Multi-Criteria Evaluation of Smart Escape and Emergency Lighting Alternatives for Offshore Platforms: Case Study of BorWin5

1. Introduction

Depending on the applicable safety standards and the specific operational requirements of the facility, a general lighting system may comprise up to three distinct networks: normal lighting, emergency lighting and escape lighting. The normal lighting network is powered by the primary electrical supply, the emergency lighting network is activated automatically if the main power fails, and the escape lighting network is designed to guide people safely out of an installation during emergencies. Escape lighting operates independently of the main and emergency networks, powered by its own power source, and activates when both the main and emergency systems fail. It is essential in facilities where the main power supply cannot be restored quickly, and where emergency lighting cannot be provided without a blackout via diesel generators or similar. It is also essential in places with high occupancy, such as public buildings. The escape lighting ensures visibility of exit routes, helps reduce panic, and remains on until normal is restored. Emergency and escape lighting play a crucial role in ensuring safety during critical situations [1]. This type of lighting is essential not only in transportation systems, such as aviation and roadways, but also in public buildings and remote areas where electricity supply may be irregular or unavailable where a blackout is a risk for the security [2]. Emergency and escape lighting is becoming increasingly prevalent in modern society and remains a mandatory requirement in all onshore buildings intended for public use, such as hospitals, educational establishments, shopping centers, educational institutions, and factories and warehouses and numerous other facilities [3].

In the offshore sector, the focus of this work, escape and emergency lighting is a critical safety mechanism designed to protect workers and personnel aboard oil rigs, drilling platforms, wind farms, and service vessels operating in remote marine environments. The offshore setting presents unique challenges, including harsh weather conditions and potential hazards such as fires or equipment failures. The complex logistics of maritime operations also present challenges. Therefore, effective lighting systems play a crucial role. Providing clear, accessible lighting during emergencies is essential for facilitating safe and timely evacuations, thereby safeguarding lives and minimizing injuries. In the offshore industry, where response times can be limited by the isolation of operations, emergency and escape lighting systems must be reliable and resilient. They must function effectively under conditions of reduced visibility, such as smoke or water ingress, and during power blackouts.

Thanks to advancements in emergency and escape lighting technology, these systems are becoming more effective and reliable. Current innovations such as smart emergency lighting solutions that utilize LEDs [4,5,6], are designed to provide more energy-efficient and sustainable lighting systems. These systems aim to address issues associated with traditional emergency lighting, such as energy inefficiency and maintenance challenges. Smart emergency and escape lighting systems are characterized by their ability to automatically activate during emergencies and verify their proper functioning even when they are not switched on. These intelligent systems ensure that escape routes are illuminated in real-time, guiding occupants to safety even in challenging visibility conditions. The integration of Internet of Things (IoT) technology is also reshaping the landscape of emergency and escape lighting. IoT-enabled emergency lighting systems can communicate with other safety devices and networks, allowing for a coordinated response during emergencies. These systems can collect data on lighting performance, usage patterns, and potential maintenance issues, ensuring that emergency and escape lights are always operational when needed. Continuous monitoring and automatic alerts for maintenance needs are crucial features that enhance the reliability of emergency lighting [7]. Smart lighting can adapt to different scenarios, such as adjusting light intensity based on occupancy or time of day, thereby maximizing effectiveness while minimizing energy consumption. Such adaptability is especially beneficial in settings where personnel work non-traditional hours, such as night shifts, enabling greater responsiveness to varying operational needs [8,9].

Research emphasizes the importance of human factors in developing these systems, since understanding how people interact with lighting can lead to more effective emergency protocols [10]. While maintenance of escape and emergence lighting traditionally involved manual testing, this is often overlooked today to reduce costs, despite its importance for safety during power outages [11,12]. In the recent years, the introduction of wireless and automatic checking mechanisms facilitates easier monitoring of light performance, ensuring that systems are functional when needed [13,14]. In this sense, the international standard IEC 62034:2012 [15] represents a significant advancement in emergency escape lighting. This standard stipulates performance and safety criteria that automatic test systems must meet, effectively ensuring that emergency and escape lighting functions reliably when it is needed most. A key benefit of implementing this standard is the reduction in preventative maintenance costs associated with these lighting systems [16]. The problem is that the implementation of IEC 62034:2012 has advantages but also disadvantages, especially in terms of cost and complex hardware and software. Accordingly, what are known as multi-criteria decision-making techniques (MCDM) are very useful in decision processes [17,18]. The application of MCDM methods allows decision-makers to systematically compare alternative options based on various criteria. These methods provide structured frameworks for ranking alternatives, grounded in both quantitative and qualitative assessments [19]. The significance of MCDM extends beyond just choosing an alternative. It enhances transparency in the decision-making process, as it accounts for subjective preferences and diverse stakeholder perspectives.

MCDM has countless applications. In this paper, it is applied to the selection of an emergency and escape system for the new building BorWin5 high-voltage direct current (HVDC) offshore converted platform. The motivation stems from the fact that designing and selecting the emergency and escape lighting system for the converter platform in question is critical to ensuring safety within industrial facilities and essential infrastructure. To the authors’ knowledge, offshore platforms constructed to date are equipped with conventional emergency and battery-powered escape lighting techniques. However, these systems have not been implemented in accordance with the IEC 62034:2012 regulation (Automatic Test Systems for Battery-Powered Emergency Escape Lighting). This paper analyses possible solutions for the emergency and escape lighting on the BorWin5 platform and confirms the solution implemented. Additionally, the recently installed emergency and escape lighting system on BorWin5 HVDC is described. This safety system has been engineered using the most advanced technologies currently available on the market to ensure maximum reliability and performance in critical infrastructure applications.

2. Materials and Methods

2.1. Case of Study

In recent years, the offshore wind energy sector has experienced exponential growth, driving significant transformation in the maritime industry and the wider renewable energy sector. This surge has initiated an unprecedented phase of technological and industrial advancement. The continuous increase in wind power generation capacity, coupled with the relocation of wind farms to greater distances offshore, has made the direct transmission of generated energy via low-voltage cables impractical. To address this, offshore electrical substations are being constructed to collect power from wind turbines and transmit it to onshore grids via HVDC systems. In this context, the BorWin5 platform, Figure 1, departed the shipyard in mid-March 2025 and is currently being installed in the North Sea.

BorWin5 is an innovative HVDC offshore converter platform designed to improve connectivity between offshore wind farms in the North Sea and the German power grid. Owned by TenneT Offshore GmbH, it was constructed at the Dragados Offshore yard in Cádiz, Spain. With a transmission capacity of around 900 megawatts (MW), BorWin5 uses voltage-sourced converter (VSC) technology to efficiently and stably convert power between alternating current (AC) and direct current (DC). The platform is part of the BorWin cluster, located off the coast of Germany.

BorWin5 platform is composed of two main elements: the offshore converter platform and the onshore converter station, Figure 2. The offshore converter transforms AC power generated by wind turbines into DC, facilitating efficient long-distance transmission via approximately 120 kilometers of subsea cable. The topsides of the platform weigh around 12,000 tons, while the supporting jacket structure weighs about 7,500 tons. The onshore converter station is located near Emden (Germany). It receives the incoming DC power and converts it back to AC for integration into the national grid. Operating at ±320 kV DC, the system ensures reliable and high-capacity energy transmission.

Not only does BorWin5 demonstrate technical excellence in HVDC transmission, it also sets a new standard in offshore energy infrastructure. It showcases the capability to integrate large-scale renewable energy sources under challenging environmental and operational conditions. Rigorously engineered and custom-designed to meet the stringent technical, environmental and safety criteria intrinsic to offshore operations, it constitutes a step forward in sustainable energy system deployment. This makes BorWin5 a pioneering application of the standard within the offshore power transmission sector.

2.2. Alternatives for the Emergency and Escape Lighting Test System

The emergency and escape lighting of the BorWin5 platform consists of approximately 1,800 luminaires from multiple manufacturers, all powered by a 3x400V+N (main network) and also 220 VDC (emergency network) based on an open-architecture design. Multiple alternatives are available for the testing system, each characterized by distinct features that require thorough evaluation in light of environmental conditions, applicable regulations, and operational demands. The following four were considered in the present work:

-

Alternative 1. Traditional Testing of Self-Contained Emergency and Escape Luminaires. This method involves scheduled manual testing of luminaires equipped with integral batteries. It offers flexibility in terms of installation and is simple to implement; however, it requires continuous human intervention for battery status checks and functional verification, increasing operational workload and the potential for human error.

-

Alternative 2. Automatic Testing (ATS) of Self-Contained Emergency and Escape Luminaires. This approach integrates automated testing and monitoring protocols directly into self-contained luminaires. By eliminating the need for manual testing, it significantly reduces maintenance demands and improves overall system reliability. Continuous monitoring ensures timely fault detection and enhances compliance with safety regulations.

-

Alternative 3. Traditional Testing of Centrally Powered Emergency and Escape Lighting Systems. Under this configuration, a centralized battery system supplies power to multiple luminaires. Manual testing is carried out periodically in accordance with established schedules. Centralized power simplifies battery replacement and lifecycle monitoring, thereby assisting and facilitating maintenance through its centralized structure.

-

Alternative 4. Automatic Testing (ATS) of Centrally Powered Emergency and Escape Lighting Systems. This solution incorporates automated testing and diagnostics within a centralized power framework. It enables real-time monitoring and streamlined fault identification, offering enhanced system performance. However, it entails greater infrastructure complexity and a higher initial investment compared to other solutions, as well as the need for skilled professionals to maintain the system.

The investment cost of each alternative was calculated taking into account the following considerations derived from the actual BorWin5 project:

-

Luminaire pricing: An average unit price was considered for both emergency luminaires (such as floodlights, technical luminaires, downlights, wall-mounted fixtures, and similar) and escape route luminaires, based on platform design standards.

-

Battery pricing: The battery cost was estimated using an average value provided by the manufacturer LIGHTPARTNER for this type of equipment which is the main supplier of luminaires of this platform.

-

A 2.5 mm² cable section was considered for each luminaire, as is standard in this type of installation. A 3G2.5 cable was used for traditional systems in platforms, while a 5G2.5 cable was used for ATS systems, assuming an average distance of 25 meters between luminaires. For centralized systems, it was further assumed that power supply from the battery panels is provided using 35 mm² cables over a distance of 60 meters. All other auxiliary system cabling (power supplies, heating circuits and automation signals) was estimated as 10% of the total value of the main cabling.

-

It was estimated that the entire project of the emergency and escape lighting system should be completed within 4 months with a team of three engineers, whereas autonomous systems are expected to require 5 months.

-

The cost of the electrical panels was estimated based on information provided by the supplier for the BorWin 5 project, considering a 30% reduction if the system does not include an automatic control system.

-

The cost of the central panels include the inverters, batteries, interconnections and distribution boxes.

-

Ilumination software: The involvement of an expert technician in programming was considered for a duration of tree months. This estimation is based on the knowledge acquired during the development of the BorWin5 project.

Figure 3 shows the diferent values of the initial investment corresponding to each alternative.

Taking into account factors beyond economic considerations, Table 1, Table 2, Table 3 and Table 4 present an analysis of the comparative advantages of each of the aforementioned four alternatives in relation to the others.

2.3. Decision Model

As indicated above, a MCDM model was developed to select the most suitable alternative among the four considered. To facilitate this selection, a decision tree was established, Table 5. As shown in this table, the model was structured around three primary criteria:

-

cost

-

system architecture and design optimization

-

operational performance

Each of these criteria was further divided into several sub-criteria. Both beneficial and non-beneficial aspects were included. Beneficial are those for which higher values are preferred, while non-beneficial are those for which lower values are desirable. The nature of each sub-criterion is specified in the last column of the table.

The values corresponding to each sub-criterion are detailed in Table 6. Most of these values were quantified using continuous measurement scales, such as monetary units (euros), time durations (hours), and other metrics. In those cases for which quantitative continuous data could not be obtained, a qualitative assessment was employed using a Likert-type ordinal scale ranging from 1 to 5, where 1 denotes the least favorable performance and 5 the most favorable.

To ensure comparability across criteria with different units and scales, a normalization process was applied to all values. Several normalization procedures can be found in the literature [20]. In the present work, the so-called linear max-min normalization was employed. In this method, the normalised values are given by the following equations:

whereby each X_i.j,k represents the value of the sub-criterion i.j corresponding to the alternative k, V_i.j,k the normalized value of X_i.j,k, X_k,max the maximum grades of the alternatives for each sub-criterion i.j and X_k,min the minimum grades of the alternatives for each sub-criterion i.j.

The linear max-min normalization method transforms all values of a sub-criterion to a [0 1] range, where a value of 0 corresponds to the least favorable outcome and a value of 1 corresponds to the most favorable outcome. Through this procedure, all values are transformed into a dimensionless scale that facilitates unbiased aggregation and comparison. The normalized values for the present work are shown in Table 7.

Several procedures can be used to evaluate and rank the alternatives. In the present work, the Simple Additive Weighting (SAW) method was employed. This is a widely used technique in MCDM approaches due to its simplicity and effectiveness. In this method, the overall performance of each alternative is calculated as a weighted sum of its normalized scores across all sub-criteria, Eq. (3). Given the hierarchical structure of the model, in which each main criterion comprises multiple sub-criteria, the adequacy index (AI) for each alternative is computed by aggregating the weighted contributions of all sub-criteria, nested within their respective main criteria, as shown in Eq. (3). When the SAW method is used with linear max-min normalized values, the adequacy index for falls within the [0, 1] range. A value close to 1 indicates a highly suitable alternative, while a value close to 0 indicates a less suitable one.

whereby AI_k is the adequacy index for alternative k, α_i is the weight (per unit basis) of the i-th main criterion, β_i.j is the weight (per unit basis) of the j-th sub-criterion under criterion i, V_i.j,k is the normalized value of the j-th sub-criterion of criterion i for alternative k, n is the number of main criteria, and m_i is the number of sub-criteria under criterion i.

3. Results and Discussion

An important aspect of MCDM is establishing the weights assigned to each criterion and sub-criterion. These weights represent the importance given to each aspect. There are numerous objective methods available in the literature for determining weights using mathematical formulas. However, subjective weighting approaches are more frequently used. Objective methods are especially advisable when there is no consensus among the expert-defined weight sets [21,22]. In the present work, the weights were established based on what is known as expert citation. The weights are shown in Table 8. It is important to note that a sensitivity analysis of the assigned weights is shown at the end of this section.

As can be seen in Table 8, a weight of 45% was assigned to cost, a 20% to system architecture and device optimization, and 35% to functionability and operability. Obviously, the sum of the weights assigned to all the criteria must add up to 100, i.e., . Similarly, weights were also been assigned to the sub-criteria. The sum of weights of the sub-criteria must also add up to 100 for each criterion, .

Based on the application of the SAW method described in the previous section, the AIs corresponding to each alternative were computed aggregating the weighted and normalized scores across all sub-criteria. The results of applying Equation (3) are shown in Table 9.

The resulting indices shown in this table provide a quantitative measure of how well each alternative satisfies the defined evaluation criteria. These values, which range between 0 and 1, allow for a straightforward comparison: alternatives with higher adequacy indices are considered more suitable according to the model. The rankings are shown in brackets in the table. The computed results reveal clear differences in performance among the alternatives, thereby supporting an informed and transparent decision-making process. As can be seen, alternative 4 resulted to be the best option with the first position, alternative 2 with the second position, alternative 3 the third position and the least recommended option, alternative 1, with the fourth and last position in the ranking.

It is worth mentioning that the AI calculated using Equation (3) can be broken down into three distinct sub adequacy indices, each corresponding to one of the key criteria analyzed: cost, system architecture and design optimization, and operational performance. In this sense, the method can enable the assessment of how each alternative contributes individually to these three pillars. As such, it is possible to derive an economic index (AI_cost), a system architecture and design optimization index (AI_system), and an operational performance index (AI_performance). These sub-indices are also determined using Equation (3), but by considering only the sub-criteria associated with each specific dimension. Table 10 presents the overall AI along with its corresponding sub-indices for the evaluated alternatives. Positions are indicated in brackets. This information is also graphically shown in Figure 4. As can be seen, alternatives 2 and 4 present notably suitability in terms of performance (AI_performance), while alternatives 3 and 4 stand out in the economic criterion. Alternative 1 has low values for the three criteria considered (AI_cost, AI_system and AI_performance).

It is important to highlight that MCDM methods are inherently influenced by subjectivity, and SAW is no exception. Subjectivity primarily arises when assigning weights. As a result, applying Eq. (3) may yield different outcomes depending on these inputs. However, while the exact numerical results might vary, the overall ranking of alternatives should remain largely consistent—provided that the weights and value functions are reasonably chosen. Shifts in ranking are typically limited to alternatives with very similar input values and performance. Likewise, the gap between the adequacy indices of two alternatives may vary with different input assumptions, although the proportional relationship between them tends to be preserved to some degree. To address these uncertainties, conducting a sensitivity analysis—such as the one carried out in this study—is a recommended approach. A weight sensitivity analysis was carried out by analysing six cases with different weight combinations, shown in Table 11. Visually, this information is represented in Figure 5. As can be seen, the adequacy indices change slightly according to the different weight distributions. Regarding the positions of each alternative, these remain practically unchanged. Alternative 4 appears as the most adequate option in the six cases analysed. This indicates that the model is robust in that small changes in the weights do not significantly affect the results.

In conclusion, based on the comprehensive analysis presented, alternative 4 emerges as the most adequate solution for emergency and escape lighting on BorWin5 offshore platform. This alternative, which combines centralized power supply with automated testing and monitoring systems, offers significant advantages in terms of reliability, maintenance efficiency, and compliance with stringent offshore safety standards. Accordingly, alternative 4 was selected and successfully implemented on the BorWin5 platform, demonstrating its effectiveness in meeting the unique technical, environmental, and operational challenges inherent to offshore installations. The details are provided in Appendix A.

4. Conclusions

This study demonstrates that the implementation of automatic testing systems for emergency and escape central lighting, in accordance with IEC 62034:2012, offers a robust and future-proof solution for offshore platforms operating in demanding environments. Through a comprehensive multi-criteria decision-making (MCDM) analysis, four alternative configurations were evaluated based on cost, system architecture, and operational performance. The results clearly identified the automatic testing of centrally powered systems as the most suitable option, offering superior reliability, maintainability, and integration capabilities.

The successful deployment of this solution on the BorWin5 HVDC converter platform validates its technical and economic feasibility. Key innovations—such as the use of a 220 VDC emergency network, open-architecture network, real-time monitoring, and DALI-based interoperability—highlight the system’s adaptability and long-term scalability. Moreover, the adoption of an open communication protocol breaks away from vendor lock-in, enabling flexible integration of multi-brand luminaires and enhancing lifecycle sustainability.

This work not only advances the implementation of the IEC 62034:2012 standard on the BorWin5 platform, but also provides a replicable framework for future offshore infrastructure projects. The BorWin5 case study serves as a reference for the modernization of emergency and escape lighting systems in critical offshore installations.