Multi-Node Thermal Network (13-state)

ADVANTAGES2 · dim 13

SolvSRK wins. At the comparison noise level, SolvSRK beats the best baseline by at least 10 percentage points of survival, or by at least 0.05 balanced score when survival is tied. Use SolvSRK for this class of problem. All verdicts →

13-state multi-node thermal network: battery, 4 motors, 4 ESCs, frame, payload temperatures plus cumulative heat. Motor winding tau ~5s vs battery tau ~500s → stiffness ~100:1. Forced convection from prop wash and airspeed.

Drone dynamics & autonomy

Problem definition

Incropera & DeWitt, 'Fundamentals of Heat and Mass Transfer'; Shahid et al., 'Thermal Analysis of Multirotor UAVs' (2022)

Canonical RHS excerpt from the registered callable used for this benchmark cell. Expand it to verify the state equations; it is not a standalone runnable fixture.

Show canonical RHS excerpt
def _mission_profile(t: float) -> tuple[float, float]:
    """Return (motor_current_per_motor_A, airspeed_ms) for each phase."""
    if t < 30.0:
        return 4.0, 0.5        # spin-up / hover
    elif t < 120.0:
        return 3.0, 8.0        # cruise
    elif t < 180.0:
        return 5.5, 2.0        # aggressive hover / filming
    elif t < 300.0:
        return 3.0, 8.0        # cruise return
    elif t < 360.0:
        return 3.5, 1.0        # hover descent
    else:
        return 1.0, 0.0        # idle

def rhs_thermal_envelope(t, y):
    """13-state multi-node thermal network for a quadrotor."""
    dy = np.zeros(13)

    T_batt = y[0]
    T_motors = y[1:5]
    T_escs = y[5:9]
    T_frame = y[9]
    T_payload = y[10]

    I_motor, airspeed = _mission_profile(t)
    I_pack = I_motor * 4

    # Convective enhancement from airspeed (forced convection scaling)
    conv_factor = 1.0 + 2.0 * math.sqrt(max(0, airspeed))

    # Battery heat generation (I^2 R)
    P_batt = I_pack**2 * _BATT_R_INTERNAL

    # Motor heat generation (copper losses, per motor)
    P_motors = np.array([I_motor**2 * _MOTOR_R_PHASE] * 4)
    # Add iron losses (proportional to speed, approximated)
    P_motors += 0.3 * I_motor

    # ESC heat generation (conduction + switching, per ESC)
    P_escs = np.array([I_motor**2 * _ESC_R_DS_ON * 2 + 0.1 * I_motor] * 4)

    # Battery thermal dynamics
    Q_batt_to_frame = (T_batt - T_frame) / _R_BATT_FRAME
    Q_batt_conv = _H_BATT_BASE * conv_factor * (T_batt - _T_AMB)
    dy[0] = (P_batt - Q_batt_to_frame - Q_batt_conv) / _C_BATT

    # Motor thermal dynamics (4 motors)
    for i in range(4):
        Q_motor_frame = (T_motors[i] - T_frame) / _R_MOTOR_FRAME
        Q_motor_conv = _H_MOTOR_BASE * conv_factor * (T_motors[i] - _T_AMB)
        dy[1 + i] = (P_motors[i] - Q_motor_frame - Q_motor_conv) / _C_MOTOR

    # ESC thermal dynamics (4 ESCs)
    for i in range(4):
        Q_esc_frame = (T_escs[i] - T_frame) / _R_ESC_FRAME
        Q_esc_conv = _H_ESC_BASE * conv_factor * (T_escs[i] - _T_AMB)
        dy[5 + i] = (P_escs[i] - Q_esc_frame - Q_esc_conv) / _C_ESC

    # Frame thermal dynamics (receives heat from all components)
    Q_in_frame = (
        (T_batt - T_frame) / _R_BATT_FRAME +
        sum((T_motors[i] - T_frame) / _R_MOTOR_FRAME for i in range(4)) +
        sum((T_escs[i] - T_frame) / _R_ESC_FRAME for i in range(4))
    )
    Q_frame_air = (T_frame - _T_AMB) / _R_FRAME_AIR
    Q_frame_payload = (T_frame - T_payload) / _R_PAYLOAD_FRAME
    dy[9] = (Q_in_frame - Q_frame_air - Q_frame_payload) / _C_FRAME

    # Payload thermal dynamics
    dy[10] = (T_frame - T_payload) / (_R_PAYLOAD_FRAME * _C_PAYLOAD)

    # Cumulative heat tracking
    dy[11] = P_batt
    dy[12] = P_batt + float(np.sum(P_motors)) + float(np.sum(P_escs))

    return dy
Parameters
  • _BATT_R_INTERNAL = 0.06
  • _C_BATT = 120
  • _C_ESC = 2
  • _C_FRAME = 50
  • _C_MOTOR = 8
  • _C_PAYLOAD = 15
  • _ESC_R_DS_ON = 0.008
  • _H_BATT_BASE = 0.1
  • _H_ESC_BASE = 0.3
  • _H_MOTOR_BASE = 0.5
  • _MOTOR_R_PHASE = 0.15
  • _R_BATT_FRAME = 8
  • _R_ESC_FRAME = 6
  • _R_FRAME_AIR = 15
  • _R_MOTOR_FRAME = 4
  • _R_PAYLOAD_FRAME = 10
  • _T_AMB = 25
Initial condition
y(0) = [25, 25, 25, 25, 25, 25, …] [shape=(13,), min=0, max=25]
Horizon
t ∈ [0, 600]

Canonical RHS excerpt captured from the same registered callable used for the published benchmark. Frozen closure values are summarized below; helper imports and solver settings are intentionally omitted.

Fingerprint

Spread: low

Default noise: low

Recommendation snapshot

Clean best: SolvSRK

Noisy best: SolvSRK

Coverage

14 solver arms · clean + 5 noise levels

Ranked on survival, precision, and speed

Versions & freeze

Methodology →
Freeze
2026-08-13
libsolvsrk
2.3.0
SciPy
1.14
SUNDIALS
CVODE (bundled backend)

20 seeds/cell default · 14 arms · TRL 4–5 · simulation-lab validated · this page: Multi-Node Thermal Network (13-state) (multi-node-thermal-network-13-state)

Governed SolvTune benchmark freeze; per-arm medians only. RHS definitions and raw trial rows are not published.

Self-reported by Resonix Labs · not independently verified

Results matrix

Pick an objective and a noise level to rank all arms on survival, median SCD, median nfev, and median wall time. Medians across seeds.

Objective

Best overall trade-off of survival, precision, and speed.

Noise level

#SolverSurvivalSCDnfevWallScore
1SolvSRK
100%
12.14,41070 ms0.907
2SciPy RadauSciPy
100%
8.95,571118 ms0.831
3SciPy LSODASciPy
100%
8.01,68020 ms0.810
4SciPy DOP853SciPy
100%
7.93,65074 ms0.806
5SciPy BDFSciPy
100%
7.62,13962 ms0.801
6CVODE BDFexternal
100%
7.31,49337 ms0.793
7CVODE Adamsexternal
100%
7.21,21531 ms0.791
8SciPy RK23SciPy
100%
7.03,47982 ms0.785
9SciPy RK45SciPy
100%
6.62,43230 ms0.776
10Tsit5external
100%
6.23,024673 ms0.767

At Clean, best balanced arm is SolvSRK.

Values are medians across seeds, measured by Resonix Labs on Resonix hardware and not independently verified; nfev and wall are on reference lab hardware (indicative). Under injected noise only SolvSRK and the SciPy arms are run. How we measure accuracy → · Verification status →

SolvScout · free

Profile your problem for free

This page shows one published benchmark cell. SolvScout fingerprints your ODE, compares it to the full corpus, and recommends a solver with the same survival / precision / speed ranking you see here — including when a SciPy arm wins.

SolvSRK · 30-day trial

Run the winner on your machine

SolvSRK is the stiffness-adaptive integrator behind the SolvSRK column in these tables. Create an account, activate a machine, and take a 30-day trial — same binary you'd ship after purchase.

Cite this page

Replace the access date. Pin the freeze ID and library versions when comparing against a later export. Cite it as what it is — a self-reported vendor benchmark, not an independently verified result. The note field says so; please keep it.

@misc{resonix_evidence_multi_node_thermal_network_13_state_2026,
  title        = {Resonix Evidence Portal: Multi-Node Thermal Network (13-state)},
  author       = {{Resonix Labs (Canada) Inc.}},
  year         = {2026},
  howpublished = {\url{https://resonix.tech/evidence/problems/multi-node-thermal-network-13-state}},
  note         = {Self-reported vendor benchmark; internally generated by Resonix Labs and not independently verified. Accessed YYYY-MM-DD. Freeze 2026-08-13; libsolvsrk 2.3.0; SciPy 1.14.}
}

Related

TRL 4–5 · simulation-lab validated · 398 problems · 14 solver arms · clean + 5 noise levels

Freeze: 2026-08-13 · scipy 1.14 · libsolvsrk 2.3.0 · Methodology

Self-reported by Resonix Labs · not independently verified · Verification status