This thesis follows the process of design and optimization of anisotropic magneto resistive sensor. Its main focus is on the design of a computational model for simulation of a sensor.
I have designed three different models, regarding the purpose. The first and most general model was designed to simulate the rough behaviour of a sensor in a rotating magnetic field, optimization of position and height and the influence of external magnetic disturbances. The second model carries out the algorithm for optimization of a shape of a deposit of AMR material to achieve the best response in a magnetic field of a magnetic scale. Third model simulates the hysteresial behaviour of a material and consequently the overall delay in sensor’s output while moving through the rotating magnetic field.
Some of the used methods are the finite elements method, interpolation, Monte Carlo method (adapted for optimization according to the chosen weighted criterial function), approximation with the least-squares method and Preisach discrete hysteresis model. To generalize the model of hysteresis to simulate the behaviour in the magnetic field with a variable direction, a generalized Preisach model for each component was developed, but it turned out to be false. Therefore, I have proposed a new experience-based model. It approximates the hysteresis with the line or hyperbola.
According to the knowledge about AMR effect, commercially available AMR products and improvements, tested by these models, I proposed a final design of a new sensor. Unfortunately at the time of writing this thesis, the product was not yet finished and therefore not evaluated.
|