An attempt to estimate the uncertainty of measurements of parameters of particulate matter oxidation kinetics
More details
Hide details
Automotive Industry Institute
Publication date: 2014-04-25
The Archives of Automotive Engineering – Archiwum Motoryzacji 2014;63(1):71-88
In this paper, an attempt has been made to estimate the uncertainty of measurements of the parameters that describe the kinetics of oxidation of the particulate matter (PM) produced during fuel combustion in compression ignition (CI) engines. The kinetics of the process is characterized by two basic parameters: activation energy Ea and PM oxidation rate constant k0. One of the methods to determine the values of these quantities is based on gravimetric measurements of the decrement of mass of the particulate matter when being burnt. The course of the PM oxidation process depends with considerable significance on the repeatability of process conditions, which affects the values of the kinetics parameters being determined. The parameters Ea and k0 are of fundamental importance at simulation tests and at analysing the PM oxidation (combustion) process; they are also vital to developing the methods of reconditioning automotive exhaust gas filters. It is essential that for the above needs, the parameters Ea and k0 should be determined with a definite level of reliability; this means that the measurement results must be subjected to a statistical analysis. A procedure of statistical analysis of measurement results and the estimated values of the above parameters, obtained from experimental investigations of the kinetics of the PM oxidation process, have been presented below. About 1 000 realizations were recorded in various process conditions on a test stand for gravimetric investigations of the PM oxidation process. For the needs of this paper, a series of measurement results was selected from that, based on predefined criteria. The selected results were subjected to a Dixon’s Q test, which is to eliminate outliers from the data set. For the normal probability distribution of data and a specific confidence level having been assumed, the uncertainty level was calculated and the measurement result was specified as an arithmetic mean of the series of individual results with the corresponding uncertainty interval. In result of the statistical analysis, reliable values of the parameters Ea and k0 and of their uncertainty were determined with the assumed confidence level. For the calculated Ea and k0 parameter values to be additionally verified, they were subjected to a statistical test in order to eliminate those being outside of the range determined for the assumed confidence level and for the standard deviation of the series of measurements. The test confirmed the uncertainty to have been estimated correctly; moreover, conclusions from this work could be formulated based on this test.
Declaration of availability
Journals System - logo
Scroll to top