Information technology of monitoring of phase delay fluctuations based on indirect measurement data processing

Abstract

The paper proposes the use of correlation analysis to process experimental data from indirect measurements to monitor phase delay fluctuations in fuel-air mixture supply to internal combustion engine cylinders. The oscillatory movements of model masses are described by a system of second-order linear differential equations, normalized using similarity theory methods. The deterministic system is solved via the Laplace transform under zero initial conditions. Frequency characteristics of cylinder torque transfer functions are analyzed in Matlab, while a measurement signal simulation scheme is developed in Mathcad. Additive noise is modeled as structured "white noise" with a frequency spectrum limited to ten harmonic components. Neural network techniques adjust the length of information links in the simulation, allowing for variation in amplification coefficients of cylinder torque amplitudes. The application for cross-correlation function calculation is implemented in Mathcad. Analysis of mutual correlation function graphs shows that its maxima correspond to the standard torque phases of the second, third, and first cylinders, respectively. It is established that a 60% uncertainty in crankshaft rotation unevenness measurement still enables unambiguous identification of these maxima. Additionally, the influence of additive "white noise" on cross-correlation function graphs is investigated under varying gain coefficients of the measurement signal modeling scheme. Results show that even with a 15% measurement uncertainty of a deterministic crankshaft rotation unevenness signal, the mutual correlation function remains effective for monitoring phase delay fluctuations in cylinder torques relative to standard engine settings.