Cycle-to-cycle combustion analysis in hydrogen fumigated common-rail diesel engine

Abstract

An experimental cycle-to-cycle analysis in multi-cylinder automotive common-rail compression ignition engine was performed to understand better the combustion in the hydrogen fuelled diesel engine. Hydrogen was fumigated in intake line at a wide-ranging from 0 lpm to 50 lpm in steps of 10 lpm. Test engine was operated at three different loads for a constant engine speed. The relative air–fuel ratios were between 1.55 and 3.31. The combustion occurred with an excess of air in all tests. Combustion data of 120 cycles were used to analyse the cyclic variations, determined by standard deviation of the cylinder pressure. Cyclic variations, coefficient of variance (CoV), standard deviation, frequency, and average value of peak combustion pressure (Pmax), maximum pressure rise rate (PRRmax), indicated mean effective pressure (IMEP), mass fraction burned (MFB), and MFB10-90 duration were analysed. Results showed that for all hydrogen addition levels, CoVs of Pmax, IMEP and MFB10-90 were always found below 3% at all loads, and PRRmax values with hydrogen operations in each cycle were under the limit of knocking combustion. Engine load was the most important factor to affect to CoV, standard deviation, and frequency of the combustion parameters. Minor cyclic variations in MFB traces were found at all engine loads and hydrogen addition levels, which agreed well with the cylinder pressure traces. Frequencies of Pmax and IMEP were moderate in low and medium loads, but reduced in high load, and also the values of Pmax and IMEP in high load were distributed in a wider range compared to low and medium loads. MFB10-90 duration increased for dual-fuel modes, its frequency was decreased, and its distribution was extended with hydrogen addition. Trend of standard deviation was mostly similar to that of CoV for the studied combustion parameters. Furthermore, variations of correlation coefficient (R) among Pmax, PRRmax, IMEP, and MFB10-90 were discussed in this paper, and the highest R value was found between IMEP and Pmax.