EP1955030A2

EP1955030A2 - Method for analyzing the noise of an internal combustion engine

Info

Publication number: EP1955030A2
Application number: EP06817472A
Authority: EP
Inventors: Franz Brandl; Werner Biermayer
Original assignee: AVL List GmbH
Current assignee: AVL List GmbH
Priority date: 2005-12-01
Filing date: 2006-11-30
Publication date: 2008-08-13
Also published as: WO2007062447A3; WO2007062447A2

Abstract

The invention relates to a method for quantitatively analyzing a noise of an internal combustion engine, especially a diesel engine. According to said method, a signal progression is determined over a predefined recording period, the determined signal is subjected to bandpass filtering in the time domain, and an envelope curve for the bandpass-filtered signal is formed for at least one frequency band. The subjectively perceived nuisance of noise containing impulses is measured objectively by carrying out the following steps: temporal masking according to human auditory perception is taken into account in the envelope curve; the remaining degree of modulation in each frequency band is calculated after taking into account temporal masking; level-dependent masking is calculated for each of all other frequency bands; the remaining audible degrees of modulation are calculated for all frequency bands; the audible degrees of modulation are weighted with a frequency-dependent weighting function; and a noise index is created based on the weighted modulation frequency spectrum.

Description

Method for analyzing a noise of an internal combustion engine

The invention relates to a method for the quantitative analysis of a noise of an internal combustion engine, in particular a diesel internal combustion engine, wherein a signal waveform over a predetermined recording period is determined, the detected signal is subjected in the time domain of a bandpass filtering and formed for the bandpass filtered signal for at least one frequency band an envelope becomes. Furthermore, the invention relates to a method for determining the forces at the force introduction centers of a vehicle body with respect to the drive unit and the suspension as a basis for determining the vibration transmission of a vehicle body.

The speed of the combustion process in internal combustion engines is relatively high at low engine and intake air temperatures due to a delayed ignition delay, which leads to rapid pressure increases and thus to impulsive engine noises. These impulsive noises are referred to as "burn nails", especially in diesel engines, and they preferably occur at idle cold engine and full load acceleration operations from a low load / RPM range. They are subjectively uncomfortable, especially in diesel engines, but also felt in direct-injection petrol engines. It is therefore one of the goals in the development of acoustics of internal combustion engines to reduce this subjectively unpleasantly felt nail noise.

For such an acoustic optimization, it is a tremendous advantage if the subjectively annoying impression of the nail noise can be determined by an objective measuring method, since this can clearly define the extent of initial states and improvement steps.

From EP 1 462 778 A1 a method for the quantitative analysis of engine noise is known. A time course of the measured noise is first subjected to a short-time frequency analysis in overlapping time windows. The short-term frequency spectrums obtained in this case are, for each carrier frequency, again taken to be a temporal carrier signal whose modulation frequency spectrum is calculated. Characteristic features for pulse-like components of the engine noise in the modulation frequency spectra are quantified by calculating a noise index. This should enable a prediction based on this to be made of the human assessment of engine noise. The disadvantage is that peculiarities of human hearing tion, such as temporal or spectral concealment are not taken into account. Therefore, the validity of the thus determined noise index for human auditory perception is not sufficient.

The vibration or force introduction and transmission in a vehicle body is usually carried out by means of the so-called "transfer path analysis" (TPA - Transfer Path Analysis). At the points of introduction of force into the body, the input intertances as well as the transfer functions from the force introduction points to microphones in the vehicle interior and vibration measurement points to the body are measured by external excitation (shaker, hammer, etc.). The effect of the real suggestions under vehicle operation on the vehicle body is determined by measuring the accelerations in the vehicle operation at the force introduction points and using the previously measured inertances and transfer functions. In the context of inertance measurements, however, metrologically caused false results at individual frequencies (in the event of inconsistencies in coherence). In addition, certain introductory phenomena such as torque introductions via the elastic elements, etc. are not sufficiently considered. This currently used TPA analysis therefore leads to unrealistic force transmission under vehicle operation via the calculation by means of the inertia matrix. This results in false results with respect to vibrations and noises in the vehicle interior in relation to the shares of the individual force discharges and transmission paths.

The object of the invention is to avoid these disadvantages and to develop a method with which the subjectively perceived annoyance of impulsive noises, in particular of the "burning manure", can be objectively measured. A further object of the invention is to enable as realistic a determination as possible of the force introduction and thus of the vibration transmission properties of a vehicle body.

According to the invention this is achieved by the following steps:

Taking into account temporal occlusion in accordance with human auditory perception of the envelope;

Calculating the residual modulation level in each frequency band after taking into account the temporal occultation;

Calculating a level-dependent occlusion for each frequency band from all other frequency bands; Calculating the remaining audible modulation levels for all frequency bands;

Weighting the audible modulation degrees with a frequency dependent weighting function;

Forming a noise index based on the weighted modulation frequency spectrum.

In order to make the subjective hearing impression objectively measurable, the processes and properties of human hearing in the time domain are largely mathematically and signal-analytically modeled in the context of the invention. The entire calculation takes place in the time domain and is largely modeled on the characteristics of human hearing.

The impulsive sound (signal) is recorded as airborne sound by means of a microphone. Thereafter, band pass filtering of the signal in the time domain is performed for all frequency bands with filters that do not produce phase shift ("zero-phase" filters), e.g. FFT filter, FIR filter or the like. For this bandpass filtered signal, an enveloping envelope is formed for each frequency band. Since rapidly successive impulsive noises from a certain pulse frequency from the human ear due to temporal occlusion can not be resolved individually, the envelopes are subjected to the temporal occlusion of human hearing. This temporal occlusion can reduce subsequent impulses in their subjectively perceived amplitude in the case of rapidly consecutive impulsive noises. Therefore, in this step, the residual degree of modulation perceived by the human hearing in each frequency band is calculated after considering the temporal occultation.

Human hearing is also limited by frequency-dependent occlusion, so the results of the different frequency bands are evaluated by frequency masking curves of human hearing. For the impulsive noise in each frequency band results in a different subjective sense of annoyance, depending on the frequency range of the respective pulse-containing signal. Based on subjective assessments of subjects, a frequency-dependent weighting function is formed, which is finally used as the frequency-dependent weighting of the signal for each frequency band.

The sum of the remaining signal amplitudes from the analysis and calculation steps for each frequency band gives an objective measure as the final result for the annoyance of impulsive noises, in particular of nail noise of an internal combustion engine. This objective measure of the annoyance of impulsive noises thus correlates particularly well with the subjectively perceived overall impression of these noises.

A realistic determination of the force introductions and the vibration transmission properties of a vehicle body is achieved by the following steps:

a. Defining the main force application points of the drive train and the suspensions in the body of the vehicle;

b. Defining a number of reference points for measuring the local vibrations on the body of the vehicle, wherein preferably the number of reference points is greater than, preferably at least twice as large, more preferably at least ten times as large as the number of main force introduction points;

c. Determining accelerations and inputances at the main power take-off points of the powertrain and the suspension by means of external excitation, for example by shakers or hammers;

d. Determining the transfer functions between the main force application points and all reference points on the body as well as microphones inside the vehicle by means of external excitation;

e. Determining so-called primary accelerations at the main force introduction points and the reference points during operation of the vehicle;

f. Estimating the forces introduced into the body at the main force application points due to the primary accelerations measured in the operating state and the detected inputances;

G. Performing a first mathematical prediction of so-called secondary accelerations at the reference points of the body based on the estimated introduced forces at the main force introduction points and the determined transfer functions between the main force introduction points and the reference points;

H. Comparing the calculated secondary accelerations with the measured primary accelerations at the reference points; i. Mathematical iteration of the forces introduced into the body to minimize the difference between secondary and primary accelerations;

j. Repeating steps h) through i) until the difference between secondary and primary accelerations becomes a minimum or less than a defined maximum permissible difference, with associated force vectors being determined;

k. Assigning real force vectors at the force introduction points from the force vectors obtained in the iteration, and

I. Calculation of the effect of all real force vectors at the force application points on the vibrations of the body, on the measured by microphones interior noise and by means of the determined in step d) transfer functions, so that the noise and vibration contributions of drive train and suspensions on the vehicle interior noise, or to determine the vibration behavior of the body.

The advantage of the present method according to the invention consists in a largely exact determination of the introduced forces in a vehicle body. In this case, a much larger number of vibration measuring points are used in the vehicle interior as force introduction points are present. Thus, individual errors in the inertance measurement and the aforementioned simplified assumptions (such as no torque introduction, etc.) can be largely eliminated by the large number of existing vibration measurement points (high overdetermination), so that the overall result of the introduced forces far better than reality the conventional transfer path analysis. Thus, far more targeted improvement measures can be taken at the noise sources, at the force application points and at the transmission paths of the body to minimize the internal noise and vibration behavior.

The invention will be explained in more detail below with reference to FIGS. Show it:

1 shows a bandpass filtered time signal of a noise.

FIG. 2 shows the time signal together with an envelope; FIG.

Fig. 3 shows the time signal with an envelope of temporal occlusion of human hearing;

Fig. 3a shows the decay time of the human ear; 4 shows a frequency-dependent masking of human hearing;

5 shows a frequency-dependent weighting function of the subjectively perceived annoyance impression of impulsive noises;

6 shows a correlation of the subjective annoyance impression with the calculated noise index;

7 shows the method according to the invention in a block diagram; and

Fig. 8 is a schematic view of the measuring arrangement.

In Fig. 1, the amplitude A is plotted against time t for a bandpass filtered time signal of a noise. As shown in FIG. 2, an enveloping curve I ^{1 is} formed for this band-pass filtered time signal for each frequency band. In a further step, the envelope of temporal masking of the human hearing of the envelope 1 'is imagined (FIG. 3). This temporal occlusion can reduce subsequent impulses in their subjectively perceived amplitude A in the case of impulsive noises following one behind the other. Therefore, the residual degree of modulation perceived by human hearing in each frequency band is calculated after taking the temporal concealment 2 'into account. The decay time t of the human ear with pulse-like noises, which is used in the calculation, is shown in Fig. 3a, where LH is the loudness.

As shown in Fig. 4, a frequency-dependent occlusion is also considered. The frequency-dependent occlusion is indicated in Fig. 4 by the curves 3 '. This takes into account that a signal from higher-frequency signals can be obscured.

For the frequency-dependent occlusion the following slope steepnesses are used.

For higher frequencies than the center frequency of the frequency band to be examined:

Occlusion links = level 27 * (b - bm)

Level .... Level in dB that leads to occlusion b ... Frequency in Bark at the level that leads to occlusion. bm ... frequency in Bark which is covered For lower frequencies than the center frequency of the frequency band to be examined:

Slope Right = Level + (-24 -0.23 / (fm / 1000) + 0.2 * Level) * (bm-b)

fm ... Frequency in Hz which is obscured

Depending on the frequency range of the respective pulse-containing signal results for the impulsive noise in each frequency band, a different subjective sense of annoyance.

The degree of modulation in each frequency band is calculated according to the following formula:

MG = LM _3X ~ maXimUm (LMjn, cover-timeΛ U / detection-frequency)

MG ... degree of modulation in each frequency band

L _Max. Maximum level of the envelope

L _min .. minimum level of the envelope

Masking time ■ level of temporal occlusion

Masking frequency - sum of all levels from frequency masking

The subjective perception of the impulsiveness is taken into account on the basis of subjective assessments of test persons in a frequency-dependent weighting function W shown in FIG. By means of the frequency-dependent weighting function W, the modulation degrees for each frequency band f are subjected to a frequency-dependent weighting.

The sum of the remaining signal amplitudes from the analysis and calculation steps for each frequency band gives as an end result an objective measure of the annoyance of impulsive noises, in particular nail sounds of an internal combustion engine. This objective measure of the annoyance of impulsive noises - the noise index CKI (Combustion Knocking Index) - correlates very well with the subjectively perceived overall impression of these noises. In Fig. 6, the correlation between the noise index CKI and the subjectively perceived annoyance L of nail sounds is plotted. This results in a high correlation of, for example, 0.991.

First, the main force application points 11 of the drive train and the suspensions are defined on the body 10 of the vehicle. It will continue Defines a large number of reference points 12 for measuring the local vibrations on the body 10 of the vehicle, wherein the number of reference points 12 is greater by an order of magnitude than the number of main force introduction points 11. By means of external excitation 13, or by a Sha- ker or hammer, accelerations and Eingangsinertanzen at the main force application points 11 of the drive train and the suspension are determined and used to determine the transfer function between the main force introduction points 11 and all reference points 12 on the body 10.

The most essential steps of the method are shown schematically in FIG.

In step 1, so-called primary accelerations Pj are determined at the main force introduction points 11 and the reference points 12 during operation of the vehicle. Based on a rough estimation 2 of the forces introduced into the body at the main force introduction points 11 due to the primary accelerations Pj measured in the operating state and the determined inputances, a first mathematical prediction of so-called secondary accelerations Si at the reference points 12 of the body 10 of the estimated one is estimated in step 3 introduced forces at the main force introduction points 11 and the determined transfer functions between the main force introduction points 11 and the reference points 12 performed. The secondary accelerations S ₁ are compared in step 4 with the measured primary accelerations P i in the reference points 12. From the difference, an error function 5 is formed. If the difference between secondary and primary accelerations Pi is greater than a permitted maximum difference, an iterative adaptation of the "real" body-initiated forces at the main force introduction points 11 is made in step 6 and the steps 3 to 6 shown in FIG. until the difference Pi - Si between secondary Si and primary accelerations Pj is smaller than a predefined maximum difference. If the difference P ₁ -Si between secondary S ₁ and primary accelerations Pj is smaller than the allowed maximum difference, the determined force vectors are regarded as "real" force vectors (step 7 in FIG. 7).

The secondary accelerations (airborne sound) Sj (ω) can be determined according to the following equation:

where Ty (ω) is the transfer function (acceleration-sound pressure / force), measured under external excitation, and f _j (co) are the secondary forces. Where i = I ₁ ..., N is the number of microphones and accelerometers, j = 1, ..., M the number of force components at all considered force introduction points 11. With ω the frequency range is designated.

The condition applies:

Σ ((P ₁ (CO) -S ₁ (P) Y = Mn. (2) i

where P | (ω) is the primary acceleration (airborne sound) measured while driving. Equation 2 thus yields an optimal set of calculated real force vectors fj, _O p _t (ω).

Claims

PATENT CLAIMS

1. A method for the quantitative analysis of a noise of an internal combustion engine, in particular a diesel internal combustion engine, wherein a signal waveform over a predetermined recording period is determined, the detected signal is subjected in the time domain of a bandpass filtering and for the band pass filtered signal for at least one frequency band an envelope is formed characterized by the following steps:

Calculating a level-dependent occlusion for each frequency band from all other frequency bands;

Calculating the remaining audible modulation levels for all frequency bands;

Forming a noise index based on the weighted modulation frequency spectrum.

2. The method according to claim 1, characterized in that the fire-pass filtering is preferably carried out with filters that do not allow phase shift for all frequency bands.

3. The method according to claim 1 or 2, characterized in that an envelope of the filtered time signals is calculated for each frequency band.

4. A method for determining the forces at the force introduction centers of a vehicle body with respect to the drive unit and the suspensions as a basis for determining the vibration transmission of a vehicle body comprising the following steps:

a) Defining the main force application points of the drive train and the wheel suspensions in the body of the vehicle; b) defining a number of reference points for vibration measurement on the body of the vehicle, wherein preferably the number of reference points is greater than, preferably at least twice as large, more preferably at least ten times as large as the number of main force introduction points;

c) determining accelerations and inputances at the main load transmission points of the powertrain and the suspension by means of external excitation;

d) determining the transfer functions between the main force application points and all reference points on the body as well as to microphones inside the vehicle by means of external excitation;

e) determining primary accelerations at the main force application points and the reference points during operation of the vehicle;

f) estimating the forces introduced into the body at the main force introduction points based on the primary accelerations measured in the operating state and the determined input inertances;

g) performing a first mathematical prediction of secondary accelerations at the reference points of the body based on the estimated applied forces at the main force introduction points and the determined transfer functions between the main force introduction points and the reference points;

h) comparing the calculated secondary accelerations with the measured primary accelerations at the reference points;

i) mathematical iteration of forces introduced into the body to minimize the difference between secondary and primary accelerations;

j) repeating steps h) to i) until the difference between secondary and primary accelerations becomes a minimum or less than a defined allowable maximum difference, the associated force vectors being determined;

k) assigning real force vectors at the force introduction points from the force vectors determined as a result of the iteration and I) Calculation of the effect of all real force vectors at the force application points on the vibrations of the body, on the measured by microphones interior noise and by means of the transfer functions determined in step d), so as to reduce the noise and vibration amounts of powertrain and suspensions on the vehicle interior noise, or to determine the vibration behavior of the body.

5. The method according to claim 4, characterized in that the force distributions at the force application points, the position of Hauptkraftein- line points, the insulating properties and / or the transfer functions of the body are selectively varied to positively influence the interior noise and the vibrations of the body.