Statistical Aspects of Cyclic Variability

Practical aspects of the measurement of cycle-by-cycle variability (CBCV) are discussed. That there is a strong ‘memory’ element in the process under certain conditions is revealed by autocorrelation techniques. The system is shown to be well represented by a statistical model involving two distinct processes, with the parameter values of this model significantly dependent on engine operating conditions. Arc duration has been found to be particularly significant in this respect.

Misfiring has the effect of destroying the chain of cause and effect between cycles, but the modelling process described here allows the combustion under extreme conditions to be investigated independently of this complication.

The degree of memory in the system is shown to be inversely related to the degree of cyclic variability.