Speaker
Description
While delayed-neutron (DN) data from neutron-induced fission have been refined for decades, photofission DN data remain comparatively sparse. Current photofission evaluations rely on sequential peeling or non-negative least squares (NNLS), which cannot deliver the full parameter covariance matrices required by modern nuclear-data libraries.
We present a Markov Chain Monte Carlo (MCMC) sampling method to jointly estimate the six relative yields $\alpha_i$. Tested on a numerical-twin benchmark, this approach recovers every $\alpha_i$ within $2\,\%$ of the input values, reducing the relative bias compared to sequential iterative methods and capturing non-Gaussian correlations.
The method is currently applied to $9\text{ MeV}$ bremsstrahlung-induced photofission of $^{238}$U. Readily transposable to other actinides, it provides the consistent covariance matrices needed to exploit next-generation photon sources.
| Session | Decay Data and Delayed Neutrons |
|---|