Machining Stress Analysis and Deformation Prediction of Connecting Rod Based on FEM and GRNN

The unequal distribution of residual stress caused by incorrect machining processes will probably distort the machined connecting rods of the diesel engines. Thus, the influence of the machining processes on the residual stress and distortion of the rod need to be studied. Firstly, the finite element model is established based on the geometry of the connecting rod. Based on the life and death unit method, the influence of material removal on the residual stress of the connecting rod is discussed. Secondly, an orthogonal test is designed. The variables of the test include the quenching temperature, tempering temperature, cooling speed and the sequence of the machining processes. The deformation of the connecting rod is analyzed and then predicted based on the generalized regression neural network. The results show that the minimum distortion is achieved when the quenching temperature is 860 °C, the tempering temperature is 620 °C, the cooling speed is 50 °C h−1, and the machining sequence is “drilling the small end hole-milling, the upper planes-sawing, the connecting rod-turning, and the big end hole”. The priority of the influence of the processes on the deformation of the connecting rod is ranked as follows: cooling speed, quenching temperature, tempering temperature, and process sequence. The tensile residual stress is produced inside the connecting rod by machining processes. The generalized regression neural network predicts the connecting rod deformation accurately and the average error is 8.1%.