565

1 INTRODUCTION

Maritime accidents are a frequent occurrence that can

cause significant losses. This is due to the increasing

complexity and dynamism of the ship environment,

resulting from the rising number and size of vessels

[1], [2], [3], [4], [5], [6]. Navigational accidents are one

of the most common types of maritime accidents, and

they can result from various factors, including human

error, adverse weather conditions, technical issues

with the ship, or a combination of these factors [7].

The effects of navigational accidents can be severe

and destructive, mainly if the collision involves ships

carrying hazardous materials or a high volume of ship

traffic. This study has two types of navigational

accidents: collision and grounding. Navigational

accidents can cause physical damage to the ships,

including their hulls and engines. They can risk

environmental damage, such as oil spills or other

hazardous substances released into the sea. Different

types of navigational accidents, such as collisions with

other ships, bridges, or docks, can result in significant

casualties, as the impact force can cause severe

damage and potentially lead to sinking. Crew

members, passengers, and those in the surrounding

area can also be endangered [8] [9].

In many cases, the costs of ship accidents outweigh

the costs of preventing them. For example, in cases of

ship sinking, the loss is not only calculated from the

loss of the ship but also from the value of the lost

cargo [10]. According to Lloyd's Register Intelligence

Casualty statistics, 3,976 maritime accidents have

Comprehensive Analysis of Navigational Accidents

Using th

e MAART Method: A Novel Examination of

Human Error Probability in Maritime Collisions and

Groundings

.P. Bowo

, A.P. Gusti

, D.H. Waskito

, F.S. Puriningsih

, A. Muhtadi

& M. Furusho

National Research and Innovation Agency, Jakarta, Indonesia

Kobe University, Kobe, Japan

ABSTRACT: Navigational accidents are one of the most common types of maritime accidents, and they can

result from various factors, including human error, adverse weather conditions, technical issues with the ship,

or a combination of these factors. In this study, navigational accidents, including collision and grounding, that

occurred in Germany were analysed using the MAART method. The novelty of this research lies in its detailed

examination of the Human Error Probability (HEP) result, which has yet to be explored in previous studies.

There are 47 collision cases and 15 grounding cases in the 13-year occurrence period. In total, 290 causal factors

were found in the analysis. Furthermore, it is found that management, media, and machines are the main causal

factors in navigational accidents in Germany. In collision accidents, management factors had the highest

number of contributing factors, followed by media and machine factors. Contrary to grounding accidents, based

on the results of the EPC, the machine factor had the highest number of contributing factors to accidents,

followed by media and management. Finally, the human error probability values for collision accidents range

from 0.06 to 1, averaging 0.54. In contrast, the HEP values for grounding accidents range from 0.0048 to 1,

averaging 0.26.

http://www.transnav.eu

the

International Journal

on Marin

e Navigation

and Safety of Sea Transportation

Volume 18

Number 3

September 2024

DOI: 10.12716/1001.18.03.1

566

occurred where a ship lost more than 100 gross

tonnes, resulting in 15,738 deaths [11].

Navigational accident risk assessment has become

a crucial aspect of maritime safety and traffic

management in reducing the number of ship

collisions. Research has indicated that navigational

conflicts are the leading cause of ship collisions [12].

To prevent maritime accidents globally, the

International Convention on Standards for Training,

Certification, and Supervision, ISM, and the

International Regulations for Preventing Collisions at

Sea have been put in place. A better understanding of

the human element and strategies to mitigate human

errors can help prevent accidents [13]. Human

reliability analysis (HRA) is a technique that predicts

the safety of particular activities involving people.

HRA considers various factors that may lead to

human errors and the potential outcomes of such

errors. HRA has been widely employed in assessing

risk and estimating the likelihood of human error in

specific activities [14] [15] [16] [17].

There are several methods for estimating the

probability of human error (HEP) in different systems,

including THERP (Technique for Human Error Rate

Prediction), ATHEANA (Analytical Technique for

Human Error Analysis), and SHERPA (Systematic

Human Error Reduction and Prediction Approach)

models [18]. CREAM is a method for measuring data

reliability that considers how often people answer

questions correctly, even when faced with time

constraints. SPAR-H is similar but focuses on how

frequently people make mistakes, while HCR assesses

how likely people are to respond correctly regardless

of time constraints. The HEART technique is used to

minimise the chances of people making mistakes.

Until recently, HRA has been used to evaluate risks

for complex systems. However, the cognitive-based

THERP (CB-THERP) method, which integrates DSA,

THERP, and HCR, has been developed to quantify

human exposure in nuclear power plants after an

accident [19].

HEART is a flexible and easy-to-use method for

analysing accidents in various industries, such as

aviation, rail, offshore drilling, and maritime

operations [20], [21], [22], dan [23], [24]. HRA is used

to assist with risk assessment for complex systems.

Islam et al. modified the HEART method to evaluate

and measure human error in maritime,

environmental, and operational conditions to enhance

the safety and reliability of maintenance and repair

practices [25]. Akyuz et al. employed the HEART and

type-2 fuzzy interval sets to assess human reliability

in cargo operations [26].

According to de Maya's research, the combination

of Hierarchical Task Analysis (HTA) and the Human

Error Assessment and Reduction Technique (HEART)

can be used to predict potential errors that may occur

when handling fires on passenger ships [27]. Another

study by W. Wang et al. demonstrated that by

modifying the HEART method with the Railway

Action Reliability Assessment (RARA) technique and

the fuzzy analytic network process (FANP), it is

possible to evaluate the likelihood of human error in

high-speed rail [28]. Bowo's study proposes a hybrid

methodology for assessing human errors by

integrating the HEART-4M method with a new

approach called Maritime Accident Analysis and

Reduction Techniques (MAART) [29]. Bowo [30] has

recently conducted an MAART study for collision

accidents. Studies with different accident cases, such

as sinking and collisions, should be performed to

enhance and delve more into MAART’s capability to

analyse the probability of human error. The new

phenomen

a could be explored within MAART by

adding different accident types, cases, and data.

This study aims to investigate the influence of

human error on collision and grounding accidents by

using the MAART method. This work is a

development of earlier MAART studies that

incorporate the grounding accident. The disparity and

correlation between the two accidents will be defined

and assessed. The novelty of this research is the

emphasis on examining HEP results, which needs to

receive more attention in previous studies.

2 DATA AND METHODOLOGY

This study's maritime navigational accident data

consists of Germany's collision and grounding

accidents from 2008 to 2020. The data will be analysed

qualitatively and quantitatively by using MAART.

2.1 Data

The official accident report is considered reliable

secondary data because it has been created by

accident investigators by interviewing and analysing

the primary data source [31,32]. The maritime

navigational accident data reports are retrieved from

the Federal Bureau of Maritime Casualty Investigation

(BSU) website in Germany. The analysed accident

reports occurred from 2008 to 2020, as shown in Table

1 below. During this period, 30 collisions were

reported, with the highest number occurring in 2008

and the most recent in 2017. Groundings were more

common than collisions, with a total of 47 incidents

reported. The highest number of groundings occurred

in 2013, while 2016 and 2017 saw the fewest incidents.

The accident reports that were collected are limited to

English-written reports. Therefore, it has been several

years with no data reports. In collision cases, if there

is information about two or more ships involved,

those ships will be analysed separately. From 31 data

reports of collision, 48 ships were analysed. Therefore,

in total, there are 48 data reports and 62 ships

analysed.

Microsoft Excel software is utilised to record the

data and tabulate it. The data that is extracted from

the data reports include accident time, type of ships,

accident locations, weather conditions, and causal

factors of the accidents. To prevent subjectivity of the

data extraction, the authors only stated the causal

factors written in the data reports; no self-opinion is

included in the analysis. The calculation of the

Human Error Probability (HEP) also uses Microsoft

Excel.

567

Table 1. Maritime Navigational Accidents Data Reports

________________________________________________

Year Collision Grounding

Data Ships

________________________________________________

2008 8 12 2

2009 1 2 -

2010 1 2 -

2011 2 3 2

2012 - - 2

2013 5 9 1

2014 7 12 1

2015 4 4 1

2016 1 1 2

2017 1 2 1

2018 - - 1

2019 - - 1

2020 - - 1

________________________________________________

Total 30 47 15

________________________________________________

2.2 Methodology

The maritime navigational data in this study is

analysed using the Maritime Accident Analysis And

Reduction Technique (MAART). MAART is a method

to find out how likely something will happen in the

maritime industry by combining the HEART – 4M

method approach and the Technique for Order

Preference by Similarity to Ideal Solution (TOPSIS) to

evaluate Human Error Probability (HEP) [33]. This

method divided the analysis into two stages: the

qualitative stage and the quantitative stage. Figure 1

below shows the framework of the MAART method

stages.

Figure 1. MAART method stages [33]

2.2.1 Qualitative stage

In the qualitative stage, the working conditions

and causal factors of the accidents are analysed.

Firstly, the working conditions when the accidents

occurred are determined. In this step, the working

condition will be matched with nine Generic Tasks

(GT) in the MAART method. These GTs are

categorised into two categories: challenging tasks and

convenient tasks. The challenging task category

consists of three working conditions, with the highest

nominal human unreliability (NHU) associated with

unfamiliarity with the working conditions.

On the other hand, the convenient task category

includes six working conditions, with the lowest

NHU associated with a highly familiar, highly

practised, routine task occurring several times per

hour, performed to the highest possible standards by

a highly motivated, highly trained, and experienced

person. The NHU values are an indication of the

likelihood of errors occurring during the task

performance. This information can be used to identify

areas where human error is likely to occur and to

develop appropriate measures to prevent accidents in

the maritime industry. Table 2 below shows the GTs

used in this study.

Table 2. Generic Tasks

________________________________________________

Code Working condition NHU

________________________________________________

Challenging Task

A Unfamiliar with the condition 0.55000

B Reinstate the system to its original state 0.26000

on a single attempt

C Complex task 0.16000

Convenient Task

D An adequately simple task 0.09000

E The routine, highly practiced, rapid task 0.02000

F Reinstate the system to its original state 0.00300

G Entirely familiar, highly practiced, routine 0.00040

task occurring several times per hour,

performed to the highest possible standards

by a highly motivated, highly trained, and

experienced person

H Respond correctly to the system instruction 0.00002

M The miscellaneous task for which no 0.03000

description can be found.

________________________________________________

Table 3. Error Producing Conditions [34]

________________________________________________

Man Factor

________________________________________________

1. Experience

EPC 1 Unfamiliarity (x 17)

EPC 12 Misperception of risk (x 4)

EPC 22 Lack of experience (x 1.8)

2. Skill and Knowledge

EPC 7 Irreversibility (x 8)

EPC 9 Technique unlearning (x 6)

EPC 11 Performance ambiguity (x 5)

EPC 15 Operator inexperience (x 3)

EPC 20 Educational mismatch (x 2)

3. Psychological

EPC 21 Dangerous incentives (x 2)

EPC 28 Low meaning (x 1.4)

EPC 29 Emotional stress (x 1.3)

EPC 31 Low morale (x 1.2)

EPC 34 Low mental workload (x 1.1)

4. Physical

EPC 27 Physical capabilities (x 1.4)

EPC 36 Task pacing (x 1.06)

EPC 38 Age (x 1.02)

5. Health

EPC 30 Ill-health (x 1.2)

EPC 35 Sleep cycle disruption (x 1.1)

Media Factor

EPC 33 Poor environment (x 1.15)

________________________________________________

Machine Factor

________________________________________________

EPC 3 Low signal-noise ratio (x 10)

EPC 8 Channel overload (x 6)

EPC 23 Unreliable instruments (x 1.6)

________________________________________________

Management Factor

________________________________________________

1. Coordination

EPC 2 Time shortage (x 11)

EPC 6 Model mismatch (x 8)

EPC 24 Absolute judgments required (x 1.6)

EPC 25 Unclear allocation of function (x 1.6)

EPC 37 Supernumeraries/ lack of human resources (x 1.03)

2. Rules and procedures

EPC 4 Features over-ride allowed (x 9)

EPC 5 Spatial and functional incompatibility (x 8)

EPC 32 Inconsistency of displays (x 1.2)

3. Communication

EPC 10 Knowledge transfer (x 5.5)

EPC 13 Poor feedback (x 4)

EPC 14 Delayed/incomplete feedback (x 3)

EPC 16 Impoverished information (x 3)

EPC 18 Objectives conflict (x 2.5)

EPC 19 No diversity of information (x 2.5)

4. Monitoring

EPC 17 Inadequate checking (x 3)

EPC 26 Progress tracking lack (x 1.4)

________________________________________________

568

Following the identification of working conditions,

the causal factors of the accidents are analysed in the

qualitative stage. To determine the causal factors, the

analysis and conclusions sections of the data reports

are scrutinised, and every causal factor that is found

in the MAART method is recorded. The MAART

method categorises causal factors as Error Producing

Conditions (EPC), of which there are 38. These factors

are classified into four categories, namely, man,

machine, media, and management factors. This

categorisation helps prioritise the mitigation and

resolution of incidents from a specific standpoint.

Additionally, every EPC has a multiplier assigned to

it to determine its weight in the quantitative stage.

The multiplier values vary for each EPC and are

indicative of the rarity of their occurrence in the cases.

Table 3 provides a comprehensive categorisation of

each EPC into the four categories mentioned above

and its multiplier. While the factors of media and

machine do not have sub-factors, the man factors have

more factors categorised and further sub-factors.

2.2.2 Quantitative stage

In the quantitative stage, all the qualitative data

that obtained in the qualitative stage will be

quantified to calculate the value of Human Error

Probability (HEP). First, the weight for every obtained

EPC, namely the Assessed Proportion Effect (APE), is

assigned by using TOPSIS calculation to calculate the

Assessed Effect (AIV) as stated in Equation 1. After

calculating the weight for every EPC, then forming

the EPC series by arranging the weightiest APE to the

less APE. By this information, the researchers may

know which factor is the main causal factor of the

accidents and what kind of series of actions that might

influence the condition to the occurrences. The

TOPSIS calculation was conducted as has been

performed by Bowo et al. [30].

(1)1



= −+





∏

AIV EPC APE

(1)

After determining the AIV value for each EPC, the

last step is to calculate the HEP value using Equation



= ⋅





∏

value i

HEP NHU AIV

(2)

3 RESULTS

The results of maritime navigational accidents

analysis are separated into three parts of explanation

as below:

3.1 Type of works

Analysing the type of work when maritime

navigational accidents occur can describe the situation

right before the accidents occur. Maritime

navigational accidents differ in two types: collision

and grounding. The results for collision and

grounding accidents are different. Whereas in the

collision accidents, more accidents occurred during

challenging tasks rather than the convenient task, and

vice versa for the grounding accident.s Furthermore,

not all types of work are applied in navigational

accidents. Table 4 shows the results of the generic task

found in the analysis.

Table 4 presents a breakdown of the type of work

being performed during maritime navigational

accidents, categorised by the severity of the task. The

table includes two categories of tasks: challenging

tasks and convenient tasks. Challenging tasks are

further classified into two subcategories: type B tasks,

which involve reinstating the system to its original

state on a single attempt, and type C tasks, which are

complex tasks. On the other hand, convenient tasks

are also classified into two subcategories: type D

tasks, which involve adequately simple tasks, and

type E tasks, which are routine, highly practised, and

rapid tasks. Additionally, there is a category called

type F tasks, which involve reinstating the system to

its original state, and there were no reported

occurrences of this type of task during collisions. The

table shows that type C tasks were the most common

tasks being performed during both collision and

grounding accidents, accounting for 25 and 5

occurrences, respectively. Type D tasks were the next

most common type of task, with 18 occurrences

during collisions and four during groundings. Type E

tasks occurred three times during collisions and five

times during groundings. Finally, there was only one

reported incident of type B tasks during collisions and

no reported incidents during groundings, and one

reported incident of type F tasks during groundings

and none during collisions.

Table 4. Generic Tasks analysis result

________________________________________________

Generic Tasks Collision Grounding Total

________________________________________________

Challenging Tasks

B 1 - 1

C 25 5 30

Convenient Tasks

D 18 4 22

E 3 5 8

F - 1 1

________________________________________________

3.2 Causal factors

There are 290 causal factors from 62 involved ships

found in the analysis of maritime navigational

accidents. All 4M factors are causal factors involved in

the accidents.

3.2.1 Man factors

Figure 3 presents the analysis of the causal factors

in collision and grounding accidents with respect to

the man factor. The analysis shows that experience,

skill, and knowledge are significant factors in both

types of accidents. However, there are some notable

differences in the contributing EPCs for each type of

accident.

For experience, EPC 1 has a higher value in

collision accidents with a score of 5, compared to only

1 in grounding accidents. On the other hand, EPC 12

has a much higher value in collision accidents, with a

569

score of 16, compared to 7 in grounding accidents.

This indicates that the importance of experience is

more significant in collision accidents, where complex

and challenging situations require high levels of

experience.

For skill and knowledge, EPC 7, which refers to

irreversibility, has a higher value in collision accidents

with a score of 8, compared to a score of 0 in

grounding accidents. This suggests that the ability to

reverse actions and decisions is more critical in

collision accidents. EPC 11, which refers to

performance ambiguity, has a higher value in collision

accidents with a score of 6, while it does not have any

score in grounding accidents. This indicates that

clarity and precision in decision-making are more

critical in collision accidents.

In terms of psychological factors, dangerous

incentives (EPC 21) have a higher value in collision

accidents, with a score of 8, compared to 0 in

grounding accidents. This suggests that external

pressures or motivations may contribute to decision-

making in collision accidents. Emotional stress (EPC

29) and low mental workload (EPC 34) also have

higher values in collision accidents with scores of 2

and 4, respectively, compared to only 1 and 0 in

grounding accidents.

Regarding physical factors, task pacing (EPC 36)

has a higher value in collision accidents with a score

of 10, compared to only 0 in grounding accidents. This

indicates that the pace of task execution is a more

significant contributing factor in collision accidents.

Finally, sleep cycle disruption (EPC 35) has the

same value for both types of accidents, with a score of

1. This suggests that sleep cycle disruption is an

important physical factor contributing to collision and

grounding accidents.

Figure 3. “Man” causal factors

3.2.2 Media and Machine Factors

Figure 4 shows the number of media and machine

factors occurrences in collision and grounding

maritime accidents. Media and machine factors were

also found to influence maritime navigational

accidents. The total number of media factors is 28

EPCs from 62 ships analysed. Collision and

grounding accidents account for about 50% of the

collected data, which shows that the media influences

accidents. The media factor covers the environmental

situation, which may influence the condition of the

seafarers in performing the task. The media factor

represented by EPC 33 shows a higher occurrence in

collision accidents, with 21 instances recorded

compared to 7 in grounding accidents.

On the other hand, the machine factor, represented

by EPC 3 and EPC 23, shows a higher occurrence of

grounding accidents relative to the machine factor in

the same category. This suggests that media factors

are more likely to contribute to collision accidents,

while machine factors are more likely to contribute to

grounding accidents.

Figure 4. Media and machine causal factors

3.2.3 Management factors

Management factors significantly influence

maritime navigational accidents, serving as both main

causal and contributing factors. Of the 172 EPCs

associated with management factors, only EPC 4 does

not contribute to these accidents. Figure 5 illustrates

the distribution of each selected EPC in collision and

grounding accidents. Communication emerges as the

most common problem faced by seafarers on the

bridge during navigational accidents. Poor feedback,

information, knowledge transfer, and diversity of

information can lead to situations where seafarers on

the bridge make the wrong decisions. As seen in these

cases, lack of monitoring is the second most common

causal factor, making accidents hard to prevent.

Communication and monitoring issues in maritime

navigational accidents, such as collisions and

groundings, are often linked to a breakdown in the

exchange of information between seafarers

responsible for navigation and communication. Poor

communication also affects coordination among

seafarers on the bridge. Coordination sub-factors also

include 29 EPCs. Additionally, there are cases where

maritime navigational accidents need proper rules

and procedures.

Figure 5. Management causal factors

570

3.3 Quantitative Results

3.3.1 Determining the APE Value

Tables 5 and 6 show the case numbers, EPC series,

TOPSIS calculation values of APE, values λ max,

consistency index (CI), and consistency ratio (CR).

λmax is the maximum eigenvalue of the pairwise

comparison matrix used to determine the weights of

the criteria, which will utilised to calculate CI and CR.

The Pairwise Comparison Matrix (PCM), which

compares each criterion with all other criteria and

assigns a relative weight to each criterion, has a better

degree of consistency when the CI values are smaller.

The CR value measures the consistency of PCM by

comparing the CI with random index (RI) values.

Greater matrix consistency was indicated by lower CR

values.

Table 5. The results of TOPSIS calculation for each EPC on

Collision accidents

________________________________________________

Case EPC Series APE λmax CI CR

________________________________________________

1 EPC 26 0.253 8.6 0.09 0.061

EPC 25 0.203

EPC 19 0.175

EPC 36 0.132

EPC 13 0.098

EPC 17 0.093

EPC 12 0.046

EPC 33 0.001

2a EPC 10 0.459 4.24 0.08 0.09

EPC 12 0.269

EPC 23 0.262

EPC 33 0.011

2b EPC 36 0.327 5.21 0.05 0.047

EPC 34 0.325

EPC 10 0.305

EPC 33 0.042

EPC 16 0.002

________________________________________________

In the table, the series of the EPC are arranged

based on the highest APE number. The highest APE

value indicates a leading factor that contributes to

human error. Since the CR values are all below 0.1, it

can be stated that the TOPSIS method for determining

APE does not have discrepancies and inconsistencies.

Table 6. The results of TOPSIS calculation for each EPC on

Grounding accidents

________________________________________________

Case EPC Series APE λmax CI CR

________________________________________________

1 EPC 35 0.270 8.97 0.14 0.098

EPC 26 0.230

EPC 22 0.145

EPC 12 0.112

EPC 10 0.089

EPC 17 0.076

EPC 16 0.076

EPC 33 0.001

5 EPC 17 0.392 4.21 0.07 0.08

EPC 26 0.325

EPC 10 0.189

EPC 33 0.093

________________________________________________

3.3.2 Main causal factors

The main causal factors in maritime navigational

accidents are determined by the highest value of the

Assessed Proportion Effect (APE) for each ship

analysed and are composed of a series of Error

Producing Conditions (EPC). Table 7 shows the

distribution of the main causal factors, which differ

from collision and grounding accidents. In

navigational accidents, the most critical error is

inadequate checking (EPC 17), followed by progress

tracking lack (EPC 26) and unreliable instruments

(EPC 23). The monitoring sub-factors, which include

EPC 17 and EPC 26, are the main causes of accidents.

The main causal factors for collision and

grounding accidents differ. EPC 23 is the main causal

factor in most grounding accidents, while it occurs

only once in collision accidents. Collision accidents

have more EPCs as the main causal factors than

grounding accidents. Management factors

predominantly dominate the main causal factors for

maritime navigational accidents. The management

sub-factors identified as the main causal factors

include monitoring, communication, coordination,

and rules and procedures.

The remaining causal factors have varying values

of APE, with some having high values for one type of

accident but not the other. Some EPCs, such as EPC 7

and EPC 12, have a value of 3 for Collision but do not

have any value for Grounding. Similarly, EPC 35 has

a value of 1 for both types of accidents, while EPC 2

has a value of 1 only for Grounding.

Table 7. Number of occurrences of EPC as the highest APE

for each ship

________________________________________________

Highest APE for each ship Navigational Accidents

________________________________________________

EPC - 4M Collision Grounding

________________________________________________

Management-Monitoring (EPC 17 and EPC 26) 14 4

Management-Communication (EPC 10, EPC 13, 13 3

EPC 14, EPC 16, EP 19)

Management-Coordination (EPC 24 and 4 1

EPC 25)

Man-Physical (EPC 36) 3 -

Man-Experience (EPC 12) 3 -

Man-Skill and Knowledge (EPC 7) 3 -

Man-Psychological (EPC 21) 3 -

Man-Health (EPC 35) 1 1

Management-Rules and Procedures (EPC 32) 1 -

Media (EPC 33) 1 -

Machine (EPC 23) 1 6

________________________________________________

In addition to the main causal factors, the analysis

of maritime navigational accidents also recorded all

contributing factors. The study found that all four

factors, namely man, machine, media, and

management, contribute to these accidents. The

subsequent paragraph will elaborate on the findings

of the contributing factors in the analysis of maritime

navigational accidents.

3.3.3 Calculate the AIV and HEP value

Table 8 shows the example of the value of AIV and

HEP from two collision accidents using Equation 2.

The results show that in case 1, the probability of

human error is 43.8%. Meanwhile, if human error is

involved in two ships, the HEP is calculated

particularly for each ship, as shown in case numbers

and 2b. The result shows that ship A has a

probability of human error of 100% and ship B has

40.37%, which means that ship A has more influence

on the accident due to adverse conditions of human

operator conditions. If the value of the HEP

calculation is more than 1, the value is rounded off to

1. For the grounding accident, Table 9 shows the HEP

value for two example cases, cases 1 and 5, which are

6.21% and 7.56%, respectively, which indicates that in

grounding accidents, the influence of human error is

considerably lower than in collision accidents.

571

Figure 5 shows the distribution of HEP values for

collision and Grounding to get an overview of the

proportion of HEP values in accident cases, especially

in navigational accidents. For collision accidents, it is

evident that there is a total of 15 ships with an HEP of

more than 75% and 12 ships with an HEP of more

than 50%, which shows that the influence of human

factors is paramount in contributing to the occurrence

of accidents. In contrast, the influence of human error

is insignificant in grounding accidents in Germany,

with only three ships out of 15 with a HEP of more

than 75%. The distinct difference can be caused by the

high involvement of the machine failure in the

grounding accidents where the value affects the APE

value in the calculation of AIV and HEP.

Figure 5. Human Error Probability (HEP) value distribution

for Collision and Grounding

4 DISCUSSION AND CONSIDERATIONS

In the previous research by Bowo 2024 [30], collision

accidents that occurred in Hong Kong were analysed

using MAART. The research focused on the use of the

MAART method. In this study, additional maritime

navigational accidents that occurred in Germany were

analysed using the same method [31]. This paper is a

development of the previous research; by adding

types of accidents other than ship collisions, the

results show that management factors are not the

main factor in grounding accidents. The novelty of

this research lies in its detailed examination of the

HEP result, which has been underexplored in

previous studies. The MAART method categorises

working situations that lead to accidents to determine

their severity and novelty, which is important in

characterising the situations that have a higher

probability of accidents. The Error Producing

Conditions (EPC) are then categorised into four main

factors: man, machine, media, and management,

which helps in understanding the main causal factors

before delving into the detailed factors for the

mitigation process. The MAART method also includes

multi-criteria decision-making, such as TOPSIS, to

make the calculation of the Human Error Probability

(HEP) more objective.

Based on the results of the EPC, the management

factor had the highest number of contributing factors

to accidents, followed by media and machine factors.

In this case, the management factor refers to bridge

resource management (BRM), which is an effort to

manage and maximise all resources on board (human

and machine) to maintain the safe operation and

passage of the ship [33]. Although there is an element

of human error, the sub-factors of the management

causal factor emphasise how the failure and

disjointedness of the BRM on the ship contribute to

accidents.

The EPC "Progress Tracking Lack" is the highest

number of EPC followed by "Inadequate Checking" in

monitoring causal factors. Progress tracking lack is the

case when the OOW (Officer on Watch) does not

frequently check the vessel tracking, and Inadequate

checking is considered as the inability of the OOW to

check vessel condition properly. Both cases are one of

the main factors that caused the ship to deviate from

the original track, leading to a collision and

grounding[35].

To analyse the accident, the two mentioned

failures occurred because of errors in conducting the

BRM technique. For example, the planning failed

when the master instructed to have two officers on

standby during the watch hour, which did not comply

with the ISM manual, and the master did not

distribute the task among officers. Due to working

overtime and not complying with STCW-2010[36], the

officer experienced fatigue; consequently, the

"Inadequate checking" cases occurred.

In addition, the crew must constantly monitor the

condition or sensors on the ship. Good coordination

between the bridge and ECR (Engine Control Room)

needs to be implemented on this occasion. For

example, an "Inadequate checking" happened when

the ECR crew did not monitor the operation of the

engine order that the master had telegraphed, which

led to an engine failure and collision. From the

previous case, it can also be concluded that the

appropriate management system is not only applied

on the bridge but also to maintain good coordination

with another department to ensure the ship's safety.

Table 8. AIV and HEP Values for two cases in Collision accidents

___________________________________________________________________________________________________

TOP BODY

___________________________________________________________________________________________________

No GT NHU EPC APE EPC APE EPC APE EPC APE EPC APE EPC APE EPC APE EPC APE HEP

___________________________________________________________________________________________________

1 C 0.16 26 0.253 25 0.203 19 0.175 36 0.132 13 0.098 17 0.093 12 0.046 33 0.001 0.438

AIV=1.101 AIV=1.122 AIV=1.262 AIV=1.008 AIV=1.294 AIV=1.186 AIV=1.137 AIV=1.000

2a C 0.16 10 0.459 12 0.269 23 0.262 33 0.011 1

AIV=3.064 AIV=1.807 AIV=1.157 AIV=1.002

2b C 0.16 36 0.327 34 0.325 10 0.305 33 0.042 16 0.002 0.4037

AIV=1.020 AIV=1.032 AIV=2.372 AIV=1.006 AIV=1.003

___________________________________________________________________________________________________

Table 9. AIV and HEP Value for two cases in Grounding accidents

___________________________________________________________________________________________________

TOP BODY

___________________________________________________________________________________________________

No GT NHU EPC APE EPC APE EPC APE EPC APE EPC APE EPC APE EPC APE EPC APE HEP

___________________________________________________________________________________________________

1 E 0.02 35 0.27 26 0.23 22 0.145 12 0.112 10 0.089 17 0.076 16 0.076 33 0.001 0.0621

AIV=1.027 AIV=1.092 AIV=1.116 AIV=1.337 AIV=1.399 AIV=1.152 AIV=1.151 AIV=1.000

5 E 0.02 17 0.392 26 0.325 10 0.189 33 0.093 0.0756

AIV=1.784 AIV=1.130 AIV=1.850 AIV=1.014

___________________________________________________________________________________________________

572

Communication problems are one of the most

frequent contributing factors to maritime accidents.

Especially in a collision accident, the inter-ship

communication problem led to a severe

accident[37][38]. Since communication in maritime

communication is related to the way to inform about

the ongoing situation or condition regarding the

vessel, there is a possibility that misinterpretation

occurs. In the maritime sector, the lines of

communication can be divided into Internal (inter-

ship) and external (Ship to ship, ship to VTS)[39]. The

inter-ship communication involves the interaction

between crew members, crew and their captain, and

captain and pilot. The problem that frequently arises

within the bridge is the different mental models

between the officer, captain, or pilot. Since they have

different views about the situation, they sometimes do

not communicate or share their thoughts for several

reasons. This situation will not create a closed-loop

communication, which will hinder successful

communication on the bridge [40].

Communication problems that occur in maritime

accidents can also be distinguished by type. This

study is called the EPC. According to the EPC

calculation, In Germany, the highest number of

communication problems that occur from a collision

accident is "Poor Feedback", while for the grounding

accident, it is "Impoverished information" and

"Knowledge Transfer." To maintain proper

communication, the sender and the receivers are

responsible for ensuring that all the information has

been received clearly or maintaining closed-loop

communication. In a "Poor Feedback" case, the

receiver does not reply to the sender with adequate

information, such as incomplete information

regarding the position, language difficulties, or

information different from the actual intention.

In grounding accidents, based on the results of the

EPC, the machine factor had the highest number of

contributing factors to accidents, followed by media

and management. In this case, the failure of unreliable

instruments had the highest number of EPCs.

Specifically, issues such as faulty navigational aids,

radar malfunctions, and defective communication

systems were identified as primary contributors to

these incidents. The prevalence of these technical

failures underscores the critical need for robust and

reliable instrumentation on board vessels.

The analysis revealed that outdated or poorly

maintained instruments often failed at crucial

moments, leading to unreliable instruments,

navigational errors, and, ultimately, grounding

incidents. These findings highlight the importance of

regular maintenance and timely upgrades of

navigational and communication equipment to ensure

their reliability and effectiveness.

Additionally, the study pointed out that media

factors, including poor environment, also played a

significant role. This includes weather forecasts and

maritime warnings that mislead the crew and hinder

effective decision-making. The integration of real-time

data and improved communication channels between

vessels and maritime authorities were suggested as

measures to mitigate these risks.

Management factors, while not the leading cause,

still contributed notably to grounding accidents.

Issues such as communication and coordination were

commonly observed. The research advocates for

enhanced training programs that emphasise the

operation and troubleshooting of navigational

instruments, as well as stricter adherence to

maintenance schedules.

On the other hand, research conducted by Bowo

[24] using the HEART method showed that the

management factor had the highest number of

contributing factors to grounding accidents, followed

by communication, human, and machine. The

different results show that the methodologies and

contexts in which these studies were conducted might

influence the outcomes. For instance, variations in

data collection techniques, sample sizes, and specific

circumstances of the accidents being analysed could

lead to different conclusions about the primary

contributing factors. Moreover, the emphasis on

management factors in Bowo's study highlights the

critical role that organisational and administrative

practices play in ensuring maritime safety. This

contrasts with this research that may focus more on

technical aspects of grounding accidents. These

discrepancies underscore the complexity of grounding

accidents and suggest that a multifaceted approach,

considering various factors and perspectives, is

essential for a comprehensive understanding and

effective prevention strategies.

Based on the classification of the EPC and TOPSIS

calculation to estimate the APE, the HEP can be

calculated for each ship. The analysis of the Human

Error Probability (HEP) results for both collision and

grounding accidents indicates some significant

differences. The HEP values for collision accidents

range from 0.06 to 1, with an average of 0.54. In

contrast, the HEP values for grounding accidents

range from 0.0048 to 1, with an average of 0.26.

The higher average HEP values for collision

accidents suggest that human errors are more

prevalent in such accidents. The most frequent HEP

values for collision accidents range between 0.4 and

0.6, indicating that human error is present in almost

half of all collisions. In contrast, the most frequent

HEP values for grounding accidents range between

0.1 and 0.2, indicating that human error is present in

about a quarter of all groundings.

Moreover, the HEP values for collision accidents

show a wider range than those for grounding

accidents, with some values as high as 1. This

suggests that human errors can be a major contributor

to the occurrence of collision accidents. On the other

hand, the HEP values for grounding accidents are

generally lower, with the highest value being 1,

indicating that the occurrence of grounding accidents

is less likely to be due to human error. However, it

should be noted that there is a difference in the

number of data used in the study between collision

and grounding accidents, which could have an impact

on the results. Nonetheless, the analysis of the HEP

results suggests that human error is a significant

factor in both collision and grounding accidents, but it

is more prevalent in collision accidents.

For further studies, it is noted that the value of

HEP does not represent the overall risk value. The

probability value should be incorporated with the

severity value of each accident. Incorporating the risk

573

rating for each EPC within the accident as the

quantified risk value could enhance the effectiveness

of reducing the risk of maritime accidents.

5 CONCLUSIONS

In conclusion, this study added grounding as the

additional maritime navigational accident that

occurred in Germany using the MAART method. The

novelty of this research lies in its detailed examination

of the HEP result, which has been underexplored in

previous studies. The analysis focused on identifying

the causal factors, categorising them into four major

factors (man, machine, media, and management), and

examining their contributions to collision and

grounding accidents as a navigational maritime

accident by using the EPC-4M and TOPSIS method.

The aim of this study is to identify what errors may

occur in the human response to marine navigational

accidents using MAART. This study used data on

collisions and groundings in Germany from 2008 to

2020. A total of 48 reports and 62 vessels were

analysed, with the limitation that only English reports

were processed. The results highlight the significant

role of management factors in influencing collision

accidents, followed by man, media and machine

factors. However, the data show notable differences in

the Error Producing Condition and Generic Task

Analysis involved in collision and grounding cases.

On the other hand, in grounding accidents, based on

the results of the EPC, the machine factor had the

highest number of contributing factors to accidents,

followed by media and management.

The novel approach in MAART methodology

incorporates the TOPSIS methodology to generate

APE that will be utilised to calculate the HEP for each

ship. The TOPSIS methodology provides a more

accurate value of APE rather than the general

judgment from the users. The more precise value of

APE will be instrumental in calculating the value of

AIV and HEP for each EPC and ship in maritime

navigational accidents. TOPSIS calculation shows

that, for the collision accidents, the EPCs related to

management have the highest occurrence of APE

value for each accident and ship case 32 times out of

47. On the other hand, EPCs related to machines have

the highest appearance, with six times out of 15. The

results of APE correspond with the HEP results.

Regarding collision accidents, 57% of the ships have

more than 0.5 -1 probability of human error, which

concluded that for collision cases, human error is

solely the most dominant factor. For the grounding,

only 4 of the ships out of 15 have more than 50% HEP,

which indicates that machine factors also play

important roles that can affect human error in

operating the equipment or instruments.

By understanding the specific factors and

estimating the pinpoint value of human error

probability that contribute to navigational maritime

accidents, this study provides valuable insights for the

development of preventive measures and the

enhancement of safety practices in the maritime

industry. It highlights the significance of effective

management, communication, monitoring, and

adherence to procedures in minimising the risk of

accidents and ensuring the safe navigation of vessels.

REFERENCES

[1] Fan L, Wang M, Yin J. The impacts of risk level based on

PSC inspection deficiencies on ship accident

consequences. Research in Transportation Business and

Management. 2019;33:100464.

[2] Bowo LP, Mutmainnah W, Furusho M. The

Development of Marine Accidents Human Reliability

Assessment Approach: HEART Methodology and MOP

Model. TransNav, the International Journal on Marine

Navigation and Safety of Sea Transportation. 2017;11:63–

68.

[3] Li B, Pang FW. An approach of vessel collision risk

assessment based on the D-S evidence theory. Ocean

Engineering. 2013;74:16–21.

[4] Weng J, Yang D. Investigation of shipping accident

injury severity and mortality. Accid Anal Prev.

2015;76:92–101.

[5] Weng J, Ge YE, Han H. Evaluation of Shipping Accident

Casualties using Zero-inflated Negative Binomial

Regression Technique. Journal of Navigation.

2016;69:433–448.

[6] Ko YG, Kim SJ, Paik JK. Effects of a deformable striking

ship’s bow on the structural crashworthiness in ship–

ship collisions. Ships and Offshore Structures.

2018;13:228–250.

[7] Kim K Il, Jeong JS, Lee BG. Study on the analysis of near-

miss ship collisions using logistic regression. Journal of

Advanced Computational Intelligence and Intelligent

Informatics. 2017;21:467–473.

[8] Du L, Goerlandt F, Kujala P. Review and analysis of

methods for assessing maritime waterway risk based on

non-accident critical events detected from AIS data.

Reliab Eng Syst Saf. 2020;200:106933.

[9] Pedersen PT, Chen J, Zhu L. Design of bridges against

ship collisions. Marine Structures. 2020;74:102810.

[10] Chen J, Bian W, Wan Z, et al. Identifying factors

influencing total-loss marine accidents in the world:

Analysis and evaluation based on ship types and sea

regions. Ocean Engineering. 2019;191:106495.

[11] Chen J, Bian W, Wan Z, et al. Factor assessment of

marine casualties caused by total loss. International

Journal of Disaster Risk Reduction. 2020;47:101560.

[12] Kim K Il, Jeong JS, Lee BG. Study on the analysis of

near-miss ship collisions using logistic regression.

Journal of Advanced Computational Intelligence and

Intelligent Informatics. 2017;21:467–473.

[13] Wang Y, Fu S. Framework for Process Analysis of

Maritime Accidents Caused by the Unsafe Acts of

Seafarers: A Case Study of Ship Collision. J Mar Sci Eng.

2022;10:1793.

[14] Sujan MA, Embrey D, Huang H. On the application of

Human Reliability Analysis in healthcare: Opportunities

and challenges. Reliab Eng Syst Saf. 2020;194:0–1.

[15] Helmreich RL. On error management: Lessons from

aviation. Br Med J. 2000;320:781–785.

[16] Kirimoto Y, Hirotsu Y, Nonose K, et al. Development of

a human reliability analysis (HRA) guide for qualitative

analysis with emphasis on narratives and models for

tasks in extreme conditions. Nuclear Engineering and

Technology. 2021;53:376–385.

[17] Mirzaei Aliabadi M, Esmaeili R, Mohammadfam I, et al.

Human Reliability Analysis (HRA) Using Standardized

Plant Analysis Risk-Human (SPAR-H) and Bayesian

Network (BN) for Pipeline Inspection Gauges (PIG)

Operation: A Case Study in a Gas Transmission Plant.

Health Scope. 2019;8.

[18] Di Pasquale V, Miranda S, Iannone R, et al. A Simulator

for Human Error Probability Analysis (SHERPA). Reliab

574

Eng Syst Saf [Internet]. 2015;139:17–32. Available from:

https://www.sciencedirect.com/science/article/pii/S09518

32015000484.

[19] Ramezani A, Nazari T, Rabiee A, et al. Human error

probability quantification for NPP post-accident analysis

using Cognitive-Based THERP method. Progress in

Nuclear Energy. 2020;123:103281.

[20] Gibson WH, Kirwan B. Application of the CARA HRA

tool to air traffic management safety cases. 9th

International Conference on Probabilistic Safety

Assessment and Management 2008, PSAM 2008.

2008;2:1312–1319.

[21] Wang W, Liu X, Qin Y. A modified HEART method

with FANP for human error assessment in high-speed

railway dispatching tasks. Int J Ind Ergon. 2018;67:242–

258.

[22] Deacon T, Amyotte PR, Khan FI, et al. A framework for

human error analysis of offshore evacuations. Saf Sci.

2013;51:319–327.

[23] Akyuz E, Celik M. A methodological extension to

human reliability analysis for cargo tank cleaning

operation on board chemical tanker ships. Saf Sci.

2015;75:146–155.

[24] Bowo LP, Furusho M. A comparison of the common

causes of maritime accidents in Canada, Indonesia,

Japan, Australia, and England. Advances in Intelligent

Systems and Computing. Springer International

Publishing; 2019.

[25] Islam R, Abbassi R, Garaniya V, et al. Development of a

human reliability assessment technique for the

maintenance procedures of marine and offshore

operations. J Loss Prev Process Ind. 2017;50:416–428.

[26] Akyuz E, Celik M. A methodological extension to

human reliability analysis for cargo tank cleaning

operation on board chemical tanker ships. Saf Sci.

2015;75:146–155.

[27] De Maya BN, Kurt RE. Application of fuzzy cognitive

maps to investigate the contributors of maritime

grounding accidents. RINA, Royal Institution of Naval

Architects - Human Factors 2018, Papers. 2018;44.

[28] Wang W, Liu X, Qin Y. A modified HEART method

with FANP for human error assessment in high-speed

railway dispatching tasks. Int J Ind Ergon. 2018;67:242–

258.

[29] Bowo LP, Prilana RE, Furusho M. A modified heart-4m

method with topsis for analyzing Indonesia collision

accidents. TransNav. 2020;14:751–759.

[30] Bowo LP, Ispandiari AR, Wibowo DA, et al. Maritime

accident analysis and reduction technique for analysing

maritime collision accidents. Australian Journal of

Maritime and Ocean Affairs [Internet]. 2024;0:1–28.

Available from:

https://doi.org/10.1080/18366503.2024.2350792.

[31] Mazaheri A, Montewka J, Kujala P. Correlation between

the Ship Grounding Accident and the Ship Traffic – A

Case Study Based on the Statistics of the Gulf of Finland.

TransNav, the International Journal on Marine

Navigation and Safety of Sea Transportation. 2013;7:119–

124.

[32] Mazaheri A, Montewka J, Nisula J, et al. Usability of

accident and incident reports for evidence-based risk

modeling – A case study on ship grounding reports. Saf

Sci. 2015;76:202–214.

[33] Bowo LP. A NEW METHODOLOGY ON HUMAN

RELIABILITY ANALYSIS FOR SHIP SAFETY: Maritime

Accident Analysis and Reduction Technique (MAART).

Kobe University; 2021.

[34] Bowo LP, Furusho M. Integrated methods for analysing

the causal factors in Australian maritime occupational

accidents. Int J Human Factors and Ergonomics.

2021;8:23–25.

[35] Kamis AS, Fuad AFA, Anwar AQ, et al. A systematic

scoping review on ship accidents due to off-track

manoeuvring: A systematic scoping review on ship

accidents due to off-track manoeuvring. WMU Journal

of Maritime Affairs. Springer Berlin Heidelberg; 2022.

[36] ITF. STCW: A Guide for Seafarers: Taking into account

the 2010 Manila amendments. International Transport

Workers’ Federation,. 2010;78p.

[37] Chauvin C, Lardjane S, Morel G, et al. Human and

organisational factors in maritime accidents: Analysis of

collisions at sea using the HFACS. Accid Anal Prev.

2013;59:26–37.

[38] Kamis AS, Fuad AFA, Saadon MSI, et al. the Impact of

Basic Training on Seafarers’ Safety Knowledge, Attitude

and Behaviour. J Sustain Sci Manag. 2020;15:137–158.

[39] Parrott DS. Bridge Resource Management for Small

Ships: The Watchkeeper’s Manual for Limited-tonnage

Vessels. McGraw Hill Professional; 2011.

[40] Salas E, Rosen MA, Burke CS, et al. Markers for

enhancing team cognition in complex environments: The

power of team performance diagnosis. Aviat Space

Environ Med. 2007;78.