Prediction Model of Net Cutting Specific Energy Based on Energy Flow in Milling

Net cutting specific energy (NCSE) reflects the actual cutting energy efficiency. Establishing a NCSE prediction model is helpful to analyze the energy consumption characteristics of machine tools. As so far, few studies have focused on the NCSE prediction in the way of energy flow. Therefore, based on the flow direction of cutting energy, a mathematical model for predicting NCSE is proposed in this paper. During milling, the energy used for cutting can be divided into forming surface energy, material removing energy and additional load energy. Thus, the NCSE model is decomposed into three sub-models. Firstly, in the cutting process of AISI 304 stainless steel, the phase transition of austenite to martensite on the machined surface is induced by parts of cutting energy, and then work hardening occur. Furthermore, the forming surface specific energy prediction model is established based on surface hardness. Secondly, the models of material removing energy and additional load energy are developed respectively with the material removal rate and spindle speed. The above sub-models are integrated into the NCSE prediction model with the determination coefficient R-2 of 0.982, and average prediction accuracy of 96.77%. Finally, the influence of input variables on NCSE and the energy consumption proportion are revealed. Among them, the forming surface specific energy, material removing specific energy and additional load specific energy account for 8.26%, 32.63% and 59.11% on NCSE respectively. The proposed model can not only predict the overall cutting energy consumption, but also predict the energy consumption of each sub-model. The research provides a new idea for analyzing cutting energy characteristics and improving processing theory.

Automated summary

This paper proposes a mathematical model for predicting NCSE based on the flow direction of cutting energy, decomposing the NCSE model into three sub-models for prediction, with an average prediction accuracy of 96.77%. The research reveals the influence of input variables on NCSE and the distribution of energy consumption proportions, providing a new idea for analyzing cutting energy characteristics and improving processing theory.