Chinese Journal of Lasers, Volume. 47, Issue 1, 0102004(2020)

Multi-Objective Optimization of Coaxial Powder Feeding Laser Cladding Based on NSGA-II

Kai Zhao*, Xudong Liang*, Wei Wang, Ping Yang, Yunbo Hao, and Zhongliang Zhu
Author Affiliations
  • Shanghai Aerospace Equipments Manufacturer Co., Ltd., Shanghai 200245, China
  • show less
    Figures & Tables(28)
    Laser cladding schematic
    Surface damaged steam turbine seat
    SEM image of spherical Inconel625 powder
    Common defects in single-pass cladding layer
    Dimensional parameters of cross section of single-pass cladding layer
    Diagram of microhardness measurement for single-pass cladding layer
    Effect of cross section of cladding layer after binarization processing
    Dilution rate measurement diagram
    Structure of BP neural network model
    Prediction results of BP neural network. (a) Microhardness; (b) HAZ depth; (c) dilution rate; (d) porosity
    Cross-section morphology of single-passcladding layer
    Comparison of prediction error of dilution rate
    Comparison of prediction error of microhardness
    Schematic of NSGA-II algorithm
    Optimized Pareto frontier solution set
    Test piece processed with optimized parameters
    Test piece after surface hardness test
    Path planning of part to be repaired
    Repaired steam turbine seat
    Surface morphology of repaired seat
    • Table 1. Performances of matrix and repaired materials

      View table

      Table 1. Performances of matrix and repaired materials

      MaterialMelting point /℃Density /(g·cm-3)Tensilestrength /MPaYieldstrength /MPaBrinellhardness /HB
      Steel 201398-14547.93410245156
      Inconel6251290-13508.4827414220
    • Table 2. Main compositions of matrix and repaired materials

      View table

      Table 2. Main compositions of matrix and repaired materials

      MaterialMass fraction /%
      CrMoCSiMnNbFe
      Steel 200.2580.17-0.240.17-0.370.70-1.003.15Bal.
      Inconel625238---4.155
    • Table 3. Cladding parameters

      View table

      Table 3. Cladding parameters

      ParameterContent
      Distance from nozzle to substrate /mm12
      Laser beam diameter at substrate /mm2.5
      Shielding gasArgon
      Powder transport gasArgon
      Powder transport gas flow rate /(m3·h-1)0.425
    • Table 4. Process parameters and corresponding coded values in CCD experiments

      View table

      Table 4. Process parameters and corresponding coded values in CCD experiments

      Parameterα=-1.5α=-1α=0α=1α=+1.5
      Laser power /W8001100140017002000
      Scanning speed /(mm·s-1)89.51112.514
      Powder flow rate /(g·min-1)10.0811.3412.613.8615.12
    • Table 5. Dilution rate predicted based on multiple regression and neural network

      View table

      Table 5. Dilution rate predicted based on multiple regression and neural network

      Single-passexperiment No.MeasuredvaluePrediction result
      MultipleregressionNeuralnetwork
      10.5200.580.54
      20.6400.600.63
      30.6010.570.63
      40.4560.490.48
      50.1500.090.14
      60.0800.070.09
    • Table 5. Variables and corresponding measured results in CCD experimental design

      View table

      Table 5. Variables and corresponding measured results in CCD experimental design

      CCDExperiment No.ParameterResponse
      Laser power /WScanning rate /(mm·s-1)Feed-powder speed/(g·min-1)Section area /mm2DilutionMicrohardness/HVHAZ depth /mmPorosity /%
      18008.015.120.880.05259.930.452.30
      28008.010.080.640.06262.670.510.27
      380014.010.080.280.34206.100.531.70
      480014.015.120.360.11258.070.470.39
      5110011.012.6.01.000.28244.030.621.17
      6140012.512.601.360.50201.230.840.36
      714009.512.601.920.44209.170.900.08
      8140011.011.341.480.50200.630.860.21
      9140011.013.861.600.45208.770.840.20
      10140011.012.601.580.48197.970.880.14
      11140011.012.601.560.50192.830.890.03
      12140011.012.601.600.47192.930.890.15
      13140011.012.601.600.46197.670.860.04
      14140011.012.601.640.45199.170.850.24
      15140011.012.601.540.47197.530.850.20
      16170011.012.602.000.55186.831.061.48
      1720008.015.123.640.52186.171.400.97
      1820008.010.083.320.63170.431.571.59
      19200014.010.082.000.68160.371.221.01
      20200014.015.122.000.58178.601.111.29
    • Table 6. Microhardness predicted based on multiple regression and neural network

      View table

      Table 6. Microhardness predicted based on multiple regression and neural network

      Single-passexperiment No.MeasuredvaluePrediction result
      MultipleregressionNeuralnetwork
      1186.43192.15184.67
      2163.70154.53167.43
      3177.70163.65171.50
      4207.47202.42206.94
      5234.00240.04257.35
      6250.47255.44258.08
    • Table 7. Comparison of response values before and after optimization

      View table

      Table 7. Comparison of response values before and after optimization

      ResponseDilutionHAZdepth /mmMicroharness /HVEfficiency /(mm2·s-1)
      Before0.5180.855186.43315.24
      After0.3200.736218.33716.17
    Tools

    Get Citation

    Copy Citation Text

    Kai Zhao, Xudong Liang, Wei Wang, Ping Yang, Yunbo Hao, Zhongliang Zhu. Multi-Objective Optimization of Coaxial Powder Feeding Laser Cladding Based on NSGA-II[J]. Chinese Journal of Lasers, 2020, 47(1): 0102004

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: laser manufacturing

    Received: Jul. 29, 2019

    Accepted: Sep. 26, 2019

    Published Online: Jan. 9, 2020

    The Author Email: Kai Zhao (zkdlut@163.com), Xudong Liang (zkdlut@163.com)

    DOI:10.3788/CJL202047.0102004

    Topics