Algorithm for signal drop-out recognition in IC engine valve kinematics signal measured by laser Doppler vibrometer

We present an algorithm for the recognition of signal drop-out developed particularly for measurements of valvetrain kinematics. This algorithm is needed in order to save data that are not affected by a drop-out phenomenon. Such an algorithm will increase the throughput of the engine test stand and decrease the time needed for the evaluation of the valvetrains of combustion engines. The work shows the most commonly encountered drop-outs and their characteristics and locations. It presents an automatic separation algorithm for the measured signal so that the drop-out recognition tests can be aimed at specific data intervals (valve opening, valve closing, etc.) with specifically set parameters of the algorithm.

► Phenomena affecting the valve kinematics signal are identified and described. ► Separation of the valve kinematics data into intervals for parameterization of the tests for drop-out recognition is described. ► The signal drop-out tests use FIR filter or DWT as a first step. The noise level is statistically determined. ► SNR of the signal is calculated and SNR thresholding is applied. ► Parameters of the test are presented.