Skip to content
TheChatArchiveThe Chat Archive
Claudeclaude-opus-4-20250514-thinking-16k· AnthropicHugging Face LMArena Human Preference 140K

Explain Discrete Element Method, complete with all the mathematical formulas used to…

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with chatgpt-4o-latest-20250326’s answer.
  1. User

    Explain Discrete Element Method, complete with all the mathematical formulas used to describe it.

  2. Claude Assistant

    The Discrete Element Method (DEM)

    The Discrete Element Method (DEM) is a numerical technique used to simulate the motion and interactions of a large number of particles. It's particularly useful for modeling granular materials, powders, rocks, soils, and other discontinuous media. DEM tracks the motion of individual particles and their interactions through contact forces.

    1. Fundamental Principles

    DEM is based on Newton's second law of motion applied to each particle in the system. The method involves:

    • Tracking individual particle positions and velocities
    • Detecting contacts between particles
    • Calculating contact forces
    • Integrating equations of motion
    • Updating particle positions

    2. Equations of Motion

    For each particle i, the translational and rotational motion are governed by:

    Translational Motion: mid2xidt2=∑jFijc+Fig+Fiextm_i \frac{d^2\mathbf{x}_i}{dt^2} = \sum_{j} \mathbf{F}_{ij}^c + \mathbf{F}_i^g + \mathbf{F}_i^{ext}

    Rotational Motion: Iidωidt=∑jMijI_i \frac{d\boldsymbol{\omega}_i}{dt} = \sum_{j} \mathbf{M}_{ij}

    Where:

    • mim_i = mass of particle i
    • xi\mathbf{x}_i = position vector of particle i
    • Fijc\mathbf{F}_{ij}^c = contact force between particles i and j
    • Fig\mathbf{F}_i^g = gravitational force on particle i
    • Fiext\mathbf{F}_i^{ext} = external forces on particle i
    • IiI_i = moment of inertia of particle i
    • ωi\boldsymbol{\omega}_i = angular velocity of particle i
    • Mij\mathbf{M}_{ij} = torque on particle i due to contact with particle j

    3. Contact Force Models

    The most common contact model is the Hertz-Mindlin model with viscous damping:

    Normal Contact Force:

    Fn=−knδn3/2n−ηnvnn\mathbf{F}_n = -k_n \delta_n^{3/2} \mathbf{n} - \eta_n v_n \mathbf{n}

    Where:

    • knk_n = normal stiffness coefficient
    • δn\delta_n = normal overlap distance
    • n\mathbf{n} = unit normal vector at contact
    • ηn\eta_n = normal damping coefficient
    • vnv_n = relative normal velocity

    The normal stiffness for Hertz contact: kn=43E∗R∗δnk_n = \frac{4}{3} E^* \sqrt{R^* \delta_n}

    Where: 1E∗=1−ν12E1+1−ν22E2\frac{1}{E^*} = \frac{1-\nu_1^2}{E_1} + \frac{1-\nu_2^2}{E_2} 1R∗=1R1+1R2\frac{1}{R^*} = \frac{1}{R_1} + \frac{1}{R_2}

    • E1,E2E_1, E_2 = Young's moduli of particles 1 and 2
    • ν1,ν2\nu_1, \nu_2 = Poisson's ratios
    • R1,R2R_1, R_2 = radii of particles

    Tangential Contact Force:

    Ft=−ktδt−ηtvt\mathbf{F}_t = -k_t \boldsymbol{\delta}_t - \eta_t \mathbf{v}_t

    Subject to Coulomb's friction law: ∣Ft∣≤μ∣Fn∣|\mathbf{F}_t| \leq \mu |\mathbf{F}_n|

    Where:

    • ktk_t = tangential stiffness
    • δt\boldsymbol{\delta}_t = tangential displacement
    • ηt\eta_t = tangential damping coefficient
    • vt\mathbf{v}_t = relative tangential velocity
    • μ\mu = coefficient of friction

    4. Contact Detection

    For spherical particles, contact occurs when: dij<Ri+Rjd_{ij} < R_i + R_j

    Where:

    • dij=∣xi−xj∣d_{ij} = |\mathbf{x}_i - \mathbf{x}_j| = distance between particle centers
    • Ri,RjR_i, R_j = radii of particles i and j

    The overlap is: δn=Ri+Rj−dij\delta_n = R_i + R_j - d_{ij}

    5. Damping Coefficients

    The damping coefficients are related to the coefficient of restitution ee:

    ηn=−2m∗knln⁡(e)π2+ln⁡2(e)\eta_n = -2\sqrt{m^* k_n} \frac{\ln(e)}{\sqrt{\pi^2 + \ln^2(e)}}

    Where m∗m^* is the reduced mass: 1m∗=1m1+1m2\frac{1}{m^*} = \frac{1}{m_1} + \frac{1}{m_2}

    6. Time Integration

    The most common integration scheme is the Velocity Verlet algorithm:

    Position Update: xi(t+Δt)=xi(t)+vi(t)Δt+12ai(t)Δt2\mathbf{x}_i(t + \Delta t) = \mathbf{x}_i(t) + \mathbf{v}_i(t)\Delta t + \frac{1}{2}\mathbf{a}_i(t)\Delta t^2

    Velocity Half-Step: vi(t+Δt2)=vi(t)+12ai(t)Δt\mathbf{v}_i(t + \frac{\Delta t}{2}) = \mathbf{v}_i(t) + \frac{1}{2}\mathbf{a}_i(t)\Delta t

    Acceleration Update: ai(t+Δt)=Fi(t+Δt)mi\mathbf{a}_i(t + \Delta t) = \frac{\mathbf{F}_i(t + \Delta t)}{m_i}

    Velocity Update: vi(t+Δt)=vi(t+Δt2)+12ai(t+Δt)Δt\mathbf{v}_i(t + \Delta t) = \mathbf{v}_i(t + \frac{\Delta t}{2}) + \frac{1}{2}\mathbf{a}_i(t + \Delta t)\Delta t

    7. Critical Time Step

    The time step must be small enough to capture the dynamics accurately. For elastic contacts:

    Δtcrit=2m∗kn\Delta t_{crit} = 2\sqrt{\frac{m^*}{k_n}}

    Typically, the actual time step is chosen as: Δt=α⋅Δtcrit\Delta t = \alpha \cdot \Delta t_{crit}

    Where α≈0.1−0.3\alpha \approx 0.1-0.3 for stability.

    8. Rolling Resistance

    To account for rolling resistance, an additional torque is applied:

    Mr=−krθr−ηrωrel\mathbf{M}_r = -k_r \boldsymbol{\theta}_r - \eta_r \boldsymbol{\omega}_{rel}

    Subject to: ∣Mr∣≤μrR∗∣Fn∣|\mathbf{M}_r| \leq \mu_r R^* |\mathbf{F}_n|

    Where:

    • krk_r = rolling stiffness
    • θr\boldsymbol{\theta}_r = relative angular displacement
    • ηr\eta_r = rolling damping coefficient
    • μr\mu_r = coefficient of rolling friction

    9. Energy Dissipation

    The energy dissipated in a collision can be calculated as:

    Ediss=∫0δmaxFndδ−∫δmax0FndδE_{diss} = \int_0^{\delta_{max}} F_n d\delta - \int_{\delta_{max}}^0 F_n d\delta

    10. Non-Spherical Particles

    For non-spherical particles, additional considerations include:

    • More complex contact detection algorithms
    • Orientation tracking using quaternions
    • Modified moment of inertia tensors

    The orientation is updated using: dqdt=12ω⊗q\frac{d\mathbf{q}}{dt} = \frac{1}{2}\boldsymbol{\omega} \otimes \mathbf{q}

    Where q\mathbf{q} is the quaternion representing particle orientation.

    11. Cohesive Forces

    For cohesive materials, additional attractive forces are included:

    Fcoh=−kcohAn\mathbf{F}_{coh} = -k_{coh} A \mathbf{n}

    Where:

    • kcohk_{coh} = cohesion strength parameter
    • AA = contact area

    12. Implementation Algorithm

    The typical DEM algorithm follows these steps:

    1. Initialize particle positions, velocities, and properties
    2. For each time step:
      • Detect particle contacts
      • Calculate contact forces for each contact
      • Sum forces on each particle
      • Update particle velocities and positions
      • Update particle orientations and angular velocities
      • Apply boundary conditions
      • Output results

    Applications

    DEM is widely used in:

    • Geotechnical engineering (soil mechanics, rock mechanics)
    • Pharmaceutical industry (tablet coating, mixing)
    • Mining and mineral processing
    • Chemical engineering (fluidized beds, hoppers)
    • Agriculture (grain storage and handling)
    • Additive manufacturing (powder bed processes)

    This comprehensive mathematical framework allows DEM to accurately simulate complex particulate systems, though computational cost increases significantly with particle number, making it suitable primarily for systems with thousands to millions of particles rather than truly macroscopic systems.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.