A calibration approach for compensating hysteresis and nonlinearity error in an MEMS based instrument for porewater pressure monitoring

Abstract

Micro-electro-mechanical system (MEMS) technology is becoming increasingly popular in geotechnical and hydrological monitoring due to their distinct features, including its small size, low cost, low energy consumption, and resistance to vibration shock. Due to the nature of design, they exhibit a series of errors, and most importantly, hysteresis and nonlinearity, which should be compensated before any practical application in order to achieve the highest degree of accuracy and precision. This article proposes a practical approach to enhance the accuracy of a newly developed MEMS instrument for groundwater monitoring. The hysteresis and nonlinearity errors were compensated using an integrated two-step approach with an optimized Preisach model and an optimized spline approximation, respectively. A hybrid differential evolution-hill climbing algorithm was utilized to achieve optimum models. This optimization algorithm integrates global and local search, offering higher accuracy and computational stability with a lower calibration point requirement. Analysing the results indicates that the nonlinearity error shows an improvement of more than 94% and hysteresis decreased up to 67%. The results illustrate that the compensation can be performed very well using the proposed methodology with lower uncertainty, mean square error, and standard deviation compared to non-optimized models.