Li Zhanfu¹ᵃ^，¹ᵇ，Dai Kewo¹ᵃ，Gao Zisheng²，Huang Qingfang²

1a.School of Mechanical & Automotive Engineering，1b. Fujian Key Laboratory of Intelligent Processing Technology and Equipment，

Fujian University of Technology，Fuzhou 350118， China；2. Fujian Golden Dragon Rare⁃earth Co.， Ltd.， Longyan 366300， China

Abstract

Objective Reliable discrete element method （DEM） analysis of neodymium iron boron （NdFeB） alloy powders requires flow-property measurements that reflect their real and highly reactive state. However， because exposure to air is restricted for NdFeB alloy powder， flowability is often measured after oxidation， which limits the reliability of DEM parameter calibration. In this study， the flowability of NdFeB alloy powder is measured under an inert gas atmosphere for the first time， and a DEM calibration method is established to obtain accurate and transferable parameters for subsequent numerical analyses.

Methods Flow-property characterization was carried out under an inert gas atmosphere to avoid oxidation， ignition， and property distortion during handling. Key physical properties required for DEM modeling， including density and elastic-related parameters， were determined through experiments， and particle morphology was analyzed to support geometric representation in simulations. To balance computational efficiency and shape fidelity， a coarse-grained particle model with an equivalent size of 8 µm was adopted.Three representative particle morphologies—triangular platelet， rectangular platelet，and rhombohedral particle—were introduced in equal proportions. To captureshort-range adhesion induced by van der Waals

attraction and the interlocking effects among micron-sized particles， the Hertz-Mindlin with Johnson–Kendall–Roberts（JKR） contact model was used. DEM simulations of angle-of-repose measurementswere established to reproduce powder discharge， pile formation， and contour extraction. Parameter calibration was then conducted using response surface methodology. A Plackett–Burman design was first used to identify significant contact parameters from the candidate variables， including restitution coefficients， static and rolling friction coefficients， and surface energy. Steepest ascent tests were subsequently employed to locate the optimal parameter interval， after which a Box-Behnken design was used to establish a second-order regression model between the significant parameters and the simulated angle of repose. Finally， the measured angle of repose obtained under an inert atmosphere was taken as the target response， and the optimal DEM parameter combination was determined by minimizing the deviationbetween simulated and experimental results.

Results and Discussion Angle-of-repose tests on ten batches of NdFeB alloy powder showed an average value of 52.45°，with relatively low variability and good repeatability， indicating that measurements under inert conditionsprovided a reliable basis for DEM calibration. The Plackett-Burman screening results demonstrated that the particle-particle coefficient of rolling friction （C ）and particle-stainlesssteel coefficient of static friction （E ）were statistically significant factors affecting flowability.Althoughthe particle-particle coefficient of static friction （B） was only marginally significant，it was retained because of its clear physical relevance to interparticle resistance. On this basis， these three parameters were selected for further optimization. Steepest ascent experiments showed that the optimal region was located between the fourth and fifth test groups， which effectively narrowed the search range for response surface modeling. The subsequent Box-Behnken analysis indicated that the quadratic regression model was highly significant with no significant lack of fit， confirming that the model could adequately describe the relationship between contact parameters and the angle of repose. The analysis further showed that parameters of B， C， and E， as well as the interaction terms （BE and CE） and quadratic terms （B² and C²）， had significant effectsonthe response. Increasing B and C enhanced resistance to sliding and rolling， leading to a more stable particle-force-chain structure and a higher angle of repose. In addition， increasing E improved basal stability and reduced particle slip on the contact surface， enhancing pile stability. The optimized parameter set was determined to be 0.428 for B，0.13 for C， and 0.258 for E. Validation simulations using this parameter set produced an average angle of repose of 52.8°，with a relative error of only 0.66% compared with the experimental value， demonstrating that the calibrated model could reproduce the actual flow behavior of NdFeB alloy powder with high accuracy under inert-atmosphere conditions.

Conclusion This study establishes an inert-atmosphere-based DEM calibration method for highly reactive NdFeB alloy powder and overcomes the limitation of conventional calibration approaches that rely on measurements performed after powder oxidation in air. By combining experimental flowability testing with Plackett-Burman screening， steepest ascent localization， and Box-Behnken quadratic regression， a reliable set of key contact parameters is obtained for the JKR-based DEM model. The calibrated model reproducesexperimentalresults with high accuracy and provides a reliable foundation for subsequent DEM simulations and optimization of NdFeB alloy powder processes， including handling， filling， molding，and related powder-metallurgy operations.

Keywords： NdFeB alloy； powder metallurgy； discrete element method； response surface methodology； parameter

Get Citation：Li Zhanfu， Dai Kewo， Gao Zisheng， et al. Parameter calibration of discrete element method for NdFeB alloy particle flowability based on JKR model［J］. China Powder Science and Technology， 2026， 32（6）： 1-11.

Received： 2026-01-16， Revised ： 2026-04-11，Online：2026-05-11.

Funding： This research was supported by the National Key R&D Program of China (Grant No.2022YFB3504602); the University-Industry Collaboration Project of the Science and Technology Office of Fujian Province (Grant No.2024H6015); the Key Technical Innovation and Industrialization Project from the Fujian Provincial Department of Education and the Provincial Department of Industry and Information Technology (Grant No. 2023XQ002); and the Project of Fujian Provincial Department of Finance(Grant No. GY-Z24187).

DOI：10.13732/j.issn.1008-5548.2026.06.010

CLC No.：TB302；TF122；TB4

Type Code：A

Serial No.：1008-5548（2026）06-0001-11