BUSINESS ANALYSIS OF COMPANY A’S MAINTENANCE ORGANISATION

(1)

MG202X – Degree Project in Production Engineering, Second Cycle

KTH Royal Institute of Technology

BUSINESS ANALYSIS OF COMPANY A’S MAINTENANCE ORGANISATION

Dennis Bremberg and Sina Soltaniah

Email: dennisbr@kth.se and sinaso@kth.se 2018-06-08

(2)

(3)

Abstract:

Existing high voltage direct current (HVDC) stations have an estimated life-cycle of 30 to 40 years.

During this time, operations will be subjected to varying disturbances, including spare part obsolescence, new cyber-security requirements, unforeseen failures as well as planned and forced outages. Until recently, HVDC station owners have managed maintenance themselves in accordance to their own experience and the documentation provided at project delivery. Lately, increasing customer interest in life-cycle cost, service agreements and novel maintenance practices have created new business opportunities such as long-term full-service agreements (FSAs), call-centre support, care packages, and more.

This master thesis strives to map Company A’s current state to assess how prerequisites for maintenance are created during greenfield¹ projects. Prerequisites for maintenance include many aspects, including HVDC station design, maintenance planning, documentation, education, safety and other aspects governing conditions for effective maintenance. The purpose is to answer how Company A may organize responsibilities during pre-tender, tendering, project execution, warranty and service commitments to ensure that prerequisites for maintenance of HVDC systems are duly considered and optimised according to the customer’s long-term expectations and satisfaction.

Qualitative data was collected through 17 semi-structured interviews across different functions within Company A’s organisation. Also, a cross-sectional survey, combining open and closed questions aimed at the maintenance situation of HVDC stations where Company A has FSAs, was conducted. The result constitutes of a current state description based on the interviews. The ensuing discussion provides recommendations based on the established needs, as well as insights provided through literature.

Conclusively, a maintenance engineering function is proposed to address task ambiguity, organisational deficiencies, and create a process for formalisation of experience, maintenance development, and continuous improvement of the maintenance process. Specific recommendations include the introduction of a new customer information channel called Incidents, where operators of HVDC stations may communicate design-flaws, suggestions on improvements and other issues that do not naturally fall within the equipment failure record (EFR) and disturbance outage report (DOR) category. Also, the maintenance activities listed in the maintenance activity list (MAL) (provided to the customer at project delivery) may be refined by clustering maintenance activities, to ensure that the maintenance time is minimised and not treated as a sequential list (which affects planned outage frequency and scheduled outage time). The survey was inconclusive due to a low response rate.

1 A project that is initialised (on a “green” field) without respect to existing facilities.

(4)

Sammanfattning

En högspänd likströmsstation (HVDC-station) har en uppskattad livslängd på 30 till 40 år, under vilken den utsätts för olika typer av driftsstörningar, exempelvis reservdelsobsolescens, nya cybersäkerhetskrav, oförutsedda haverier såväl som planerade och oplanerade strömavbrott. Fram tills nyligen har ägare av HVDC-stationer egenhändigt underhållit sina anläggningar utifrån tidigare ackumulerad erfarenhet och den dokumentation som inkluderats vid projektleverans. På senare tid har dock kundintresset för livscykelkostnader, serviceavtal och moderna underhållstekniker skapat nya affärsmöjligheter för HVDC-leverantörer i form av långsiktiga serviceavtal.

Det här mastersarbetet avser kartlägga Company A:s nuläge för att utvärdera hur förutsättningar för underhåll skapas under ett nyanskaffningsprojekt. Förutsättningar för underhåll inkluderar ett flertal aspekter som sammantaget styr effektiviteten i stationsunderhållet, däribland stationsdesign, underhållsplanering, dokumentation, utbildning och säkerhet. Arbetets syfte är att besvara hur Company A kan organisera underhållsansvaret under offerering, projektexekvering, garanti och serviceåtaganden för att säkerställa att förutsättningar för underhåll av HVDC system är tillbörligt beaktade och optimerade utifrån kundens långsiktiga förväntningar.

Genom en intervjuserie bestående av 17 semistrukturerade intervjuer med olika funktioner inom Company A insamlades kvalitativa data. Därutöver granskades interna processer och databaser vartefter en enkätundersökning genomfördes. Undersökningen var baserad på både öppna och slutna frågor och avsåg utröna underhållssituationen för HVDC-stationer där Company A har ett omfattande och långsiktigt serviceåtagande (FSA). Resultatet består av en nulägesbeskrivning utifrån det insamlade informationsunderlaget. Den efterkommande diskussionen föreslår rekommendationer utifrån identifierade behov och insikter hämtade ur litteraturen.

Slutligen föreslås en underhållsteknisk funktion för att adressera tvetydighet i befintliga arbetsuppgifter, organisatoriska brister, samt en process för formaliserandet av erfarenheter, underhållsutveckling, och förutsättningar för kontinuerlig förbättring av underhållsprocessen.

Specifika rekommendationer inkluderar introduktionen av en ny (kundstyrd) informationskanal (Incidents) genom vilken operatörer av HVDC-stationer kan kommunicera designbrister, förslag på förbättringsåtgärder och andra händelser som normalt inte faller inom de befintliga equipment failure record (EFR) och disturbance outage report (DOR) kategorierna. Vidare föreslås att underhållsaktiviteterna listade i underhållsaktivitetslistan (MAL), som tillfaller kund vid projektleverans, bör ”klumpas”. Detta bör resultera i en minskning av rekommenderad underhållstid och säkerställa att aktiviteter som kan ske parallellt inte sker sekventiellt, vilket i sin tur påverkar både driftstoppsfrekvens och varaktighet. Avslutningsvis kan det konstateras att enkätundersökningen var in-konklusiv på grund av låg svarsfrekvens.

(5)

Acknowledgements

This thesis was conducted on behalf of Company A as part of their developmentary maintenance efforts.

We would like to give our special thanks to our supervisors; Dr Jerzy Mikler at the IIP department of KTH Royal Institute of Technology and our two supervisors from Company A, all of whom have been a tremendous help and, on multiple occasions, have offered great advice, conversation and guidance.

Also, hats off to our steering committee who put up with a continuous onslaught of questions at the most inconvenient of times, but never failed to provide great answers and enthusiasm.

Lastly, our sincere thanks go to all interviewees whom have taken the time to patiently enrich this thesis with their experience, to Company A and Company B for an excellent recipience and to all the other people at Company A whom have facilitated this work.

Authors

Dennis Bremberg and Sina Soltaniah

(6)

Nomenclature

4Q: Four questions AC: Alternating current CAPEX: Capital expenditure

CBM: Condition-based maintenance CM: Corrective maintenance DC: Direct current

DOR: Disturbance outage report EFR: Equipment failure record

EPC: Engineering procurement and construction

FACTS: Flexible Alternating Current Transmission Systems FMEA: Failure mode and effect analysis

FMECA: failure mode effect and criticality analysis FOR: Forced outage rate

FSA: Full service agreement

HVAC: High voltage alternating current HVDC: High voltage direct current IGBT: Insulator-gate bipolar transistor IAEA: International Atomic Energy Agency LCC: Line communicated converters MAL: Maintenance activity list

MEF: Maintenance engineering function MTBR: Mean time before failure

MTTR: Meant time to repair

NASA: National Aeronautics and Space Administration OEM: Original equipment manufacturer

OPEX: operating expenditure P&C: Protection and control PE: Project execution

PM: Preventive maintenance PMP: Preventive maintenance plan PT: Pre-tender

QIT: Quality improvement tool R&D: Research and development

RAM: Reliability, availability, maintainability

RAMS: Reliability, availability, maintainability, safety RCM: reliability centred maintenance

RPN: Risk priority number S: Service

SSC: Systems, structures and components T: Tender

TFR: Transient fault recorder VSC: Voltage source converter W: Warranty

(9)

1 (60)

1. Introduction

The market conditions by which high voltage direct current (HVDC) systems are developed are changing. The global competitive climate is intensifying and the need for power transmission increases (Siemens AG, 2011, TechNavio, 2017). Presently, HVDC is a well proven solution of electric bulk power transmission and built as point-to-point links, grids or back-to-back installations (Cigré, 2017a, IEC, 2017). The main purpose of point-to-point systems and HVDC grids is to transfer electricity across long distances in an efficient way, which becomes economically viable for distances over roughly 400 km, whereas back-to-back HVDC systems allow interconnection of asynchronous (different frequencies) alternating current (AC) networks (Cigré, 2017a).

There are several benefits of using HVDC in comparison to high voltage alternating current (HVAC). For instance, “HVDC systems can transfer more electrical power over longer distances than similar AC transmission systems” which allows a smaller cable cross-section (which saves material and cost), requires fewer transmission lines (saving material, cost and the environment) and features lower losses (Cigré, 2017a). On the other hand, an HVDC station is (in a rudimentary sense) an expanded version of an HVAC station, where the technology for direct current (DC)/AC conversion has been added (Company A, 2018). This inherently raises the capital expense and complexity of an HVDC station (hence the breaking point of roughly 400 km, which of course varies according to a multitude of parameters, such as topography, regulations, performance requirements, market forces, etc.) (Cigré, 2017a).

The first commercial HVDC link was built by ABB (previously known as ASEA) in 1954 featuring an underwater cable connecting mainland Sweden and Gotland, providing 100 kV (Cigré, 2017a).

Currently, there are several large companies delivering HVDC solutions, including ABB, GE Grid Solutions, Hitachi, Mitsubishi Electric Corporation, Siemens AG and others. Also, due to the increasing demand of power transmission in China, Chinese HVDC suppliers are gaining market-shares (TechNavio, 2017).

In general, HVDC systems are subject to high availability requirements, often around 98% (Roberto, 2000), meaning that planned outage frequencies must be low and forced outages scarce. Generally, scheduled unavailability is below 0.5% and service intervals may be set at 2 years (ABB, 2017).

Consequently, in the case of equipment failures, the right competence needs to be available on short notice, and maintenance procedures must be clearly defined to allow rapid mitigation (in case the utility TRIPs). In order to avoid forced outages and facilitate planning of scheduled outages, maintenance strategies governing HVDC systems rely heavily on inspections and redundancy.

Therefore, maintenance may sometimes be deterred (by allowing redundant components to fail and postponing mitigation until the next scheduled outage), since mitigation itself may require an outage.

Major market trends affecting the electric power transmission industry include changing regulations, challenged monopolies, globalisation (which pushes efficiency requirements higher), electricity trading between countries, urbanisation (which increases the local power load and demand), tougher requirements on quality, availability, security of supply and lead times, etc. (TechNavio, 2017).

Together, these forces necessitate continuous improvement of value-creating processes. One such process that, lately, has gained increasing focus within Company A, is the maintenance engineering process.

According to De Carlo and Antonietta Arleo, maintenance engineering is an organisational process strictly related to maintenance management and maintenance implementation. The maintenance engineering function receives information from maintenance management which is analysed and used

(10)

2 (60) in development of maintenance plans that are subsequently reiterated to maintenance implementation via management (De Carlo and Antonietta Arleo, 2017). Moreover, Hale et al., in a safety evaluation of the chemical process industry, tentatively concluded that a lack of a strong (and explicit) maintenance engineering function has caused a failure, partly in the incorporation of maintainability considerations in new designs and modifications, but also in the utilisation of feedback

“from maintenance and safety experience” (Hale et al., 1998). Ultimately, not only safety suffered due to this lack, but the entire maintenance process (Hale et al., 1998). Conclusively, both Hale et al., De Carlo and Antonietta Arleo recognises the maintenance engineering function as an important part of maintenance development.

To optimise system availability, create market opportunities and minimise operative expenditures, Company A has expressed a desire to investigate what such a maintenance engineering process or function may entail, in respect to their current practices. In particular, the question of maintainability has been raised, in terms of ensuring that the right prerequisites (such as tools, spare parts and documentation) are present as needed to enable effective maintenance.

1.1. Background

This thesis work is commissioned by Company A’s service department, which is responsible for upgrades, warranty, and service agreements for HVDC systems. Company A’s base of operation is for reference placed at Location A, whereas Company B (see Appendix B), is situated at Location B Within Company A, greenfield projects are divided into five major phases; pre-tender, tendering, project execution, warranty (initiated by a provisional acceptance certificate and concluded by a final acceptance certificate) and service commitments. During the pre-tender, marketing, standardisation of design, supplier qualification and procurement discussions take place. During tendering, the goal is to develop a concept which fulfils customer requirements to a competitive price. If the bid (project) is won, the project is advanced to project execution. During project execution, the initial concept is refined and realised. At project delivery, the warranty is taken into effect. After the warranty period, Company A may be involved as a service provider and/or a potential supplier for an upgrade.

According to Cigre’s (2016) report on life extension of existing HVDC systems, early design stages must be engineered so that capital and maintenance expenditures may be accurately captured (to a reasonable degree) and all design options should be clearly documented and accepted by critical stakeholders (Cigré, 2016). As a well-recognised practitioner of maintenance development, National Aeronautics and Space Administration (NASA) further reinforces this point by recognising that design for maintenance and maintainability (the ease by which something is maintained), from a financial perspective, need to be approached early during the conceptual phase, as the cost of changes increase rapidly over time (NASA, 1994). Furthermore, Mital et al. stresses that maintainability must be recognised as an integral part of the design process, and that it should not be treated as a separate issue (Mital et al., 2008). Consequently, Mital et al. also points out that the design specification review is an important project gateway to ascertain that maintainability requirements have been met. The design specification review was defined by Thompson as a “quantitative and qualitative examination of a proposed design to ensure that it is safe and has optimal performance with respect to maintainability, reliability and performance variables needed to specify the equipment” (Thompson, 1999).

According to the IEC’s (2017) technical report on HVDC installations, there is an absence of standardise asset management (including maintenance) of HVDC systems which have resulted in utilities, globally, practising asset management based on individual interpretation and experience which, naturally, may not be the best way. It is noted that a separate asset management policy and strategy may be

(11)

3 (60) appropriate if operational performance and cost structures are dissimilar. However, it is also recognised that information exchange concerning processes and methodologies between original equipment manufacturers (OEMs) and utilities is desirable from a performance, life-cycle cost and developmentary perspective. According to the IEC, several important aspects should be included in the asset management strategy in order to minimise the total life-cycle cost of an asset and ensure delivery of the required service level, including (but not limited to) (IEC, 2017):

• “Operation and maintenance”.

• “Development”.

• “Review and evaluation of operational performance data compared to target values”.

• “Review and evaluation of scheduled and forced outage trends and impacts”.

• “Development of maintenance planning”.

• “Determination of applicable and useful maintenance methods”.

HVDC systems are, per Cigré, inherently complex and constructed based on unique premises.

Consequently, the need for documentation, such as maintenance procedures, risk analyses, single-line diagrams, etc. is more extensive in comparison to that of HVAC systems, which poses comparatively higher requirements on “staff training and acceptance” (Cigré, 2016). This may also be a complicating factor in respect to the standardisation of asset maintenance.

Historically, owners of HVDC systems have themselves managed or performed maintenance based on documentation provided by the supplier, as well as internal experiences and procedures. As a supplier of HVDC systems, Company A has in turn focused on winning tenders by minimising capital expenditures (CAPEX) in relation to stipulated requirements, as this has been a rewarding and prudent strategy. Recently, customers’ interest concerning operative expenditures (OPEX) has grown and Company A has responded with new services, for example providing call-centre support, care packages and service agreements to various effects, including full maintenance ownership service agreements.

As part of this development, new responsibilities recently befell Company A’s Service department (from now on referred to as Service), including the commercialisation of maintenance services and development of maintenance activity lists (MALs). Also, the question of how maintenance may be approached to ascertain that prerequisites for maintenance are present in time of need was raised, which, in turn, led to the formulation of this thesis work.

1.2. Aim

This thesis work aims to, through investigative efforts of various kind, identify possibilities of improvement in the current maintenance organisation (including work practices, information flow and policies) and to ultimately establish a process and/or function by which Company A may continuously improve their maintenance services further. The main target of aforementioned investigation is the issue of maintainability, that is, the ease by which a component, piece of equipment or system is maintained, which in turn relate to how the facility layout is planned, what spare part policies are used, how certain pieces of equipment from the onset are designed, or what external means (such as fork- or scissors-lifts) are necessary in order to enable effective maintenance. As such, objectives are formulated as follows:

• To identify potential problems and/or opportunities for improvement in the way maintenance work is understood, planned, performed and documented.

• To suggest a process and/or function by which maintenance improvement of HVDC systems may be pursued through pre-tender, tendering, project execution, warranty and service commitments.

(12)

4 (60)

• To suggest a framework of responsibilities addressing potential shortcomings in the present organisation related to maintenance engineering.

1.3. Research question

Based on the background concerning Company A and the proposed aim, the following research question is formulated:

How may Company A organize responsibilities during pre-tender, tendering, project execution, warranty and service commitments to ensure that prerequisites for maintenance of HVDC systems are duly considered and optimised during said project phases according to the customer’s long-term expectations and satisfaction?

In order to address any unclarities, the research question is further dissected into sub-questions, which will be discussed in chapter 4.9. These sub-questions are:

• What are customer expectations?

• When is the customer satisfied?

• Why should customers be satisfied?

• What is an HVDC system?

• What is maintenance?

• How long is “long-term”?

• What does “duly” entail?

• What are “prerequisites for maintenance”?

• How do other industries address maintenance and maintainability?

1.4. Methodology

This thesis work is predicated on a business analysis of Company A’s maintenance organisation. The methodology is divided into three main components, that is, information gathering, analysis and recommendations, which are presented in detail below.

1.4.1. Information gathering

Initially, a literature review was conducted to gather relevant theory and knowledge concerning information gathering, existing industrial maintenance practices, relevant background and definitions and, lastly, organisation of maintenance responsibilities, activities and processes.

Aside from the literature review, specific information relating to Company A was collected through qualitative research, including:

1. Face-to face interviews with various Company A functions.

2. Steering-group meetings with representatives from Service.

3. Reviewing existing documentation relating to processes, organisation, responsibilities, etc within Company A.

4. A survey targeting existing HVDC stations where Company A has full ownership of the maintenance through a full service agreement (FSA).

Firstly, face-to-face interviews were conducted to establish an information-basis of opinions, experiences, insights, values and work practices, as well as to direct investigative efforts. Both structured, semi-structured, and unstructured interviews took place. Unstructured meetings were characterised by open questions, often without the respondent knowing that any particular information was sought, whereas structured interviews were conducted with a clear statement of purpose and a predefined set of questions. Semi-structured interviews constitute the majority of the

(13)

5 (60) interviews compiled in Appendix C and were based on predefined questions but allowed follow-up questions.

Interviews took place at two companies; Company A at Location A and Company B at Location B. Also, Skype-interviews with company A functions located outside of Location A were conducted. The interviews are presented as statements from each interviewee in chapter 3.

The overall purpose of conducting interviews was to establish the current state of Company A’s maintenance processes and organisation, as well as to identify possible best-practices from Company B. To remain consistent in the presentation of information (gathered through different forms of interviews), all questions, as posed to the interviewee, have been omitted. Instead, the interviewees’

answers are presented as statements in Appendix C.

Inherently, there is a risk of inaccurately representing information when summarising interviews. For instance, meaning may be ascribed or inferred from a statement that was not the interviewee’s original intent. To extenuate this, both authors consistently verified the derived statements towards the original protocol.

The semi-structured interviews were carried out in different rooms with two interviewers (the authors of this thesis). Interviewer 1 (lead interviewer) made the majority of posed inquires whereas interviewer 2 acted as a typist and kept a record of what was being said on a computer in a word- processor program, in this case Microsoft Word 2016. The typist would however occasionally intervene and ask follow-up questions.

To avoid issues of selective memory, that is, the risk of potentially forgetting details that, at the time, did not seem as important as others, the summary was written as fast as possible in wake of the interview. However, the construction of a summary itself implies that some information was omitted, which is true to some degree (but only in the sense that it was consciously deemed unimportant, in contrast to forgotten in benefit of more “attractive” pieces of information). The summary also served to condense the communication. In addition to addressing the previously described issue of inaccurate representation, prompt summarisation also served to pre-empt telescoping, which is the risk of misremembering and mixing times when statements were made. Such errors may for instance result in loss of context. Occasionally, different interviewees responded to similar or identical questions, which in turn reinforced certain facts. In cases were accounts did not match, the discrepancies were investigated.

Secondly, two formal steering-meetings were held with an appointed steering-group. The steering- group consisted of four members. The first meeting served to discuss and clarify the scope of the thesis work and the latter to validate the work in respect to stipulated expectations and to decide how to best proceed, although several informal meetings were held continuously during the course of the thesis.

Thirdly, relevant documentation such as failure mode effect analyses (FMEAs), MALs, responsibility specifications, internal processes, etc., were accessed via Company A’s intranet, mail correspondence and through various internal databases. This information-basis served to delineate organisational relationships and decision-making procedures.

Lastly, a survey targeting existing HVDC stations where Company A has full ownership of the maintenance function (through an FSA) was conducted. The purpose was to make a current state cross- section with respect to spare part-, tool- and information management, as well as to map key maintenance metrics used at site during operations. The aim was to identify potential improvements

(14)

6 (60) in the current maintenance organisation and, ultimately, to provide a pre-emptive basis for station requirement specification. The respondents were informed that the survey would provide a foundation for forthcoming development and resource allocation recommendations, that answers would be kept confidential (meaning that all answers will be anonymized) and that the surveyors were bound by a non-disclosure agreement. The survey consisted of 36 questions and may be found in its entirety in Appendix A.

Due to a low response frequency, no results could be derived from the survey. Hence, no procedure concerning its analysis is provided.

1.4.2. Analysis

The interview results were summarised into 162 statements (see Appendix C) for Company A, and 49 statements for Company B (see Appendix B). In total 17 interviews were conducted within Company A and 3 within Company B. Based on prominent trends and the previously defined aim, 11 categories were developed by the interviewers. Subsequently, each statement was mapped to applicable categories. If several statements within one interview repeatedly scored a particular category, that result was measured as one “hit”. Both authors independently categorised the statements to ensure consistency. Subsequently, both authors’ scoring was compared, and coincident statements were retained.

1.4.3. Recommendations

Based on the interview results, literature review and general understanding of Company A’s business, recommendations, aimed at existing problems and opportunities identified in the analysis, were provided.

1.5. Delimitations

This thesis work only investigates projects in which Company A was the sole provider of maintenance, in contrast to arrangements were the customer per agreement is responsible for some, or all, maintenance activities (which is relevant in respect to the survey). Furthermore, the work is limited to greenfield projects and does not consider brownfield² projects to any significant extent.

2 Modification (upgrading, refurbishment, installation, etc.) of already existing facilities.

(15)

7 (60)

2. Literature review

Literature relating to the performance optimisation and specific parts of an HVDC system is plentiful, whereas recommendations concerning specific processes, aimed at optimising and creating prerequisites for effective maintenance of HVDC stations, appears sparse. Below follows a compilation of relevant background and theory which introduces the reader to the specific system at hand (HVDC systems), its variability, maintenance practises of industries with comparable requirements on safety, availability and reliability, as well as an outlook of different maintenance regimes that may be applicable or of interest to Company A to understand.

2.1. Methodologies

This section covers a literature background to the methods in this report. Presenting interview structure and strategic dependency modelling.

2.1.1. Interview structures

As a large part of the information in this report originates from interviews with different functions within Company A, the differentiation and application of relevant interview techniques (unstructured, semi-structured and structured interviews) are outlined below.

Unstructured interviews are of an exploratory nature where the interviewer does not wish to pose restrictions on the ensuing dialog. Unstructured interviews may be appropriate to gather rich, in-depth data from stakeholders or other capacities of interest (Wilson, 2014).

Semi-structured interviews combine the exploratory nature of an unstructured interview with the more fixed and precise questions of a structured interview. The semi-structured interview is often based on an “interview guide”, which consists of a set of questions or topics. The semi-structured interview is normally opened with a statement of purpose and closed by some closing comment. Semi- structured interviews may be considered appropriate to gather “facts, attitudes, and opinions”

(Wilson, 2014).

Structured interviews make use of closed or open questions. The interview is conducted by means of a verbally communicated questionnaire based on predefined questions, which should be followed with as little deviation as possible. Questions are formulated in advance and no spontaneous questions are asked. Structured interviews may be considered appropriate whenever there is a predefined issue at hand, which is to evaluate purposely (without simultaneously assessing other matters). According to Wilson, structured interviews are especially suitable to advance the results or insights gained by unstructured or semi-structured interviews. Structured interviews require the interviewer do act consistently (Wilson, 2014).

2.1.2. Strategic dependency modelling

To represent relations between different departments or company functions, the strategic dependency model illustrates dependencies between two entities. In a strategic dependency model, relationships are called dependums and inherently consists of one depender and one dependee (Yu et al., 1996). There are four types of dependums in a strategic dependency model, which are illustrated in Figure 1. The top part of the figure shows how the depender and dependee are represented, the relationships are explained below (Yu et al., 1996).

(16)

8 (60)

Figure 1. Dependencies in a strategic dependency model (Yu et al., 1996).

1. Task dependency: one actor depends on the other to perform a task.

2. Goal dependency: one actor depends on the other to achieve a predefined result or goal, the goal can be quantified.

3. Resource dependency: one actor depends on the other to deliver certain resources. The dependency may be both that of physical resources and information.

4. Soft goal dependency: this dependency is similar to a goal dependency; however, the goal is not predefined in detail and is harder to measure since it is not completely quantifiable.

Strategic dependency models are developed based on different scenarios, after the models are developed, they are compared to each other. To determine the best solution, the strategic dependency models are used as basis for a strategic rationale model, this model allows to explore different options and find the most optimal solution (Yu et al., 1996).

2.2. HVDC systems

This section contains fundamental information about HVDC systems. The information is mainly based on Cigré reports. Cigré is the “international council for large technical systems” which collect, compile and publishes reports on HVDC link performance all over the world (Cigré, 2018).

As stated in the introduction HVDC systems are currently an important (and growing) solution of electric bulk power transmission. HVDC systems are mainly built as “point to point”-links, which means that the link transfers electricity between two geographical areas (Cigré, 2017a). The main purpose is to transfer electricity across long distances or to connect two alternating current (AC) networks (Cigré, 2017a). Benefits of HVDC are:

1. Having the same current flow over a smaller cable cross section area compared to AC.

2. Has lower losses and no need of charging current in the cables, compared to AC, where a charge is required to change polarity in the capacitance in the cable.

3. Allows connection of AC networks with different frequencies.

4. Supports power-transfer between AC networks.

An HVDC point-to-point link consists of two converter stations with DC lines in-between. A transmitting converter station converts AC to DC and a receiving station vice versa. There are two main types of HVDC systems; line commutated converters (LCC), which are thyristor based, and voltage source converters (VSC), which utilise insulator-gate bipolar transistor (IGBT) valves. VSC feature

(17)

9 (60) comparatively higher transfer losses, although VSC converter stations are smaller in size and can be used to energise an AC grid (Cigré, 2017a). There are five major point-to-point setups:

1. Monopolar earth return system. This variant requires one converter at each HVDC station (end-point) and two conductors to interconnect the stations (Hamzehbahmani et al., 2015).

The monopolar earth return system is the least costly variant for underwater power transmission since it only requires one ground cable. However, if one of the two converters fail, the whole system will be down (which is denoted as an outage). Another drawback concerning the monopolar earth return variant is that underwater cables may cause chemical pollution (Hamzehbahmani et al., 2015).

2. Monopolar metallic return system. This variant is similar to the monopolar earth return system but without the associated risk of chemical pollution. Here, a neutral or low voltage cable is used to return the current, in comparison to earthing both stations separately. If the return cable is out for maintenance, this solution operates as a monopolar earth return system (Hamzehbahmani et al., 2015).

3. Bipolar system with neutral point earthing of a single terminal. In a bipolar system, each station consists of two converters operating with opposing polarity in respect to a common neutral point, which is located in one station (Hamzehbahmani et al., 2015). Benefits of a bipolar system is that it can operate as a monopolar system during limited durations (which may be practiced during maintenance activities).

4. Bipolar system with neutral point earthing both terminals. This variant is similar to a the one described in paragraph 3, but there is a neutral point located at each station respectively.

5. Bipolar system with earthing of the metallic return. Metallic return is described in paragraph 2 and bipolarity in paragraph 3.

As stated previously, the demand for HVDC point-to-point links is increasing. This is particularly clear in China, where developmentary efforts are made to build a ±1100 kV link spanning 3324 km. Presently, the largest LCC HVDC links in terms of transferred power are built in China. Examples of HVDC links are presented in Table 1 and Table 2 (Cigré, 2017a).

Table 1. Examples of point-to-point LCC projects outside of China, modified from Cigré (Cigré, 2017a).

Year Country Project Voltage

(±kV) Power (MW) Length (km)

2012 India Mundra-Mohindergarh 500 2500 970

2010 India Ballia-Bhiwadi 500 2500 780

2003 India Takher-Kolar 500 2000 1450

1998 India Chahdrapur-Padghe 500 1500 752

1990 Brazil Itaipu 600 6300 790

1990 India Rihand-Delhi 500 1500 814

1986 USA Intermountain 500 1600 787

1982 Republic of

Congo Inga-Shaba 500 560 1700

1979 South Africa Cabora-Bassa 550 1930 1420

1965 USSR Volgograd-Donbass 400 720 470

Table 2. Examples of point-to-point LCC projects in China, modified from Cigré (Cigré, 2017a).

Year Country Project Voltage

(±kV) Power (MW) Length (km)

(18)

10 (60)

2014 China Haininan-Zhengzhou 800 8000 2210

2013 China Xiluodu-Guangdong 500 6400 1286

2012 China Jinping-Sunan 800 7200 2095

2010 China Yunnan-Shanghai 800 5000 1373

2010 China Xiangjiaba-Shanghai 800 6400 1907

2006 China Three Gorges-Shanghai 500 3000 1040

2004 China Three Gorges-

Guangdong 500 3000 975

2004 China Guizhou-Guangdong 500 3000 980

2003 China Three Gorges-Guandong 500 3000 860

2001 China Tianshengqiao-

Guangzhou 500 1800 960

1989 China Gezhouba-Shanghai 500 1200 1045

Apart from point-to-point links, there are also multiterminal links, which Cigré defines as a network with more than two connected HVDC terminals. Such multiterminal links may be connected in series or as a grid. Regardless, both LCCs and VSCs can be used to create multiterminal links (Cigré, 2017a).

2.3. Maintenance

Maintenance is defined as “the combination of all technical, administrative and managerial actions during the life-cycle of an item intended to retain it in, or restore it to, a state in which it can perform the required function” (EN13306, 2010).

According to Mikler, the aim of the maintenance function is to continuously identify, plan, schedule and execute maintenance activities. There are two types of maintenance strategies; corrective and preventive maintenance. Corrective maintenance (CM) may be classified as maintenance carried out on an item after fault recognition, in order to restore an operational state. Conversely, preventive maintenance (PM) is carried out prior to the failure of an item in order to (Mikler, 2015):

• Reduce the likelihood of failure.

• Detect and prevent impending failures.

• Extend the life of the item (by for example, refurbishment or replacement).

Corrective maintenance can be subdivided into emergency and deterred maintenance, where emergency indicates that failures are addressed immediately upon detection and deterred that the rectification is postponed. Preventive maintenance can in turn be subdivided into predictive and planned maintenance. Predictive maintenance includes condition-based maintenance (CBM) and reliability centredmaintenance (RCM). In predictive maintenance, scheduling occurs adaptively, whereas planned maintenance relies on a fixed schedule (Ben-Daya, 2009).

According to Blischke, relying on preventive maintenance is only feasible if (Blischke, 2003):

a) The rate by which components fail is increasing or will increase if no preventive replacement occurs.

b) The cost of preventive replacement is lower than the cost of corrective replacement.

2.3.1. Maintenance organisation

In the book, Strategic Maintenance Planning, Kelly investigates the maintenance function from a systems perspective and defines an organisation as an open system containing multiple subsystems such as maintenance, production, capital asset acquisition, design, etc. In turn, each subsystem is said to be responsible for different functions. For example, corporate management may oversee

(19)

11 (60) organisational goals and strategies, as well as control other subsystems. Conversely, capital asset management may be responsible of buying, installing and commissioning physical assets. Regarding the maintenance function, Kelly states that its purpose is to sustain the function of physical assets through repair, modification and/or replacement (Kelly, 2006).

Kelly further recognises that each subsystem requires information from one or several different subsystems. For instance, marketing may develop forecasts which are used in production planning to create aggregate plans, equipment data (utilisation, failure rate, performance, etc.) generated during production is likewise translated into maintenance actions. An illustration of the maintenance subsystem is presented in Figure 2, which indicates the subsystem’s boundary and dependency based on inputs, outputs and function. Notably, the maintenance subsystem controls plant safety, condition and quality by managing workforce, spare parts and tools based on resources and information (Kelly, 2006).

Figure 2. Maintenance subsystem (Kelly, 2006).

In order to minimise the overall life-cycle costs of each asset, Kelly recognises that a coordinated decision-making process between the procurement-, design- and maintenance subsystem is required in order for procurement-, reliability- and maintainability requirements to be congruent (Kelly, 2006).

According to Mikler, the scope of the maintenance function differs between organisations, although a shared main goal commonly is to “maintain facilities and production equipment so they can be operated safely, efficiently and economically” (Mikler, 2015). Mikler also states that systems and organisational means by which maintenance activities are controlled, in turn identified and planned for in relation to some requirement, need to be developed and established. Management needs to define a maintenance strategy by which the maintenance function achieves its objectives and is systematically developed to ensure effectiveness of work-practices. Moreover, the core process of the maintenance function “is the continuously ongoing process of identifying, planning, scheduling and executing of maintenance work” (Mikler, 2015). Figure 3 outlines factors that are frequently mentioned in the literature in association with an efficient maintenance organisation (Mikler, 2015).

(20)

12 (60)

Figure 3. “Factors related to maintenance organisation frequently occurring in the literature” (Mikler, 2015).

Improvement of the maintenance functions, as shown by Mikler, may be divided into four “cycles”

which are the continuous improvement processes of resource planning, work, organisation and the maintenance function (Mikler, 2015).

According to Linnéusson et al. effective maintenance development may be achieved by understanding results at the operational level in relation to the customer’s needs. As shown in Figure 4, operational results are retrieved and used as a decision-making basis at the strategic level. Furthermore, in order to facilitate the development of strategies, the cost of maintenance performance must be visualised and short-term budget decisions must be replaced by a systems thinking (Linnéusson et al., 2018).

Figure 4. Feedback behaviour between strategic and operational levels in maintenance development (Linnéusson et al., 2018).

(21)

13 (60) 2.3.2. Maintenance engineering

On the 31^st of January 2017, CEN (the European Committee for Standardisation) agreed to develop a new standard addressing maintenance engineering with a publication date set in 2020. Presently, there are two draft definitions of maintenance engineering, which are provided below (SN, 2017);

maintenance engineering is a:

• “Staff function whose prime responsibility is to ensure that maintenance techniques are effective, that equipment is designed and modified to improve maintainability, that ongoing maintenance technical problems are investigated, and appropriate corrective and improvement actions are taken.”

• “Combination of theories, practices and applications of all engineering disciplines to do maintenance on physical assets and its components, in order to:

o Measure and evaluate the integrity of the status and technical conditions

o Maintain and improve the functionality characteristics using the criteria of reliability theory as: safety reliability, maintainability, supportability and availability

o Extend the life of assets optimising preventive maintenance and appropriate technical improvements

o Carry out the necessary work execution according to the best technical practices in terms of safety, quality, productivity and effectiveness.”

2.3.3. Maintenance strategies

Customers are increasingly requiring more extensive maintenance programs to be delivered alongside their ordered HVDC system. In a few projects, Company A has been required to provide new forms of maintenance analyses, which was discovered through the interview process at site in Location A. In order to facilitate, and possibly improve, this process in the future, a brief outline of RCM and a more comprehensive review on FMEA and failure mode effect and criticality analysis (FMECA), is provided.

2.3.3.1. Reliability centred maintenance

RCM is a systematic way of choosing a safe, suitable and cost-effective maintenance strategy for a specific component (Mikler, 2015). The goal of an RCM program is not to bring an item to its ideal condition, rather, it should be restored to a predefined state of functionality (Ben-Daya, 2009). It was noted by Mikler that RCM focuses attention on the maintenance activities that have the largest impact on the organisation’s performance and diverts energy from less important areas. Furthermore, the use of RCM provides justification for maintenance activities and allows budgeting based on planned activities, rather than relying on an estimate of “last year’s expenses” (Mikler, 2015).

RCM used reliability estimates of equipment, components and systems to create a cost-effective schedule. This is achieved by pre-empting failures, which in turn reduce downtime and cost (Ben-Daya, 2009). The work stipulated by an RCM program is two-fold. It consists partly by an FMEA and partly by an evaluation of maintenance schedules effects on reliability.

For each failure that occurs, RCM logic dictates a number of questions to be answered. These questions were formulated and issued in the SAE JA-1011 Standard by the Society of Automotive Engineers in an attempt to structure and generalise the RCM process. The question were taken as follows (Ben-Daya, 2009):

1. “What are the functions and associated performance standards of the asset in its present operating context?

2. In what ways can it fail to fulfil its functions?

3. What causes each functional failure?

(22)

14 (60) 4. What happens when each failure occurs?

5. In what way does each failure matter?

6. What can be done to predict or prevent each failure?

7. What should be done if a suitable proactive task cannot be found?”

2.3.3.2. FMEA and FMECA

According to IEC 60812 (IEC, 2006), FMEA and FMECA are methods used to identify potential failure modes, failure causes and resulting effects on a system. A system is represented as either hardware, software or processes. Moreover IEC 60812 states that if an FMEA is conducted early in a development cycle, insufficiencies detected by the FMEA can lead to cost savings, provided the results are based on assessments of qualified experts. Ultimately, FMEA and FMECA are methods able to identify the severity of potential failures, prioritize solutions and provide mitigating measures in order to reduce risk and optimise cost and downtime.

The FMEA is initiated by a hierarchical breakdown of a system into its basic elements (often components), producing a block diagram. According to IEC 60812, an FMEA consist of four main stages (IEC, 2006), namely:

1. Establish basic rules for performing the FMEA, including a time-plan to ensure that the right competence and sufficient time are allocated to complete the task.

2. Perform the FMEA using an appropriate worksheet.

3. Summarise and document the FMEA in order to conclude specific actions.

4. Update the FMEA continuously over time with new data.

FMECA is an extended FMEA and includes ranking of severity to the failure modes to enable prioritization of actions (criticality). Criticality is based on the probability of occurrence, failure mode and severity of effects. Historically, the criticality assessment has been based on either a risk priority number (RPN) or item criticality number (Bowles and Peláez, 1995).

RPN is mainly used in automotive industry and utilise linguistic terms to rank the probability of a failure-mode occurrence, severity of the failure’s effect, and probability of the failure detection on a scale from 1-10 (Bowles and Peláez, 1995). The rankings are multiplied to provide the RPN value. A high RPN value is prioritized higher than low RPN value (Bowles and Peláez, 1995). Decision-criteria for occurrence frequency is presented in Table 2, severity criteria in Table 3 and detectability criteria in Table 4, note that the tables are from IEC standard 60812 and used in the vehicle industry (IEC, 2006).

(23)

15 (60)

Table 2. Frequency of occurrence evaluation criteria (IEC, 2006).

Rating Failure mode occurrence Frequency

1 Remote: Failure is unlikely ≤ 0,010 per

thousand vehicles/items

2 Low: Relatively low failures 0,1 per thousand vehicles/items

3 0,5 per thousand vehicles/items

4 Moderate: Occasional failures 1 per thousand vehicles/items

5 2 per thousand vehicles/items

7 High: Repeated failures 10 per thousand vehicles/items

9 Very high: Failure is almost inevitable 50 per thousand vehicles/items

10 ≥100 per thousand vehicles/items

Table 3. Severity evaluation criteria (IEC, 2006).

Rank Severity Criteria

1 None No discernible effect.

2 Very minor Fit and finish/squeak and rattle item does not conform. Defect noticed by discriminating customers (less than 25 %).

3 Minor Fit and finish/squeak and rattle item does not conform. Defect noticed by 50 % of customers.

4 Very low Fit and finish/squeak and rattle item does not conform. Defect noticed by most customers (greater than 75 %).

5 Low Vehicle/item operable but comfort/convenience item(s) operable at a reduced level of performance. Customer somewhat dissatisfied.

6 Moderate Vehicle/item operable but comfort/convenience item(s) inoperable.

Customer dissatisfied.

7 High Vehicle/item operable but at a reduced level of performance.

Customer very dissatisfied.

8 Very high Vehicle/item inoperable (loss of primary function) 9 Hazardous

with warning Very high severity ranking when a potential failure mode affects safe vehicle operation and/or involves non-compliance with government regulation with warning.

10 Hazardous without warning

Very high severity ranking when a potential failure mode affects safe vehicle operation and/or involves non-compliance with government regulation without warning.

(24)

16 (60)

Table 4. Detectability evaluation criteria (IEC, 2006).

Rank Detection Detection: Likelihood of detection by Design Control

1 Almost certain Design Control will almost certainly detect a potential cause/mechanism and subsequent failure mode

2 Very high Very high chance the Design Control will detect a potential cause/mechanism and subsequent failure mode

3 High High chance the Design Control will detect a potential cause/mechanism and subsequent failure mode

4 Moderately

high Moderately high chance the Design Control will detect a potential cause/mechanism and subsequent failure mode

5 Moderate Moderate chance the Design Control will detect a potential cause/mechanism and subsequent failure mode

6 Low Low chance the Design Control will detect a potential cause/mechanism and subsequent failure mode

7 Very low Very low chance the Design Control will detect a potential cause/mechanism and subsequent failure mode

8 Remote Remote chance the Design Control will detect a potential cause/mechanism and subsequent failure mode

9 Very remote Very remote chance the Design Control will detect a potential cause/mechanism and subsequent failure mode

10 Absolutely

uncertain Design Control will not and/or cannot detect a potential cause/mechanism and subsequent failure mode; or there is no Design Control

Criticality number is mainly used in the aerospace and nuclear industries and it categorises the severity of failure modes which provides a criticality ranking. It consists of failure-effect probability, failure mode ratio (between occurrence of the failure modes of components), part failure rate and operating time (Bowles and Peláez, 1995).

(Bowles and Peláez, 1995) states that both approaches have several limitations regarding the calculations and the interpreted results. An example is stated in Bowles report (Bowles and Peláez, 1995) where a failure mode with a high severity, detectability and low rate of occurrence may have a lower RPN than one with all parameters moderate, even though it should have a high priority for corrective actions. Moreover, (Bowles and Peláez, 1995) states that qualitative assessments are quantified and based on “guesses” or assumptions, which makes them less accurate from an analytic perspective (Bowles and Peláez, 1995).

A literature review written by Liu et al., consisting of 75 research articles published between the years of 1992 and 2012 concluded that there are weaknesses in the RPN value system. The most frequently observed weaknesses are listed below (Liu et al., 2013):

• The relative importance of severity, occurrence and detection decisions variables are not considered

• Different combinations of severity-, frequency- and detection-values may produce the same RPN, hiding consequences that may differ between events.

• It is hard to assign accurate values to the decision variables, which complicates their evaluation.

(25)

17 (60) Liu et al. observed that the most common method of determining RPN values utilises artificial intelligence (AI) and the second most frequent method is based on multi criteria decision-making.

Furthermore, Liu et al. concludes that risk factors must be differentiated to minimise imprecision, one example would be to split severity into three factors, say, damages, production and maintenance costs and consider all three factors separately when determining severity. As a final statement, Liu et al.

concludes that risk priority numbers is an appropriate method for specific risk evaluation problems (Liu et al., 2013).

2.4. Maintenance of HVDC systems

HVDC systems consist of a large number of components which, throughout an HVDC system’s life- cycle, may necessitate changes of spare part suppliers (Cigré, 2017a). Consequently (and particularly in DC grid applications), an HVDC system should have functional interfaces towards other producers’

systems (Cigré, 2017a). However, Cigré does not specify how this should be achieved.

One of the bigger challenges in maintaining HVDC stations (the end-points of an HVDC link) is to match all planned maintenance activities to available time slots (which may be scheduled in relation to customer agreements, from the perspective of the utility owner or operator), which is slightly less problematic in the case of a bipole, since maintenance of valves, transformers and switchgear may be performed on one pole at the time, keeping one pole energised (in service).

Another complicating fact from a maintenance planning perspective is that different components or systems have varying maintenance intervals and needs, for instance, a transformer may require annual service whereas valves may only require service every two years (Cigré, 2017b).

For some components, CBM is used to predict failures, which pre-empts unnecessary outages (a scheduled maintenance outage may be refuted since the actual condition of the component is good enough). Also, redundancy (such as duplicates of critical systems) are built in to provide a failsafe and to create maintenance opportunities (while the system is kept in service) (Cigré, 2017b). Life-cycle estimates of key HVDC components are presented in Table 5.

Table 5. HVDC grid components lifespan considerations (Cigré, 2017b).

HVDC grid component Approximate life span Life-cycle consideration

HVDC overhead lines >50 years Generally, maintenance and repair considerations only for the life of the HVDC grid asset

HVDC disconnects &

switches >50 years

HVDC cables >40 years

Converter transformers >40 years Changeout method is key along with online CBM and end of life prediction Converter valve equipment 20-40 years High probability of one replacement

during HVDC grid life-cycle Protection & control (P&C)

systems 10-20 years Must be planned for replacement multiple times during HVDC grid life-cycle

Aside from the importance of maintenance planning, spare part management is an important aspect of maintenance, as insufficient spare part levels, or poor management, may disrupt effective maintenance and cause delays if acquisition lead times are anything but short. According to Cigré 713, spare parts are mainly held using four policies (Cigré, 2017b):

(26)

18 (60) 1. Redundancy: spare parts kept in the form of redundancy may be activated immediately upon

failure.

2. On-site storage: spare parts are stored on-site (close-by or in annex to the HVDC station).

Unlike the redundancy policy, spare parts are disconnected from the HVDC system and may require certain tools to be changed.

3. Off-site storage: spares are stored off-site at a separate storage location or, for instance, at a supplier’s factory. This method may necessitate special transportation equipment to, and on site.

4. Spares clubs: spares are kept off-site. A number of different operators has access to the spares.

This policy may require standardise interfaces and plug-and-play capabilities.

Cigré 713 recommends including an operations manager and key staff in the project development phase (of the HVDC system) in order to create the right maintenance prerequisites and to train maintenance staff efficiently (Cigré, 2017b). In the beginning of a service agreement, the buyer’s maintenance management system should be updated with OEM maintenance instructions. Moreover, Cigré recommends customers to address these issues already during tendering (Cigré, 2017b).

According to Cigré 713, the largest challenges, in terms of maintenance, are changes in personnel and technology. To manage personnel changes, it is recommended to thoroughly document training programs and to update all relevant documentation continuously, since technology may change throughout the HVDC station’s life-cycle (Cigré, 2017b). As technology advances during the HVDC system life-cycle, interfaces between components or subsystems are changing due to the lacking standardisation, therefore, spare part suppliers and HVDC suppliers must cooperate to support different generations of equipment. One option for the HVDC system’s owner is to sign a long-time service agreement where the HVDC supplier will provide repairs and technical support (Cigré, 2017b).

2.4.1. Reporting operational performance of HVDC systems

Cigré, published a protocol for reporting operational performance of an HVDC installation (Cigré, 2014). Measurements of performance indicators such as utilisation, availability, reliability and severity of outage are considered, and the result of the protocol will prescribe actions taken whenever a failure occurs (Cigré, 2014).

Apart from Cigré, there are other initiatives for gathering operational performance from HVDC systems. In a survey sent to all HVDC-stations globally at the time (1996), which received 38 respondents, Cochrane et al. found compelling arguments to believe that locally operated HVDC stations have higher availability than remotely operated stations. Apparently, both locally and remotely operated stations (despite an anticipated bias to the contrary in the latter case) responded that they thought local operations would yield a higher availability, which also seemed to be the case to a significant degree when looking at reported station availability. Furthermore half of the respondents reported that they would hire more maintenance personnel if there were no budget constraint (Cochrane et al., 1996).

Cochrane et al. also reports that 15 out of 38 respondents would make significant changes to the maintenance requirement section (if given the opportunity to redo them) and, moreover, that 25 stations did not consider themselves subjected to conditions demanding particularly difficult maintenance. Lastly, no correlation was found between a high number of maintenance work-hours per year and station availability (Cochrane et al., 1996).

(27)

19 (60) In a more contemporary paper, treating refurbishment strategies for HVDC projects, Kirby et al.

concludes that the variety of constraints governing each system and owner, as well as specific project design features, render any refurbishment process exclusive to the extent that it cannot be applied to all HVDC installations (Kirby et al., 2012).

2.4.2. Risk assessment of the HVDC system

During tendering or development of HVDC systems, the conceptual design is reviewed to ensure that it fulfils all customer requirements, including performance rates, availability, buffers for outages, etc., the process is called risk review. According to IEC TR 62978, risk reviews are executed to provide proactive actions to address risks with the design of the developed HVDC project (IEC, 2017). According to IEC 62978, risk is considered as probability multiplied with consequence, the risk assessment covers both factors, the risk assessment should be applied and adequately defined throughout the organisation. To summarise the risk assessment parameters, the IEC 62978 description is presented in Table 6 (IEC, 2017).

Table 6. Risk assessment parameters as defined by the IEC.

Risk identification SWOT analysis

Hazard and Operability studies (HAZOP) Risk assessment workshops

Industry benchmarking Incident investigation Auditing and inspection Risk analysis Threat analysis

FMEA FMECA

Root cause analysis Event tree analysis Fault tree analysis

Risk controls Reliability centred maintenance Condition based maintenance Risk based inspection

Instrument protective function

2.5. Outlook on industrial maintenance practices

This section briefly investigates maintenance programs from different industries with comparable conditions to HVDC systems. These industries rely on equipment with high capital investments, customised solutions and significant safety requirements. The aim is to identify best practices that may be translatable to Company A’s business and show how these industries approach maintenance and the issue of maintainability. Investigated industries include the space (NASA) and nuclear energy industries through the International Atomic Energy Agency (IAEA). Additionally, a case study by Barabadi investigates the mining industry. Barabadi’s case-study features a step-by-step method for the development of a maintainability management program.

2.5.1. Space industry (NASA)

NASA performs maintenance on equipment both “on-ground” systems and in-orbit systems.

Equipment must be designed to avoid the necessity of “delivering” equipment to satellites and telescopes. Presently, NASA has ongoing projects within their space programs (throughout this section referred to as “program”).

(28)

20 (60) 2.5.1.1. Benefits of implementing maintainability in NASA’s space programs

NASA defines maintainability as “a measure of the ease and rapidity with which a system or equipment can be restored to operational status following a failure” (NASA, 1994). Maintainability will increase system availability through adherence to certain design aspects, such as visibility, accessibility, testability, simplicity, and interchangeability of the systems being maintained (NASA, 1994). At NASA, program management is responsible for implementing maintainability early in a project’s life-cycle. To that end, maintainability is studied using trade-off investigation on a certain design’s impact on life- cycle cost. An example is the development of the Hubble Space Telescope, where maintainability concepts were introduced early in the design of the program, which led to significant operational cost savings. NASA programs has five major phases, namely, conception, validation, full scale development, production and operation. An illustration of costs as a function of design manoeuvring (in terms of maintainability) during different program phases is presented in Figure 5 (NASA, 1994).

Figure 5. Effect of implementing maintainability during different program phases. Program refers to NASA's different space programs (modified) (NASA, 1994).

NASA’s system engineering process requires that the system is designed for “easy maintenance” in its specific working environments, additionally, the process is developed by personnel (experts) within design and operations, additional funds are invested in order to generate maximum program benefit.

NASA recognises that lacking system maintainability will affect the project scheduling since maintenance times and number of activities will increase, which in turn lowers system availability (NASA, 1994).

As mentioned earlier, NASA recognises the importance of developing prerequisites for maintainability early in the program, since maintenance will support the end-item once it is operational. Apart from system design, maintenance prerequisites (in the sense of logistics support) should span the requisite

(29)

21 (60) test equipment, facilities, spare parts, as well as special tools. Development of logistics support early in the program will allow optimisation of maintenance planning and solidify operational aspects of the program (NASA, 1994).

Design aspects affecting system availability are presented in Table 7 (NASA, 1994).

Table 7. Maintainability design aspects (NASA, 1994).

Design aspect: Description

Visibility Visual access to a system or a component, including “short duration tasks”, since downtime can be increased when the view is blocked.

Accessibility Ease of access during maintenance. Poor accessibility can lead to failures of other components caused by removal/disconnection of such equipment.

Testability Speed of fault diagnosing and ability to detect system faults as well as isolating them to the lowest replicable component (to not replace more than strictly necessary). Can be done using built-in, automatic test equipment or offline tests.

Simplicity Regulate the number of parts/subsystems within a given system. Simpler systems lead to lower spare part investments, more efficient maintenance overhauls and reduced overall cost. Moreover, simplicity may reduce training costs. Even if personnel are sufficiently skilled to perform maintenance on complex systems, a simpler system may lead to an increased availability.

Interchangeability Ability to replace a component with a similar component without recalibrating the system. NASA recommends standards or common end-items Human factors Identify features that prevents operators from working in ergonomic

positions. This aspect can be investigated when the system is represented as detailed drawings. However, an earlier evaluation may also prove valuable.

According to NASA, implementing maintainability during the design phase can save operational costs for both manned and unmanned systems. Since NASA operates in hostile micro-gravity environments (for example, Hubble telescope), minimal exposure to this environment can be achieved through designing for maintainability (NASA, 1994). Moreover, NASA developed remote system restoration by redundancy management and contingency planning to assure successful space missions (NASA, 1994).

During program development, program management is responsible for integrating maintainability in the early program stages to gain proper control of the maintainability discipline throughout the program (NASA, 1994). Figure 6 represents NASA’s maintainability program (NASA, 1994). A brief description of the stages in NASA’s program development is presented below (NASA, 1994):

BUSINESS ANALYSIS OF COMPANY A’S MAINTENANCE ORGANISATION