*3.1. Utility of MEL*

Munro and Kanki [26] described the MEL as an operational document for air operators, which has direct implications on the safety and airworthiness of an aircraft. Pierobon [27] shared a similar opinion by naming MEL as a barrier against aircraft dispatch with unauthorised defects. Grüninger and Norgren [28] believe that MEL serves two different purposes. Its first purpose is to identify aircraft systems/components that could be inoperative while keeping the aircraft airworthy for dispatch. The second purpose of MEL is to serve as a reference document for engineers and pilots to conduct MEL-specified tasks/procedures before aircraft dispatch. Interestingly though, Hertzler [29,30] suggested that the name MEL is unsuitable for the specific document as it could be possibly interpreted as a list of equipment/items/systems that must be installed on an aircraft. Thus, he suggested a name change to "permissibly inoperative instruments and equipment or stuff that doesn't have to work".

Following a report published by the Accident Investigation Board of Norway [31], Herrera, Nordskaga, Myhreb and Halvorsenb [32] researched the impact of changes and developments in the Norwegian aviation industry. Amongst other findings, the authors found that, although MEL provides information about the overall status and serviceability of an aircraft, it was not considered as a parameter in the holistic assessment of safety and a leading indicator of safety performance. Consequently, Herrera et al. [32] recommended the development of a risk assessment model that would integrate safety indicators such as the MEL to facilitate the identification of maintenance-related airworthiness and safety issues by air operators proactively.

### *3.2. Di*ff*erences in MEL Standards and Requirements*

In his review, Feeler [33] confirmed differences in MEL standards worldwide. For instance, he compared MEL standards for corporate/business jets published by Transport Canada (TC) and the Federal Aviation Administration (FAA), and he observed that MEL provisions and applications differ in these regions. In Canada, both business and commercial operators can operate without an approved MEL, and the decision to fly is ultimately based on the pilot in command by considering, amongs<sup>t</sup> other factors, applicable Airworthiness Directives and aircraft equipment/systems required for the intended flight route and conditions (e.g., day or night flight, operating under visual flight rules or instrument flight rules) [34,35]. Even where there is an approved MEL, the ultimate dispatch decision is still made by the pilot in command [33,34].

However, in the United States, non-commercial, business/corporate operators (a.k.a. Part 91 operators) are not permitted to operate an aircraft without a MEL [33]. Nonetheless, compared to their commercial counterparts, Part 91 operators enjoy some leniency. For example, commercial operators or operators with MELs approved under the Federal Aviation Regulations (FAR) 135, 129, 125 or 121 must comply with the repair intervals specified in MMEL/MEL, whereas Part 91 operators are not obligated similarly. Hertzler [29,30] compared the use of MEL under two distinct types of regulations, FAR Part 135 and 91, and, amongs<sup>t</sup> others, found that MEL is not a technically approved document for Part 91 operators because they are authorised to use the MMEL as MEL, subject to FAA approval. Moreover, for the same category of operators concerned, compliance with deferral categories/repair intervals specified in MMEL was not mandatory. It is noted that, according to the advisory circular published by FAA [36], MELs approved under Part 135 apply even when the operator conducts operations under Part 91. According to FAA [36] "to provide relief to operators under the MEL concept, some operators may find it less burdensome or less complicated to operate under the provisions of FAR 5 91.213(d)".

On the other hand, in Australia, aircraft are not allowed to operate without an approved MEL or a manual for permissibly unserviceable components/systems under the provisions of CAR 37 [37]. In addition, in the United Kingdom (UK), aircraft are not permitted to commence a flight with inoperative equipment governed by Commission Regulation (EU) No 965/2012 [5,38]. Therefore, although the institutionalisation of MMEL/MEL has been promoted since the 1960s [2], some countries still exempt operators from the scheme and the aviation industry lacks a uniform approach.

### *3.3. Issues in MEL Development and Application*

In the late 1980s, FAA launched a special inspection program to evaluate compliance of commuter air carriers with FARs with the participation of fifteen FAA inspection teams, consisting of six airworthiness inspectors each. A total of 35 air carriers were thoroughly inspected with a focus on 13 different areas, the MEL included [39]. The teams identified a total of 87 findings relevant to MEL, and they highlighted the following ones:


The overall conclusion of the FAA airworthiness inspectors was that there was a need for MEL compliance training as managemen<sup>t</sup> personnel were not familiar with the MEL. Interestingly, over 30 years later, similar concerns were raised by Airbus [8] whose report restated the correct application/use of the MEL with a focus on compliance with its provisions as well as the principles and best practices when deferring and dispatching aircraft according to MEL. Furthermore, Pierobon [27] stressed the need for pilots and engineers to have prior knowledge and experience in airworthiness managemen<sup>t</sup> for proper interpretation of the MEL document because the short familiarisation training received by pilots is insufficient. For instance, as part of the MEL application process in Canada and Australia, there is a requirement for operators to have MEL training programmes in place before approval and commencing operations with the MEL [37,40]. On the other hand, in the USA and UK, MEL training requirements are not defined in their respective guidance documents [5,36,41].

Similar findings were revealed by the study of Munro and Kanki [26], who reviewed 1140 maintenance reports issued between 1996 and 2002 and detected 143 reports relevant or related to the use of MELs. Their research identified improper deferral of MEL defects, failure to accomplish MEL specified tasks due to errors of omission, placarding issues and unrecorded MEL defects in technical logbooks. However, Munro and Kanki [26] also revealed contributory factors to MEL-related incidents including time pressure, unclear MELs, lack of familiarity, misinterpretation of the MEL, and communication flaws regarding the applicability of the MEL to aircraft status.

Moreover, Pierobon [27] stated that there is no standard MEL development methodology. After collecting opinions from industry professionals, especially those experienced in the MMEL/MEL process, he concluded in the necessity for NAAs to publish more specified guidelines to help air operators develop their MELs. This position and urge agree with the work of Feeler [33]—as presented in Section 3.2 above, the concerns of Hertzler [29] about the difficulty in interpreting the FAR Part 91.213 MEL regulations, and the observations of Yodice [42] regarding the ambiguity in FAA MEL regulations. To overcome these inherent challenges, the professionals interviewed by Pierobon [27] recommended the following improvement points:


### *3.4. Published Studies on MEL-Related Accidents*

Grüninger and Norgren [28] analysed the Spanair's McDonnell Douglas MD-82 fatal crash. The aircraft crashed shortly after take-off because the flaps/slats were not deployed due to a series of omissions/mistakes [43]. The investigation concluded that the take-off warning system did not activate, leaving the crew clueless about the incorrect configuration of the aircraft. The flight had been dispatched according to the MEL due to a faulty ram air temperature probe heater. Albeit the 'ground sensing relay' controls for both the take-off warning system when on the ground and the ram air temperature probe heater when airborne, the MEL did not require a detailed inspection to determine if the source of failure was, in fact, the inoperative temperature probe or a defective ground sensing relay. Had the MEL specified this, then perhaps the inoperative take-off warning system, which is a 'No Go' item if faulty, could have been discovered. Furthermore, the report also highlighted other instances of MEL misuse days before the accident; in one case, the crew had used the MEL during taxiing while it should be consulted only on ground before taxiing or take-off, and, in another case, the crew used the MEL without consulting with technicians/engineers.

Grüninger and Norgren [28] asserted that the interconnectedness and complexity of modern aircraft systems require a detailed understanding of the failure modes of each component/equipment within each system because a system's malfunction can be caused by several failures, but a single point of failure can also have several different effects. The authors above pointed out that the MMEL does not cover all conceivable scenarios during the operational phase of an aircraft, and they stressed the importance for engineers and crew members to maintain their 'mindfulness' since they comprise the last line of defence when dispatching an aircraft under the provisions of a MEL. Nonetheless, as Thomas Fakoussa, founder of Awareness Training Fakoussa (cited in Pierobon [27]) articulated, even with the MEL, pilots need advanced troubleshooting skills to tackle failures/defects under the provisions of the document. This is because flight crews are more aware of the type of operation and condition ahead of them (weather, terrain, region, etc.) as well as the required components/systems for aircraft dispatch.

Pierobon [27] analysed the Air Canada Boeing 767 event in 1983 where the aircraft was dispatched with inoperative fuel tank gauges and ran out of fuel while airborne [44]. Although the aircraft landed safely, the investigation report states that the captain had "consulted the MEL in a very cursory way" before the flight [44], suggesting that the MEL was not thoroughly reviewed by the pilot. Another concern raised by the investigators was the fact that the maintenance control centre on another occasion had granted an aircraft release against the restrictions of the approved MEL. However, the MMEL should not be consulted once a MEL has been customised for an aircraft; before this accident, an illegal relief had been granted based on the MMEL [44]. Pierobon [27] believes that the dispatch was based on (mis)perception rather than the use of the MEL, which must be the ultimate decision-making tool for aircraft dispatch for both pilots and engineers.

### **4. Review and Analysis of MEL-Related Events**

Considering the standards, guidelines, reports and literature reviewed above, the authors aimed to complement the work of Munro and Kanki [26] and examine the types of MEL-related issues emerging from aviation safety events to detect necessary intervention areas. It is noted that the goal of the research was not to derive epidemiological data and rates across and within the regions of the investigation agencies. Instead, we aimed to map types of MEL issues that have contributed to incidents and accidents, generate an overview and compare the findings against relevant literature to detect possible development gaps in the particular area. The first step was to identify the sources of safety investigation reports and define proper keywords to conduct the search. The criteria for selecting sources were the availability of reports online and publicly in the English language from regional agencies that are responsible for a large volume of aviation operations. The databases identified were the ones maintained by National Transportation Safety Board (NTSB) of the United States, Transportation Safety Board of Canada (TSBC), Air Accidents Investigation Branch (AAIB) of the United Kingdom and Australian Transport Safety Bureau (ATSB). All of these repositories relate to the respective country's registered aircraft and include occurrences that have been investigated by the respective agencies.

Relevant incidents and accidents were identified by utilising the search option on the AAIB and NTSB websites and using 'Maintenance' and 'Minimum Equipment List' as keywords. This way, the research strings covered also MMEL-related records since the term MEL is a subset of MMEL. The keywords above were also used on the ATSB and TSBC websites, but the search did not result in a substantial number of reports to review; TSBC's website produced 0 and 4 reports, respectively, and ATSB's website produced 0 and 1 reports correspondingly. Consequently, considering research time limitations, we decided to review the 400 most recently published incident/accident investigation reports on both ATSB and TSBC websites.

The search described above resulted in 1323 investigation reports, the synopsis/summary and findings/conclusions of each were studied to identify and analyse MMEL/MEL related events. The analysis included reports where MEL-related issue was identified regardless of its attribution as a contributory factor or not to each event. In addition, cases where unreported or unrecorded defects contributed to the events were reviewed and analysed. The latter cases were consulted based on the view of Airbus [8] on the importance of recording defects to, amongs<sup>t</sup> others, allow the retrofit of the MEL and assess their criticality in conjunction with other possible system and component malfunctionings or failures.

The filtering process described above resulted in the identification of 52 investigated events, 42 of which were directly related to MEL and ten regarded maintenance issues not covered by MEL but indirectly and likely a ffecting safety. In addition, although the search strategy followed is seen as comprehensive enough to generate a representative sample of relevant safety investigations, there might be reports that were unintentionally excluded from this study (e.g., input/typing errors of operators when populating the fields of the databases searched) as well as MEL-related events that have happened in other regions. Nonetheless, we believe that the final sample analysed represents the

best-case picture adequately when considering the intensive safety initiatives of the specific agencies and the overall developments in the aviation industry of the respective regions.

Table 4 presents the results of each of the search steps, and Table 5 summarises the classification of the analysed events. The full list of the reports included in our study is available to the reader upon request to the corresponding author of the paper.

Regarding the accident/incident reports reviewed, AAIB and NTSB provided a detailed discussion of the MEL-related issues identified via a dedicated MEL section/paragraph in their reports, especially in the cases where MEL-related issues directly contributed to the particular events On the other hand, although ATSB and TSBC highlighted MEL-issues in di fferent sections of their reports such as findings, analysis and conclusions, we did not find dedicated parts in the reports addressing MEL-related findings cumulatively. Nonetheless, all the reports reviewed provided enough information to enable the authors of this paper understand the situation and circumstances around the MEL-relevant events. On a side note, only 8 of the 29 reports including MEL-related contributory factors addressed the respective findings through specific recommendations. Although it was outside the scope of our research to evaluate the quality of the investigation reports and the degree of coverage of MEL-related issues through targeted measures, the picture above indicates that investigating agencies did not focus on the resolution of such problems even when they were detected in the course of investigations.

As Table 5 suggests, 50% of the MEL-related events regarded aircraft dispatch against MEL requirements, followed by cases where operations were conducted before MEL was approved (about 17%); this suggests non-compliance with MEL to the level of 67%. Cumulatively, the issues concerning MEL development (i.e., exhaustiveness, completeness and clarity) accounted for 21% of the cases, whereas cases relevant to human error/decision-making when applying MEL were jointly at the level of 12% of the events analysed (i.e., delayed rectifications or misused MEL item).


**Table 4.** Results of the search strategy and filtering of reports.


