in the monthly reporting rate to 23. This increase coincided with the introduction for that year of the Progress Integrity Initiative. The figures for the latest year, 2001/2002, show a fall in the total number of releases to 241, a decrease of 11% on the previous year. There was also a fall in the monthly average back to 20. There is quite convincing evidence that the increase in 2000/01 was due to improved reporting performance and improved awareness of the reporting requirements by offshore personnel, and increased emphasis on reporting by oil company senior managers. The fall in 2001/02 may be confirmation that there is now more or less full reporting of hydrocarbon releases and the work done to reduce the incidence of major and significant releases is giving a benefit across the range. However further observation of the data over the next 4 quarters will be needed to confirm whether this is an overoptimistic interpretation of the situation.
All releases have been classified according to agreed severity criteria (see Appendix 2). Table 1 in Section 5.0 shows that, of the 2312 reported releases to 31 March 2002, 147 were classified major (6.4% of all leaks), 1284 were significant releases (55.5%) and the remaining 881 were minor (38.1%).
Figure 3 in section 6.0 shows the reporting patterns by year for major, significant and minor releases in both tabular and line graph form. The overall number of major releases in 2001/2002 continued the decrease shown in last year's figures (down from 8 to 4), and the number of significant releases also decreased (from 117 to 109) over the same period. This means that the reduction in the number of major and significant releases has now reached 19% (139 to 113) against an overall target of 50% (i.e. down to 70) by end March 2004. The overall number of minor releases reported in 2001/2002 has reversed the increase - quite possibly due to improved reporting performance by offshore personnel - experienced last year, down from 145 to 128.
Figure 4 in section 6.0 shows the reporting patterns by year for major, significant and minor gas releases in both tabular and line graph form. The upward trend in the annual number of reported gas releases ceased in 2001/2002, with a decrease of 17% over the previous year (down to 117 from 141). The number of major gas releases also went down, decreasing to 4 in 2001/2002 from 8 in 2000/2001. Similarly, significant gas releases dropped from 74 to 69 in the same period. The combined number of major and significant gas releases has decreased to 73 in 2001/2002 from 81 in 2000/2001. This is the lowest full-year figure for combined major and significant gas releases since the start-up of the HCR database. The number of minor gas releases has also shown a decrease for this year, to 44 from 60 in the previous year.
Figure 5 in section 6.0 shows the reporting patterns by year for major, significant and minor liquid releases in both tabular and line graph form. It can be seen from these that the annual numbers of reported liquid releases (i.e. oil, condensate and non-process combined) have fluctuated over the years since database start-up, with no discernible trend. Although 4 less incidents than last year, the number of minor releases at 75 is still higher than all other earlier years, apart from 2000/01. As reported earlier, time will tell if the increase was due to improved reporting performance and therefore the current fall is symptomatic of an improvement in hydrocarbon containment across the board.
Figure 6 in section 6.0 shows the reporting patterns by year for major, significant and minor 2-phase releases in both tabular and line graph form. From these, it can be seen that the numbers of reported 2-phase releases have also fluctuated over the years since database start-up. As with liquid releases, there is no discernible trend.
The barchart in figure 7, section 6.0, shows the failure rates in leaks per system year for the highest ranked 15 release rates out of a total of 52 system types.
The system with the highest failure rate is gas compression with a failure rate of 2.90 x 10-1 leaks per system year. High operating temperatures and pressures, vibration, and the consequent effects of these on vulnerable equipment items such as seals, instruments, and small bore pipework items etc. are considered to be major contributors to the frequency of hydrocarbon releases from this system type, and these factors have been taken into account in the current Process Integrity initiative (Appendix 1).
Next highest ranking system failure rates were oil export (1.56 x 10-1 leaks per system year), and fuel gas (1.38 x 10-1 leaks per system year).
The barchart in figure 8, section 6.0, shows the failure rates in leaks per equipment year for the highest ranked 15 release rates out of a total of 119 equipment types.
The major equipment items with the highest failure rates are dual fuel turbines at 7.24 x 10-2 leaks per equipment year, and reciprocating compressors at 6.52 x 10-2 leaks per equipment year, both of which indicate strong links with the high gas compression system leak rates shown above.
It should be noted that the equipment items having the highest numbers of reported leaks were instruments (370 total), and pipework items such as valves, flanges and piping. Their respective population numbers are also very large and thus their respective failure rates in the rankings are reduced. However, the problems (with small bore fittings in particular) have been recognised and guidelines have been produced for addressing loss of containment, etc.
Table 2 in section 5.0 summarises the breakdown of all reported releases by hydrocarbon type versus installation type, i.e. Fixed which includes floating production systems (FPS), Mobile including both drilling and accommodation units, and Subsea.
Each installation also has an associated location in either the Northern (above 59° Latitude North), Central (56° to 59° Latitude North), or Southern (below 56° Latitude North, including Irish Sea and English Channel) areas of UKCS. The total current population and the estimated number of installation years (as at 31 March 2002) is also given for each installation type on the UKCS by area.
The breakdown by installation type showed that fixed installations were responsible for 2229 (96.4%) of releases reported. In contrast, only 83 releases (3.6%) were reported by mobile drilling and accommodation installations. These proportions are not considered unusual, because there is constant exposure to hydrocarbons on fixed installations whereas the exposure on mobile installations is intermittent.
Failure rates have been calculated in terms of leaks per installation year, for each type of installation, by dividing the total number of leaks for that type by the total number of installation years.
Floating Production Systems (including FPSO and FSU) had the highest overall leak rate of 2.57 leaks per installation year, with 5.3% major releases. Fixed attended installations in the Northern North Sea had the next highest leak rate of 2.57 leaks per installation year, with 4.5% of these being major releases. Central fixed attended installations were next, with 2.26 leaks per installation year, with 4.9% major releases. Southern fixed attended installations had a leak rate of 0.37 leaks per installation year, with 8.1% major, and Southern normally unattended installations (NUI) leak rate was 0.15 leaks per installation year, but with 20.8% major releases.
The production equipment on the larger attended oil production platforms in the Northern and Central areas is more complex and subject to much greater fluctuation of operating parameters than on the simpler and smaller gas production platforms in the Southern area. It is reasonably foreseeable, therefore, that there may be a greater propensity for hydrocarbon releases of all types on the larger installations.
Using the population data held in the HCR database, leak rates have been calculated in terms of leaks per installation year, for each type of mobile installation. However, in the absence of accurate rig years data, it was necessary to assume that each unit was present on the UKCS for the entire period since their arrival, and so these leak rates must be considered very approximate (and tending toward best case) as a result.
Combining Southern and Central release figures gave an overall leak rate for jack-up mobile drilling and accommodation units of 0.34 leaks per installation year, with 7.0% major releases. Northern and Central release figures combined gave an overall leak rate for semi-submersible mobile drilling and accommodation units of 0.25 leaks per installation year, but with 17.5% major releases.
Table 3, section 5.0, summarises the numbers of ignitions for each year versus the number of releases for each type of hydrocarbon. It is important to note that the percentages shown are ignition rates (i.e. percentage of releases ignited) and NOT ignition probabilities (i.e. the likelihood of a particular release igniting).
There was a grand total of 143 reported ignitions (i.e. 6.2% of all releases) over the 9½ year period, none of which were major.
There were 42 gas ignitions (29.4% of all ignitions), of which 17 were classed as significant releases (ignition rate 2.1%), and 25 as minor (ignition rate 7.0%). There were no major ignited gas releases.
There were 101 liquid ignitions (69.5% of all ignitions) of which 24 were classed as significant releases (ignition rate 7.3%), and 77 as minor (ignition rate 15.5%). There were no major ignited liquids releases.
There have been no reported 2-phase ignitions to date.
The higher ignition rates in the case of minor releases, and liquids releases in particular, is considered to be due to the fact that reports of unignited minor releases cover only those reportable under RIDDOR definitions based on the potential for fire/explosion, whereas ALL ignitions are reportable no matter how minor.
Table 4, section 5.0, comprises three parts, which contain full details of every ignition reported in the 9½ years to 31 March 2002.
The three parts cover the following details :
Each individual record has a unique number and so it is possible to read all three parts for one ignition incident to obtain the full picture. The purpose of this report is to provide statistics only, so any analysis of ignitions has been excluded here.
Table 5 in section 5.0, summarises the modes of detection employed during an incident versus the type of hydrocarbon and severity of release. The modes comprise heat, smoke, flame, gas and 'other'. It should be noted that more than one means of detection may be reported for any one release, and so the totals for each column may be more than the total number of releases reported.
There were 2471 detection modes connected with the total 2312 reported releases, more than one mode being effective on some releases. Gas detectors detected 41.6% of all releases (75.9% of gas releases), and the remaining releases were mainly detected by means other than equipment designed for the purpose (see 3.8.2 below).
A breakdown of the 'other' category which includes detection by visual means, by sound, by smell etc., is shown in table 6, section 5.0. It is considered important to have a separate table for these, since they have made a substantial contribution to the overall detection of offshore releases.
Of the total 2471 modes, 1411 (57.1% of modes) involved means of detection other than by dedicated detection systems. Of these 1411 'other' means of detection, 1089 (77.2%) were visually detected, of which 51 (4.7% of visual detection) involved major releases, 569 (52.2% of visual detection) involved significant releases, and 747 (43.3%) minor releases.
Specific detection details for all reported ignitions are included in table 4 (c).
Table 7, section 5.0, shows a breakdown of the various types of emergency actions taken against the types and severity of releases involved. These actions include either automatic or manual initiation of shutdown, blowdown, deluge, and/or CO2/Halon systems, plus whether any musters and/or any other emergency actions were carried out.
Two-thirds of releases (1523) required shutdown action (461 automatic, 1062 manual), of which 107 (7% of shutdowns) were for major releases. Blowdown operated on 724 (295 auto, 429 manual = 31.3% of all releases) of which 66 (2.9% of blowdowns) were for major releases.
There were 625 musters (582 at stations, 43 at lifeboats = 27.0% of all releases), 60 of which (2.6% of musters) were for major releases.
In contrast, deluge only operated on 50 releases (2.2% of all releases), of which 13 (26% of deluge) were for major releases, and only 59 (2.6% of all releases) involved the use of CO2/Halon systems, of which only 1 was for a major release.
Table 8, section 5.0, summarises the breakdown of all types of release into the four main causation categories (i.e. design, equipment, operational and/or procedural faults) plus details of the mode of operation in the area at the time of the incident. It should be noted that more than one category may be involved in any one incident, and so caution should be exercised when interpreting the figures shown.
The most frequently reported factor was equipment fault at 1532 (66.3% of incidents). Operational faults were next with 1227 (53.1% of incidents), procedural faults 640 (27.7% of incidents), and design faults 348 (15.1% of incidents).
The proportion of releases occurring during normal production was 49.1% compared to that during intervention type activities such as start-up/reinstatement (17.9%), drilling/workover (9.5%), and maintenance/construction (9.5%). These proportions have generally been maintained throughout each year since database start-up in October 1992.
The causation factor involved in most incidents was 'mechanical failure' which occurred on 1034 occasions (44.7% of all incidents), followed by 'improper operation' that occurred on 549 occasions (23.7% of all incidents).