Parametric modelling of knock intensity data using a dual log-normal model
| Autor: | Jesse Frey, James C. Peyton Jones, Saeed Shayestehmanesh |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
Stochastic process
020209 energy Mechanical Engineering Mathematical analysis Pearson's chi-squared test Aerospace Engineering Ocean Engineering Probability density function 02 engineering and technology Mixture model Intensity (physics) Dual (category theory) symbols.namesake 020303 mechanical engineering & transports Distribution (mathematics) 0203 mechanical engineering Automotive Engineering Log-normal distribution 0202 electrical engineering electronic engineering information engineering symbols Mathematics |
| Zdroj: | International Journal of Engine Research. 21:1026-1036 |
| ISSN: | 2041-3149 1468-0874 |
| DOI: | 10.1177/1468087418796335 |
| Popis: | The Pearson test is used to confirm that knock intensity data closely approximate a cyclically independent random process which is therefore fully characterized by its probability density function or cumulative distribution function. Although these distributions are often assumed to be log-normal, other results have shown that the data do not conform to a log-normal distribution at the 5% significance level. A new dual log-normal model is therefore proposed based on the assumption that the data comprise a mixture of two distributions, one knocking and one non-knocking. Methods for estimating the parameters of this model, and for assessing the quality of fit, are presented. The results show a significantly improved model fit. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|