Diesel Injector Dynamic Modelling and Estimation of Injection Parameters from Impact Response Part 2: Prediction of Injection Parameters from Monitored Vibration
Summary (3 min read)
INTRODUCTION
- The injection process plays one of the most influential roles in the performance and emission control of a diesel engine.
- Obtaining accurate values for injection parameters is the key to the specification, set-up, adjustment and diagnosis of injection equipment.
- This in turn permits enhanced performance such as reductions in fuel consumption and pollutant emissions, and increases in power output.
- A non-intrusive measurement approach, which does not influence injection characteristics, is attractive because it may permit more accurate estimates of injection parameters.
- The layout of the paper is as follows: Section 2 analyses injector vibration; Section 3 describes the test methods used in the vibration measurement; Section 4 details the.
2 INJECTOR VIBRATION
- There are two sources of injector body vibration associated with the injection process: impacts due to the needle hitting its backstop and seat (that is mechanical excitation), and the flow of high-pressure fuel within the galleries and chambers of the injector (that is fluid flow excitation).
- Aside from the nature of the excitation sources, the monitored vibration is also influenced by the dynamic properties of the injector body.
2.1 Mechanical excitation
- Part 1 developed a theoretical correlation between the impacts and fuel injection parameters.
- The amplitude of the first collision in the opening impact series was shown to be related to the fuel injection pressure, and the energy of the opening impact series was shown to be related to the fuel injection rate.
2.2 Fluid flow excitation
- The presence of flow induced vibration response is confirmed in Fig. 1 which overlays monitored body vibration upon cylinder pressure, line pressure and needle lift traces.
- From this figure it can be seen that there is a vibration response commencing prior to the opening of the needle (at a time corresponding to the onset of high-pressure fuel supply) and continuing until the needle is fully retracted, at which point it becomes swamped by the high-amplitude impact response.
- There are no mechanical impacts, only the flow of fuel within the injector.
3 INJECTOR VIBRATION MEASUREMENT
- Due to the complex nature of the injector body, its mounting assembly and its interaction with the cylinder head, it is difficult to derive a mathematical model relating the internal vibration sources to an external surface-mounted monitoring transducer.
- For this reason, an experimental investigation of the relationship between monitored vibration and fuel injection was adopted in this study.
- Two types of experimental set-up were used: a working engine test rig, and a bench-top rig providing an out-of-the-engine injection test facility.
- The test engine was a Ford FSD-425 production unit fitted with a Bosch V injection system and coupled to a hydraulic dynamometer.
- The purpose of the bench-top rig was to isolate those components of monitored vibration which were due to the injection process from those components associated with combustion, piston slap and other engine sources.
3.1 Bench-top injector tests
- Only one injector (the test injector) is mounted in the bench-top cylinder head.
- The other three injectors, used to maintain representative loading of the injection pump, are mounted together in a dummy cylinder head with associated fuel collection vessels.
- This separation of the test injector from the load injectors is further to ensure its vibration isolation.
- Fuel levels can be adjusted via the injection pump rack setting in the usual way, and the pump speed can be varied by the closed-loop.
I I
- Two-channel data acquisition from the bench-top rig was made with a computer-controlled 12-bit device sampling simultaneously at 65 kHz.
- From the bench-top tests, several aspects of injector vibration response were confirmed.
- This conclusion is reinforced by the directly measured needle acceleration traces shown in Figs 3(a2) and 3(b2).
- The inserts within Figs 3(a2) and (b2) show the acceleration waveforms as predicted by the model developed in Part 1.
- These can be seen to exhibit quite a good correlation with their measured equivalents, further validating the accuracy of the model.
3.2 Engine injector tests
- Before conducting a test, the engine was run to normal operating temperature, and then a constant speed test was performed at four load settings between 35 and 130 Nm.
- A wide range of operating conditions, and consequently of fuel injection characteristics, was covered.
- In the same manner as for the bench-top rig, the incoming data was gated in software to give 1024-point segments.
It can be seen that during injection, vibration energy
- The opening impact responses are seen to attenuate is spread over a wide range of frequencies from around gradually with time and the two higher frequency response ridges have all but disappeared by the time that the injector closing impact occurs.
- Figure 6 shows WVD analysis results for differing engine loads at a speed of 3000 r/min.
- Furthermore, it is more difficult to distinguish the injector closing response in the lowfrequency region.
- The needle opening is governed by a large and repeatable fuel pressure front, but the needle closing is governed by the residual pressure in the injector, and this is likely to vary quite considerably from injection to injection.
- If the residual fuel pressure is low, it is likely that a large amplitude body response will result.
5 ANALYSIS OF THE INJECTION PROCESS
- Having identified within the WVD analysis those regions of frequency response which contain injection related information, bounds can be placed upon the frequency range of interest.
- With these bounds it is possible to set a bandpass filter so that envelope analysis can be used as an alternative means of extracting the appropriate high-frequency information from the monitored vibration data.
5.1 Extraction of timing information from monitored
- It can be seen that the values of the injection events obtained from the monitored vibration time signature are highly consistent with those obtained by direct needle lift mea- surement.
- The vibration-derived timing of the needle opening and fully open events is nearly identical to the equivalent needle lift information, and the vibrationderived band of the closing event timing neatly contains the directly measured trace.
- It can be seen from the closing impact bands depicted in Fig. 8 that averaging of the band limits does, in the majority of instances, give a timing value for the needle closing event which is very close to that obtained by needle lift measurement.
5.2 Extraction of fuel pressure information from
- Injector vibration is caused by a combination of mechanical impacts and high-pressure fluid flow within the injector body.
- This vibration response can be detected using an accelerometer mounted either on the injector body, or on the injector fixing saddle.
- An engine-isolated bench-top rig was used to demonstrate that injector vibration response is not contaminated by engine-related vibration sources such as combustion noise and piston slap.
- Wigner-Ville distribution analysis was applied to the highly non-stationary vibration signals monitored on an injector body and this showed that the highfrequency body response contains regions of information which describe the injection process.
Q IMechE 1996
- The proposed vibration-based non-intrusive approach to the analysis of diesel fuel injection systems could provide a powerful alternative to the conventional intrusive fuel line pressure and the needle lift measurement techniques that are commonly used.
- In addition, this technique has important implications in the field of diesel engine condition monitoring.
- Deviations from the demonstrated good condition relationships between monitored vibration and fuel injection parameters will be likely to indicate a change in condition.
Did you find this useful? Give us your feedback
Citations
2 citations
2 citations
2 citations
Cites methods from "Diesel Injector Dynamic Modelling a..."
...According to the dynamic injector performance principle and literature study [40], [41], the AE technique can provide an accurate real-time characteristics monitoring for the transient injection process, which can clearly record events when fuel is pumped in the injector nozzle - the needle arises by increasing fuel pressure, fuel injection is sprayed through the orifice and then the needle goes back to the nozzle seat....
[...]
2 citations
Cites background from "Diesel Injector Dynamic Modelling a..."
...Additionally, research on vibratory analysis of individual elements of ICE has been carried out to predict relevant parameters, such as injector vibratory models (Gu and Ball, 1996; Gu et al., 1996), piston slap vibratory response (Richmond and Parker, 1987) and indirect measurement of main bearings load (Leclère et al....
[...]
1 citations
References
974 citations
706 citations
573 citations
4 citations
2 citations