[Paper Review] On linear hypersingular Boltzmann transport equation and its variational formulation

Jouko Tervo|arXiv (Cornell University)|Aug 29, 2018

Numerical methods in engineering18 references3 citations

TL;DR

This paper presents a rigorous variational formulation for the linear hypersingular Boltzmann transport equation (BTE) arising in charged particle transport, particularly for Møller-type scattering where differential cross-sections exhibit hyper-singularities. It shows the exact BTE contains first-order energy derivatives coupled with Hadamard finite part integral operators and second-order angular derivatives, and derives a CSDA-Fokker-Planck-type approximant via variational methods, enabling existence proofs via Lions-Lax-Milgram theory.

ABSTRACT

For charged particle transport the linear Boltzmann transport equation (BTE) turns out to be a partial hyper-singular integro-differential operator. This is due to the fact that the related differential cross-sections $σ(x,ω',ω,E',E)$ may have hyper-singularities. In these cases the energy integral appearing in the collision terms must be interpreted as the Hadamard finite part integral leading to hyper-singular integral operators. The article considers a refined expression for the exact transport operator and related variational formulations of the inflow initial boundary value problem for one particle equation containing hyper-singularities. We find that the exact BTE contains the first-order partial derivatives with respect to energy combined by partial Hadamard (first-order) singular integral operators. In addition, it contains the second-order partial derivatives with respect to angle and some mixed terms. The analysis will be carried out only for the so called Møller-type interaction (scattering) which is a kind of prototype of hyper-singular interactions. The generalizations to other type of collisions, such as to Bremsstrahlung, go analogously. We also expose a weak form (the variational formulation) of the hyper-singular transport problem. Another variant variational formulation decreases the level of singularities in the integration (appearing in the due bilinear form) containing only singularities of order one that is, singularities like ${1\over{E'-E}}dE' dE$. The variational formulation is an essential step in order to show the existence of generalized solutions e.g. by Lions-Lax-Milgram Theorem based methods (proceedings for solution spaces and existence theory are omitted here). The corresponding approximative transport operator is deduced. It turns out to be a CSDA-Fokker-Planck type operator.

Motivation & Objective

• To analyze the exact structure of the linear Boltzmann transport operator when differential cross-sections contain hyper-singularities, particularly in Møller scattering.
• To develop a variational formulation for the inflow initial-boundary value problem of the hypersingular BTE without approximations.
• To reduce the singularity level in the bilinear form to enable numerical treatment via Galerkin finite element methods.
• To establish a theoretical foundation for existence of generalized solutions using variational methods and the Lions-Lax-Milgram theorem.
• To derive an approximative transport operator that matches the structure of CSDA-Fokker-Planck models, suitable for numerical computation.

Proposed method

• The paper identifies that hyper-singularities in differential cross-sections lead to first-order partial derivatives with respect to energy combined with Hadamard finite part integral operators.
• It formulates the exact transport operator as a partial integro-differential equation containing second-order angular derivatives and mixed terms, with the collision integral interpreted as a Hadamard finite part integral.
• A weak form (variational formulation) is derived using test functions in appropriate Sobolev-Slobodevskij spaces, reducing the singularity level to order one in the bilinear form.
• The variational formulation is shown to be amenable to existence theory via the Lions-Lax-Milgram theorem, enabling generalized solution existence proofs.
• The approximative transport operator is derived as a CSDA-Fokker-Planck-type operator, obtained through Taylor expansion and angular approximations of primary particles.
• The analysis is restricted to Møller-type scattering but generalizes to other interactions like Bremsstrahlung via analogous methods.

Experimental results

Research questions

• RQ1 How does the presence of hyper-singularities in differential cross-sections alter the structure of the linear Boltzmann transport equation?
• RQ2 What is the precise form of the exact transport operator when collision integrals are interpreted as Hadamard finite part integrals?
• RQ3 Can a variational formulation be constructed for the hypersingular BTE that reduces singularity order and enables existence proofs?
• RQ4 What is the structure of the approximative transport operator derived from the exact hypersingular operator?
• RQ5 How does the variational formulation support numerical solution strategies such as Galerkin finite element methods?

Key findings

• The exact linear BTE for Møller scattering contains first-order partial derivatives with respect to energy, combined with Hadamard finite part integral operators, forming pseudo-differential-like terms.
• The transport operator also includes second-order partial derivatives with respect to angular variables and mixed derivative terms, with the angular part resembling the Fokker-Planck operator.
• The variational formulation reduces the singularity level in the bilinear form to order one, specifically of the form $ \frac{1}{E' - E} dE' dE $, enabling numerical stability and convergence analysis.
• The derived approximative transport operator is of CSDA-Fokker-Planck type, matching conventional approximations used in radiation therapy dose calculations.
• The variational formulation provides a rigorous mathematical pathway to prove existence of generalized solutions via the Lions-Lax-Milgram theorem, even for hyper-singular operators.
• The analysis confirms that the inflow boundary value problem has variable boundary multiplicity, complicating well-posedness, but the variational framework offers a viable path to handle such cases.