Davide Cherubini, Alessandra Fanni, Augusto Montisci and Pietro Testoni
To present a neural network‐based approach to the design of electromagnetic devices.
Abstract
Purpose
To present a neural network‐based approach to the design of electromagnetic devices.
Design/methodology/approach
A neural model is created which reproduces the relationship between the design parameters of the device and the performance parameters, typically field values.
Findings
The neural model is a single hidden layer MLP network, trained by using a set of cases calculated, for example, by means of a finite element analysis. The design problem can be solved by fixing the performance values at the output of the network and by calculating the corresponding input values. The relationship between the input and the output of the neural network is represented by three equations systems. By means of these three systems, we can forward the domain of the input, and we can back propagate the desired output throughout the network layers. In such a way, both the domain of the design parameters and the domain of the desired performances values can be projected in the same space. Whatever point inside the intersection between the two projected domains corresponds to a solution of the design problem.
Originality/value
Presents a procedure which is able to find a point belonging to such an intersection.
Details
Keywords
Alessandro Bonito Oliva, Andrea Gaetano Chiariello, Alessandro Formisano, Raffaele Martone, Alfredo Portone and Pietro Testoni
The purpose of this paper is to analyze the impact of different current representation models in the high field magnets characterization. Inverse source methodology used for…
Abstract
Purpose
The purpose of this paper is to analyze the impact of different current representation models in the high field magnets characterization. Inverse source methodology used for current reconstruction is discussed. The regularizing effect of successive field map computation in different regions is also assessed.
Design/methodology/approach
Under suitable hypotheses, the resulting inverse source problem is linear, and different current representation bases are used to assemble different matrices. Properties of matrices are then assesses using SVD. The following field computation problem is also formulated using a projection matrix, and the properties of combined matrix operators are analyzed and compared to the inversion matrix.
Findings
The characteristics of the inverse matrix depend on the choice of the current representation basis, but in any case the application of the further projection matrix has a relevant regularizing effect.
Research limitations/implications
The method is intrinsically tied to the linearity assumption, and the regularizing effect of the projection operator is stronger for further field regions.
Practical implications
The accuracy in the current reconstruction procedure can be reduced if data will be used only to compute field in distant regions.
Originality/value
The paper casts the problem of field computation in distant regions from magnetic measurements in the language of direct and inverse operators, allowing to assess its properties and fine tune the procedure parameters to achieve satisfactory results with minimum effort.