Session: Government Agency Student Posters

Paper Number: 173295

A Blockchain Enabled Privacy-Preserving Framework for Cross-Industrial Data Collaboration in Smart Manufacturing 

The advancement of smart manufacturing technologies has made operational data more accessible and readily available across manufacturing processes. Sensors, connected equipment, and industrial IoT devices are continuously generating large volumes of time-series and condition-monitoring data that can be leveraged for intelligent decision making. While this data holds potential for process optimization through Machine Learning (ML) and advanced analytics, effective utilization often requires aggregating data across multiple business units, organizations, and supply networks. However, concerns regarding data privacy, ownership, competitive advantage, and trust have significantly hindered such collaborative efforts, especially in environments where data sensitivity and intellectual property protection are critical.

In this work, we proposed a blockchain-enabled, privacy-preserving data framework that facilitates secure and auditable exchange of encrypted sensor data among manufacturing stakeholders, enabling them to collaborate without compromising proprietary information. By integrating a permissioned blockchain network with the Interplanetary File System (IPFS) for decentralized data storage, the framework ensures both traceability and data integrity. Smart contracts embedded within the blockchain are employed to enforce data-sharing policies, ensure compliance with ownership rights, and enable transparent access control mechanisms. Furthermore, an off-chain analytics agent is introduced in our framework that retrieves encrypted data via Content Identifiers (CIDs) from IPFS. This agent trains a ML model, which is then stored on the blockchain. The trained model parameters are accessible to all participating organizations, enabling them to benefit from globally learned insights without directly sharing sensitive raw data. This approach not only preserves data confidentiality but also enhances the accuracy and generalizability of predictive models by leveraging diverse data sources.

A prototype was implemented within a simulated multi-organizational manufacturing network to validate the feasibility of the proposed framework. The Hyperledger Fabric network was configured with multiple peer organizations and a single channel to support data visibility policies. Sample sensor data was encrypted and uploaded to IPFS, with metadata recorded on the blockchain ledger. Preliminary tests showed that data could be securely shared and retrieved using smart contracts, and ML models trained on aggregated encrypted data achieved high accuracy in prediction.

In conclusion, the framework demonstrates a practical and scalable solution for enabling secure, collaborative analytics in cross-industrial smart manufacturing. By addressing key challenges such as data sovereignty, trust management, and privacy-preserving model training, this work contributes to the advancement of decentralized industrial intelligence. The proposed framework enables collaborative approaches to predictive maintenance, supports joint quality assurance efforts, and drives digital transformation across interconnected manufacturing systems.

Presenting Author: Md Irfan Khan University of South Carolina

Presenting Author Biography: I am a Graduate Research Assistant and Ph.D. student at the University of South Carolina, specializing in smart manufacturing. My research focuses on developing secure, data-driven infrastructures for industrial collaboration, integrating technologies like blockchain and federated learning for predictive maintenance and decentralized analytics.

Authors:

Md Irfan Khan University of South Carolina
Mojtaba A. Farahani University of South Carolina
Thorsten Wuest University of South Carolina