Industrial Applications of Accident Causation Management System
HYUCKMYUN KWON AND HYUNGJOON YOON
Center for Chemical Plants Safety, Korea Occupational Safety and Health Agency, Inchon, South Korea
IL MOON
Department of Chemical Engineering, Yonsei University, Seoul, South Korea
kACMS (KOSHA Accident Causation Management System) has been developed to control human errors in Korean chemical industries. kACMS is a safety manage- ment system using the Korean GFT (general failure type) methodology, which has been found a good approach to eliminating, or at least minimizing human errors. To observe the trend of human errors in the chemical industry, about 5500 near-miss cases have been collected from a Korean chemical plant. The analysis of the col- lected cases shows that the removal of human errors is the key to preventing these near-miss cases that have the potential to lead to actual accidents.
A Korean petrochemical company applied kACMS in its 9 chemical plants. Fifty-five employees participated in the survey and 12,000 safety data were collected based on a questionnaire. As a result of each survey, the average, best, and worst scores were 85.0, 90.6, and 79.6, respectively. These results led to a thorough inves- tigation of the safety systems of the worst scored plant and directions for improving safety.
Keywords Human error; Near miss; Korean GFT; kACMS; Risk causation
Background
Hundreds of thousands of incidents in chemical industries occur every year all over the world, often incurring devastating human and economic costs (U.S. Chemical Safety and Hazard Investigation Board, 1999). The effects are indiscriminate. Until now, with few exceptions, chemical incidents have gone largely unnoticed, perhaps due to the lack of definitive shared knowledge of previous analysis of chemical acci- dents in different countries. Theoretical models have evolved from investigations into the ‘‘why’’ and ‘‘how’’ of case histories. These insights, so gained, have made possible better explanations of incident causation. According to the incident prone- ness theory, incidents are a result of individual differences (International Labour Office, 1998).
Address correspondence to Il Moon, Center for Chemical Plants Safety, Korea Occu- pational Safety and Health Agency, 34-9 Kusan-dong Bupyung-gu, Inchon 403-120, South Korea. E-mail: ilmoon@yonsei.ac.kr
Chem. Eng. Comm., 193:1024–1037, 2006 Copyright # Taylor & Francis Group, LLC ISSN: 0098-6445 print/1563-5201 online DOI: 10.1080/00986440500352089
1024
A classical example is H. W. Heinrich’s theory of causation, which has signifi- cantly influenced practical investigation of process safety incidents (Heinrich, 1980). Currently, the most widely accepted and adopted theories rely on the system theory. According to this theory, an incident is regarded as an abnormal effect of the technological or management system.
A fundamental principle in modern incident investigation is to look for the underlying causes behind an incident. Furthermore, one axiom of systematic incident investigation is that human factors play an important role in incident causation (Peterson, 1984). Human error is the most common cause, accounting for at least 90% of all industrial accidents (U.K. Health and Safety Executive, 1991).
While purely technical errors and=or uncontrollable physical circumstances may also contribute to accident causation, human error is the largest source of failures. The increased sophistication and reliability of machinery means that the relative pro- portion of accident causes attributed to human error increases as the actual fre- quency of accidents decreases. Changes and new technologies always introduce hazards as well as benefits. Computers are no exception. Many computer experts find it hard to write in a language that can be readily understood by the operating staff. As far as computer controls are concerned, the errors are most often human failures such as a failure to foresee or allow for faulty equipment or software bugs, failure in understanding what the system could and could not do, or failure to realize how people respond to displays. As we shall see, certain characteristics of computer- controlled systems tend to induce human errors (Center for Chemical Process Safety, 1993a, 1998). INPO (Institute of Nuclear Power Operation) made a significant announcement that ‘‘till now, statistically, cause factors and situation factors when nuclear power station accidents happened, they are similar to those of near-miss inci- dents relatively to the comparison of the two factors [INPO-84-027]’’ (Genizzi, 1998). There is no difference between the cause of actual accidents and the cause of near-miss incidents (Yoon et al., 1999a).
Fifty-five hundred pieces of accident data have been collected from a petro- chemical plant and analyzed by ySIMS (Yonsei Safety Information Management System) (Yoon et al., 2000). The survey indicated that human factors in general are the major contributors to near-miss incidents. In response to a request for an efficient and forceful method of human factor analysis kACMS (Korean Occu- pational Safety and Health Agency Accident Causation Management System) was developed to reduce and eliminate human error and ultimately near-miss incidents.
Methods of Human Factor Analysis
The development of kACMS includes various analytical methods for predicting and reducing human errors. They are classified in the following four groups.
. Techniques for the acquisition of information about the worker’s actions and the chain of events in an accident
i) DTE (discussions and interviews with experts) technique: The analysis of com- plex tasks is usually best done in collaboration with a task expert (Bainbridge, 1987). They are useful in checking the accuracy of the information that has been collected (Bainbridge, 1974).
Accident Causation Management System 1025
ii) CI (critical incident) technique: This technique is used to collect data about near-miss incidents and critical events that were experienced by the operating team (Flanagan, 1954).
iii) AA (activity analysis) technique: Data about the plans and routines that are used by workers in controlling a process are obtained by means of an activity analysis, a type of input and output analysis (Crossman et al., 1974).
. Various task analysis techniques
i) HTA (hierarchical task analysis) technique: A systematic method of describing how the job process is organized to suit the overall objective of the job (Shepherd, 1985).
ii) OAET (operator action event trees): Tree-like diagrams that represent the sequence of various decisions and actions that the operating team is expected to perform when confronted with a particular event (Kirwan and Ainsworth, 1993).
iii) OSD (operational sequence diagram): Flow-charting techniques that represent any sequence of control movements and activities concerning information collec- tion that are executed in order to perform a particular task (Kirwan et al., 1988).
. Approaches to quantification
i) THERP (technique for human error rate prediction): Identical to the aforemen- tioned event tree method (Swain and Guttmann, 1983).
ii) SLIM (success likelihood index method): The chemical, transportation, and various other industries utilize this technique. Tasks within the SLIM technique are numerically rated on the influence and the probability of error, these rat- ings being combined for each task to give an index called the SLI (success like- lihood index) (Embrey, 1986; Kirwan, 1990).
iii) IDA (influence diagram approach): A technique used to evaluate human error probabilities as a function of the complex network of organizational influences, among others, that have an impact upon these probabilities (Embrey, 1992).
. Various checklists of factors that can influence human reliability: It is important to identify the human component using the general failure type (GFT) analysis for effective safety management systems and risk identification programs, wherein various checklists are used to identify general failures potentially hidden in the working procedure, design, facilities, etc. (International Labour Office, 1998).
i) PIFs (performance influencing factors): PIFs are defined as the factors that determine the likelihood of error or effectiveness in human performance. PIFs, such as quality of procedures, level of time stress, and effectiveness of training, will vary on a range from the best practicable to the worst possible (Center for Chemical Process Safety, 1994). General failure type used in kACMS has a similarity to this method.
ii) MORT (management oversight and risk tree analysis): A logic tree that assist in providing analysts with a disciplined method for accident investigation, use- ful in safety program evaluation and applied to operational readiness reviews. MORT is a useful aid to the previously mentioned checklist type of analysis techniques that assists the analysts in visualizing the needed hardware and work procedures to match personnel capabilities at all levels of the organiza- tional hierarchy (Gertman and Blackman, 1994).
1026 H. Kwon et al.
iii) Contextual Level Classification: Many miscommunications and pen taxo- nomies are constructed at this level and include references to contextual trigger- ing features such as anticipation and preservations. Such categorizations are valued since they help in focusing attention on the complex interaction between local triggering factors and the underlying error tendencies (Reason, 1998).
Formulation of General Failure Type
After years of experience in improving safety techniques and process design, many organizations discover that accident rates, processing plant losses, and profitability reache a plateau beyond which further improvement is impossible to achieve (Center for Chemical Process Safety, 1993b). Another discovery is that even in organizations with generally good safety records, occasional large-scale disasters occur, which invariably shake the public’s confidence in the chemical processing industry.
The common factor in both discoveries is human error. Errors are viewed as the natural outgrowth of unfavorable combinations of people and the working situa- tions. Simply put, an error is a human output outside the tolerances established by the system requirements where the person operates (Process Safety Institute, 2000).
Human errors represent a major target in prevention and are becoming more important (Center for Chemical Process Safety, 1995). A rigorous analysis of human errors shows that they might be relative to the human management system (Stellman, 1984; Goh et al., 1998). So a more effective safety management system has to be applied to manage operators systematically and to identify causes of human errors efficiently. In order to get more insight into the controllable parts of the accident causation process, an understanding of the possible feedback loop in a safety control system is necessary. Figure 1 shows the component structure of a safety control sys- tem that can form the basis of managerial control of human error.
The GFT (general failure type) of Gop Groeneweg’s accident causation model was used when we designed the safety control and risk management systems that concern human errors (International Labour Office, 1998). kACMS can report the weak points of the management of a company by collecting, classifying, and analyz- ing data based on GFT. GFT is defined as the factors that cause substandard acts and situations in the generating mechanism of an accident. GFT was modified to adapt it to petrochemical plants by including new classifications of 11 fields by risk causation, as shown in Table I (Yoon et al., 1999b). The 11 risk causation fields reported are major areas that may be greatly influenced by human factors in terms of implementing safety management. The 11 types of GFT consist of hardware, training, incompatible goals, and eight others. Each type of GFT contains 20 related questions and answers to investigate working conditions of people involved. Refer to Table II for GFT sample questions; the appendix also gives examples of 20 related questions for a single GFT sample question.
These questionnaire sheets are distributed to the employees to let them diagnose the status of their safety level. The final 20 questions were prepared by processing several steps in order to check the reliability of the questions. In the Center for Chemical Plant Safety of the Korean Occupational Safety and Health Agency, 12 executive engineers are working towards the implementation of the PSM (process safety management system) in Korea. Each engineer, who has an average of 15 years experience in different backgrounds, such as plant design, operation, instrumentation,
Accident Causation Management System 1027
electronics, hazard analysis, fire fighting, piping, and material selection, and also possesses a professional engineer certificate in his own field, proposed for the ideas in preparation of the questions (20 questions�11 types).
When the engineers prepared the draft questions, their main considerations were the following classifications of human error:
. Errors due to a slip or a momentary lapse of attention: the intention is correct but the wrong action is taken.
. Errors due to poor training or instructions: someone does not know the correct procedures or worse, thinks he knows but does not. Some analysts note these mis- takes to emphasize that the intention was wrong.
. Errors due to a lack of physical or mental ability; thus, the abilities of the person and the situation match poorly.
. Errors due to a lack of motivation or a deliberate decision not to follow instruc- tions or expectations.
. Errors made by managers, often due to a lack of comprehension of the part they should play.
From a certain perspective, almost all accidents are due to management errors. If the management had ensured that the plant was better designed, the training and
Figure 1. Flow of safety control system (Reason et al., 1998).
1028 H. Kwon et al.
instructions were better implemented, or previous violations were addressed, most likely the accident would not have occurred (Keltz, 1991).
Through several meetings, a tentative list of 220 of the most appropriate ques- tions were selected. After that the questionnaire sheets were circulated to 20 chemical companies to check the reliability of the questions. KOSHA updated the list of ques- tions by making amendments based on the comments from industry. About 15% out of the total questions were revised through this process. Then the final questionnaire sheets were distributed within the industry. We collected the answer sheet and ana- lyzed the data using a computer program. The program, which uses Excel, is com- posed of an input file, a table of correct answers, and a result file. The table of correct answers contains the predefined answer to each question as Yes or No. The answers from the industry were classified into a relative risk level with scores from 0 to 100, where 0 signifies highly dangerous and 100 means perfectly safe. By analyzing the score of each person, each department, and each GFT, the weak areas of each plant can be identified and this result will be used to guide safety enhancements. The weak points in the company’s management will be reported to
Table I. Definition of Korean GFT
Type of GFT Definition
Design (DE) Failures due to poor design of a whole plant as well as individual items of equipment
Hardware (HW) Failures due to poor state or unavailability of equipment and tools
Procedures (PR) Failures due to poor quality of the operating procedures with respect to utility, availability, and comprehensiveness
Error enforcing conditions (EC) Failures due to poor quality of the working environment, with respect to circumstances that increase the probability of mistakes
Housekeeping (HK) Failures due to poor housekeeping Training (TR) Failures due to poor or inadequate training
or insufficient experience Incompatible goals (IG) Failure due to the poor way safety and internal
welfare are defended against a variety of other goals like time pressure and a limited budget
Communication (CO) Failure due to poor quality or absence of lines of communication between the various divisions, departments, or employees
Organization (OR) Failure due to the way the project is managed and the company is operated
Defenses (DF) Failures due to the poor quality of the protection against hazardous situations
Maintenance management (MM) Failure due to poor quality of the maintenance procedures regarding quality, utility, availability, and comprehensiveness
Accident Causation Management System 1029
the managers to serve as a guideline for further actions to be taken and new invest- ments to make to eliminate safety hazards.
Case Study: Administration of GFT in a Company
After finalizing the question and answer sheets, we applied them to a chemical com- pany that has nine operating units such as styrene monomer, and tere-phthalic acid plants.
The surveying information is as follows:
. Period: September 1998–February 1999
. Surveying target: A large Korean chemical company
. Number of employees requested to answer: 62
. Number of employees who answered: 55 (percentage of reply: 88%)
. Number of items in the analysis: 12,100 items (55 employees�11 GFT�20 questions=GFT)
Table II. Examples of questions for each type of GFT
Type of GFT Sample questions for each type
Design (DE) Have you ever participated in risk assessment of relevant equipment when modifying or installing equipment?
Hardware (HW) Have you ever stop production work due to mechanical problem during the past four weeks?
Procedures (PR) Have you ever found anything wrong in operation procedures?
Error enforcing conditions (EC) Have you ever not used PPE that is provided when you handle hazardous material because of its discomfort?
Housekeeping (HK) Are drains or water pipes well maintained in the workplace?
Training (TR) Have you ever felt that your trainer is incapable of training?
Incompatible goals (IG) Have you ever received any order from your manager to shorten your production time that might cause your plant operation unsafe?
Communication (CO) Has every result from a recent safety meeting been reported to a manager in your department?
Organization (OR) Do you know the reporting procedures when any accident occurs?
Defenses (DF) Do you know your duty and action in an emergency?
Maintenance management (MM) When you conduct maintenance work do you start your work after consulting a permit-to-work sheet?
1030 H. Kwon et al.
The answers from nine operating units were classified into relative risk levels as scores of 0 to 100, where 0 signifies highly dangerous and 100 signifies no hazard present, based on the probability of the answers. The analysis results are shown in Figures 2 and 3, and Tables III and IV. The abscissa of Figure 2 is the probability
Figure 2. Results of analysis on GFT (company level).
Figure 3. Result of analysis on accident causation model (company level).
Accident Causation Management System 1031
T ab leII I.A na lysi ssu mm ary of acc iden tca usa tio nm od el(c om pa ny lev el) G FTP roce ss(D ept. )
D EH WP RE CH KT RIG CO OR DF MM A
v era ge SD aA 75 79 93 85 90 84 79 87 74 95 93 84 .91 2.3 4
B 84 82 98 93 92 93 89 88 83 98 96 90 .55 1.7 5
C 73 63 84 86 87 84 84 90 69 85 93 81 .64 2.7 9
D 81 75 91 85 87 83 76 86 70 97 91 83 .82 2
.4 0
E 68 87 93 78 88 81 86 84 72 97 89 83 .91 2.6 1
F 76 80 89 85 94 79 86 93 74 95 92 85 .73 2.2 7
G 83 88 96 96 91 90 91 88 69 99 93 89 .45 2.4 5
H 71 80 97 86 90 76 67 90 84 95 10 08 5.0 93 .23
I6 86 99 77 08 87 86 48 07 59 59 47 9.8 23 .6
0 Av era ge 75 .44 78 .11 93 .11 84 .89 89 .67 83 .11 80 .22 87 .33 74 .44 95 .11 93.4
D a2 .04 2.6 91 .53 2.5 40 .81 2.2 13 .19 1.2 91 .86 1.3 61 .05a Sta nd ard dev iati on .T ab leIV .L evel so fre spo nd ers (Process ‘‘I’ ’)R esp on der emp loy eeP a
W Yb CS cD EH WP RE CH KT RIG CO OR DF MME
mp l.1 I2 11 00 79 73 95 83 82 89 81 89 01 00 88
Em pl. 2I 41 10 05 06 79 36 21 00 67 60 78 50 10 01 00
Em pl. 3I 31 10 0657 41 00 76 88 93 70 83 82 88 86
Em pl. 4I 81 10 06 57 19 44 16 96 05 07 6 79 59 2
E mp l.5 I8 11 00 74 57 10 08 41 00 75 56 70 88 89 10 0A ver ag e6 66 89 66 98 87 76 38 05 79 59 3a Pro cess .b Wo rkin gy ears .c Co mp an ysi ze.1032
of the score and the ordinate is different types of processes such as naptha craking center (NCC), poly propylene (PP), linear density polyethylene (LDPE), purified tere-phthalic acid (PTA), etc. The process exposed to the most incident conditions was ‘‘I’’ with a score of 79.82, the safest process was ‘‘B’’ with a score of 90.55, and the average score was 84.99. Table III represents the score of 11 GFTs for each process in abscissa, by which the safety manager can search and analyze the weak and strong points of each process from the management’s point of view. ‘‘B’’ has higher scores than average for all 11 types, but ‘‘I’’ has lower scores than average for almost every type. Figure 3 displays the same data in Table III graphically in order to help managers understand the trend of the GFT scores more easily.
For the lowest scored processes, ‘‘I,’’ in Table III, it is recommended that more detailed analysis should follow and immediate remedial action be requested.
Table IV represents the 11 GFT scores rated by each employee for process ‘‘I.’’ As the score for design, hardware, error enforcing conditions, incompatible goals, and organization are found to be very low, being 66, 68, 69, 63, and 57 respectively, safety personnel should primarily concentrate their efforts on these areas. The atten- tion of special task management ought to be given for the organization areas with a 57 GFT score. If data represented in Table IV are not sufficient, the safety manager must review the questionnaire records and find the root causes that result in such low scores.
Through the review of this analysis, the weak areas of each unit were identified and this result was used as a guide to safety enhancements. The weak points in the company’s management were reported to the managers to serve as a guideline for further actions to be taken and new investments to make to eliminate safety hazards.
The risk level of each department and unit was also analyzed based on 12,100 items (11 GFT�20 sub-items). kACMS managed the survey efficiently and system- atically using the new questions and classifications.
Table V gives the results of the two-way analysis, which clarifies the score vari- ation shown in Table III. The two-way analysis of variance was performed to deter- mine the effects of the chemical process and GFT, which are two nonmetric independent variables, on the score, which is a single-metric dependent variable. The analysis gave statistically significant results (F value ¼ 12.0771, P ¼ 0.0001). In the analysis of the primary effect of the independent variables, the first inde- pendent variable, process, showed a statistically significant difference (F value ¼ 4.7268, P ¼ 0.0001), indicating that the averages of the first dependent variables for the nine process groups were different. The second independent variable, GFT, also made a statistically significant difference (F value ¼ 17.9573, P ¼ 0.0001),
Table V. Two-way analysis of score variance by chemical process and GFT
Source of variation Sum of squares DFa Mean square F value Prob. > F
Main effects 5840.1414 18 324.4523 12.0771 0.0001 Process 1015.8990 8 126.9874 4.7268 0.0001 GFT 4824.2424 10 482.4242 17.9573 0.0001
Error 2149.2121 80 26.8652 — — Total 7989.3535 98 — — —
aDegree of freedom.
Accident Causation Management System 1033
representing that the averages of the second dependent variables for the 11 GFT groups were diverse.
To conclude, the dependent variable, score, which is calculated by each inde- pendent variable, showed statistically significant differences.
Conclusion
Human error is probably the most significant contributor to loss of life, personal injury, and property damage in the chemical industry. We developed a safety man- agement system called kACMS by using a Korean GFT to examine the role of humans in the accident causation process, help safety managers to find unsafe areas of their plant, and advise them to rectify any weakness. ‘‘Defined Safety’’ seized by controlling the controllable is the key principle of the kACMS. The kACMS pursuit concentrates upon systematic factors and the way in which management decisions can be performed in unsafe conditions at a workplace. kACMS may comprehen- sively define and regulate any unsafe work process factors by providing a unique and systematic definitive question methodology for GFT type, process type, and employee type, each constituted of 20 specialized safety evaluation GFT questions. Furthermore, a close analysis may be made on the lowest GFT score type and its questions to define and settle the ‘‘real’’ problem on a detailed regulatory level; an interview with the corresponding (GFT type) responsible employee would also be helpful. kACMS then would give a clear and determinative safety evaluation report for the corresponding workplace. A computer program supporting kACMS reports the problem areas by classifying and analyzing the data efficiently. We applied this kACMS to a Korean chemical company, where 55 employees of nine operating units participated and 12,100 pieces of data were collected and analyzed. Based on the indications from the kACMS application, the company was able to find the mana- gerial weak points, suggest the priority of each rectification, check the result of investments, and finally, prevent accidents in petrochemical processing.
kACMS is used for (1) the assessment of safety level by person, department, and company, (2) the utilization of GFT data for accident prevention, (3) the implemen- tation of countermeasures to prevent similar accidents, and (4) the application of cost-benefit analysis, which can indicate the relationship between the investment and its effectiveness.
kACMS already proved to be highly efficient in domestic (Korean) chemical industries. A Korean company that has been applying the kACMS recently reported that it was getting favorable outcomes with an improved safety environment. It added that by eliminating human errors in the work process (through the efficient use of kACMS), it reduced real accidents. As adopted in Korean chemical industries and plants, ‘‘Controlling the Controllable’’ is obtained through a systematic evalu- ation methodology pursuit, kACMS.
Appendix Indication of the Score Level of Control
A. General
In this Appendix a list of 20 questions is presented that serves as an example of one of the GFT’s type questions to comprehensively show how the kACMS tool works. Employees and respondents would have to answer the question anonymously; they
1034 H. Kwon et al.
should also give extended answers for later evaluation purposes. For instance, it is not sufficient to answer ‘‘Yes’’ on a GFT type question like ‘‘Have you ever been ordered by your manager to shorten your production time, which might cause any unsafe factor for the plant operation process, in the past six months?’’ The employee would have to indicate which order it was and under what conditions it had to be applied. This serves two goals: it increases the reliability of the answers and it provides management with information it can act upon.
B. How to Measure the ‘‘Level of Control’’: Instructions for Respondents
Answer all 20 indicated questions, corresponding them to your own situation. Be punctual with the time limits in answering questions. Some of the questions might not be applicable for your situation; answer them with ‘‘n.a.’’
If the indicated question is somewhat difficult for you to answer, enter a ques- tion mark ‘‘?’’ After you have answered all the questions, compare your own answers with the reference answers and check your score. You should get a point for each correctly answered question, i.e., answered the same as the reference answers. Add your points together, then calculate the scale percentage by dividing the points by the number of questions you answered ‘‘Yes’’ or ‘‘No.’’ Note that questions answered with ‘‘n.a’’ and ‘‘?’’ are not taken into account. The scale percentage should lie on a 0 to 100 scale.
C. Twenty Questions About the GFT ‘‘Incompatible Goals’’
Choose your answer: Y ¼ ‘‘Yes,’’ N ¼ ‘‘No,’’ ‘‘n.a.’’ ¼ not applicable, ‘‘?’’ ¼ don’t know
1. Have you ever violated the specified work hours in the past month? 2. Have you ever exceeded the standard work hours in the past month? 3. Have you ever worked more than 12 hours in a day in the past month? 4. Have you ever been ordered by your manager to shorten your production time,
which might cause any unsafe factor for the plant operation process in the past six months?
5. Have you ever been pushed for working hour compare with your work in the past six months?
6. Do you think that the zero accident target day of your company is reasonable? 7. Have you ever been informed of a reduced production schedule over that
determined originally in the past month? 8. Have you ever been ordered to reduce a budget in the past three months? 9. Have you ever received any changed orders from your manager in the past
month? 10. Have you ever thought that the orders from your manager are difficult to under-
stand and unreasonable in the past month? 11. Have you ever been unreasonably blamed by your manager in the past month? 12. Have you ever been confused due to different orders from managers in the past
month? 13. Have you ever felt any discrepancy between company’s management policy and
orders from your managers in the past 12 months? 14. Have you ever quarreled with your coworkers over your job in the past month?
Accident Causation Management System 1035
15. Have you ever been ordered to actually sell goods produced by your company in the past six months?
16. Have you ever received any orders from your managers that are difficult to implement in the past month?
17. Have you ever felt your work capability is underestimated in the past three months?
18. Have you ever felt your salary is low considering the nature of your work in the past 12 months?
19. Have you ever felt that production scale increase or plant expansion is not poss- ible considering the financial status of your company in the past 12 months?
20. Have you ever been denied a purchasing requisition for tools or materials neces- sary to the work due to budget shortage in the past three months?
D. Reference Answers
1 ¼ N, 2 ¼ N, 3 ¼ N, 4 ¼ N, 5 ¼ N, 6 ¼ Y, 7 ¼ N, 8 ¼ N, 9 ¼ N, 10 ¼ N, 11 ¼ N, 12 ¼ N, 13 ¼ N, 14 ¼ N, 15 ¼ N, 16 ¼ N, 17 ¼ N, 18 ¼ N, 19 ¼ N, 20 ¼ N.
E. Scoring GFT ‘‘Incompatible Goals’’
Percent score ¼ (a=b)�100 where a ¼ no. of questions answered correctly (answered the same as the reference answers) and b ¼ no. of questions answered ‘‘Y’’ or ‘‘No.’’
F. Application of GFT Results
The case study illustrated in this article deals with a Korean chemical company oper- ating its own similar plants managing nine work=operation processes. Using kACMS, the GFT calculated score is regarded as an absolute value even though it could be difficult to decide the relative value of each score for each process (ques- tion). However, the GFT definition and the score difference for each process still serves a significant role. They are determinative in focusing and controlling any poss- ible problems of low scored processes or GFTs by effectively using resources.
References
Bainbridge, L. (1974). Analysis of Berbal Protocols from a Process Control Task, 24, Taylor & Francis, Washington, D.C.
Bainbridge, L. (1987). Ironies of Automation, 167, John Wiley, New York. Center for Chemical Process Safety. (1993a). Guidelines for Safe Automation of Chemical
Processes, 23, American Institute of Chemical Engineers, New York. Center for Chemical Process Safety. (1993). Guidelines for Engineering Design for Process
Safety, 5, American Institute of Chemical Engineers, New York. Center for Chemical Process Safety. (1998). Guidelines for Improving Plant Reliability through
Data Collection and Analysis, 78, American Institute of Chemical Engineers, New York. Center for Chemical Process Safety. (1994). Guidelines for Preventing Human Error in Process
Safety, 104–106, American Institute of Chemical Engineers, New York. Center for Chemical Process Safety. (1995). Plant Guidelines for Technical Management of
Chemical Process Safety, 73–201, American Institute of Chemical Engineers, New York. Crossman, E. R. F. W., Cooke, J. E., and Beishon, R. J. (1974). Visual Attention and the
Sampling of Displayed Information in Process Control, 136, Taylor & Francis, Washington, D.C.
1036 H. Kwon et al.
Embrey, D. E. (1986). Chem. Ind., 7, 456. Embrey, D. E. (1992). Reliab. Eng. Syst. Saf., 38, 201. Flanagan, J. C. (1954). Psychol. Bull., 51, 331. Genizzi, T. (1998). The 19th Japanese Safety Engineering Seminar, Section for the Prevention of
Fire and Explosion, 5, Tokyo. Gertman, D. I. and Blackman, H. S. (1994). Human Reliability and Safety Analysis Data
Handbook, 242, John Wiley, New York. Goh, S., Chang, B., Jeong, I., Kwon, H., and Moon, I. (1998). Comput. Chem. Eng., 22, 532. Heinrich, H. W. (1980). Industrial Accident Prevention, 27, McGraw-Hill, New York. International Labour Office. (1984). Encyclopaedia of Occupational Health and Safety, 6–57,
International Labour Office, Geneva. International Labour Office. (1998) Encyclopaedia of Occupational Health and Safety, 4th ed.,
Vol. II, 57, International Labour Office, Geneva. Kirwan, B. (1990). Human Reliability Assessment, 56, Taylor & Francis, Washington, D.C. Kirwan, B. and Ainsworth, L. K. (1993). Guide to Task Analysis, 63, Taylor & Francis,
Washington, D.C. Kirwan, B., Embrey, D. E., and Rea, K. (1988). in: Human Reliability Assessors Guide, 128,
UKAEA Safety and Reliability Directorate, Warrington, U.K. Kletz, T. (1991). An Engineer’s View of Human Error, 4, Taylor & Francis, New York. Petersen, D. (1984). Human-Error Reduction and Safety Management, 2nd ed., 172, Aloray,
New York. Process Safety Institute. (2000). Human Reliability Analysis, 12, EQE International,
Warrington, U.K. Reason, J. (1998). Human Error, 11, Cambridge University Press, Cambridge. Shepherd, A. (1985). Program. Learn. Educ. Technol., 22, 165. Swain, A. D. and Guttmann, H. E. (1983). Handbook of Human Reliability Analysis with
Emphasis on Nuclear Power Plant Applications, 132, U.S. Nuclear Regulatory Com- mission, Washington, D.C.
U.S. Chemical Safety and Hazard Investigation Board. (1999). The 600 K report: Commercial Chemical Incidents in the United States, 1987–1996: Special Congressional Summary, U.S. Chemical Safety and Hazard Investigation Board, Washington, D.C.
U.K. Health and Safety Executive. (1991). Successful Health and Safety Management, 32, HMSO, London.
Yoon, H., Lee, H., Kwon, H., and Moon, I. (1999a). Paper presented at European Symposium on Computer Aided Process Engineering, ESCAPE9, Budapest.
Yoon, H., Lee, H., Kwon, H., and Moon, I. (1999b). Paper presented at AIChE Annual Meeting, Dallas.
Yoon, H., Lee, H., Kwon, H., and Moon, I. (2000). Hydrocarbon Process. Sept., 84. 79, 9.
Accident Causation Management System 1037
Do you need a similar assignment done for you from scratch? Order now!
Use Discount Code "Newclient" for a 15% Discount!
