Characteristics of Rail Pressure Fluctuations under Two-Injection Conditions and the Control Strategy Based on ANN University of Bedfordshire

Format:

Conference/Event

Author/Creator:

Peng, Peng, author.

Contributor:

Bai, Fu-qiang

Du, Qing

Ma, Mingyang

Weizhi, Wang

Zhang, F.

Conference Name:

International Powertrains, Fuels & Lubricants Meeting (2017-10-16 : Beijing, China)

Language:

English

Physical Description:

1 online resource

Place of Publication:

Warrendale, PA SAE International 2017

Summary:

AbstractHigh-pressure common rail (HPCR) fuel injection system is the most widely used fuel system in diesel engines. However, when multiple injection strategy is used, the pressure wave fluctuation is un-avoided due to the opening and closing of the needle valve which will affect the subsequent fuel injection and combustion characteristics. In this paper, several parameters: injection pressure, injection intervals, the main injection pulse widths are investigated on a common rail fuel injection test rig with two injection pulses to explore their effect on the fuel injection rate and fuel quantity. The result showed that the longer injection interval between the pilot and main injections will lead to a rail pressure drop at the beginning of the main injection so that a smaller fuel quantity will be delivered. The main injection pulse width also influences fuel injection rate and the main fuel quantity. The fuel injection rate plays a dominant role on fuel quantity when the main injection pulse width is less than 600 us in the current case. For the rather complicated injection system, a control strategy based on ANN (artificial neural network) with the BP algorithm is adopted to predict intervals, fuel injection pulse width and thus the fuel quantity. First, the injection interval is predicted by the required injection pressure and then the injection pulse width is predicted by injection rate and fuel injection quantity. The model is tested by experiments and the results show that the error is less than 5% between prediction and measurement

Notes:

Vendor supplied data

Publisher Number:

2017-01-2212

Access Restriction:

Restricted for use by site license