Sample Size Effect on Combustion Analysis

Engine parameters vary from one cycle to the other and this makes engine analysis with data from a single working cycle insufficient in capturing or modelling an engine behaviour. The variation observed in engine has necessitated the use sample sizes of data obtained during an engine operation to obtain results that are representative of the engine being investigated. Research has shown that the use of very large data sample size increases the storage needed and processing time and does not necessary give better results over results obtained with lesser sample sizes. The number of sample size to use for analysis remains a subject of debate and investigation with researchers proposing the use of varying sample sizes for combustion analysis in engines. There is a need for the selection of an optimum sample size for engine analysis.
Engine data were obtained from a spark ignition engine which operated on gasoline and varying degree of blend of gasoline and biofuel. The effects of the use of sample sizes of 20, 40, 60 and 100 on the result of the analysis were determined. The percentage difference and the mean percentage difference for each of the sample sizes tested relative to the maximum available sample size were determined too.
Based on results from the analysis, it was suggested that sample sizes that gave mean percentage difference values within the range ±1.5 relative to the maximum available samples size are appropriate for use in combustion analysis in engines.