On the Memory of Measurements in Spray Diagnostics

In spray diagnostics, a measurement is usually treated as the accumulation of validated droplets until the statistics of the size, velocity, flux, and derived quantities can be estimated. This implicitly treats the measurement as a sampling process. However, if the spray or measurement context changes during acquisition, the droplet-event sequence may retain traces of previously sampled events. In this sense, a measurement may have memory: the present record carries forward information about past event organization and changes in the measured local spray state. This study introduces an infodynamic framework for time-ordered droplet-event data, where each validated droplet carries morphological, kinematic, and arrival information through its size, velocity, and inter-arrival time. Droplet-events are mapped into a fixed quantized alphabet, and sliding time windows are used to compute a memory trace from the mutual informature between past and future event blocks. A permutation baseline is subtracted to obtain a corrected memory trace that separates the resolved event-order memory from the finite sampling effects. When applied to Time-Shifting Time-Of-Flight measurements of a flat-fan spray, the method showed that stable local measurements remained close to the finite-sampling baseline, supporting a distributional-sampling interpretation. When the injection pressure and/or measurement position varied, the infodynamic descriptors revealed measurement morphing through changes in event diversity, alphabet redistribution, and intermittent predictive ordering. The proposed approach extends conventional spray characterization by distinguishing repeated sampling of a local spray state from the monitoring of an evolving measurement state.