Large-diameter Thin-wall Tube Thinning Spinning Process Research And Wall Thickness Error Prediction

The large-diameter and ultra-thin shielding is kernel main pump's key component of the third generation nuclear power plant system, and its dimensional accuracy, mechanical properties and corrosion resistance of nuclear power plant directly affects the work efficiency and reactor safety of nuclear power plant. Therefore, studying the forming method of shielding and processing high-quality shielding that meets the size requirements to ensure the safety of the reactor and raise the reactor power is of great significance.Based on the power spinning technology of ultra-thin tube and the material properties of Nickel-based alloy C-276, this paper firstly established the finite element model of the heavy-caliber and ultra-thin tube and made use of MSC.marc to simulate the spinning process. And the simulation results revealed the law of wall thickness errors due to the distribution of stress and strain of the workpiece and the roller shape, thinning rate, feed rate, friction coefficient, the gap between workpiece and mandrel and other parameters. on the amount of errors of law of the wall thickness. Then used the orthogonal experimental design method to get parameters of primaries of the power spinning experiment.Combined with the numerical simulation and orthogonal experimental results, the Nickel-based alloy C-276 ultra-thin wall and large diameter tube power spinning experiments were carried out to obtain the cylindrical part with wall thickness of 0.42mm, diameter of 557mm, length of 1200mm, thickness error of less than±0.03mm, the surface roughness of 6-10μm, diameter error of±0.1mm, which was consistent with the demand accuracy. Meanwhile the tensile test, optical inspection and corrosion test were carried out with the workpiece, and it turned out that corrosion resistance and mechanical properties of the workpiece met the operating conditions of the shield.Finally this paper established a BP neural network model which used power spinning process parameters as input and the wall thickness error as output, the experimental data and numerical simulation data set were used to train the network to access the prediction error model of the Nickel-based alloy C-276 ultra-thin wall and large diameter tube power spinning experiments to realize the error prediction of the wall thickness, validated the model with high prediction accuracy by comparison with the experimental results.What this paper studied enriched the content of power spinning technology of the ultra-thin wall and large diameter tube and the establishment of wall thickness errors prediction model has a certain engineering applications value.