A two-order finite-time-convergent differentiator with application to nuclear reactors

Abstract

Signal differentiation is widely used in system monitoring and control. The key problem in differentiator design is how to obtain a satisfactory tradeoff between differentiation exactness and robustness to uncertainties and noises. Based upon proposing a novel finite-time stabilizer for 2-order integrator chain, a new finite-time-convergent 2-order differentiator with bounded estimation error of both the signal and its derivatives is newly developed. This differentiator is then applied to estimate the derivative of the neutron flux of nuclear reactors. Numerical simulation results not only verify the theoretical results but also show the satisfactory performance of this newly-built differentiator.