Acoustic emission mapping of diesel engines for spatially located time series—Part II: Spatial reconstitution
Introduction
Acoustic emission (AE) can be used as a non-intrusive method to monitor the mechanical events, components and processes in diesel engines [1], [2], [3], [4], [5] and can be used to diagnose faults and running conditions. AE has been used successfully to detect exhaust valve leakage [1], fuel injection behaviour [2] and various aspects of the combustion process [3], [4]. One major advantage of using AE is that, at the high carrier frequencies from 0.1 to 1 MHz, there is a very high signal-to-noise ratio. There are a number of AE sources in engine operation such as valve impacts, fluid flow/injection, combustion and sliding and rolling contact, which it is very useful to be able to monitor non-intrusively. However, these sources can be complex, generating a number of wave modes at various positions on the engine. The propagation of the AE waves through an object like an engine is also very complex, with wave attenuation/dispersion, transmission/reflection needing to be considered. To optimise monitoring and diagnostic capabilities it is necessary to be able to identify each source of the AE and extract information specific to that source, a process which is described here as spatial reconstitution of the acquired AE time series. To enable any kind of spatial reconstitution, a sensor array and information about the mechanical operation (timing) of the events is needed. In order to choose a suitable sensor array, studies of AE wave propagation in engines must be carried out.
In Part I of this study [6] a simplified model of AE wave propagation through various objects was developed. The same modelling techniques are applied here where, firstly, simulated sources are used to generate AE at various locations on the cylinder head where actual sources are expected and data is acquired from an array of sensors. The simplified model of AE wave propagation [6] is then applied to the data to find appropriate attenuation coefficients and path lengths. Then, under normal engine operating conditions, AE data are acquired, signals are mapped, events identified and a spatial reconstitution technique is developed. The sensor array is limited to positions around the cylinder head in this work.
In general, AE wave propagation in a real structure attenuates along a transmission path. Other studies [6], [7], [8] have identified internal damping, reflection, refraction, mode conversion and diffraction as factors which affect AE transmission/attenuation. Other work, mostly relating to structural integrity monitoring (e.g. [9], [10]), has identified important wave components. Nivesrangsan et al. [6] have presented a preliminary study of AE wave transmission on relatively simple cast iron objects, including an engine base, a surface table, a narrow strip and a cylinder block. They found that there are two dominating modes of AE wave propagation (as did Holford and Carter [10]); a low amplitude, fast wave, which may itself consist of one or a number of modes, precedes the main signal (slow wave), which carries the peak amplitude of the signal. The AE wave propagates through these cast iron structures mainly in two frequency bands around 100–200 kHz (low frequency) and 300–350 kHz (high frequency). In Part I [6], attenuation of AE energy and the effects of geometry, transmission and reflection along each transmission path was considered and, rather than attempt to describe the wave mathematically (which is probably intractable), a simplified exponential attenuation relationship was proposed. Good correlation was found between this model and AE wave propagation and this successfully confirmed estimates of transmission path length using a time of flight technique [10]. The results of that work together with mapping techniques developed by El-Ghamry et al. [4] are applied here to spatially reconstitute AE source signals associated with fuel injection and exhaust valve opening using a sensor array on the cylinder head. Source location using time of flight and energy based techniques will be considered in Part III of this work [11].
Section snippets
Acoustic emission energy
In order to study the propagation of AE around the cylinder head using the techniques developed in Part I [6], it is necessary to determine the energy in the acquired AE signal. For simulated sources, this is a simple matter of determining the energy in a specific time or frequency window for a single event. For signals acquired from a running engine, it is possible to map the engine cycle onto the AE time series and identify individual events within a cycle. It is then possible to select
Experiment
Two sets of experiments were carried out on the cylinder head of a Perkins 4-stroke, four cylinder, 74 kW diesel engine using a sensor array illustrated in Fig. 1. For both sets of experiment a four-channel PCI-6115, National Instruments data acquisition (DAQ) card was used. For simulated source tests four Physical Acoustics, micro-80D broadband AE sensors with preamplifier (including band-pass filter between 0.1 and 1 MHz) were used with one positioned beside the source and three positioned on
Simulated source results
The data acquired from the simulated sources was used to determine the effective source–sensor distance and the attenuation factors, which could then be compared with results found in Part I [6] and used for the engine running test results in the next section. It is important to be able to estimate reliably the source–sensor distance because this will then be used to study the attenuation of the AE energy with distance and calculate the attenuation factors (using Eq. (4)). A computer generated
Engine running tests
Examples of the AE signals acquired from one cycle of the running engine are shown in Fig. 4 against crank angle for each of the sensor positions P1–P9. The mapping of the mechanical events onto crank angle is shown along the bottom of Fig. 4. As can be seen, the largest AE events are associated with injection (INJ) around top dead centre (TDC) for each cylinder and these are observed at all sensor positions to a greater or lesser extent. Smaller amplitude events can also be seen for exhaust
Crank angle—frequency analysis
Fig. 9 shows an example contour plot of crank angle against frequency for receiver sensor at P1 on the cylinder head for the engine running at 800 rpm with no load. The frequency spectrum has been determined for the time period associated with each degree of crank angle. The maximum value of each spectrum is then used to normalise the data so that spectral energy can be compared across the crank angle domain despite the variations in signal energy with crank angle. It can be seen that there are
Spatial reconstitution
Fig. 12 shows normalised raw AE signals for various sensor positions where the source was emitted from injector 4 at low-speed and no load. Each signal was normalised using its maximum amplitude and, as an example, the results shown for sensors 4, 5 and 6 were acquired for the same engine cycle. The energy content shown for each signal () was determined using Eq. (2), and the arrival time shown for each sensor () was determined (referred to zero at the sensor closest to
Conclusion
The propagation of AE signals has been studied using a multiple AE sensor array on the cylinder head of a diesel engine. Two sets of tests were carried out where simulated source data and engine running data were collected. AE events associated with fuel injection showed attenuation factors of around 8–9 and good correlation with the exponential absorption model. The measured attenuation factors of exhaust valve opening events were more variable, between 5 and 7, with slightly poorer
References (18)
- et al.
The development of automated pattern recognition and statistical feature isolation techniques for the diagnosis of reciprocating machinery faults using acoustic emission
Mechanical Systems and Signal Processing
(2003) - et al.
Acoustic emission mapping of diesel engines for spatially located time series
Mechanical Systems and Signal Processing
(2005) - et al.
Exhaust valve leakage detection in large diesel engines
Condition Monitoring and Diagnostic Engineering Management COMADAM, Clayton, Australia
(1998) - J.D. Gill, R.L. Reuben, J.A. Steel, A study of small HSDI diesel engine fuel injection equipment faults using acoustic...
- J.D. Gill, R.L. Reuben, M. Scaife, E.R. Brown, J.A. Steel, Detection of diesel engine faults using acoustic emission,...
The role of acoustic emission in industrial condition monitoring
International Journal of COMADEM
(1998)Stress Waves in Solids
(1963)- A.A. Pollock, Classical wave theory in practical AE testing, in: Progress in Acoustic Emission III—Japanese Society of...
- et al.(1987)
Cited by (29)
Sheath-outside AE monitoring method for multiple sources of damage to cables considering the effect of an HDPE sheath on AE propagation
2024, Measurement: Journal of the International Measurement ConfederationStudy on propagation mechanism and attenuation law of acoustic emission waves for damage of prestressed steel strands
2023, Measurement: Journal of the International Measurement ConfederationResearch on the generalisation method of diesel engine exhaust valve leakage fault diagnosis based on acoustic emission
2023, Measurement: Journal of the International Measurement ConfederationStandard deviation based acoustic emission signal analysis for detecting valve internal leakage
2018, Sensors and Actuators, A: PhysicalCitation Excerpt :When the fluid medium is liquid, the frequency of AE signal of valve leakage is below 200 kHz. The AE signal of valve leakage is a random signal, and its waveform is very complex because it is affected by some factors, such as the propagation path of AE signal, transmission medium and environmental noise [10–14]. In order to obtain the coefficients in Eq. (11), the experiments of valve leakage detection were conducted in Chongqing Chuanyi Automation Co., Ltd., in Chongqing city, China.
Acoustic emission characteristics of a single cylinder diesel generator at various loads and with a failing injector
2017, Mechanical Systems and Signal ProcessingCitation Excerpt :However, in the case of engine diagnostics this relieves the challenge of poor signal to noise ratio in vibration signals noted earlier where dynamic phenomena from other components and vehicle motion complicate vibration analysis. Characterization of the AE transmission characteristics through the engine block and representative structures was conducted in a series of papers by Robertson et al. [17] and Nivesrangsan et al. [18–20], which calculated wave speeds and attempted to determine the spatial location of the source based on the arrival times of AE events. Many of these engine AE diagnostic studies are listed in chronological order in Table 1.
Normalization and source separation of acoustic emission signals for condition monitoring and fault detection of multi-cylinder diesel engines
2015, Mechanical Systems and Signal ProcessingCitation Excerpt :Nevertheless, the nonlinear frequency response of AE sensors remains a challenge in sensor calibration to provide a meaningful measurement. It also poses a problem in AE data analysis when multiple sensors are needed in a multi-cylinder diesel engine such as in Refs. [13–15]. Under such circumstances, extensive expert knowledge is needed to correctly interpret the information conveyed in the AE signals and the analysis of AE data can only be carried out in a qualitative manner.