Overview Evidence tour Methodology Dead zones Problems

HEL Thermal Kill (30kW.PL.800)

PARITYS2 · dim 8

No clear winner. The survival gap is under 10 percentage points and the balanced-score gap is under 0.05, so neither SolvSRK nor the best baseline clears the win threshold. Either works - choose on cost, licensing, or integration effort. All verdicts →

HEL Thermal Kill (30kW.PL.800) benchmark in the defense-autonomy domain.

Defense autonomy

Problem definition

Canonical benchmark implementation

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 rhs(t, y):
    T_tgt = max(y[0], T_ambient)
    T_bat = max(y[1], T_ambient)
    T_pc  = max(y[2], T_ambient)
    sig_d = y[3]
    R     = max(y[5], 1.0)
    eps   = max(abs(y[7]), 1e-10)

    d = np.empty(8)

    # Atmospheric transmission (Beer-Lambert)
    tau_atm = np.exp(-beta_ext * R / 1000.0)

    # Tracking efficiency (Gaussian beam, fraction on target)
    spot_var = 2.0 * (eps * R) ** 2
    eta_track = 1.0 - np.exp(-A_target / max(spot_var, 1e-20))

    P_beam = P_laser * tau_atm * eta_track

    # Target surface heating
    Q_laser = alpha_abs * P_beam
    Q_conv = h_conv * A_t * (T_tgt - T_ambient)
    Q_rad = epsilon_t * _SIGMA_SB * A_t * (T_tgt**4 - T_ambient**4)
    d[0] = (Q_laser - Q_conv - Q_rad - k_cond_bt * (T_tgt - T_bat) - k_cond_pt * (T_tgt - T_pc)) / C_t

    # Battery heating (conduction from target + self-heating)
    d[1] = (k_cond_bt * (T_tgt - T_bat) + I_batt**2 * R_internal - h_b * (T_bat - T_ambient)) / C_b

    # PCB heating
    d[2] = (k_cond_pt * (T_tgt - T_pc) + P_electronics - h_pcb * (T_pc - T_ambient)) / C_pcb

    # Arrhenius damage accumulation
    d[3] = A_arrhenius * np.exp(-E_activation / (_K_B * max(T_tgt, 200.0)))

    # Beam power state (for diagnostics)
    d[4] = (P_beam - y[4]) / 0.01

    # Range decreasing
    d[5] = -V_close

    # Track angle rate (simplified LOS)
    d[6] = V_close * 0.01 / max(R, 1.0)

    # Tracking error settling
    d[7] = -(y[7] - sigma_jitter * 0.5) / tau_track

    return d

Parameters: A_arrhenius = 1e+08
A_t = 0.02
A_target = 0.05
C_b = 200
C_pcb = 80
C_t = 500
E_activation = 1.2e-19
I_batt = 20
P_electronics = 15
P_laser = 30000
R_internal = 0.05
T_ambient = 293
V_close = 20
_K_B = 1.38e-23
_SIGMA_SB = 5.67e-08
alpha_abs = 0.9
beta_ext = 0.3
epsilon_t = 0.9
h_b = 10
h_conv = 25
h_pcb = 8
k_cond_bt = 5
k_cond_pt = 3
sigma_jitter = 2e-05
tau_track = 0.1
Initial condition: y(0) = [293, 293, 293, 0, 0, 800, 0, 2e-05]
Horizon: t ∈ [0, 10]

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: high

Default noise: high

Recommendation snapshot

Clean best: SciPy Radau

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: HEL Thermal Kill (30kW.PL.800) (hel-thermal-kill-30kw-pl-800)

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

#	Solver	Survival	SCD	nfev	Wall	Score
1	SciPy RadauSciPy	100%	13.8	2,279	56 ms	0.947
2	SolvSRK	100%	12.5	1,484	27 ms	0.916
3	SciPy LSODASciPy	100%	9.3	615	5 ms	0.841
4	SciPy RK23SciPy	100%	9.2	3,521	47 ms	0.839
5	CVODE BDFexternal	100%	9.2	399	14 ms	0.838
6	SciPy BDFSciPy	100%	9.0	744	28 ms	0.834
7	CVODE Adamsexternal	100%	8.8	547	18 ms	0.828
8	Tsit5external	100%	8.4	2,412	1000 ms	0.819
9	SciPy RK45SciPy	100%	8.3	2,180	25 ms	0.816
10	SciPy DOP853SciPy	100%	7.8	3,050	32 ms	0.805

At Clean, best balanced arm is SciPy Radau · SolvSRK survival 100%, SCD 12.5.

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 →

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_hel_thermal_kill_30kw_pl_800_2026,
  title        = {Resonix Evidence Portal: HEL Thermal Kill (30kW.PL.800)},
  author       = {{Resonix Labs (Canada) Inc.}},
  year         = {2026},
  howpublished = {\url{https://resonixusa.com/evidence/problems/hel-thermal-kill-30kw-pl-800}},
  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.}
}

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

Problem definition

Versions & freeze

Results matrix

Cite this page

Related