DE102008046111A1 - Method for continuous determination of acoustic quality of road from noise into toroidal area of smooth tire in flowing traffic area, involves determining outdistancing and texture levels as acoustic quality characteristics of road lining - Google Patents

Abstract

The method involves determining a 7.5 m outdistancing level and a texture level with a series of m-digital single mass-resonators with resonant frequency and a stage ratio as parallel resonators for a torus-noise, and with spectral transfer functions as acoustic quality characteristics of a road lining. A time discrete torus signal given as a road section is acted at a spring (C) of the single mass-resonators. A deflection, momentary resonant energy and a momentary power are determined by digital integration and assigned to spectral power density of the torus signal.

Description

The task is to determine the acoustic quality of roads. The method should be applicable in flowing traffic and also in residential areas and the ISO standard 11819 "Acoustics: Method for measuring road surface on traffic noise". Specifically, it is about the continuous, continuous determination of the 7.5 m pass-by level and the dB (A) -weighted texture level with an exact assignment of measured value and measuring location and the fluctuations of these parameters. This is intended to detect the noise effect of fluctuating and impulsive noises.

To evaluate the rolling noise, the internationally standardized 7.5m pass-by level is used. Such a measurement is relatively cumbersome and time consuming, provides only punctual values and is not applicable to streets in residential areas because of reflective house walls. Measuring pendant after the ISO standard 11819 (2) with several microphones near the tire provide continuous values, the environmental noise is problematic here, especially due to wind and oncoming traffic. Despite standardization, the results of measurements with different measuring pendants are only partially comparable. The measurement of road textures with the laser profilometer provides only a geometric, linear image. Also, a tire has only a contact surface portion of only 20-60% to the road when rolling, so that the laser measurement reproduces the acoustically effective texture only incomplete.

In DE 197 26 608 For example, a measurement method has been proposed to use the noise in the torus space of a profileless tire to determine the acoustic quality of a road. When unrolling, unevennesses in the tires are reflected and produce noise levels in the tire torus of more than 130 dB. The torus noise can be measured continuously, even in flowing traffic without interference and with a repeatability better than 0.2 db (A). From the torus spectrum in a profileless tire, the associated 7.5m pass-by level and texture level are directly determined by linear transfer functions. With this method, a road length of approximately 80 km per hour can be measured. This allows a simple acoustic quality control both after installation and after prolonged rest of a road surface. Such a process ultimately serves to exhaust the large noise reduction potential on roads over 5 db (A).

• (i) Die Ungenauigkeit der herkömmlichen satellitengestützen GPS-Navigation resultiert aus der bis heute ungeklärten sog. Pioneer-Anomalie. Dieses Phänomen ist offensichtlich entfernungsabhängig und wurde erst bei den extrasolaren, den Anziehungsbereich der Sonne verlassenden Pioneer-Sonden bemerkt. Innerhalb deren 15-jährigen Missionszeit und bei einem Abstand von 10¹⁰ km wuchs die Fehlerdifferenz von gemessener und gerechneter Position auf 1 Mill. km an. Andere Satellitenmissionen, insbesondere Flyby-Manöver zeigten ebenfalls Diskrepanzen. Bei den Navigations-Satelliten mit einer mittleren Entfernung von 20 000 km ist die Pioneer-Anomalie entsprechend geringer, beeinträchtigt aber die Triangulation.

Unresolved here is the problem of (i) exact assignment of a sound measurement to the measurement location. Sol / actual trays of the track are extremely sensitive. With retracted ruts a tire width is up to 3 db (A). Other (ii) deals with the recording of changes over time. The problem is the natural, statistical fluctuation of narrow-band spectra and the process-related leveling of real fluctuations by the cyclic operation in the FFT analysis.

• (i) The inaccuracy of conventional satellite-based GPS navigation results from the so far unexplained so-called Pioneer anomaly. This phenomenon is obviously distance-dependent and was first noticed in the extrasolar Pioneer probes leaving the Sun's gravitational field. Within its 15-year mission period and at a distance of 10 ¹⁰ km, the error difference between measured and calculated position increased to 1 million km. Other satellite missions, especially flyby maneuvers, also showed discrepancies. In the navigation satellites with an average distance of 20 000 km, the Pioneer anomaly is correspondingly lower, but affects the triangulation.

• (ii) Um aus der aktuellen Messung des Torus-Geräusches die maßgebenden, gewichteten Kennwerte einer Straße, z. B. den Vorbeifahrt- oder den Texturpegel zu bestimmen, sind jeweils Spektralanalysen notwendig. Mit der Fast-Fourier-Analyse steht zwar ein ausreichend schnelles Verfahren zur Verfügung, dieses arbeitet aber nicht unmittelbar zeitgleich, sondern im Taktbetrieb. Gerade beim Fourier-Verfahren ist ein Signalimpuls innerhalb eines Taktes nicht lokalisiert. Einer Verkleinerung der Taktlänge zur besseren Ortsauflösung steht aber eine höhere Streuung gegenüber. In der Quantenphysik wird diese Abwägung als grundsätzliches physikalisches Phänomen, als Unschärferelation behandelt. – Auch der Übergang zur Wavelet-Analyse erscheint wenig Erfolg versprechend, so verhält sich das gebräuchliche Haar-Wavelet komplementär zu Fourier, es ist zwar ausreichend ortsgenau, ist dafür aber im Frequenzbereich schlecht lokalisiert.

The Pioneer anomaly is attributed here to a Doppler disposition. Thus, the Doppler effect in the general case can not be expected with a constant wave carrier. In the case of airborne noise, the wind speed must be considered as a priority. A temporally changing air movement induces besides the normal, in addition an abnormal Doppler shift; in particular, an expansion of the air has a similar anomaly pattern to the Pioneer measurements. To simulate such anomaly was after DE 10 2008 036 812.2 introduced in the optical Doppler measurement the cosmological expansion of space. Their kinematics is characterized by the single Hubble parameter H, after which an expansion speed u = Hr is induced at a distance r. The Hubble constant H = 1.2 · 10 ^-14 [1 / s] determined from the Pioneer measurements refers to the dark matter as a carrier of the Doppler waves and is 10 ⁴ times larger than that for the visible masses, astronomical Hubble factor H ₀ . A Doppler measurement at an airspeed v and with the frequency f provides the frequency shift Δf _meas = (-v / c + u / c) f in this model. (c = speed of light). In addition to the regular Doppler shift Δf _Norm = -fv / c, an anomalous frequency shift Δf _Anom = fu / c = fHr / c occurs in addition. With the correction _rule Δf _standard = Δf _measuring Δf _Anom , the true Doppler value Δf _{norm is} formed from the raw value Δf _measurement . The simplest correction Δf _Anom = fH <r> / c consists of _calculating the distance from the satellite to the earth's surface flat with a mean distance <r>. In the nearest best approximation, the average latitude <r> takes into account the geographical latitude of the navigation device. Finally, in the case of exact correction, the orbits of the satellites are stored in the navigation device chert, so that the actual distances r are known together with the measuring location and can be taken into account according to Δf _Anom = fHr / c.

• (ii) To obtain from the current measurement of the torus noise the authoritative, weighted characteristics of a road, e.g. For example, to determine the pass or texture level, spectral analyzes are necessary. Although fast-Fourier analysis provides a sufficiently fast method, it does not work immediately at the same time, but rather in clock mode. Especially in the Fourier method, a signal pulse is not localized within a clock. A reduction of the cycle length for a better spatial resolution is contrasted with a higher dispersion. In quantum physics this consideration is treated as a fundamental physical phenomenon, as a uncertainty principle. - The transition to the wavelet analysis appears to be less promising, so behaves the common hair wavelet complementary to Fourier, although it is sufficiently accurate, but it is badly localized in the frequency domain.

In contrast, analogue data processing offers simple solutions for simultaneous playback of frequency-weighted characteristic values. If z. B. a resonator with the resonant frequency f ₀ and a loss factor η acted upon by the signal to be analyzed, this synchronously registers the lying in the frequency range .DELTA.f = ηf ₀ signal energy. A further advantage is that the resonator with respect to the arithmetic frequency gradation in Fourier directly provides a geometric gradation according to the third-octave analysis used in practice. Nevertheless, in order to be able to retain the advantage and versatility of digital data processing, digital resonators are introduced here. These resonators can be realized very closely f ₀ = {f _oi } and with different loss factors η = {η _k }, so that in forming a stable average not only over time, but also over the neighboring frequencies {f _oi }, loss factors {η _k }, different types of resonators and different types of excitations can be averaged.

The model resonator in 1 establishes a relationship between the analysis signal and the resonator signal and is able to supply a geometrically stepped, continuous frequency spectrum from a signal given as a train train. In 2 is the signal to be analyzed as a staircase curve - as a piecewise constant function - before. In both cases, for the sake of clarity, the signal s = s (t) was mapped as path excitation. In both cases the calibration is done via defined noise.

1 : Given is a resonator with the mass M, the spring C and the damping D and have the complex natural frequency f ₀ (1 + iη). This resonator experiences a path excitation s (t) = (s _n ) given by a time series with the signal values s _n at the time interval Δt. The calculation of the temporal oscillation profile y (t) = {y _n } of the resonator mass M for a given path profile s (FIG. t) is a classical basic task for which there are a number of programs In a time-delayed calculation, at the current time t _n = nΔt with the deflection y _n not only the path disturbance s _n but already the following disturbances s _{n + 1} , s _{n + 2} ... Thus, faster convergent algorithms can be used with which the course of the oscillation energy A = A (t) of the resonator of interest as a function of the path excitation s in the time domain t can be calculated.

In frequency range f, the resonator has the resistance R = R (f) and absorbs the power W. R (f) = 2πηMf 0 f 2 / [(1 + f 0 / F) 2 (f - f 0 ) 2 + η 2 f 2 dW = w (f) R (f) df

The fast intensity w = w (f) can be attributed to the Fourier spectrum s (f) of the signal s (t), w = [2πfs (f)] ² . For integration, the variable z = f - f _{0 is} introduced, which has a pronounced, finely limited maximum at z = 0, ie in the case of resonance f = f ₀ . With good approximation, the intensity w (f) → w (f ₀ ) and the frequency f → f _{0 can} be assumed to be constant as slowly changing quantities in this range. This gives the power W and the stored energy A of the resonator. W → π 2 w (f 0 ) f 0 2 M [W] A → W / 2πf 0 η [Ws]

The method exemplified here with the resonance frequency f ₀ can be expanded with f _{0 -f 0m} to other resonance frequencies f _0m . A geometric frequency _gradation f _0m = f ₀ q ^m yields an octave analysis with q = 2 and a ^third octave analysis with q = 2 ^1/3 . (m = 1, 2, 3 ...). Grading q and loss factor η are independently selectable.

The time average A = <A _{n> T} from the A _n values is as shown above, is directly related to the spectral intensity, w (f ₀₎ at the resonance frequency f ₀ and can be calibrated with a defined noise. Averaging time T = 25 μs gives the pulse and T = 125 μs a fast rating

2 : Again, a resonator of the mass M, the spring C and the damping D is given and is here stimulated by a step function s = {s _n } with the steps Δs _{n + 1} = {s _{n + 1} - s _n }. With the exception of the discontinuities, the resonator performs a force-free oscillation with the complex resonance frequency ω ₀ = ω '+ iω''and the real amplitude a = a (t). It is expedient to replace the amplitude a by the total oscillation energy A = A _kin + A _pot = 2A _{pot of} the resonator. With A _pot = ½a ² C the amplitude becomes a = √ (A / C) and we get y = y (t) and the fast y ° = y ° (t) for the rash y = √ (A / C) sin (ω 0 t + φ) y ° = ω 0 (√ (A / C) cos ω 0 t + φ)

Without limiting the generality, at the time t = 0, due to the step-shaped path excitation Δs, a jump of the deflection y takes place by Δy = Δs. The fast remains with Δy ° = 0 erhal th, therefore also the kinetic energy, ΔA _kin = 0. Thus, in a linear approximation, the total vibration energy A undergoes a change ΔA and the phase φ a change Δφ ΔA = 2√ (AC) sinφΔs Δφ = √ (C / A) cosφΔs

After the discontinuity n the resonator has the oscillation energy A _n and the phase φ _n . During the time .DELTA.t _n performs a damped oscillation, the resonator of the energy corresponding to the damping A _n ω 'decreases exponentially and the phase increases by Δφ _n. The new state before the next jump is A _n 'and φ _n '. A n '= A n exp (-ω''Δt n ) φ n '= φ n + ω'Δt n

After the jump Dy _{n + 1} = .DELTA.s _{n + 1} = s _{n + 1} - s _n is the state of A _{n + 1} and φ _{n + 1} A n + 1 = A n '+ 2√ (A n 'C) sinφ n '.DELTA.s n + 1 φ n + 1 = φ n '+ √ (C / A n ') Cos n '.DELTA.s n + 1

The sudden energy change ΔA _n applied to the time interval Δt _n results in the 1 analog, absorbed by the single-mass resonator oscillation output W _n = .DELTA.A _n / _n and .DELTA.t is in equilibrium with the attenuation loss (A _n - A _n ') / _n .DELTA.t.

This completes one cycle and repeats the process. The jump values ΔA and Δφ can be read directly from a two-dimensional table with the parameters Δs / a and φ without time-consuming iteration calculation. Since the resonance frequency ω ₀ and the time step Δt remain constant during an analysis, the change during the undisturbed oscillation in the time interval Δt _n can also be integrated into the table. The main advantage, however, is that instead of the linear calculation, the table is filled with the exact values, thus saving computation time.

In the 1 and 2 the signal to be analyzed s = {s _n } was interpreted as a path. In the same way s can be realized as a force in a digital resonator with force excitation at the mass M. Similarly, oscillators with complex mass M = M '+ iM''and / or complex spring C = C' + iC '' can be used as a frequency analyzer. To replicate weighted frequency responses, it is possible to expand to two-mass and multi-mass oscillators instead of a one-dimensional resonator. The characteristic of a bandpass filter are the so-called tuned mass damper with coinciding resonance frequencies. High- and low-pass characteristics are made up of dampers D and springs C and aggregates and vibration absorbers with cut-off frequencies. Since there is a causal relationship between resonator energy A and the spectral exciter signal s = s (f), these arrangements can also be used for frequency analysis.

In the example of 1 the time-discrete A _n and W _n values were averaged according to the standardized pulse, fast or slow evaluations. The smaller the averaging time, the more disturbing are the statistical level fluctuations and the greater the more short-term sound events are leveled. The quick table reading after 2 allows averaging process to detect even short-term signals with arbitrary accuracy and time resolution. For this purpose, parallel to the resonator with the resonant frequency ω ₀ = ω '+ iω''additional resonators in the - small - distance Δω ₀ = Δω _i ' + iΔω _k '' operated. The greater the number of parallel resonators used, the more accurate and unobtrusive is their mean value.

QUOTES INCLUDE IN THE DESCRIPTION

This list The documents listed by the applicant have been automated generated and is solely for better information recorded by the reader. The list is not part of the German Patent or utility model application. The DPMA takes over no liability for any errors or omissions.

Cited patent literature

• - DE 19726608 [0003]
• - DE 102008036812 [0005]

Cited non-patent literature

• - ISO standard 11819 [0001]
• - ISO standard 11819 [0002]

Claims (13)

A method for continuously determining the acoustic quality of roads from the noise in the torus space of a flat tire, characterized in that with a series of m single-mass digital resonators with resonant frequencies f _0m = f ₀ · q ^m and the step ratio q as, parallel spectral analyzers for the torus Noise and with spectral transmission functions of the 7.5m pass-by levels and the texture levels as acoustic quality features of a road surface. Method according to Claim 1, characterized in that the time-discrete torus signal s = {s _n }, which is given as a path train, acts as path excitation on the spring of a single-resonator of the resonant frequency f _m , from which the deflection and the instantaneous resonator energy A = {A _n } and the currently converted power W = {W _n } is determined by means of digital integration and assigned to the spectral power density of the torus signal. Method according to Claim 1, characterized in that the digital torus signal s = {s _n } given as a staircase curve acts as path excitation on the spring of a single-resonator of the resonant frequency f _m , from which the deflection and the instantaneous resonator energy A = {A _n } from the jumps Δs _{n + 1} = {s _{n + 1} - s _n } determined according to the rule A ' n = A n exp - (ω''Δt n ) φ ' n + 1 = φ n + ω'Δt n A n + 1 = A ' n + 2√ (A n C) sinφ ' n .DELTA.s n + 1 φ n + 1 = φ ' n + √ (C / A ' n ) Sinφ ' n .DELTA.s n + 1 and the power converted at the nth jump point W = {W _n } W n = 2√ (A n-1 C) sinφ n-1 / .DELTA.t is assigned to the spectral power density of the analysis signal. Method according to claim 3, characterized that from a two-dimensional table with the parameters Δs / a and φ the state ΔA, Δφ after a jump Δs is read. Method according to claims 1 to 4, characterized in that to improve the time resolution parallel to a resonator with the frequency ω ₀ = ω '+ iω''further resonators at a distance Δω _0m = Δω _i ' + iΔω _k '' are operated and their Signals over i, k, m = 1, 2, 3 ... are averaged. Method according to claims 1 to 5, characterized in that in order to improve the time resolution in parallel with excitation-excited resonators with the frequencies ω _0m = ω _i '+ iω _k ''force-excited resonators with the distance Δω _{0m *} = Δω _{i *} ' + iΔω _{k *} '' and be averaged over i *, k *, m * = 1, 2, 3 .... Process according to claims 1 to 6, characterized characterized in that for the reproduction of frequency responses digital two- and multi-mass resonators are used. Process according to claims 1 to 7, characterized characterized in that for the realization of bandpass filters digital Tuned mass damper used with coincident natural frequencies become. Process according to claims 1 to 4, characterized characterized in that for the reproduction of high and / or low-pass filters digital units consisting of springs C and dampers D are used. Process according to claims 1 to 4, characterized in that for the reproduction of high and / or Low pass filters digital vibration absorbers with cut-off frequencies be used. Process according to claims 1 to 10, characterized in that the digital resonators, aggregates and vibration absorber with defined sound and / or noise signals be calibrated. Method according to claims 1 to 5, characterized in that for the purpose of improving the GPS localization, the true Doppler frequency shift Δf _{norm is determined} from the measured Doppler frequency shift Δf _measurement according to Δf _Norm = Δf _{measurement -Δf Anom} wherein the correction _amount Δf _Anom = fu / c = fHr / c, which depends on the Doppler frequency f, on the distance r between the satellite and the navigation device, the speed of light c and the Hubble expansion factor H = 1,2 · 10 ^-14 [1 / s]. A method according to claim 11, characterized in that the correction _amount Δf _Anom = fHr / c approximately the average value <r> of the satellite distances is used.