Blockchain represents a trusted source of data for process mining where the benefit is mutual since process mining provides blockchain applications with helpful insights on user behavior, performance and security. However, blockchain data is not adequate for process mining and this issue has been studied in several works. The focus of existing works was on extracting event logs from activity-centric applications, excluding blockchain artifact-centric applications which are more common. To address this limitation, we propose an approach for extracting blockchain data from artifact-centric applications in a format that is suitable for process mining techniques. To do so, we introduce ACEL (Artifact-Centric Event Logs), an extension of the OCEL (Object-Centric Event Logs) standard for artifact-centric processes. We also propose an algorithm to extract ACEL logs from Ethereum, a blockchain platform. We implemented the approach and compared its results with related standards using as case study a popular Ethereum decentralized application.