A Conceptual Costing Software Tool Based on Machine Learning for Turbomachines

The contemporary design landscape emphasizes the importance of product cost alongside performance, sustainability, and quality. Early determination of production costs, particularly in preliminary design phases like conceptual design, is crucial for maintaining competitiveness. Parametric cost estimation methods are preferred during these stages, leveraging identifiable relationships between design variables (cost drivers) and costs. Industry 4.0 innovations offer solutions to challenges posed by traditional Cost Estimation Relationship methods, with machine learning (ML) applications emerging as efficient tools for industrial processes. In the scientific literature, several approaches aim to create cost models using regression analysis, neural networks, and decision trees. The methods are implemented using data analysis tools that are not entirely suitable for product designers or cost engineers. The potential benefits of ML methods are not wholly exploited due to the limited effectiveness of the software tools. Addressing this gap, the paper introduces a software tool with an administrative module enabling cost engineers to develop parametric cost models using ML algorithms based on the CRISP-DM method. The tool facilitates data collection, model training, evaluation, and benchmarking. Cost models are stored in a database accessible to design engineers, allowing quick and accurate cost estimations by providing relevant cost drivers. Feature importance algorithms highlight critical cost drivers. The tool was tested in collaboration with a company that designs turbomachinery. During the experimentation, cost models were developed for one of the main rotor parts of an axial compressor (discs and spacer). The tool is a prototype to be further developed to manage an entire bill of materials, integrating risk management concepts.