Condition-based monitoring techniques and algorithms in 3d printing and additive manufacturing: a state-of-the-art review

Additive manufacturing (AM) has revolutionized the manufacturing processes across various fields like aerospace, bio technology, etc. AM is now gaining popularity in industries for mass scale production for parts due to its numerous advantages over subtractive manufacturing techniques. AM gives freedom to designers and manufacturers for producing complex parts with extremely high precision. Despite its advantages, AM production techniques come with several issues especially when it is shifted from prototyping to mass production. To detect the fault in printed part, post-print analysis techniques are not that effective, because mechanical properties change during layer-by-layer construction of parts as they are direction dependent. Additionally, most parts have hollow sections which are not easy accessible for inspection. To resolve this, real-time condition-based monitoring techniques are developed to monitor the health of the part that is being printed by monitoring the parameters of the machine. These are then fed to machine learning algorithms that identify the issue and perform correction in real time to produce the part with no defect. This research paper reviews the most used AM techniques and condition-based monitoring techniques and algorithms that are most suitable for each additive manufacturing technique. A comprehensive literature view was conducted in which various solid, liquid, and powder-based additive manufacturing techniques were discussed in detail. Various real-time and post-print CBM techniques were discussed in detail like acoustic monitoring, vibration monitoring, ultrasound monitoring, etc. and which one is more suitable for each AM process. Data acquisition and data processing methods were discussed in detail, and then, algorithms and predictive models were reviewed in detail and which ones are suitable for application in real-time condition-based monitoring for AM processes. This review study is extremely significant as there are very little data present in which all these AM processes and their relevant CBM techniques are discussed in such detail and furthermore types of algorithm and predictive models that are suitable for each AM process. This review paper is also very useful for industrialists and designers that are in prototyping and mass production phase and want to review the options that which CBM technique to be used for monitoring real-time AM process and which data acquisition and data processing technique to be used for real-time monitoring and which type of algorithm will give best results in real-time monitoring of the process.