Simulation Based Fault Detection and Diagnosis for Additive Manufacturing

Objectives: To develop a mechanism for error detection and diagnosis in additive manufacturing system by analysing generated data for reducing manual efforts. Methods/Analysis: Data generated from the actual run is more and currently this data is analysed manually for error detection and diagnosis. We proposed developed a software tool which will help for sorting, filtering and analysing data for error detection and these data are fed to Bayesian network which help to find the cause of that error and help for error diagnosis. Bayesian network with probability reasoning is used and it requires historical data and experience person for efficient prediction. Findings: This mechanism is efficient compared to manual system. It takes lesser time to sort, filter and analyse the data for error detection. We get the idea about scenario and parameters which causes the error; hence we can limit those parameters and can resolve those issues quickly. As this mechanism helps to set a proper relationship between cause and error, hence it reduces the uncertainty of the system in fault finding. As this is automated process hence saves too much time of error detection and diagnosis, hence this proposal will speed up the fault detection and diagnosis and enable the developer to solve issue quicker and indirectly help to optimize the process and cost. Novelty: Statistical tool used for this is a Bayesian network with Probability Reasoning which has ability to infer and self-learning.