A Repurposable Attribute Basis for Identifying Repurposing Opportunities in Decommissioned Products

Multiple life cycle products are essential for the efficient use of natural resources [1]. Strategies such as remanufacturing can enable a second life cycle for a product that has reached its end-of-life. However, returning the product to the same value stream might not always be viable. For example, the product could be outdated [2], the remanufactured product could cannibalize the new product [3], or critical safety and performance requirements could hinder remanufacturing [4]. The next convenient options would be to reduce the product to material by recycling, energy recovery by incineration, or disposal. These methods eliminate the embedded values of the product while using energy and capital. However, the product in its end-of-life state could be repurposed to create value in a different application [5]. That way, the product has multiple life cycles before the recycling or disposal stages.

Repurposing is challenging because the end goal, i.e., what to repurpose the product for, is unknown. We need to decide which attributes (“a quality or characteristic that something has” [6]) of the product to be repurposed and which changes or modifications to be made so that the waste product creates value. Therefore, a starting point to analyze a product in the context of repurposing is required. We aim to fill this gap by developing a repurposable attribute basis (RAB). The basis is a set of attributes that encompass the possible pathways or actions for identifying repurposing opportunities. We hope designers and practitioners can use the RAB as a foundation for enabling repurposing as an effective decommissioning strategy.

Repurposing can preserve a product's remaining attributes by extending the lifecycle while delaying the disposal stage. It could also generate additional revenue, which can be utilized later for recycling and sustainable disposal activities [7]. According to Cusenza et al. [8], repurposing can postpone an electric vehicle (EV) battery's recycling or disposal stages by ten years. Even though recycling is essential for recovering the material, delaying it is currently advantageous for hard-to-recycle products such as wind turbine blades and electric vehicle batteries [9,10] since it can provide time for improving recycling technologies and infrastructure [8]. For example, most wind turbine blades are currently used as low-value filler material in the cement industry [11].

A repurposed product can prevent a new product from being manufactured, providing environmental and economic benefits. For example, electric vehicle batteries can be repurposed for stationary energy storage applications, eliminating the need for purpose-built stationary storage. Kamath et al. [12] showed that using repurposed batteries for stationary energy storage can reduce the levelized cost of electricity and carbon emissions. Furthermore, Steckel et al. [13] showed that the upfront cost is approximately 30% less for a repurposed system than for a new battery energy storage system. According to Kamran et al. [14], by 2050, all the energy storage systems in the UK could be built from repurposed electric vehicle batteries without hindering the material supply for new batteries. Additionally, the repurposed product could have higher performance than the product it replaces. For example, repurposed wind turbine blade parts can replace lower-performance conventional construction materials like wood and steel, as confirmed by a structural analysis conducted by Joustra et al. [15].

Through a literature review, Hewa Witharanage et al. [16] identified repurposing as a potential strategy for decommissioning large-scale mechanical systems in a circular economy. They suggested approaches such as designing the systems to be utilized in their local ecosystem or designing to be cascaded between industries after decommissioning. Therefore, repurposing can be considered an advantageous decommissioning strategy.

Multiple studies have investigated methods to enable repurposing as a decommissioning strategy. Kwon et al. [17] used the concept of visual similarity to find repurposing opportunities for decommissioned wind turbine blades. They utilized image recognition to find visually similar images of wind turbines to extract concepts and compared them with a concept generation activity. They concluded that visual similarity is a promising method to pursue, which produces correctly scaled repurposing ideas. Studies conducted by Arabian et al. [18,19] and Sisson et al. [20] investigated the effects of recalling episodic memories [18], the inclusion of scale [19], and induced gratitude [20] on generating repurposing ideas. The ideas generated by the participants with and without the interventions were compared. The total number of idea categories and the level of detail in the ideas were improved after the episodic memory induction intervention. Including the scale as part of the intervention improved correctly scaled repurposing ideas even though the number and categories of ideas were reduced. Induced gratitude intervention did not produce significant differences in the ideas. However, the initial evidence shows promise in incorporating psychology research into repurposing idea generation. The participants of all the studies mentioned above were given instructions to aid in generating repurposing ideas, such as “You should maximize the reuse of each part, which are made of fiber-reinforced composites that have high strength and stiffness and low density” [17] and “You may further cut the parts” [17]. The instructions seem to cover the product's characteristics to be repurposed or an activity that can be performed to enable repurposing.

Kim et al. [21,22] investigated how users utilize product elements when repurposing. They identified that the product's form, material, and manipulability influence the repurposing ideas, and affordances are crucial for repurposing. Affordances are “action possibilities of a user when the user interacts with an object” [23]. “Store-ability, hang-ability, cover-ability, and decorate-ability” [21] were the most common affordances identified in this experiment. Furthermore, around 50% of the ideas were based on hidden affordances. They argued that prior experience with the product, gained due to the few weeks of use, and the experimental setting could be the reasons. However, identifying the hidden affordances could create new repurposing opportunities.

Deeney et al. [24] conducted a multiple-criteria decision analysis to determine the best decommissioning strategy for wind turbine blades. They compared two repurposing applications, building bridges, and furniture alongside landfills, incineration, and co-processing in cement kilns. They concluded that the most sustainable options are building bridges, followed by making furniture. Therefore, repurposing is considered more sustainable in this example.

The above studies utilize characteristics such as visual similarity between two products, material properties, manipulability, form, or affordances of a product, and actions such as cutting to find repurposing opportunities. Even though the results are promising, the question remains on what other characteristics or actions can be used to find repurposing opportunities and improve the results of these methods. Furthermore, the repurposing examples used in the decision analysis were static applications, even though the blade had a dynamic application. Additionally, making furniture using high-performance blade material could be considered a significant downgrade in value. Therefore, support is needed to identify more repurposing opportunities to maximize the impact of the sustainable decisions made. Furthermore, the designers need support in identifying the repurposing opportunities to design a product capable of being repurposed in the local ecosystem or a different industry as suggested by Hewa Witharanage et al. [16]. Therefore, a basis that supports the understanding of repurposable attributes is necessary to enhance the impact of repurposing.

We investigated the current repurposing examples to determine the underlying attributes that enabled repurposing. During the analysis, we identified two types of repurposable attributes: common and transformer. A common attribute is a quality or characteristic both products have (e.g., material property) that can be repurposed. A transformer attribute is an action that helps one product be transformed into another product (e.g., deform), creating a repurposing opportunity. Then, we generalized the identified repurposable attributes to create the RAB, which contains a set of generic phrases representing all the repurposable attributes identified under the two types.

We gathered repurposing examples through a literature review and Internet sources such as Do-It-Yourself instruction websites. Search terms such as repurpose, repurposing, alternative use, reuse, and reusing were used. The databases for the literature review were Scopus and Web of Science.

The inclusion criterion was repurposing examples of decommissioned products. We collected repurposing examples based on a wide range of products, from complex systems (e.g., wind turbines) to household items (e.g., chairs). However, we excluded examples such as repurposing construction materials, food, and industrial byproducts. Repurposing examples that reduced the product to the material level, such as melting a metal component to create a new product, were also excluded.

Seventy-six unique repurposing examples were identified, 37 from the literature review and 39 from the internet sources. After the attribute tagging, seven examples were excluded (see Sec. 3.2); hence, 69 examples were included in the final analysis. See Table 1 for the list of examples.

The concept of inductive thematic saturation was applied to confirm that the RAB is exhaustive, i.e., new examples were collected by repeating the search with gradually less strict conditions. For example, examining beyond the top web search results or looking at blog posts and personal websites. These examples were then analyzed until no new RAB phrases were identified from the attributes tagged from them (see Sec. 3.2.1 for attribute tagging). We identified that 82% of the RAB phrases were discovered in the first 43% of examples, and no new RAB phrases were discovered after the first 75% of examples. As shown in Fig. 1, there are signs of saturation. However, since the field is yet to mature, the RAB is open to improvement through future repurposing examples.

Analyzing the repurposing examples requires product or domain-specific knowledge. Therefore, we utilized TechNet² as a support tool. TechNet is a publicly available semantic network based on patent data. It outperforms general knowledge-based semantic networks such as WordNet and ConceptNet for engineering and technology-related knowledge retrieval tasks [94]. All the terms in TechNet vocabulary have a unique vector value (word embedding) derived through natural language processing techniques (TechNet was based on Word2Vec [95]). The vector values are based on the similarity and proximity of those words to other words in the corpus, which allows words with similar meanings to have similar vector representations [96]. This enables the mathematical analysis of words using vector calculations. For example, the similarity between two words can be calculated based on the cosine similarity of their vectors.

Each repurposing example involves two products. For example, repurposing waste sheet metal from automobiles to create metal meshes [49] involves two products: sheet metal and metal mesh. The similarities in the knowledge domains between these two products could be analyzed to find repurposable attributes. Therefore, using the Gensim model [97] as the cosine similarity calculator, we gathered the top 100 most similar terms to each repurposing example, i.e., what are the 100 most similar terms to the two terms sheet metal and metal mesh? These 100 terms represent the common knowledge domain between the two products. We gathered a total of 7600 terms to be reviewed (76 examples × 100 most similar terms for each example). The number of most similar terms was limited to 100, the same as the method followed by the creators of TechNet [98], to avoid reviewing an unrealistic number of terms. Limiting the scope to 100 most similar terms proved sufficient since 80% of the tagged repurposable attributes were in the top 72 most similar terms (see Sec. 3.2.1 for tagging repurposable attributes). Additionally, no new repurposable attributes were discovered after the top 59 most similar terms, i.e., all the attributes tagged after the top 59 similar terms in each example did not reveal new RAB phrases.

We reviewed the 7600 terms and tagged the common and transformer attributes. Terms that are synonyms to either product in the repurposing example were excluded. For example, the term sheet-metal is the same as the product sheet metal. Additionally, terms that are out of scope or have no meaning in the context of the repurposing example were excluded. For example, the term capstan housing assembly has no meaning in the context of repurposing sheet metal to metal mesh. Table 2 shows examples of each attribute type and the two exclusion criteria. Only the top five ranked results of the sheet metal-to-metal mesh example are shown. The similarity column shows the cosine similarity value of the term to the repurposing example. The rank column shows the position of the term within the 100 in terms of similarity. Seven of the 76 collected examples were excluded since no repurposable attributes were tagged. A total of 189 attributes and an average of 2.4 common and 1.7 transformer attributes were tagged for the other 69 examples.

Weighted kappa (K_w) analysis with linear weights was conducted to ensure reliable repurposable attribute tagging. Two terms (2 out of 100 most similar terms) each were picked from 10 examples (20 total terms). Two raters tagged the 20 terms in a four-point ordinal scale from 0 (definitely not a repurposable attribute) to 3 (definitely a repurposable attribute). Two rounds of tagging showed that there was a statistically significant agreement between the two raters; Round 1: K_w = 0.744 (95% CI, 0.528–0.960), p < 0.001 and Round 2: K_w = 0.755 (95% CI, 0.515–0.996), p < 0.001.

The tagged 189 common and transformer attributes were then generalized, creating the RAB phrases. During the generalization, we used the definitions of the function-behavior-structure (FBS) ontology [99,100] to support the differentiating between product functions (what the product is for) and behaviors (what the product does). For example, a function of a wind turbine blade is to capture wind energy to create rotational motion. A behavior to achieve that function could be creating a lift force. It helped us identify functions and behaviors among the tagged attributes.

In Table 3, we present generalizations of ten tagged attributes, including five common and five transformer attributes, chosen from ten repurposing examples. We selected these attributes so that each one represents a different repurposing example and a different RAB phrase, providing the reader with a broader understanding of the generalization process. Additionally, Fig. 2 provides an overview of all the steps involved in the method. From the figure, it is visible that a single repurposing example could contribute to multiple common and transformer attributes. Furthermore, different tagged attributes from different examples could be generalized into a single RAB phrase. Figure 3 displays the number of tagged attributes contributing to each RAB phrase.

Analyzing the 69 repurposing examples resulted in 189 tagged attributes. Out of the 189 attributes, 132 (70%) represented common attributes, which were generalized into 13 common attribute categories (RAB phrases). The other 57 (30%) attributes represented transformer attributes, which were generalized into 15 transformer attribute categories. They are discussed in Secs. 4.1 and 4.2 in alphabetical order. Figure 3 presents the occurrence of the RAB phrases among the 189 tagged attributes.

 C1. Repurpose the ability to withstand forces

Some products are designed to withstand forces such as normal, tension, friction, and spring [101]. That ability can be used in a different application. For example, a wind turbine blade and a bridge both need the ability to withstand bending loads. Therefore, the blade could be repurposed as a small bridge [29].

C2. Repurpose the behaviors

A product's behaviors represent what the product does to achieve its functions as per the FBS ontology [100]. These behaviors can be repurposed to achieve new functions. For example, a wine cork needs gripping behavior to seal the wine bottle. A bathmat needs gripping behavior to provide a safe and slip-free surface to stand on. Therefore, the gripping behavior can be utilized by repurposing wine corks to build a bathmat [39].

C3. Repurpose the container

The product that acts as a container can be repurposed. For example, a water bottle is a container for water. It can be repurposed as a planter [41], and then the bottle becomes a container for soil and plants.

C4. Repurpose the contradicting attributes

The attributes that were minimized during design can be repurposed where they should be maximized. For example, a wind turbine blade needs to maintain a smooth airflow around it to maximize the wind energy harvested, i.e., the blade needs to have minimal drag [102]. However, a noise barrier could reduce the sound wave strength by increasing the airflow resistance [103], i.e., increasing the drag. Therefore, wind turbine blade parts can be repurposed as noise barriers by employing them in high drag inducing positions or angles [31,32,42].

C5. Repurpose the element/ingress protection capabilities

Capabilities such as dust, liquid, debris, and weather resistance of a product can be repurposed. For example, the weather-resistance capabilities of wind turbine blades can be utilized to build shelters [30,31].

C6. Repurpose the features unique to the location

Products can be repurposed within their local ecosystem to serve a new purpose. For example, the foundation of an off-shore wind turbine can be repurposed as an artificial reef to promote marine life [33].

C7. Repurpose the force transfer methods

Some products have methods to transfer forces between components, and that ability can be repurposed. For example, wind turbine blades have shear connections to transfer shear forces between structural components. Repurposing the blade to build a bridge can utilize this ability since bridges also require shear connections [31].

C8. Repurpose the functions

The functions of a product represent what the product is for as per the FBS ontology [100]. All or some of a product's functions can be repurposed. For example, an electric vehicle battery stores energy. This function can be repurposed by using the battery in stationary energy storage applications such as residential solar energy storage [70].

C9. Repurpose the geometric features

Geometric features of a product such as length, width, curvature, flatness, etc., can be repurposed. For example, the pull tab of a soft drink can and a picture hook have similar geometric features even though they have different functionalities. Therefore, a pull tab can be repurposed as a picture hook [26].

C10. Repurpose the material properties

The product's material properties, such as mechanical, chemical, thermal, and electrical properties, can be repurposed. For example, the composite wind turbine blade material has high stiffness. Therefore, the blade can be repurposed to build walls [32] or transmission poles [31], which also require high stiffness.

C11. Repurpose the product category characteristics (substitution)

Products can be repurposed within their product category. For example, an offshore wind turbine foundation and an artificial reef could be categorized as artificial offshore structures. Therefore, the foundation can be repurposed as an artificial reef [33].

C12. Repurpose the structural elements

Structural elements within products, such as beams, columns, struts, ties, diaphragms, etc., can be repurposed. For example, the structural elements of a wind turbine blade can be repurposed to build bridge girders since both structures act as I-beams [29].

C13. Repurpose the subcomponents

The subcomponents (structure) help achieve a product's behaviors (FBS ontology). These subcomponents can be repurposed to achieve new behaviors. For example, desktop computers and programmable logic controllers (PLCs) need CPUs to achieve different behaviors. Therefore, the CPU of a desktop computer can be repurposed as a PLC CPU [90].

 T1. Repurpose by adding behaviors

Behaviors can be added to a product to enhance its functions, enabling repurposing. For example, fabric scraps can be used to create artificial flowers [27]. The function of the artificial flowers could be to decorate/improve the aesthetics of a location. Adding the behavior of fragrance will enhance that function.

T2. Repurpose by adding functions

Adding functions can enable new repurposing opportunities for a product. For example, offshore pipelines can be repurposed as energy storage through hydro-pneumatic principles [92]. The original function of the pipeline could be to transport oil or gas. Adding the functionality of storing a pressurized mixture of seawater and compressed gas enables pressure energy storage. Therefore, the pressure energy storage function is added to the current transportation function, creating a new repurposing opportunity.

T3. Repurpose by altering the stability

The stability of a product can be increased or decreased to create new repurposing opportunities. For example, a ship's stability can be decreased to have a higher motion response to waves. That motion can be used to harvest wave energy. Therefore, the ship can be repurposed as a wave energy extractor [45].

T4. Repurpose by changing the contents

A product's contents can be changed or replaced to create a new repurposing opportunity. For example, repurposing suitcases to create chairs [25] requires changing the usual contents of a suitcase (clothes, travel items, etc.) to cushions.

T5. Repurpose by changing the fixed end

Some products have a fixed end to a stationary or moving object. Changing that fixed end could create new repurposing opportunities. For example, the fixed end of a wind turbine blade is its root, which connects to the rotor hub. The blade can be repurposed as a fence or wall by changing the fixed end to the leading edge [31,32].

T6. Repurpose by changing the orientation

Some products have an orientation in which they operate. For example, a ladder is used in a vertical orientation for a person to climb up. Changing that orientation could create new repurposing opportunities. For example, a ladder can be repurposed as a shelf if used in the horizontal orientation [25]. Therefore, the usually horizontal ladder rungs are now vertical and used as separators. The side rails, usually vertical, are now horizontal and used as the shelf for storage and weight bearing.

T7. Repurpose by deforming

The elastic or plastic deformation capabilities of a product can create new repurposing opportunities. For example, to repurpose waste sheet metal as metal meshes [46–49], we need to expand and deform the sheet metal to create mesh patterns.

T8. Repurpose by eliminating movement

Dynamic products can be repurposed for stationary applications. For example, a wind turbine blade requires motion to create wind energy. However, it can be repurposed as a transmission pole, which is a stationary application [31]. Even though the movement is eliminated, the wind turbine's other attributes can be utilized as a transmission pole.

T9. Repurpose by employing multiple instances of the same product

A product can be repurposed by using multiple instances of the same product to create a new application. For example, waste laptops can be repurposed as thin client computers [36]. Thin client computers are used in a network connected to a central server. The central server does the main processing and storage activities. Therefore, multiple low-performance laptops with limited functionalities can be repurposed as thin clients. Even though a single waste laptop has little value, it creates value in this example as a collection.

T10. Repurpose by performing a manufacturing operation

Standard manufacturing operations such as molding, forming, machining, joining, and additive manufacturing can transform a product and create a repurposing opportunity. For example, waste sheet metal can be repurposed to create metal meshes [47] by performing manufacturing operations such as perforation.

T11. Repurpose by re-programming

Programmable devices can be re-programmed to serve a different purpose. For example, waste cell phones can be repurposed as a transport shuttle monitoring/tracking system using the phone's sensor functionalities [91].

T12. Repurpose by separating a section

If the whole component lacks repurposing opportunities, a section of it can be repurposed. For example, the wind turbine blade skin/shell can be separated and repurposed to build shelters [30,31] or road signs [32].

T13. Repurpose by substituting the behaviors

A product can be repurposed by substituting its behaviors to achieve the same function. For example, the function of a light bulb and an oil lamp could be to illuminate. A light bulb can be repurposed as an oil lamp by removing the filament parts and filling the bulb with oil [25]. Therefore, the filament glowing behavior is substituted by the burning of oil behavior to achieve the same function.

T14. Repurpose by using in a less demanding application

Some products are decommissioned when they cannot perform at the required level. For example, an EV battery is decommissioned when its capacity reaches 80% [76]. The battery can be repurposed for a less demanding application, such as stationary energy storage [70], even though it cannot meet the performance requirement of an EV.

T15. Repurpose by utilizing the aesthetics

A product's attractive and artistic features can be repurposed. For example, repurposing teacups as candle holders could create an aesthetically pleasing product [27].

Figure 4 presents the repurposable attribute generalizations discussed in Secs. 4.1 and 4.2 as the RAB. The phrases are presented in no particular order.

This section presents two example applications to illustrate the usefulness and potential applications for the new repurposable attribute basis.

Figure 5 shows a desk phone and its components after disassembly. Figure 6 presents an example of using the RAB to find repurposing opportunities for the handset of that desk phone. The repurposing opportunities were inspired by actual products in the market. The common attributes of the handset are shown in dark gray circles. The light gray circles show the transformer attributes. Five potential repurposing opportunities are shown. The speaker is an example of a repurposing opportunity found using only the common attributes. The sound output behavior can be repurposed by replacing the handset speakers with higher quality speakers, and the handset container's attributes match a small speaker. The cleaning brush example uses only the transformer attributes. The cleaning brush can be built by separating the outer shell of the handset and adding bristles to its inner surface. The baby monitor, the wall mountable grab handle, and the armrest examples show the combined use of common and transformer attributes. For example, the handset and the wall-mountable grab handle have similar geometric features. Adding the ability to be attached to a wall through vacuum suction enabled repurposing.

The desk phone handset example shows the RAB's potential as a creativity support tool. The RAB has the potential to act as an idea promoting mechanism, specifically as a stimulus for creativity as per the creativity mechanisms identified by Kirjavainen and Hölttä-Otto [104]. Future work could test the RAB's creativity stimulation capabilities in an experimental setting.

With the support of the RAB, we can leverage databases, semantic networks (e.g., TechNet), and large language models (LLMs) to identify repurposing opportunities. Prompt engineering addresses structuring text inputs efficiently so that they can be understood by an LLM [105]. The RAB can be used as prompts to find repurposing opportunities using LLMs efficiently. Table 4 shows how the RAB can improve the prompt quality to find repurposing opportunities for a wind turbine blade using OpenAI's chatgpt 3.5 [106]. The first prompt does not utilize the RAB. The second prompt uses the common attribute, repurpose the ability to withstand forces. The third prompt uses the transformer attribute, repurpose by changing the contents. The results of the first two prompts are similar to the identified examples in the literature review. We can see a significant improvement in the third prompt, where chatgpt produces innovative and more detailed results, such as aquaponics systems and solar-powered charging stations, which were not identified in our literature review. However, these results were obtained using a single default RAB phrase for illustrative purposes. Customized and detailed prompts utilizing the entire RAB could provide better results. Future work could follow similar methods presented by Ma et al. [107] for generating repurposing concepts using LLMs.

The common attributes can be used to deconstruct a product's repurposable attributes. Products with similar attributes could be potential repurposing opportunities. We can focus on one common attribute or a combination of attributes. Therefore, the most valuable attribute/s to repurpose can be identified. The transformer attributes support finding repurposing opportunities beyond the common attributes. Repurposing opportunities can be found using only the common or transformer attributes or by combining the two.

The repurposable attributes can be further divided into categories. For example, the common attribute, repurpose the material properties, can be presented at a more detailed level, such as repurpose the mechanical properties or further detailed as repurpose the toughness. Similarly, the transformer attribute, repurpose by deforming, can be further detailed as repurpose by plastic deformation. Therefore, the RAB can be presented with a detailed hierarchy for specific cases.

Upon revisiting the studies discussed in Sec. 2.2, we can see the visual similarity method of Kwon et al. [17] exemplified the use of a common repurposable attribute. The material properties given in the instructions, such as high strength and stiffness, are also common repurposable attributes. Furthermore, instructing the wind turbine blade to be cut into pieces is an example of a transformer attribute. Expanding the scope and possibilities through the RAB could improve the results of the concept generation activities and introduce new methods similar to visual similarity. Kim et al. [21,22] showed the influence of repurposable attributes, recognized as form, material, manipulability, and affordances. Furthermore, understanding the repurposable attributes can help uncover hidden affordances. While the studies conducted by Arabian et al. [18,19] and Sisson et al. [20] show promising results and intriguing benefits, combining the benefits of the interventions with the RAB could maximize its influence. For example, the instructions given to the participants were mainly limited to the material properties of the blade. For example, “high strength/strong, high stiffness/brittle, low density/light” [18]. Adding comprehensive instructions based on the RAB could improve the results.

Identifying a product's repurposable attributes will support the decommissioning decision-making processes. Repurposing opportunities can be compared with other options such as remanufacturing, recycling, and incineration. For example, if a product is not suitable for remanufacturing due to the risk of cannibalization, repurposing could create better value than recycling or disposal. Therefore, the product's embedded value is utilized to the maximum before disposal. Understanding the repurposable attributes of the wind turbine blade will enable the multiple criteria decision analysis of Deeney et al. [24] (discussed in Sec. 2.2) to consider more repurposing options. They only looked at furniture and bridge manufacturing as the two repurposing options. The bridge utilizes the blade's ability to withstand forces (C1) and the furniture is made by separating small sections of the blade (T12). However, using the other RAB terms could result in more options to be considered. For example, the hollow and large interior (C9) and the element/ingress protection capabilities (C5) of the blade could potentially be repurposed as an underground/underwater storage facility (C3).

Similar to the example above, the RAB can support expanding on the current repurposing examples. If we consider the wine cork to bathmat [39] example discussed under C2 in Sec. 4.1, it utilizes the behaviors of the wine cork to repurpose it as a bathmat. However, we can expand the repurposing opportunities by considering other RAB phrases, such as utilizing its buoyancy (C10) to repurpose it as a fishing bobber (T2), changing its orientation (T6) and fixed end (T5) to attach it as a door stopper/bumper on a wall or as a drawer knob due to its shape (C9), and using multiple wine corks (T9) as a shock-absorbing structure utilizing its material properties (C10). Therefore, the RAB can support expanding the benefits of current repurposing examples.

Using repurposed products could result in function losses and function gains. Let us consider the same wine corks to bathmat example. Two primary functions of a bathmat could be to dry the user's feet and provide a non-slip surface to stand on. Using wine corks will retain the non-slip functionality but lose the drying function. Revisiting the shock-absorbing structure example mentioned above, the wine cork's water-resistant capabilities (C5) could result in function gains. Therefore, the opportunities identified through the RAB should be evaluated from a function gain and loss point of view. Furthermore, the function gain/loss analysis could be supported by already established bases, such as the functional basis by Hirtz et al. [108]. For example, once the repurposing ideas are shortlisted, we can develop functional models to analyze the product to be repurposed and the new products (repurposing ideas) in detail for function gains and losses.

The RAB phrases could be modified to be the edge terms that connect nodes of a knowledge graph specifically developed to find repurposing opportunities. Here, the nodes would be products and components. Therefore, we can find new repurposing opportunities by analyzing the missing links in the knowledge graph. Additionally, having a knowledge graph with product components connected through repurposable attributes could also enable “reverse repurposing,” where we can find components that can be repurposed to design a new product. For example, we can take inspiration from drug repurposing research, which utilizes knowledge graphs to find existing drugs to treat new diseases [109,110].

Even though the RAB could support improving current repurposing opportunities and finding new ones, future work should address their feasibility. Bauer et al. [5] identified product, process, and business model related factors that should be addressed for a repurposing opportunity to be feasible. Safety issues, reverse supply chain, legal aspects, market, and partnerships are a few example factors.

The above discussions were focused on repurposing an already decommissioned product, i.e., design “from” repurposing. The RAB can also support design “for” repurposing, which is designing a product that can be repurposed after decommissioning. We can include design data, labels, or markings to identify common and transformer attributes. For example, if a product can be repurposed by deforming, the optimum points for deforming and limits can be embedded within the product.

A repurposable attribute basis was developed, analyzing 69 unique repurposing examples. We identified 28 attributes for repurposing, divided into 13 common and 15 transformer attributes.

The common attributes are qualities or characteristics a product has that can be repurposed. They are material properties, structural elements, subcomponents, ability to withstand forces, force transfer methods, element/ingress protection capabilities, features unique to the location, product category characteristics (substitution), geometric features, container, functions, behaviors, and contradicting attributes.

The transformer attributes are actions that help one product to be transformed into another product for repurposing. We can transform a product for repurposing by altering the stability, eliminating movement, changing the contents, separating a section, employing multiple instances of the same product, changing the fixed end, changing the orientation, utilizing the aesthetics, deforming, performing a manufacturing operation, re-programming, adding functions, adding behaviors, substituting the behaviors, and using in a less demanding application.

The RAB will keep updating and improving with new repurposing examples. The RAB could act as a foundation for developing creativity-based or data-driven methods for identifying repurposing opportunities. The generated ideas could support the decommissioning decision models and frameworks by providing repurposing options alongside remanufacturing, recycling, incineration, and disposal.

This study would not be possible without research funding from the University of Melbourne.

There are no conflicts of interest.

The datasets generated and supporting the findings of this article are obtainable from the corresponding author upon reasonable request.