Developed by foundries for foundries as part of a consortium, DEROC is self-learning process optimisation software. It combines the latest in data analysis tools with cutting edge artificial intelligence to convert production data into a simple process optimisation report. This allows process operators to quickly identify the most probable causes of wasted material, energy and time, and take prompt, well informed corrective action.
A DEROC analysis relies on the concepts of Responses, Factors and Levels. The goal of process optimisation is to improve the process in terms of a response in the system. The response is the resulting behaviour in the process, which we wish to alter or optimise. This is achieved by finding the optimal settings (levels) for the process parameters (factors). DEROC uses similar terminology to DoE, but this optimisation software is sensitive enough to work with up to five levels for each factor, distributed within the upper and lower limits of process specification. This gives the reports generated by DEROC, the ability to focus on the fine-tuning of process parameters as well as defect/waste reduction.
The software works with process data to search out correlations between different process parameters (factors) and the quality of the end product (responses). This is done by optimising the shape of the knowledge Hyper-surface constructed from the factor-response data.
The paper will first explain the unique features of a DEORC analysis and then presents a case study undertaken with Rolls-Royce Plc as part of our recent research project.
Keywords: Process Optimisation, Design of Experiments (DoE), Receiver Operating Characteristics, Six Sigma, X1Recall Process Improvement Solutions, DEROC Process Optimisation Reports.