Application of Differential Evolution to Mono-objective Tuning of Vibration Spectrum Analyzers based on Microelectromechanical Systems

The vibration spectrum is a frequency-domain characteristic used to monitor many systems and can be calculated using twin-microaccelerometers with nonlinear electromechanical positive feedback. This strategy has its greatest limitation on the accelerometers parameters differences due to the manufacturing process, such that the spectrum distortion may be considerably high. To overcome these differences, in this work a meta-heuristic algorithm to tune the spectrum analyzer microdevice is proposed by adjusting the accelerometers actuation voltages amplitudes. To perform the tuning, the Differential Evolution (DE) is used to solve the mono-objective optimization problem related to the sensitivity maximization constrained by a maximum distortion level. The objective and constraint functions are based on the components of the spectrum analyzer system closed-loop gain Fourier series – which depends on the actuation voltages. The results – specially the distortion-sensitivity compromise – are demonstrated and discussed. The proposed tuning strategy is able to determine the voltage amplitudes to be applied to the micro spectrum analyzer to attend the distortion level and sensitivity requirements.