This is a self-archived version of an original article. This version may differ from the original in pagination and typographic details.

This paper introduces a geometric solution strategy for Laplace problems. Our main interest and emphasis is on efficient solution of the inverse problem with a boundary with Cauchy condition and with a free boundary. This type of problem is known to be sensitive to small errors. We start from the standard Laplace problem and establish the geometric solution strategy on the idea of deforming equipotential layers continuously along the field lines from one layer to another. This results in exploiting ordinary differential equations to solve any boundary value problem that belongs to the class of Laplace's problem. Interpretation in terms of a geometric flow will provide us with stability considerations. The approach is demonstrated with several examples.Here, the Laplace problem has become formulated as a system of ODEs. The field lines 'begin' on @ and they are forced to 'end' to @ † with condition ' D 1. Evidently, such a solution strategy for the Laplace problem is not that competitive compared with finite element kind of methods. However, our aim is not to suggest a new approach to solve the standard Laplace problem but instead to highlight the common geometric properties behind Laplace type of boundary value problems. § An example of such routine is MATLAB function bvp4c. Figure 17. The modeled phone connector in cylindrical coordinates in meters. We assume that the field lines are tangential to the surface Z D 0, which corresponds to a case when two connectors glued together through this surface are electroplated simultaneously. Figure 18. The solved field lines and equipotential surfaces.

Purpose The purpose of this paper is to introduce a reverted way to design electrical machines. The authors present a work flow that systematically yields electrical machine geometries from given air gap fields. Design/methodology/approach The solution process exploits the inverse Cauchy problem. The desired air gap field is inserted to this as the Cauchy data, and the solution process is stabilized with the aid of linear algebra. Findings The results are verified by solving backwards the air gap fields in the standard way. They match well with the air gap fields inserted as an input to the system. Originality/value The paper reverts the standard design work flow of electrical motor by solving directly for a geometry that yields the desired air gap field. In addition, a stabilization strategy for the underlying Cauchy problem is introduced.