Research on robot grinding material removal rate prediction and model optimization based on NRBO-XGboost

Aiming at the problem that the material removal rate (MRR) is difficult to accurately predict during the grinding process, an MRR prediction model based on the extreme gradient boosting (XGBoost) algorithm was constructed. By analyzing the vibration signals generated during robot grinding under different process parameters, the characteristic parameters that can reflect MRR are extracted. Utilizing a Newton-Raphson-based optimizer (NRBO), the initial parameters of the XGBoost model are optimized to address parameter adjustment and resolve local optimization issues, ultimately constructing an NRBO-XGBoost model. Comparing this model with traditional XGBoost, particle swarm optimization (PSO)-XGBoost and long short-term memory (LSTM) algorithms, the results show that the prediction results of the NRBO-XGBoost model are better than other models, the prediction evaluation index is significantly improved, and the initial solution optimization speed is faster than the PSO optimization algorithm. The fitness value is smaller. The research results can provide reference for robot grinding MRR prediction and process optimization.