Errors from discretization and large data volume of field maps is a concern for beam dynamicssimulations w.r.t. achievable accuracy and to the required amount of time.
High-order Singular Value Decomposition (HOSVD) has recently emerged as simple, effective,
and adaptive tool to extract the essentials from multi-dimensional data. This method is on the compression
and noise reduction of electromagnetic field map data with HOSVD. The method has been applied
to an electric field map of a DTL cavity with 11 m in length comprising 55 rf-gaps.
The original field map data of 220 MB was converted into practically noise-free data of just 20 KB.
Noise was reduced by 95% as demonstrated using a cubic cavity for which the analytical field map is available.