Trends in risk levels - Norwegian Continantal Shelf

Transcription

1 Trends in risk levels - Norwegian Continantal Shelf SUMMAY EPOT PHASE OD-03-08

2

3 Trends in isk Levels on the Norwegian Continental Shelf Summary eport Phase

4 (Blank page)

5 apport EPOT TITLE Trends in isk Levels on the Norwegian Continental Shelf Phase 3 Summary eport 2002 CLASSIFICATION Public Publication exempted Limited Confidential Strictly confidential EPOT NUMBE OD AUTHO/COODINATO Norwegian Petroleum Directorate OGANISASJONAL UNIT D-isk SUMMAY APPOVED BYDATE Øyvind Tuntland Director The purpose of the project is to establish and evaluate status and trends for the levels of risk on the Norwegian Continental Shelf. The project is based on two complementary evaluation processes: The registration, analysis and evaluation of data relation to defined situations of hazard and accident and the efficiency of barriers The performance of sociological analyses using free text analysis of questionnaires, workshops and interviews On the basis of the available data and the indicators used in this project, the increase in risk levels witnessed in 1999 and 2000 would seem to have been reinforced in 2002, following a temporary improvement in There are few indicators which point to a clear and positive reduction, and a number of important contributory factors to major accident risk which show a significant increase. The frequency of serious cases of injury on the installations remains at the average level over the last 10 years. KEYWODS isk, HMS, NCS POJECT NUMBE NUMBE OF PAGES EDITION POJECT TITLE Trends in isk Levels on the Norwegian Continental Shelf

6 (blank page)

7 Foreword Trends in risk levels on the Norwegian Continental Shelf are not only a matter of concern to everyone involved in the petroleum industry but are also of interest to the public at large. It was therefore a logical and important step for us to establish a system for measuring the impact of the total safety effort in the industry. Against this background, the Norwegian Petroleum Directorate launched a project in 1999/2000 to evaluate trends in risk levels on the Norwegian Continental Shelf. The initial phases of the project proved that the chosen method was capable of establishing a picture of the situation. Phase 3 of the project has continued the work of previous phases and extended the limits of the elements measured. Phase 3 includes for example expanded risk indicators for helicopter transport of personnel. These indicators cover all relevant transport of personnel and have been drawn up in close cooperation with the Civil Aviation Authority Norway and the helicopter companies operating on the Norwegian Continental Shelf. Our industry has a high level of competence in the field of safety. We have endeavoured to benefit from this competence by making the process an open one and inviting key experts from oil companies, consultancy firms, research and education to contribute to the project. Objectivity and credibility are key words in promulgating any serious views on safety. The project results have been presented to the Safety Forum in which both employees and employers organizations are represented. Their response so far has been positive and constructive and hopes have been expressed that this work will contribute to the building of a common platform for improving safety. We must bear in mind that the project is still in an early phase and that to the best of our knowledge it is the first attempt to measure risk levels for an entire industrial sector in this way. We have also been subject to constraints in regard to time and availability of information. The results must therefore be approached with a degree of caution. It is our intention, however, that the project should lead to the regular measurement of trends, with a resultant improvement in quality. There are many people both inside and outside the project group who have contributed to the completion of the project. It would take too long to list them all but I should like to mention the positive response the project has received in all its activities and communications aimed at gathering views and examples. I should also like to take this opportunity to thank our consultants and the HES expert group for their particularly valuable contribution. Stavanger, 24 April 2003 Øyvind Tuntland Assistant Director

8 (Blank page)

9 Table of contents PAT 1: PUPOSE AND CONCLUSIONS PUPOSE AND LIMITATIONS PUPOSE OBJECTIVES IMPOTANT LIMITATIONS CONCLUSIONS QUESTIONNAIES AND INTEVIEWS ISK INDICATOS ISK INDICATOS, SEIOUS OCCUPATIONAL ACCIDENTS ON INSTALLATIONS QUALITATIVE ANALYSES GENEAL CONCLUSION... 4 PAT 2: IMPLEMENTATION AND SCOPE IMPLEMENTATION MAIN PINCIPLES IMPLEMENTATION OF PHASE 3 OF THE POJECT USE OF ISK INDICATOS TENDS IN ACTIVITY LEVEL POJECT DOCUMENTATION SCOPE INTEVIEWS WITH KEY INFOMANTS ANALYSIS OF STATISTICAL ISK HELICOPTE TANSPOT EXPANSION IN DEPTH... 9 PAT 3: ESULTS FOM PHASE 3 OF THE POJECT STATUS AND TENDS DFU12, HELICOPTE EVENTS OVEVIEW OF NEW INDICATOS ACTIVITY INDICATOS EVENT INDICATOS STATUS AND TENDS INDICATOS FO MAJO ACCIDENTS ON INSTALLATIONS DFUS ELATED TO MAJO ACCIDENT ISK ISK INDICATOS FO MAJO ACCIDENTS TOTAL INDICATO FO MAJO ACCIDENTS STATUS AND TENDS BAIES AGAINST MAJO ACCIDENTS BAIES IN THE POCESS AEA GENEAL EVALUATION BAIES AGAINST STUCTUAL FAILUE INTEVIEWS, WOKSHOPS, QUESTIONNAIES SUMMAY OF INTEVIEWS WITH KEY INFOMANTS WOKSHOPS AND ANALYSIS OF FEE TEXT IN QUESTIONNAIES STATUS AND TENDS OCCUPATIONAL ACCIDENTS OCCUPATIONAL ACCIDENTS, PODUCTION INSTALLATIONS OCCUPATIONAL ACCIDENTS, MOBILE UNITS OTHE FACTOS OTHE INDICATOS DFU21 FALLING OBJECT NOTIFICATION OF UNPLANNED EVENTS OTHE DFU S ECOMMENDATIONS FO FUTHE ACTION DEFINITIONS AND ABBEVIATIONS DEFINITIONS ABBEVIATIONS EFEENCES...33

10 Overview of tables Table 1 Overview of DFUs and data sources... 7 Table 2 Emergency preparedness aspects Overview of figures Figure 1 Trends in activity level, production... 7 Figure 2 Trends in activity level, exploration... 8 Figure 3 Volume of crew change traffic, flight hours and person flight hours, Figure 4 Volume of shuttle traffic, flight hours and person flight hours, Figure 5 Event indicator 1, per person flight hours, Figure 6 Event indicator 2, per person flight hours, Figure 7 All reported DFU s divided by category Figure 8 Number of hydrocarbon leaks exceeding 0.1 kg/s Figure 9 Trend, leaks, normalised against installation age, all production installations...16 Figure 10 Average leak frequency per installation year, Figure 11 Contribution to leaks from normal operation and process system intervention, Figure 12 Well kicks according to scale of gravity per 100 wells drilled, for exploration and production drilling Figure 13 Number of major events and damage to structures and marine systems Figure 14 Major events for mobile installations per 100 million manhours Figure 15 Total indicator, production installations, normalised in relation to manhours Figure 16 Total indicator for production installations, events susceptible of influence and external threats Figure 17 Total indicator, mobile production units only, normalised against manhours Figure 18 Total indicator, mobile units, normalised against manhours Figure 19 The barrier prevent ignition Figure 20 Number of faults for selected barrier elements Figure 21 Serious injuries to personnel on production installations related to manhours Figure 22 Serious injuries to personnel per million manhours, mobile units Figure 23 Overview of barrier breaches for DFU21 Falling object, Figure 24 Trends in the number of reported events per year... 32

11 1 1. Purpose and limitations Part 1: Purpose and conclusions 1.1 Purpose In the letter of 2002 approving the project, its scope and purpose is formulated as follows: The NPD shall seek to establish a realistic and jointly agreed picture of trends in levels of risk in the HES area and support the efforts made by the industry to improve the level of HES in the petroleum industry. This is intended to result in: 1) a clear picture of the risk situation 2) a reduction in the probability of accidents and their consequences 3) the development of reliable measurement tools 1.2 Objectives The aim of the project is to: measure the impact of safety-related measures in the petroleum industry. help to identify areas which are critical for safety and in which priority must be given to identifying causes in order to prevent unplanned events and situations. improve understanding of the possible causes of accidents and unplanned situations together with their relative significance in the context of risk, in order to create a reliable decision-making platform for the industry and authorities which will enable them to direct their efforts towards improving conditions of health, safety and the environment. The work will also help to identify potential areas for making regulatory changes and for conducting research and development. 1.3 Important limitations The project focuses on risk to personnel and covers major accidents, occupational accidents, work environment factors and a number of minor accidental events. New indicators have been used for major accidents in addition to existing databases. For occupational accidents, existing databases only have been used. Until 2002, the project was limited to factors falling under the NPD s area of responsibility in relation to safety and the work environment. See also subsection 4.3. From Phase 3 it has been extended to include all forms of personnel transport by helicopter, in cooperation with the Civil Aviation Authority Norway and helicopter operators on the Norwegian Continental Shelf. The project therefore now covers: all production and mobile units on the Norwegian Continental Shelf transport of personnel by helicopter between helicopter terminal and installation (point of departure to point of landing on installation/arrival at heliport). the use of vessels inside the safety zone around the installations. 2. Conclusions 2.1 Questionnaires and interviews As stated in subsection 1.3, DFU 12 Helicopter event now includes all personnel transport by helicopter relating to petroleum activities on the Norwegian Continental Shelf. The decision has been made to remove

12 2 helicopter-related events from the major accident indicator, because the data are not directly comparable. Helicopter-related risk has been studied in the light of three event indicators and two activity indicators for the period 1999 to If we look at a given period as a whole, a reduction (event indicator 1) may be observed in the number of events, but this reduction is not statistically significant. elatively speaking, it is the most serious events of which there are fewer, a fact that may indicate a positive trend. The frequency of events per hour flight time related to shuttle traffic is higher than the corresponding frequency for crew change traffic. Activity level in the period is relatively constant with a peak for both shuttle and crew change traffic in 2001 and a minor decrease in isk indicators None of the DFUs indicating the risk of major accident on installations has involved fatalities in the period. The last time there were fatalities in connection with one of these DFUs was in 1986, with a shallow gas blowout on the mobile drilling unit West Vanguard. The three event types in 2002 which have made by far the greatest contribution to loss of life in major accidents are hydrocarbon leaks, well-kicks and structural failure. Together these made up more than 70 % of the total risk indicator Indicators pointing to an increase A high number of hydrocarbon leaks were registered in 2000, while there was a significant reduction in 2001 in relation to the average for the period In 2002 there was almost a doubling in the number of hydrocarbon leaks from 2001, but because 2001 was low, the 2002 value is not a significant increase, compared with the average for the previous six years ( ). When the number of leaks is normalised in relation to the number of installation years, it is only in relation to fixed installations that there is a significant increase. Hydrocarbon leaks have shown an unfortunate trend for several years. Insufficient attention seems to have been paid to reducing the frequency of such leaks. All parties must accept responsibility here. The Norwegian Petroleum Directorate has decided, for example, that the most serious cases of hydrocarbon leaks shall be the subject of investigation by NPD. Other initiatives are also required. For exploration drilling there was a significant reduction in the number of well-kicks per drilled well in the period , but a sharp increase in Three of the six well-kicks in the course of exploration drilling occurred during HPHT drilling, meaning that there would have been a significant increase even if HPHT wells had been kept out of the picture. It is particularly the more serious well-kicks which have increased in number, making the picture even more serious. A shallow gas blowout also occurred in 2002, but blowouts of this nature occur so seldom that there are too few data on which to draw any conclusion about trends. For production drilling there was a significant increase in the number of well-kicks per drilled well at the end of the 1990s. The level flattened out in the period , but increased significantly in 2002, and is now at its highest level since The frequency per well drilled 2002 is almost as high as the average for exploration drilling in the period Even though there are several barriers intended to prevent a well-kick from developing into a blowout, such a clear increase should call for measures aimed at reducing this frequency. Structural damage to mobile installations has shown a marked increase over the last three years and is the most obvious reason for a total increase in the risk of major accident related to mobile installations.

13 3 For all indicators taken together, there is also a statistically significant increasing trend, for installations as a whole, for floating production units and mobile installations. This refers to normalised frequencies, in which changes in activity level have been taken into account. The increase is most marked for DFUs with causes relating to the installations, meaning that these are situations on which the installation itself can exercise an direct influence Indicators pointing to a reduction While the frequency of collisions involving field-related traffic increased significantly up to year 2000, there was a significant reduction in years 2001 and 2002, especially in regard to mobile units. An analysis of collected data indicates that the Statoil project aimed at improved vessel safety ( bedre fartøysikkerhet ) is a telling factor in this picture. Situations involving drifting vessels and objects on a collision course also show a reduction since It should be noted, however, that none of these DFUs makes a significant contribution to risk to personnel. DFU 12 (helicopter event), event indicator 1 (see subsection 2.1) shows a reduction in the period from Indicators pointing to a stable trend Fires unrelated to hydrocarbon leaks in process plants (DFU4) have shown a constant level for several years. This too is a DFU which plays only a limited role in the risk to personnel. The number of registered vessels on collision course has increased since The reason for the increase is probably better registration after Statoil s traffic monitoring centre came into operation in November There is reason to believe that there was previously a significant degree of under-reporting. A supplementary analysis was carried out in 2002, with the conclusion that the level has remained more or less stable since The description of events indicates that a number of installations were well prepared to cope with the contingency of vessels on a collision course, until the situation was clarified. On other installations there is still potential for improvement. See also section 2.2.4, in relation to indicators where it is not possible to point to any trend Indicators pointing to no particular trend There were 3 cases of leaks from risers and pipelines in 2001, while in 2002 there was only 1 such leak. Because these are rare events, it is impossible to show any particular trend. A new factor in 2002 is the significant amount of work the project has done on barriers, reflecting the new regulatory focus on this area. This has been followed up by the industry in an overall positive and active way. Taking into consideration that this is demanding work, both conceptually and in relation to the effort required, the pilot project carried out in 2002 must be characterised as promising. As early as 2002 significant quantities of test data on safety and preparedness barriers have been registered.. The levels of reliability indicated by these data are generally speaking on a par with the level specified by the industry for new installations. It will not be possible to show any trends in these data until data from more years are available. The industry was asked to perform general analyses of the efficiency of all barriers, as the regulations require. There was a mixed response to this, ranging from companies which have documented these studies very efficiently to those which have not documented them at all. The official requirements do not appear at the moment to have been met satisfactorily in the industry as a whole. Other indicators where to date it has not been possible to show trends:

14 4 ignited hydrocarbon leaks, none registered in the period leaks from risers, because of the low number of events serious structural damage to production installations, because of the low number of events. 2.3 isk indicators, serious occupational accidents on installations The frequency of serious injuries to personnel for the entire shelf is now back at a level on a par with the average for the period. It must be emphasised, however, that this positive trend is not reflected in the number of fatalities resulting from occupational accidents in 2002, the highest since The frequency of serious injuries on production installations from the mid-1990s shows an increase with a frequency peak in year In subsequent years a reduction has been observed. In 2002 we are back at the same level as the average for the entire period. This decrease is mainly due to a halving of the frequency in drilling/well operations. The frequency of serious injuries on mobile installations shows a peak in years 2000 and 2001 with about 2.8 cases of serious injury per million manhours. In 2002 a clear decrease is noted in relation to the average for the period. Injuries relating to drilling/well operations are still those dominating on mobile installations. 2.4 Qualitative analyses The main impression from the interviews is that risk levels are perceived as being the same as in All interviewees expressed clear disappointment over trends in crane and lifting operations. A certain concern is felt over the fact that all the efforts made over the last year have not led to any concrete results. Further concern was expressed at a possible loss of competence resulting from more rigs being laid up. Everyone interviewed believes that a good HES culture is important, but there are also divergent opinions among the parties as to what is meant by a good HES culture and what kind of measures are necessary to create such a culture. 2.5 General conclusion The conclusions relating to the risk indicators from Phase 3 (2002) of the project are, with few exceptions, negative. The same was true of the Pilot Project ( ), whereas there was greater variation in the conclusions following the Phase 2 project ( ). On the basis of the data and indicators used in this project, the increase in levels of risk in 1999 and 2000 would seem to be further strengthened in 2002, after a temporary improvement in There are few indicators which show a clear and positive reduction, and other important factors relating to the risk of major accident which show a clear increase. The frequency of serious injuries on the installations is on the same level as the average over the last ten years. The result of the work done in the field of barriers is seen as promising. In 2002 significant quantities of test data relating to safety and preparedness barriers have already been registered. The level of reliability indicated by these data is generally on a par with that specified by the industry for new installations. On the other hand, the regulatory requirements for an overview of the overall performance of barriers are not currently assessed as being fully satisfied in the branch as a whole. Mutual trust and cooperation are characterised as good, but the parties emphasise that this cooperation remains fragile. Several interviewees express concern that the general slump in which the industry currently finds itself may put a stop to many of the positive HES measures being worked on. In 2002 there were two fatal accidents within the Norwegian Petroleum Directorate s area of authority, both caused by falling loads. This underlines the need for the industry to prioritise measures aimed at reducing the frequency of accidents and incidents with falling objects.

15 5 3. Implementation Part 2: Implementation and scope 3.1 Main principles Phase 3 is a continuation of earlier phases of the project completed in , see NPD (2001) and NPD (2002). In Phase 3 the same general principles have been followed and reporting has been further extended with particular focus on the following points: The work of analysing and evaluating data relating to defined situations of hazard and accident has been continued and partly expanded with a number of new indicators. New indicators for helicopter transport (expansion in breadth) have been developed and analysed. Significant quantities of experience data have been gathered for barriers against major accidents. Interviews with key informants have been continued. The work of expanding the project in breadth to include vessels is still ongoing and it is hoped that this will eventually result in new indicators. 3.2 Implementation of Phase 3 of the project Phase 3 of the project began in summer 2002, taking as its starting point the experience gained from Phase 2. Early in October, a seminar was arranged in the Norwegian Petroleum Directorate for the parties involved, at which some of the experience from Phase 2 was presented, together with plans for Phase 3. The following participants were involved in Phase 3: The Norwegian esponsible for annual reporting and continuation of the project Petroleum Directorate Operator companies Provide data and information on activities on the installations Civil Aviation Authority Norway esponsible for reporting of public data of helicopter activities and quality assurance of data, analyses and conclusions Helicopter operators Provide data and information on activities in the helicopter transport sector HES expert group (selected specialists) The Safety Forum (representatives from unions, employers and authorities) Evaluate methods, databases, views on development, evaluate trends, propose conclusions Comment on methods and procedures used in the project, make recommendations for further work. General presentation of the project The Norwegian Petroleum Directorate has had support from external experts during part of the work. The following persons were responsible for specific tasks in connection with Phase 3: Jan Erik Vinnem, Preventor Odd J. Tveit Terje Aven, Stavanger University College (HiS) Jorunn Seljelid, Safetec Knut Haukelid and Lars Eklo, University of Oslo Espen Olsen, Jorunn-Elise Tharaldsen, Åshild Bakke, Kari Kjestveit and Kirsten Allred, ogaland esearch Institute (F)

16 6 The following persons from the Norwegian Petroleum Directorate have participated: Einar avnås, Øyvind Lauridsen, Hilde B. Haga, Mona Haugstøyl, Arne Kvitrud, Sigvart Zachariassen, eidar Lindeberg and Torleif Husebø. In addition, the following persons have contributed to the work by establishing indicators for helicopterrelated risk: Olai. Hjetland, Civil Aviation Authority Norway Jon Sneltvedt, Civil Aviation Authority Norway Harry A. Larsen, CHC Helicopter Service Torbjørn Amundsen, CHC Helicopter Service Inge Løland, Norsk Helicopter Per Skalleberg, Norsk Helicopter 3.3 Use of risk indicators isk indicators for major accidents and occupational accidents intended to cover two sets of factors: the occurrence of accidents, incidents, near-accidents and injuries the function of the barriers installed to protect exposed objects In Phase 3, data have been registered for: defined situations of hazard and accident relating to major accidents, with the following main categories: uncontrolled release of hydrocarbons, fires (i.e. process leaks, well-kicks/shallow gas, riser leaks, other fires) structure-related events (i.e. structural damage, collisions, threat of collision) experience data relating to the efficiency of barriers against major accidents on installations accidents, events and significant operational disruption in helicopter transport occupational accidents diving accidents other DFUs with minor consequences Data collection for the DFUs relating to major accidents rests partly on existing databases in the Norwegian Petroleum Directorate (CODAM, DDS, etc.), but also to a significant extent on data collection carried out in cooperation with the operator companies, including the database HCLIP for hydrocarbon leaks. All external events data have been quality-assured by, for example, checking them against the events register and other databases in the Norwegian Petroleum Directorate. Table 1 shows an overview of the DFUs included in Phase 3 (21 of a total of 24, the same as in Phase 2), and the data sources used. For the new DFUs introduced in 2001, feedback from the industry is that reporting will only be complete from The industry has used the same categories for registration of data through the database Synergi. 3.4 Trends in activity level Figure 1 and Figure 2 show changes over the period for production and exploration activities, focusing on the parameters used for normalisation against activity level. Annex A to the Phase 3 eport (NPD, 2003) presents the supporting data in detail. Errors in the data material used in previous reports have been corrected.

17 7 Table 1 DFU no. Overview of DFUs and data sources DFU description Data sources 1 Non-ignited hydrocarbon leaks HCLIP via data acquisition* 2 Ignited hydrocarbon leaks HCLIP via data acquisition* 3 Well kicks/loss of well control DDS (NPD) 4 Fire/explosion in other areas, flammable liquids Data acquisition* 5 Vessel on collision course Data acquisition* 6 Drifting object Data acquisition* 7 Collision with field-related vessel/installation/shuttle tanker CODAM (NPD) 8 Structural damage to platform/stability/anchoring/positioning failure CODAM (NPD) + industry 9 Leaking from subsea production systems/pipelines/risers/flowlines/loading buoys/loading hoses 10 Damage to subsea production equipment/pipeline systems/diving equipment caused by fishing gear CODAM (NPD) CODAM (NPD) 11 Evacuation (precautionary/emergency evacuation) Data acquisition* 12 Helicopter crash/emergency landing on/near installation Data acquisition* 13 Man overboard Data acquisition* 14 Injury to personnel PIP (NPD) 15 Occupational illness MOAS (NPD) 16 Total power failure Data acquisition* 17 Control room out of service Data acquisition* 18 Diving accident DSYS (NPD) 19 H 2 S emission Data acquisition* 20 Lost control of radio-active source Data acquisition* 21 Falling object Data acquisition* * Data acquired with the cooperation of operator companies 1,2 1,1 elative activity level 1,0 0,9 0,8 0,7 0,6 0,5 No of prod. install. Manhours, prod. Produced volume No of prod. w ells Pipeline length (km) 0, Figure 1 Trends in activity level, production

18 8 elative activity level 2,0 1,8 1,6 1,4 1,2 1,0 0,8 0,6 No of MODUs Manhours, MODUs No of expl. Wells 0, Figure 2 Trends in activity level, exploration The changes in activity level are relatively limited, except for drilling of exploration and production wells. The choice of absolute or normalised values to present DFUs or risk will therefore not result in any substantially different picture. Normalised values are nevertheless those consistently shown. A corresponding activity overview for helicopter transport is shown in subsection Project documentation Analyses, evaluations and results in connection with the project are documented as follows: Phase 3 summary report (Norwegian and English versions) Phase 3 project report for 2002 Phase 3 methodology report These reports can be downloaded free of charge from the Norwegian Petroleum Directorate s home pages ( 4. Scope 4.1 Interviews with key informants As part of the qualitative survey, further interviews have been conducted this year with representatives of the parties involved but with limited numbers compared with In Phase 3 there were also two activities based on the questionnaire survey presented in the report from Phase 2. These activities were workshops on specific topics for representatives from the petroleum industry and a qualitative analysis of free text from the questionnaire (Phase 2, NPD 2002). 4.2 Analysis of statistical risk The pilot project contained a detailed description of the method used to analyse the statistical level of risk, including the use of: DFUs reflecting early-stage accidents normalisation weighting of DFUs the need to identify future threats and trends trends relating to statistical risk over a long-term perspective

19 9 This approach has been continued in Phase 3, with only minimal changes. The most important changes relate to DFU12, helicopter transport and barriers, see subsections 4.3 and 4.4 respectively. The plan for Phase 3 was that the project should be extended in breadth to cover helicopter transport and vessels relevant to the industry while these are operating outside the safety zone but engaged on petroleumrelated activities. Helicopter transport is discussed in subsection 4.3 and section 5 in its entirety. In the case of vessels, no unified data source was found to exist for these operations, either activity level or experience data. It will therefore take longer to establish indicators for this area of activity. 4.3 Helicopter transport Close cooperation has been established with the Civil Aviation Authority Norway with a view to including data on helicopter transport of personnel between the shore base and the installations (crew change traffic), and between installations (shuttle traffic). During this process the Civil Aviation Authority Norway approached the two helicopter operators, an initiative which resulted in positive cooperation between the Norwegian Petroleum Directorate, the Civil Aviation Authority Norway and the two helicopter operators CHC Helicopterservice and Norsk Helikopter. The project has obtained detailed activity data directly from the helicopter operators. Air traffic data was supplied by both the Civil Aviation Authority Norway and the helicopter operators. This cooperation has also resulted in completely new indicators for major accident risk relating to helicopter transport. Helicopter-related risk is illustrated by three event indicators and two activity indicators. Data proved to be consistently available only for the period 1999 to 2002, and the analysis is therefore limited to this period. 4.4 Expansion in depth Barrier indicators Barriers are a focal point in the new HES regulations, and emphasis has therefore been put on establishing indicators for barriers in addition to indicators for the first signs of major accidents. In Phase 2 a method was developed (see Annex B to the report from Phase 2, Norwegian Petroleum Directorate, 2002) with a view to measuring the performance of barriers. The method is two-fold, the intention being that the two parts should complement each other to give the best possible overall picture without the use of too many resources in the companies. The companies were asked to report data and general qualitative analyses in accordance with this method Indicators for factors which may lead to occupational illness The project has developed proposals for indicators for the two factors noise and chemical work environment, the former with indicators based on exposure to noise, while the indicators for chemical work environment are based on managerial aspects. These indicators will be tested in the course of 2003.

20 10 Part 3: esults from Phase 3 of the project 5. Status and trends DFU12, helicopter events In previous phases of the NNS project, DFU 12 Helicopter event covered transport of personnel by helicopter inside the installation s safety zone. From now on, DFU 12 will be taken to cover all personnel transport by helicopter in connection with petroleum activities on the Norwegian Continental Shelf. The last major accident resulting in fatalities occurred in September 1997 in connection with a helicopter accident outside Brønnøysund. This was the last major accident on the Norwegian Continental Shelf. Cooperation has been established between the Civil Aviation Authority Norway and the Norwegian Petroleum Directorate in accordance with the intentions set out in NOU 2002:17 Helicopter Safety on the Norwegian Continental Shelf eview 2: Trends, objectives, risk influencing factors and measures (Statens forvaltningstjeneste, 2002). The helicopter operators have also been actively involved in this process through the provision of air traffic and production data as well as the evaluation of established event and activity indicators. Air traffic data have been collected from the relevant helicopter operators, and the scope has been decided in cooperation with the Civil Aviation Authority Norway. Air traffic data include type of event, risk category, severity, type of traffic, phase, helicopter type and departure/arrival information. The main project report (Norwegian Petroleum Directorate, 2003) contains further information on scope, limitations and definitions. A review of events submitted in the current phase of the project points to a partly unclear practice of the guidelines for classification, as there is no correspondence in classification made by individual parties for individual events. In this phase of the project it has been decided to use the helicopter operators classification of events. Production data have been collected from the relevant helicopter operators, and the scope has been decided in cooperation with the Civil Aviation Authority Norway. Production data have been divided into type of traffic (crew change traffic and shuttle traffic) and include flight hours, person flight hours, number of flights, number of passengers and number of landings. Passengers and crew have been taken together in the analyses. With the data sources used in the project, only events and production data were accessible for the period Overview of new indicators Three event indicators and two activity indicators have been established to give the best possible picture of helicopter risk. The event indicators provide a greater degree of differentiation than the indicator previously used for DFU12. The activity indicators show the trends for exposure to helicopter risk, and are therefore more proactive as indicators. The event indicators may also be seen as proactive, since they now reflect less serious events, without these necessarily resulting in accidents or corresponding outcomes. The following indicators have been established: Event indicator 1 covers: o air traffic accident with severity classed as high, medium and low o serious air traffic event with severity classed as high, medium and low o air traffic event with severity classed as high, medium and low o operational disruption with severity classed as high

21 11 o all events with the exception of events during the phase parked, i.e. discovered while the helicopter is parked on the helideck. Event indicator 2 covers: o the same type of events as for indicator 1 o operational disruption with severity classed as medium and low Event indicator 3 covers: o the same type of events as for indicator 1 o events relating to the parked phase are included whereas they are omitted from indicator 1 Activity indicator 1 covers: o volume of crew change traffic (personnel transport where the helicopter s departure and final point of arrival is on an onshore base. The exception is Frigg, see subsection 4.2 in the main report from Phase 3 of the project) Activity indicator 2 covers: o volume of shuttle traffic (personnel transport where the helicopter s point of departure and arrival is on an installation) 5.2 Activity indicators Activity indicator 1 Figure 3 shows activity indicator 1, volume of crew change traffic, as the number of flight hours and number of person flight hours per year in the period There was an increase in volume up to 2001, followed by a reduction in The number of flight hours is virtually constant in the period compared with the number of person flight hours, with a peak in Person flight hours Flight hours Person flight hours Flight hours Figure 3 Volume of crew change traffic, flight hours and person flight hours, Activity indicator 1, volume of crew change traffic per year, must be seen in the context of activity level on the Norwegian Continental Shelf. There was a certain decrease in the total number of work in However, it may be noted that in 1999 there were 21 person flight hours flown per 1000 manhours, while this figure decreases to 20.4 in This may be due to fewer visitors but may also to a certain degree reflect changes in work schedules on the installations. The change-over to a regular shift of two weeks on the installation and four weeks onshore (from 2-3 to 2-4 ) will lead to a reduction in the number of flights per employee.

22 Activity indicator 2 Figure 4 shows activity indicator 2, volume of shuttle traffic, number of flight hours and number of person flight hours per year in the period There was an increase in volume up to 2001, and thereafter a reduction in 2002 compared with previous years Person flight hours Flight hours Person flight hours Flight hours Figure 4 Volume of shuttle traffic, flight hours and person flight hours, Activity indicator 2 must be seen in the context of activity level on the Norwegian Continental Shelf. The reduction in the number of person flight hours in shuttle traffic is greater than the reduction in volume of crew change traffic, and also greater than the reduction in the number of person flight hours in In recent years, there has been a focus in various fora on the reduction in shuttle traffic because of the increased risk, particularly in relation to the phases departure and arrival. The volume of shuttle traffic is an indicator which shows the exposure to such risk. In 2002 a normally unmanned installation was removed from the Norwegian Continental Shelf, which may be one of the possible reasons for the reduction in Event indicators Event indicator 1 Figure 5 shows the number of events included in event indicator 1 normalised in relation to the number of million person flight hours per year. In the main report, the corresponding trend is also shown per flight hours. There is a consistent falling trend throughout the entire period , while the value in 2001 is lower than the trend. It also seems that the severity of registered events has been reduced in the form of a stabilised number of operational disruptions and a reduced number of air traffic events. The discussion around the low number of events in 2001 has not revealed any obvious reason, nor has any general fall been shown in the degree of reporting in In the main report from Phase 3 it is shown that the reduction in Figure 5 is not statistically significant in relation to the methodology used in the project to measure trends.

23 13 25 Events/ person flight hours Serious air traffic incident H Air traffic incident H Air traffic incident M Operational disruption H Figure 5 Event indicator 1, per person flight hours, Event indicator 2 Figure 6 shows the number of events included in event indicator 2, normalised in relation to the number of million person flight hours in the period (In the main report a corresponding trend is also shown per flight hours.) There seems to be an increase in the number of events relating to crew change traffic up to 2002, when a small reduction is shown. For the number of events relating to shuttle traffic there are minor variations round a stable level, with no clear trend. 400 Events/ person flight hours Shuttle traffic Crew change traffic Figure 6 Event indicator 2, per person flight hours, It is obvious that a considerably greater number of events relating to crew change traffic are reported annually compared with events relating to shuttle traffic. Normalisation of events in relation to volume of activity indicates, however, that the frequency of events with a corresponding severity is rather higher for shuttle traffic than for crew change traffic. This can give an indication that risk is higher in connection with shuttle traffic. The exception is the number of events normalised in relation to exposure data in 2001, which also in terms of other event indicators has distinguished itself as a special year.

24 14 In shuttle traffic, a permanently stationed helicopter and crew are generally used. This means that the pilots are more familiar with local conditions and procedures and can more readily detect nonconformities Event indicator 3 Event indicator 3 shows (not shown here, see subsection in the main report) the same events which are included in event indicator 1, with the addition of events in the parked phase, i.e. events identified while the helicopter is standing on the helideck. These are events without major accident potential Events discovered on the helideck show an increase, especially in 2002, a fact that may be due to great attention being paid to such situations. Events detected in the transit phase, i.e. in transit between heliport and installation, show a certain reduction in the period. None of these trends is statistically significant because of the small number of events concerned. 6. Status and trends indicators for major accidents on installations In the current phase of the project, the indicators developed in previous phases for major accident risk have been followed up further. The main emphasis is still on indicators capable of showing major accident risk on installations. Indicators for major accident risk with helicopter are discussed in section 5. There have been no major accidents on installations on the Norwegian Continental Shelf in the last ten years. None of the DFUs showing major accident risk on installation have involved fatalities in the period. The last time there were fatalities in connection with one of these major accidents DFUs was in 1986, with a shallow gas blowout on the mobile installation West Vanguard, see also page 10 with reference to the helicopter accident off Brønnøysund. Nor has there been any case of ignited hydrocarbon leaks from process systems since 1992, except for the occasional quite minimal leak without the potential for leading to major accidents. The most important individual indicators for production and mobile installations are discussed in subsection 6.2, for leaks, well-kicks and collisions/structural damage. The other DFUs are discussed in the Phase 3 report. Helicopter accidents are discussed separately in section 5 for production and mobile units together. The indicator for total risk is discussed in subsection DFUs related to major accident risk Figure 7 shows the trend in the number of reported DFUs on installations in the period It is important to emphasise that these DFUs make a greatly varying contribution to risk. DFU5 (vessel on collision course) in particular would appear by our reckoning to have been underreported in previous years. This is less true for those DFUs relating to hydrocarbon leaks and loss of well control. Figure 7 shows that these are dominant in number in all years with the exception of 2001, when they stood for approximately 50 %. The increase in DFU5 (vessel on collision course) in Figure 7 is not a good indication of risk trend (see discussion in subsection 6.2.3). Figure 7 shows however that the increase for DFUs 1 and 3 is so clear that the picture remains largely the same, even if we exclude DFU5.

25 15 Number of DFU occurences i Evac/muster Dam. subs inst Subs equipm leak Struct. damage Coll. attend vess. Drif ting object Ship on coll course Other fire/expl Kick Ign HC leak Unign HC leak Figure 7 All reported DFU s divided by category 6.2 isk indicators for major accidents Hydrocarbon leak in process area Figure 8 shows the total number of leaks over 0.1 kg/s in the period Up to 1999 there was a downward trend, then 2000 and 2002 with a higher number, while 2001 was on a par with In the current phase hydrocarbon leaks are still classified by leak rate in rough categories as shown in Figure 8. In the main report there is also a finer subdivision, for 2001 and Number of leaks >10 kg/s 1-10 kg/s 0,1-1 kg/s Figure 8 Number of hydrocarbon leaks exceeding 0.1 kg/s Only 1 leak over 10 kg/s was registered in 2002, the first since those two in The figure shows a sharp increase in the number of significant hydrocarbon leaks in 2002, compared with the number for 2001, virtually a doubling. There is no information which can give any reliable explanation for this trend, but there would appear to be a need for a significantly increased emphasis on measures for reducing the number of hydrocarbon leaks. An important change has been introduced in relation to the weighting of hydrocarbon leaks, resulting in leaks with a rate over 10 kg/s being given an individual weighting assessment. This weighting is determined from a detailed description of the leak and the circumstances around it. The leak in 2002 was reported at 22 kg/s,

26 16 and as such is the biggest registered in the period from1996. The leak occurred in the flare system aft on the stern of a production vessel. There is good ventilation in this area, and little likelihood of explosion since the area is open. In general there is less probability of ignition than is typical for process leaks. In addition, this is an area where there are not many people to be found, except during shuttle tanker hook-up. This leak is assessed as having a significantly lower weighting than others of more than 10 kg/s in the period. Figure 9 shows the trend for leaks over 0.1 kg/s, normalised in relation to installation year, for all production installations. The figure illustrates the technique used throughout the entire project to assess the statistical significance of trends. Figure 9 shows that the increase in the number of leaks per installation year is statistically significant in 2002, in relation to the average for the period This is shown by the height of the column for 2002 falling in the red field (see subsection in the pilot project report). In the main report, leaks may come to be normalised in relation to both manhours and the number of installations. 9 8 elative risk indicator Int Year Figure 9 Trend, leaks, normalised against installation age, all production installations There are significant variations between installations in regard to the frequency of leaks over 0.1 kg/s. 12 installations have not experienced leaks over 0.1 kg/s at any time during the period Approximately 50 % of all installations have less than 1 such leak every second year, 12 % of the installations have more than 1.5 leaks per year over 0.1 kg/s. There are examples on comparable installations with widely varying frequency, with the same operator company. This underlines the fact that there is significant potential for improvement. The potential for improvement is also underscored by Figure 10 which show the average leak frequency per installation year for anonymised operators on the Norwegian Continental Shelf. It should be emphasised that the number of leaks over the seven year period is so high that the greatest differences in Figure 10 are statistically significant differences. For leaks in the order of kg/s, there may be some reporting uncertainty. The figure shows, however, that those companies with the highest overall leak frequency also have the highest leak frequency in terms of leaks greater than 1 kg/s taken alone.