predicting maximum process temperature in cortical bone milling: an xgboost approach with sensitivity insights

Bone milling, a crucial biomechanical process in medical engineering, finds applications in dentistry, orthopedic surgery, and bone-related treatments. the utilization of computer numerical control (cnc) surgical mills has significantly enhanced this process, but it comes with challenges such as elevated temperatures that induce thermal necrosis in bone tissue. this study examines key inputs, including tool diameter, feed rate, rotational speed, and cutting depth, conducting a detailed experiment to predict maximum process temperature using the xgboost machine learning algorithm. the xgboost model consistently demonstrated exceptional predictive accuracy, yielding high determination coefficients of 0.99 in training and 0.94 in testing. accurate predictions were evident through close alignment between model-predicted and actual values, with mean absolute percentage error (mape) values of 0.33% and 3.38% for training and testing, respectively. rotational speed emerged as a critical factor, indicating a key point where temperature trends shift. higher speeds are correlated with lower temperatures due to enhanced chip removal and reduced bone heat conductivity. elevated feed rates were associated with increased bone temperature, emphasizing the intricate interplay between frictional forces and heat production. additionally, often-overlooked factors like cutting depth and tool diameter substantially influenced process temperature, impacting surgery recovery times. sobol sensitivity analysis identified cutting depth, rotational speed, tool diameter, and feed rate as primary factors influencing maximum process temperature fluctuations, with effectiveness percentages of 46.7%, 36%, 13.2%, and 4.1%, respectively. this comprehensive analysis sheds light on optimizing bone milling processes and mitigating thermal risks in medical applications.