On the Velocity of Light Signals in the Deep Underwater Neutrino Experiments

L.A.Kuzmichev

DOI: https://doi.org/10.1016/S0168-9002%2801%2901679-5

2000-05-25

Abstract:During the last few years deep underwater neutrino telescopes of a new generation with dimensions close to 100 m or more were taken into operation. For the correct track reconstruction and for the interpretation of light pulses from calibration lasers one has to use the group velocity for light signals. The difference between group velocity leads to an additional delay of about 10 ns for a distance of 100 m between light source and photjmultiplier. From the time of the appearance of the first projects of deep underwater neutrino telescopes in the middle of 70th this fact was never mentioned in the literature.

High Energy Physics - Experiment

What problem does this paper attempt to address?

This paper attempts to solve the problems encountered when using the finite element method for approximation on quadrilateral meshes. Specifically:

1. **Differences in approximation properties**: The paper explores the approximation properties of the finite - element space when the reference element is transformed from a square to an arbitrary convex quadrilateral through a bilinear mapping. It is known that for affine isomorphisms, in order to achieve the approximation orders of \(O(h^{r + 1})\) in the \(L^2\) space and \(O(h^r)\) in the \(H^1\) space, the function space on the reference element must contain all polynomials with a total degree not exceeding \(r\). For bilinear isomorphisms, the same estimates hold if the function space contains all polynomials with the degree of each variable not exceeding \(r\).

2. **Proof of necessary conditions**: The author proves that the above conditions are necessary through counter - examples. That is, if the function space on the reference element does not contain all polynomials \(Q_r(\hat{K})\) with the degree of each variable not exceeding \(r\), then on a general quadrilateral mesh, the approximation order will decrease. For example, for serendipity finite elements and various mixed and non - conforming finite elements, the convergence order on quadrilateral meshes is lower than that on rectangular meshes.

3. **Application examples**: The paper shows the degradation of the convergence order when using serendipity finite elements and other types of finite elements on quadrilateral meshes. In particular, for serendipity elements, when the mesh changes from a rectangle to a trapezoid, the convergence orders of the solution and its gradient decrease from \(O(h^{r+1})\) and \(O(h^r)\) to \(O(h^{\lfloor r/2 \rfloor + 1})\) and \(O(h^{\lfloor r/2 \rfloor})\) respectively.

### Summary

The main objective of the paper is to prove that on quadrilateral meshes, in order to maintain the same high - order approximation accuracy as on rectangular meshes, the function space on the reference element must contain all polynomials \(Q_r(\hat{K})\) with the degree of each variable not exceeding \(r\). Otherwise, the approximation order will decrease significantly, especially when using specific types of elements such as serendipity finite elements. This conclusion is of great significance for the application of the finite element method to complex geometries.