Leveraging Blockchain Usage to Enhance Slag Exchange

Abstract

The slag generated as a by-product of the steelmaking process can be used to manufacture cement, reducing the generated waste and contributing to the circular economy. Currently, steelmaking companies promote long-term bilateral deals with one or few cement companies where the price is fixed, and the slag is a treated as commodity. We propose a new solution, which promotes slag reuse through its differentiation with a composition-based grouping and an auction. This process is carried out in a blockchain network, which increases trust in the system, provides guarantees about the slag composition to cement companies and helps external regulators to reliably extract circularity indicators.

1. Introduction

Steelmaking, a pivotal industrial process, transforms iron ore and scrap into versatile steel. This production occurs in two stages: primary steelmaking which involves obtaining a semi-product from a converter, electric arc furnace or induction furnace; and secondary steelmaking, or the refining process, which refines the steel using processes like ladle refining furnaces or vacuum degassing furnaces. Both primary and secondary steelmaking processes occur in the presence of slag. These by-products result from the interaction of flux and impurities during the smelting and refining of steels. These slags refine, protect and interact with steel and inclusions, ensuring high-quality steel production. The composition of steelmaking slags varies due to different facilities and the variability of produced steel grades.

During the refining process in the ladle furnace, various materials are added to the molten steel for desulfurization, degassing, and adjustment of its chemical composition. Ladle furnace slag is formed as a result of the reactions and interactions between these additives and impurities in the liquid steel. Once the refining process is complete, the slag is separated from the steel. The manufacture of each ton of steel via this method involves the generation of roughly 20 to 30 kg of ladle slag. Globally, in 2023, projections indicate that the production of ladle slag exceeds 20 million tons [1].

Ladle furnace is a solid, glassy material with a range of particle sizes. It typically contains oxides of elements like calcium, magnesium, aluminum and silicon. Disposal methods may include landfilling or other methods that comply with local environmental standards. It can also be used as a raw material for the cement industry [2]. However, steelmaking companies may prefer to send slag batches to landfill either because there are no cement companies interested in it or because the offered price does not compensate the shipment costs. On the other hand, regulations and political actions like the European Union’s green deal [3] promote circularity to reduce carbon emissions of such by-products, making their disposal more expensive.

Current long-term bilateral deals do not optimize the interest of both cement and steelmaking companies. While the cement company is interested in certain elements contained in the slag like the calcium, aluminum, or silicon, the steelmaking company is interested in simplifying the process while minimizing costs for getting rid of the slag.

This paper explores an auction system which can help to set a fair price for both types of companies more accurately. A blockchain network lies at the core of our proposal to enforce trustworthy data sharing and agreed procedures. Our solution promotes a more dynamic market where any cement company can check the composition of a certain slag aggrupation and offer a price for it. The composition of each slag and its aggrupation in bigger batches is tracked by blockchain. This provides guarantees of the slag quality (i.e., a certainty of its minimum composition) to the cement companies. Promoting a trustworthy market, we expect to increase the slag sold and recycled by the steelmaking companies. This alone can constitute a benefit for the latter and promotes circularity better than any existing regulation.

The goal of the solution proposed in this paper is therefore threefold: (1) to track the composition of the slag while it is grouped in the steelmaking process, (2) to help in the price definition through an auction where any interested cement company can participate and (3) to ease circularity metrics extraction for public administrations and regulators.

The rest of the paper is organized as follows. Section 2 sets the theoretical background which helps to understand both the technology and the application domain. Section 3 presents the methodology used in this research. Section 4 describes the implementation details of the tool developed. Section 5 discusses the strengths and flaws detected during the validation of the implementation presented and proposes several future improvements. Finally, Section 6 concludes our work.

2. Background

This section provides an overview of the theoretical foundations and concepts relevant to the research, including slag circularity and the application of blockchain technology to similar contexts.

2.1. Slag Circularity

The European Union and other governments are promoting circular economy to reduce carbon emissions in multiple sectors. The steelmaking industry has huge potential to get closer to the ambitious “zero-waste” goal [4] by recycling many of the by-products generated [5].

Steel slag is one of those by-products and it is often used as a recycled raw material for various construction products such as roadstone, asphalt or cement [6]. Particularly, the cement-making industry can benefit from the calcium oxide (CaO), silicon dioxide (SiO₂) and aluminum oxide (Al₂O₃) in the slag [7], but often, recycled rates are low, and sending it to landfill or holding the slag in a facility is preferred [8].

With the advent of the mentioned regulations, landfill usage is being penalized by public administrations, encouraging steelmaking companies to recycle their slag and minimize landfill disposal [9]. In Europe, the approximate cost for ladle slag disposal ranges from EUR 20 to EUR 50 per metric ton [10]. The slag is inexpensive. Its price typically ranges from EUR 5 to EUR 15 per metric ton for construction purposes. In addition, it is a bulky and heavy material, resulting in high transportation costs and low profit margin [11]. This stresses the importance of local actors and markets where the slag is consumed near to where it is produced.

Currently, slag composition (i.e., its quality) is ignored and slag batches from different steelmaking processes are mixed and sold to one or few cement companies [12]. In this scenario, the slag can be seen as a commodity that is essentially uniform across producers.

2.2. Blockchain Applications

Blockchain technology [13] is the key enabler for the solution proposed in this paper, where parties with opposing interests must interact over common data. It shines at maintaining an agreed state in all the participants of the network [14]. This state is composed by data structures called blocks. Each block groups a series of transactions sent and signed by individual users and contains a hash of the previously mined block. This structure ensures data integrity, as any change to a past block alters the hash in the latest block. New blocks are added via a consensus algorithm agreed by all participants. Peers usually contain a copy of all the blocks previously agreed. Some networks also allow transactions to include program deployments or calls, known as smart contracts, which are verified by the same consensus mechanism.

Thanks to the decentralization of the state (i.e., any party can locally access to it without relying on other participants), its immutable nature and to the fact that it is signed by the sender, blockchain and traceability are often seen as a perfect match [15,16,17,18]. Rejeb et al. [19] examine how blockchain technology has been applied to boost circular economy.

Belt and Kok [20] analyze the benefits obtained thanks to blockchain in different phases of commodity trading and how it affects different markets. This study highlights how a fairer price can be achieved and can reduce price-setting threats. Unlike more mature markets, the slag market lacks established IT systems to help in its exchange, so blockchain and the solution we propose could fill this need.

However, the cement generated with the slag is consumed by the construction industry, and many authors have explored how to use blockchain in this sector. Perera et al. [21] describes the different approaches in which the construction industry could benefit from using blockchain. Some of the relevant applications it analyzes are supply chain management, asset management, construction management, building maintenance or waste management. Our solution sits at the beginning of the supply chain and helps to certify the origins of the cement [22].

Shi et al. [23] provide a comprehensive description of various auction applications that leverage blockchain networks. Using this survey as a reference, we can position our work in comparison to others. When it comes to the application domain, energy auctions are the most popular ones, and as a consequence have been thoroughly discussed in other reviews [24,25,26]. To the best of our knowledge, this is the first proposal that takes advantage of blockchain-based auctions to promote the circularity of by-products. As indicated by these reviews, setting up auction schemes requires a thoughtful analysis of several market-specific factors. These factors include the needs and preferences of those participating in the auction, the existing rules and standards, and the unique features of the market. The auction scheme we present in this paper is secure, transparent, and efficient, and it considers these factors specifically for the steel slag market.

In addition to auctions, there are other ways in which blockchain technology has been utilized to improve the circular economy. A recent bibliometric analysis [27] provides insights on these applications. They primarily focus on conventional applications such as asset traceability in the supply chain applied to waste management. It is worth noting that the solution proposed in this paper can be extended to be used in the following steps of the supply chain too. For instance, an eco-cement produced with slag could be backed up with the trustworthy provenance information which comes from our blockchain-based solution. Similarly, when this cement was used in a construction site, one could track the origins of the building back to the slag.

3. Materials and Methods

This section outlines the research methodology and the design considerations that guided our solution.

3.1. Research Methodology

Figure 1 defines the methodology followed in this research. As Section 1 presents, we first identified the current problems slag reuse faces (problem identification). Afterwards, we defined the objective our solution must achieve (objective definition). That is, promote slag reuse and increase its perceived value. To achieve this goal, we identified that (1) cement companies need trust guarantees on the composition of the slag they bid for, (2) public administrations must be able to trustfully compute the amount of slag reused, (3) and that the auction system should not be manipulated by any of the participants (requirements identification).

After analyzing different technologies, blockchain arose as the perfect match for such an scenario where the different participants have competing interests (technology selection). Over such a network, we need a framework to trace the composition of the slag and the bidding of the participants.

Afterwards, we formally defined the use case to be implemented (experimental design). In this use case, we identified the participant actors, the operations that they should perform and the data models used to express their needs. With this design, we implemented a proof of concept by adapting the underlying traceability framework to the use case and deploying it in a blockchain network (implementation). The complete solution has to then be validated in terms of functionality and impact on the circular economy approach inside the HyperCOG project (functional validation).

3.2. Actors

Our solution considers three types of organizations: steel companies, cement companies and public administrations. The steel company generates different types of slag as a by-product. The company wants to get rid of the white slag and has several options depending on its quality. It can either sell it to a cement company in case of the high-quality slag or ship it to a landfill at zero or higher cost otherwise.

Cement companies use slags to produce cement. A cement company can bid for the slag to the steel company when its composition meets a minimum quality criteria.

Finally, a public administration seeks to promote the reuse of slag and the progressive reduction of the amount of slag sent to the landfill. For this, it uses indicators that allow them to assess the current situation and its evolution over time. With an accurate view of what is happening, the public administration can take more appropriate measures and propose more useful regulations.

3.3. Data Model

The traceability platform manages the life cycle of the slag. It considers the following slag states and aggregations, each of them with its own information: slag, slag batch, stocked, sold and discarded.

The slag is the material generated in the steel manufacturing process. Slags from several heats are put together in a cone. The resulting material is considered a slag batch. Categorized batches are sent into the corresponding stock silo depending on their composition.

Finally, when a bid for a certain stock amount from a cement company is accepted by the steel company, the batch is marked as sold and it is transferred to the cement company. If no high enough bid is received for a stock amount, the steel company can discard certain amount to free stock space.

3.4. Use Case

The organizations described above interact with each other exchanging information which is altered through the operations shown in Figure 2. These operations are performed sending transactions to the smart contract deployed in the underlying blockchain network.

The steelmaking process is held in the ladle furnace. To avoid impurities to obtain the desired steel quality, the ladle is covered with reducing slag. The steel company creates basic information of a slag in the blockchain network using a register operation. Afterwards, the composition of this slag can be measured and registered with a composition operation.

Once the heat has finished, the slag is poured to a cone, and this fact is recorded in the system using the to-cone operation. The cone can store a slag of up to three heats, so once it is full, it is grouped with slag batches of similar qualities (i.e., compositions) in a slag stock.

Cement companies can bid for any amount of slag, providing a price per the desired volume and using the bid operation. Thanks to blockchain and its properties, any participant can know when a slag was generated, modified or grouped and who made each change. Furthermore, the composition of each stock is expected to give them a good indicator of the quality of the slag (i.e., better slag optimizes the cement-making process). This information should help both parties to adjust the price of the bids.

At any point, steel companies can either accept (bid-accept) or reject (bid-reject) any of the bids. After acceptance, the appropriate volume is disaggregated from the stock and sent to the cement company. Steel companies can also get rid of a certain volume of stock if they do not meet an appropriate purchaser with the dispose operation to mark that they were sent to the landfill.

The proposed auction scheme does not have time restrictions. Any cement company can see the composition of a slag when they are in the stocked status and bid for it. Similarly, the steel company can accept or reject bids anytime. However, the logic of the smart contract could be easily extended to define the period when the bids can be received and another time to start the decision process. In fact, this decision process could be coded in the smart contract if the steel company has a predefined acceptance criteria, e.g., the bid with the best unitary price or the bid with the larger volume of slag.

Finally, public administrations can use the stats operation to obtain indicators of the amount of slag recycled by each company.

4. Results

This section presents the implementation and deployment details of the solution, including the data model or smart contract operations.

4.1. Implementation

To implement the use case described, we extended the Traceblock traceability platform [28]. Traceblock is a framework written in Golang [29] which helps to create smart contracts that can be deployed in a Hyperledger Fabric network [30] to trace assets for different use cases and industries.

Figure 3 shows the base data model that Traceblock offers to track an asset. Each asset is individually identified and has a type, a quantity associated to the specified units, composition information (i.e., parents and children), a location, the current owner and some status information about its creation, last modification or removal (if it applies). This model can be extended through the fields field, which can contain any arbitrary data.

In particular, we use it to include a customized status (slag, batched, stocked, sold or discarded), the composition (as a range of maximum and minimum $CaO$ percentage in the stock) and a list of the bids received for the slag stock. Each Bid contains a quantity, a unitary price, bidder information (automatically recorded by the smart contract) and its status (i.e., proposed, accepted or rejected).

Traceblock offers some off-the-shelf operations which are used in the implementation like $transfer$, $join$ or $split$. $Join$ and $split$ allow one to aggregate and separate slags in batches and stocks to the owner of the asset (i.e., the steel company in the beginning). $Transfer$ is used to change the owner of an asset.

Apart from these generic operations, we implemented $bid$, $bid-accept$, $bid-reject$ and $stats$. $Bid$ is the operation used by cement companies to bid for an specific slag stock. If the amount bid is greater to the available stock, the bid is automatically rejected. Otherwise, this operation appends a $Bid$ object to the $bids$ array inside the $fields$ field. $Bid-reject$ simply marks a given bid as $rejected$. Algorithm 1 describes the $bid-accept$ operation in pseudo-code.

Algorithm 1: Function bid-accept.

The $registerSale$ function registers statistical information about the operation in the smart contract. This information is later used to provide aggregated information to public administrations as a result of calling the $stats$ operation. Figure 4 shows the data structures used to store this information. These data structures are not linked to the $TraceableAsset$ model. The list of steel and cement companies are stored in the $SteelOrganizations$ and $CementOrganizations$ as arrays. Each steel and cement company have their raw stats stored in $RawSteelOrgStats$ and $RawCementOrgStats$ respectively. Whenever the $stats$ operation is invoked, raw data are refined into $SteelStats$ and $CementStats$ for each organization.

In Hyperledger Fabric, the certificate authority (CA) of each organization issues credentials to its users. The smart contract can read the user who signed each transaction and it can access the attributes associated to her certificate. These attributes are taken into account to control the access to the different operations as described by Figure 2. Specifically, it considers the type of organization and the role of the user.

To ensure both conciseness and transparency, the code developed for this paper has been publicly shared [31]. This code was analyzed with Sonarqube [32]. No bugs or vulnerabilities were detected. The code had 23 code smells, and only 3 of them major. None of the major code smells affected the chaincode.

4.2. Deployment

As explained in the previous section, Traceblock framework helps to develop smart contracts to trace different types of assets in an Hyperledger Fabric network [30]. Hyperledger Fabric is a well-established framework which allows more expressive smart contracts, better scalability and higher throughput than other popular solutions like Ethereum.

Hyperledger Fabric requires from each organization to maintain several components like a peer, a membership service (MSP) and an orderer. The peer has the ledger where information is written immutably and the smart contracts with the business logic to be executed. The MSP allows to manage the different users in each organization that can operate in the network. The orderer orders transactions ensuring a consensus in the new blocks stored in network. Figure 5 depicts the Fabric deployment configured to support the use case and the different organizations which maintain the network.

Apart from these components, we developed a Gateway API to intermediate between the HTTP clients (e.g., web browsers) and the peers. This API was written in NodeJS [33] using the Express framework [34]. On the left-hand side of Figure 5, we can appreciate the different applications made to allow end users to interact with the functionalities provided by the smart contract. Each actor has its own command line application (CLI) written in NodeJS to specifically cover the types of actions that concern him/her. These applications translate more abstract and human-friendly operations in one or more calls to the smart contract functions deployed in the blockchain network.

Besides the individual CLIs, all actors have access to a web dashboard written in JavaScript using the Vue.js framework [35], which allows them to track and inspect the assets exchanged using the web browser (see Figure 6). Note that each organization maintains both a Gateway API and a dashboard instance to avoid introducing a centralization point in the architecture. That is, a cement company should not rely on the information provided by an application fully controlled by the steelmaking company who is trying to sell the slag. The whole purpose of the blockchain network is not having to trust on the honesty of any of the other actors once the information has been recorded in such a network.

5. Discussion

The HyperCOG project focuses on transforming the process industry through cyber-physical systems and data analytics. By implementing our prototype and evaluating it within the project, valuable insights were gained. This section describes these lessons along with identified areas for improvement.

Initially, one might argued that the proposed auction-based market could be applicable within a business-to-business framework [36]. However, the consistency of the data shared within such a system could be compromised if it is distributed among various parties. More alarmingly, if a single party assumes full control over these data, there is a risk of it being manipulated or falsified at any given time. For instance, they could alter the quantity of a slag stock sold in the previous month. Moreover, each party would need to place their trust in the others to apply an agreed-upon logic to the data. The use of a blockchain network, with its inherent agreement and immutability of data and logic, is anticipated to alleviate these issues.

To trust data shared by a blockchain network, participants must trust the organizations managing its infrastructure. Our implementation uses a permissioned network where only participants authorized by other participants can join. This type of network is particularly suitable for long-term and stable business relationships due to its superior performance and ability to safeguard shared data privacy [37]. However, it also requires each participant to either fully trust in another organization or to deploy and maintain several components. In a scenario like the one presented, the conflicting economic interests of all participants makes trusting another organization unfeasible. In practice, this forces each participant to deploy and maintain its own blockchain components (see Figure 5).

In contrast to permissioned blockchain networks, public blockchain networks are open to any new participant. As a result, a larger number of parties become involved in managing these networks. Cement companies can place trust in other organizations within the network for two key reasons: (1) network protocols incentivize honest behavior, and (2) these organizations have no conflicting interests related to the use case. Consequently, these trusted entities can handle transaction processing, relieving cement companies from the burden of maintaining their own infrastructure.

However, companies might have concerns about data privacy and business confidentiality in such public networks. To make a blockchain system appealing to steel companies, privacy measures such as implementing bidding thresholds and selective data reveal strategies are suggested to prevent the indirect revelation of the real business volume of steel companies. To achieve bid privacy during the auction, a commit-and-reveal pattern [38] could be adopted. Lastly, an additional layer of security can be offered encrypting data on top of blockchain [23] so it can only be deciphered by authorized users or by asserting data properties without revealing the actual data with zero-knowledge proofs [39].

The deployment and maintenance of the solution presented might involve significant costs, which can represent an adoption barrier for companies. We refer to the work of [40,41,42], who conducted comprehensive cost analyses of blockchain-based systems. In this regard, the analysis of the economic costs associated with blockchain-based networks demonstrates that these kinds of systems should not be a cost barrier once they are correctly tailored for a need. We propose to extend our study by researching the cost factors from the perspective of resource-based view theory [43,44] for costs associated to resource management and consequent adoption barriers from companies, thereby providing a more realistic assessment of the feasibility of our proposed system.

Another noteworthy concern for blockchain-based applications is their performance in terms of scalability and throughput. Like in any blockchain-based solution, its scalability will primarily be influenced by the blockchain network used, its configuration and how the data are stored in the smart contract. For performance evaluations of the underlying Hyperledger Fabric platform and their configurations, we direct the reader to Refs. [45,46,47].

The manner in which data are stored by the smart contract is the only aspect where our design can influence the performance and scalability of the entire solution. We extended a traceability framework called Traceblock with the operations described above. This framework provides a basic and extendable asset model which includes aspects like the location and other details. However, the smart contract presented in this paper does not use these attributes in its business logic. Only web visualizations make use of them. As a consequence, these types of data can be more efficiently stored in an off-chain database to decrease the payload of blockchain transactions. This strategy alleviates the load that the blockchain peers need to handle, and as long as the hash of these details is stored in the smart contract, the system would keep ensuring its immutability.

However, it is important to note that in the context of the use case described, the throughput might not be a primary concern. According to the USGS, up to 290 million tons of steel slags were produced worldwide during 2023 [48]. In Europe, an estimated 11.4% of this slag was repurposed for cement production in 2021 [49]. This translates to a total market volume of 33 million tons per year, or approximately one ton per second. If we consider each ton as an individual slag in our use case, and given that Hyperledger Fabric can comfortably process hundreds of transactions per second [46], it becomes evident that even under the most optimistic scenarios, our solution is unlikely to encounter high-throughput issues.

Finally, the connection between the real world and the virtual world is also a crucial design factor. Our solution requires the unambiguous identification of physical objects (specifically, slag at all stages) and storing their identifiers in the blockchain. Regrettably, current slags lack tags or any identification mechanism. Thus, once batched or stocked, individual identification is lost, and we can only guarantee the original slags’ composition by providing minimum and maximum composition ranges for these groupings.

Moreover, when a stock is partially sold, it results in a smaller stock, complicating the inference of the original composition (i.e., which slags it originated from). To mitigate these issues, steel companies could opt for smaller groups or implement some form of identification mechanism. While both options would increase the cost of slag storage, at least one of them should be adopted to enable use cases like the one proposed here.

The assumption that both steel and cement companies will readily adopt the proposed system is indeed a simplification to keep the focus on the enhancement method for the slag exchange though blockchain technology. We recognize that there could be resistance due to various factors such as technological readiness, regulatory concerns, and perceived risks. Taherdoost [50] and AlShamsi et al. [51] explain the most common adoption barriers detected in equivalent industrial sectors. Therefore, we can assume that the lower entry barriers for new participants may foster a more dynamic auction market, encouraging market acceptance and addressing the concerns of the industry stakeholders.

6. Conclusions

This paper presents a blockchain-based model to promote slag reuse in the cement manufacture. The model offers guarantees about the composition of the slag sold to cement companies to increase their interest in such by-product and, through the auction, we ensure that steelmaking company can select the best offer every time. Blockchain plays a vital role in this solution since it helps to increase trust not only in the data shared but also in the correct execution of the bidding algorithm.

The presented solution also digitalizes the selling process, making it easy to extract lessons in subsequent phases. While the steelmaking company can learn the most desirable properties for the slag and try to group slag batches in different heaps to maximize those properties, public administrations can easily extract circularity metrics from the sector to adjust regulations or landfill disposal fees.

Our future work will not only involve translating the presented model to a public blockchain network and creating a lighter data model that aligns more closely with existing standards, but also embarking on a comparative performance analysis and measuring its impact in pilot studies. The dual goal is to reduce entry barriers for new auction participants and to foster interoperability with existing complementary contracts, while simultaneously benchmarking our model against other functionally equivalent solutions that may emerge. This comprehensive approach will provide valuable empirical insights into the efficiency, effectiveness and attractiveness of our solution, further guiding its refinement and development.