JP2001194224A - Method for predicting road noise while taking reflected sound into account - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a predicting method for road noise which is applicable without any restriction and takes into account a reflected sound enabling road traffic noise to be predicted as an absolute quantity. SOLUTION: This method consists of a 1st step stage for finding a sound pressure level by a border element method and the energy sum of all frequency components by dividing the part between object frequency bands by specific frequency division width Δf and then finding the insertion loss ΔLP (dB) of a noise propagation model taking a structure as a sound reflecting surface into account on the basis of a direct propagation model omitting the mentioned structure at a noise prediction point and a 2nd step stage for finding a direct propagation noise level LF (dB) propagated from a road noise source directly to a noise prediction point while taking only distance attenuation into account and also finding the total noise level LR taking the reflected sound into account from the direct propagation noise level LF (dB) and insertion loss ΔLP (dB).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for predicting road noise in consideration of reflected sound using a boundary element method, which is applied to, for example, a multi-story road structure having an elevated vehicle exclusive road on a general road. About.

[0002]

2. Description of the Related Art In recent years, in the metropolitan area, general roads (hereinafter referred to as "general parts") have been disposed under elevated roads of automobile-dedicated roads (hereinafter referred to as "dedicated parts") in terms of utilization efficiency of road land. A so-called two-layered road structure has been widely adopted.

[0003] In recent road maintenance projects, sound insulation measures such as installation of sound insulation walls or sound absorbing plates are taken so as to satisfy the environmental standards in the law by sufficiently considering the influence of automobile noise on roadside areas. .

The planning and design of such noise countermeasures include:
At present, at the time of environmental assessment, it is generally performed by a so-called acoustic society formula (ASJ Model 1975, ASJ Model 1993) based on prediction calculation. However, the calculation method for the two-story road structure is not specifically specified in the formula of the Acoustic Society, so the specified plane, embankment, cut,
It was often replaced by superimposing general correction values for elevated roads. However, with such an approach,
The effect of noise propagation due to the interaction between the general part sound source and the dedicated part, such as the reflected sound below the digit, that is, the reflected sound is not faithfully evaluated, so the noise level that is much higher than the predicted value is two layers. It was observed along the road with a road structure, and in some cases, there were many problems that complaints were generated from residents along the road.
Especially when the sound insulation wall is high, the diffraction attenuation of the direct sound increases, and the reflected sound below the diffraction sound is more significant than the diffraction sound, there is a large difference between the predicted value and the measured value. In some cases exceeded 10 to 15 dB.

Therefore, in the noise prediction of a two-layered road structure, an effective noise prediction method that can correctly evaluate the reflected sound at the planning / design stage is required. Several methods have been proposed, such as the application of the cosine rule. However, these methods have limitations such as being able to consider only relatively low-order reflected sound, and having a narrow application range for the road cross-sectional shape that can be handled, and also consider the reflection between sound insulation walls in addition to the underside reflection under the girder. Has a problem that the calculation procedure becomes too complicated to be practical.

In view of such a background, the present inventors have disclosed in Japanese Patent Application Laid-Open No. 10-206227, a two-layer road structure using a Monte Carlo simulation which is one of general-purpose numerical analysis methods using random numbers. A simple noise calculation method (hereinafter referred to as Monte Carlo method) was proposed. In the case of this method, it is relatively easy to evaluate by comprehensive model formation regarding viaducts, sound insulation walls, positional relationship of road surface, girder shape, surrounding terrain, arrangement position of sound absorbing material, etc., and multiple reflection can be considered. The predicted value and the measured value are well matched, and it has been confirmed that the method is practically effective as a noise calculation method in a two-layer road structure.

[0007]

However, even in the case of the above-mentioned Monte Carlo method, since the analysis principle uses a geometric acoustic method, for example, the diffraction sound of multiple reflection sound becomes a dominant factor. The problem is that the error increases in principle under the conditions where the effect of the sound wave phenomenon is remarkable, such as when the amount of opening between the viaduct girder and the sound insulation wall is narrow and the diffuse radiation at the opening is remarkable. was there. For this reason, in the Monte Carlo method, a correction coefficient using the opening amount D as a parameter is provided based on the result of noise measurement of an actual two-layer road structure, but the applicable range is limited.
A condition of ≧ 2.0 m is provided.

On the other hand, in recent years, as one of effective methods for analyzing a complex noise propagation model, there have been some research reports on applying a numerical analysis method using a boundary element method to a sound field analysis of a road structure. . However, these analysis methods are only used for limited sound field analysis such as sound insulation walls, trench structures, low-noise pavements, and reflection characteristics under viaducts. Ratio).

Therefore, a main object of the present invention is to eliminate the limitation of application, expand the application range, improve the prediction accuracy, and consider a reflected sound that can predict the road traffic noise as an absolute amount. An object of the present invention is to provide a noise prediction method.

[0010]

According to a first aspect of the present invention, there is provided a method for predicting road noise when a structure serving as an acoustic reflection surface is provided around a road surface. In each of the noise propagation model that takes into account the structure that becomes the sound propagation structure and the direct propagation model that omits the structure that becomes the acoustic reflection surface, a predetermined reference sound source is set at the noise generation position, and the boundary condition is set for each boundary element. Is set, and the frequency band to be processed is divided into predetermined frequency intervals Δf, and the sound pressure level at the noise prediction point for the sound source is determined for each frequency interval Δf by the boundary element method. At the same time, the sound pressure levels for each frequency step Δf are integrated to obtain the energy sum of all the frequency components, and then the energy sum of all the frequency components in the noise propagation model, Based on the energy sum of all frequency components in serial direct propagation model, the insertion loss of noise propagation model for the direct propagation model in the noise prediction point ΔL
A first step obtaining a _{P (dB),} with obtaining a direct propagation noise level L _{F (dB)} from considering attenuation only the road noise source is propagated directly to the noise prediction point,
This direct propagation noise level L _{F (dB),} the total noise level in consideration of reflected sounds in the noise prediction point based on the insertion loss of noise propagation model for the direct propagation model in the noise prediction point ΔL P _(dB) And a second step of obtaining _LR .

Next, a second invention according to a specific aspect is a method for predicting road noise when a structure serving as an acoustic reflection surface is provided around a road surface, wherein the structure serving as the acoustic reflection surface is considered. In each of the noise propagation model and the direct propagation model in which the structure serving as the acoustic reflection surface is omitted, a predetermined reference sound source is set at the noise generation position, and a complex acoustic impedance is set as a boundary condition for each boundary element. The sound frequency level at the noise prediction point for the sound source is determined by the boundary element method for each frequency step width Δf, and the obtained noise prediction point is obtained. Is corrected by the representative spectrum of the vehicle noise, the audibility of the A characteristic and the set frequency characteristic value of the sound source, and the corrected sound pressure level for each frequency step Δf is obtained. In each case, the corrected sound pressure level for each frequency step Δf is integrated to obtain the energy sum of all frequency components, and then the energy sum of all frequency components in the noise propagation model and the energy sum of all frequency components in the direct propagation model based on the sum, a first step the step of determining the insertion loss of noise propagation model for the direct propagation model ΔL P _(dB) in the noise prediction point, taking into account the distance attenuation only in the noise prediction point from the road noise source together directly obtained the propagated directly propagated noise level L _{F (dB),} and the direct propagation noise level L _{F (dB),} insertion loss of noise propagation model for the direct propagation model in the noise prediction point ΔL P _(dB) second obtaining from Doo following formula (I) the total noise level L _R considering reflected sound in the noise prediction point based on _{L R = L F + ΔL P} (dB) ... (I) And step process, is characterized in that consist.

[0012] In these cases, the direct propagation noise level L _F on the basis of the Acoustical Society of formula _{(ASJ Model 1975), L W} (dB) noise power levels of the sound source, the distance l between the sound source and the noise prediction point (m), as the average headway time of vehicle running sound source d (m), the following formula _{(II) L F = L W} -8-20 · log 10 l + 10 · log 10 (π (1 / d) tanh2π (1 / d)) (dB)… (II) or the Acoustical Society of Japan (ASJ Mo
del 1993), the A-weighted noise power level per vehicle after human hearing is corrected is set to L' _W (dB), and the number of sound source positions is set to m at intervals of Δl in the direction of the road. _{III) LF = L 'W -8-20} · logl i (dB) ... (III) here, _{l i} is the distance from the sound source ^{i *} th to noise prediction point, the following formula (IV) _{l i} = √ ( (l ₀ ² + (i ^* · Δl) ² )) (i ^* = 0, 1, 2,... m) (IV)

[0013] Even when the sound wave phenomenon is remarkable, it is effective to perform wave acoustic analysis in order to accurately and accurately perform noise prediction. However, to perform a wave acoustic analysis of road traffic noise, the wave equation is solved under the set boundary conditions, but the boundary conditions become very complicated in the target two-layer road structure. For,
It is difficult to solve analytically. In such a case, a numerical analysis method using the boundary element method is practically effective. The boundary element method has the advantages that the element division only needs to be performed on the boundary surface, and that consideration for the infinite region is naturally incorporated into the formulation, and it is much better than other numerical analysis methods such as the finite element method. It has the advantage that the calculation time and memory can be reduced.

In the actual road noise prediction, it is effective in a predetermined case to know the relative value before and after the sound insulation measure, but from the viewpoint of environmental standards and laws and regulations, the absolute value of the sound pressure level It is necessary to know.

Therefore, in the present invention, a sound field analysis is performed on a two-layered road structure using the boundary element method, and ASJ Model 1975
Or by combining with ASJ Model 1993, it is possible to calculate the noise level median L _A50 or equivalent sound level L _Aeq is an absolute value. In particular,
A noise propagation model that takes into account the structure that will be the acoustic reflection surface,
The analysis by the boundary element method is performed for each of the direct propagation models omitting the structure serving as the acoustic reflection surface, and the insertion loss (insertion loss) of the noise propagation model with respect to the direct propagation model at the same noise prediction point is calculated. If this insertion loss is reflected on the noise level (dB) directly propagated only by the distance attenuation at the noise prediction point, the total noise level including the reflected sound is calculated based on the direct noise level. According to this method, it is possible in principle to apply even a model having a large wave acoustic factor, which exceeds the application limit in the conventional method such as the Monte Carlo method.

In the present invention, in particular, in the analysis by the boundary element method, the frequency band to be analyzed is divided into predetermined frequency steps Δf, and each frequency step Δf is used at the noise prediction point for the sound source. The sound pressure level is determined by the boundary element method. As shown in FIG. 4, the sound pressure level of the noise fluctuates greatly in each frequency band, and if a specific frequency is analyzed in a spot manner, the noise cannot be evaluated faithfully. Therefore, the frequency step width Δf is set, the sound pressure level for each frequency is obtained, and these are integrated and evaluated as the energy sum of all the frequency components to improve the prediction accuracy.

Further, in the present invention, the sound pressure level at each frequency is corrected by the representative spectrum of the vehicle noise, the audibility of the A characteristic and the set frequency characteristic value of the sound source to improve the accuracy. ing.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below by taking a two-layer road structure as an example, formulating a sound field analysis of a two-layer road structure by a boundary element method, and a two-layer road by a boundary element method. This will be described in detail separately from the calculation method of the noise level of the structure.

[Formulation of sound field analysis of two-layer road structure by boundary element method] It is assumed that the cross-sectional shape does not change in the traveling direction of the road, and a viaduct 1 and left and right sound insulation walls 2 as shown in FIG. , 2 and a road surface 3 with reflection propagation (hereinafter referred to as a noise propagation model). Here, Ω _A represents a sound absorbing type boundary condition of a sound absorbing plate or the like, and Ω _R represents a high reflection type boundary condition of a concrete or steel material. Predicted position P ([p]) ([p] = (x, y))
<Hereinafter, [] notation represents a vector for convenience, and when it is expressed by the following mathematical expression, an arrow is displayed above the character. 〉 Time t
If the velocity potential at is represented by Φ ([p], t), the two-dimensional wave equation which is a basic equation is expressed by the following equation (1).

[0020]

(Equation 1)

From the velocity potential Φ ([p], t),
Particle velocity v ([p], t), sound pressure P at predicted position P ([p])
([p], t) is represented by the following equations (2) and (3).

[0022]

(Equation 2)

[0023]

(Equation 3)

Here, c is the transmission speed of sound in the air, and ρ is the air density. If Φ ([p], t) for separating the space coordinate [p] and the variable of the time t is represented as a complex sine wave of each frequency ω with a complex execution value Φ, the following equation (4) is obtained.

[0025]

(Equation 4)

The wave equation of the above equation (1) can be rewritten as a Helmholtz equation as shown in the following equation (5).

[0027]

(Equation 5)

Here, ω / c is the wave number of the sound, j (= √ (−1))
Is an imaginary number symbol. In FIG. 1 (A), the running sound of a general vehicle is set by the line sound sources S _{1 and} S ₂ at a position 30 cm from the road surface in the same manner as the sound source setting by ASJ Model 1975. Each sound source is set by the following formula using a cylindrical sound source.

[0029]

(Equation 6)

Here, P ₀ is the sound pressure at a distance ζ from the center of the cylinder, and A (= 1.0 Pa) (reference sound source) is a unit radius ζ (= 1.0 Pa).
m), the sound pressure H ₀ ⁽²⁾ () is the 0th-order Hankel function of the second kind, and is inversely proportional to ω to the 乗 power. Therefore, the sound pressure of this sound source has a frequency characteristic that decreases in inverse proportion to the frequency. The prediction point P ([p]) is set to an arbitrary point in the model space. For noise, it is assumed that the sound of the general part only reverberates in a semi-enclosed space below the dedicated part girder, and the sound source of the dedicated part is omitted in the numerical analysis model of the boundary element method. Set the sound source in the academic style and consider it in the calculation).

FIG. 1A shows a sound reflecting surface of a two-layer road structure constituted by boundaries Ω (= Ω _R ＲΩ _A ). Each boundary Ω
Is defined as the complex acoustic impedance Z ([r], ω) by the following equation (7).

[0032]

(Equation 7)

Here, R ([r], ω) is a complex sound pressure reflection coefficient. The relationship with the sound absorption coefficient α ([r], ω) is expressed by the following equation (8).

[0034]

(Equation 8)

Of the main boundaries Ω to be constructed,
Since the high reflective boundary Omega _R of road surface which changes due the state of the surface, although strict provisions rather difficult, since basically the properties similar to the fully reflective, usually in this example simply fully reflective wall (Α (ω) = 0) is set.

On the other hand, the absorption type boundary condition Ω _A of the sound absorbing plate of the sound insulating wall, the back surface sound absorbing plate, etc. is the standard α (ω) of the sound absorption coefficient of the reverberation room method.
0.8 or the standard α (ω) = 0.9 of the oblique incidence sound absorption coefficient is set. Since the imaginary part of the complex acoustic impedance determines the phase difference of the reflected wave with respect to the incident wave, it affects the interference characteristic at the predicted point position in the diffraction propagation and the reflection propagation.
For this reason, the peak (particularly the minimum value) frequency of the frequency characteristic at the position of the prediction point slightly shifts, but when considering the overall value (the energy sum of all frequency components),
Since this influence is small, in this example, the calculation was performed with the imaginary term set to 0 (meaning that the phases of the incident wave and the reflected wave match). This impedance boundary condition is defined by the following equation (9) as a Robin type boundary condition in numerical analysis by the boundary element method.

[0037]

(Equation 9)

In the boundary element method, the Helmholtz equation of the above equation (5) is developed as follows in the form of a boundary integral equation in consideration of all the boundaries Ω. Here, the boundary integral at the point at infinity converges to 0 and can be omitted.

[0039]

(Equation 10)

Where [r] is an observation point on the boundary element, [r ']
Is the position vector of the source point, and G ([r]-[r '], ω) is a two-dimensional Green's function for the Helmholtz equation and is given by the following equation (11).

[0041]

[Equation 11]

The boundary surface Ω is the maximum frequency f to be analyzed.
It is subdivided into about 1/6 of the wavelength of the sound at _max , and a boundary element Ω _j is obtained. If the above equation (10) is discretized, the following equations (12) and (13) are obtained.

[0043]

(Equation 12)

[0044]

(Equation 13)

Where N is the number of divisions of the boundary element, H
_{i, i '} and Gi _{, i'} are the boundary integrals in the Neumann and Direlet type boundary conditions, respectively. From the above equation (9), either the potential value Φ ([r '], ω) at the source point or the flux value ∂Φ ([r'], ω) / ∂ [n] in the normal direction is obtained. Since it can be specified, the boundary element method calculates the effective value Φ ([r], ω) of the velocity potential on all the elements at each frequency by solving the linear equation of the above equation (13). The effective value Φ ([r], ω) of the velocity potential at an arbitrary prediction point P ([p]) in space is its value Φ on the obtained boundary.
([P], ω) is obtained by the following equation (14).

[0046]

[Equation 14]

From the above, the sound pressure P at the predicted position P ([p])
([p], omega) and sound pressure level _{L P ([p], f} ) (dB) of the following formula (15)
To (17).

[0048]

(Equation 15)

[0049]

(Equation 16)

[0050]

[Equation 17]

[0051] Here, _{P R} (= 20μPa) is a reference sound pressure. In this example, the frequency interval Δ
For each f, the sound pressure levels LP _{, 1} ([p], f), LP _{, 2} ([p], f) at the predicted point P ([p]) for the sound sources S ₁ , S ₂ are calculated using the boundary element method. Calculated by numerical analysis according to For the obtained sound pressure levels LP _{, 1} ([p], f), LP _{, 2} ([p], f) at the predicted positions, the representative spectrum of the vehicle noise considering the weight of the road traffic noise (Equation 18 below),

[0052]

(Equation 18)

Further, the auxiliary sense correction W of the A characteristic_A(f) and set
Correction value W of frequency characteristics of cylindrical sound source _s(f) (= 10logf)
Calibrate and correct sound pressure level L at each frequency^* _{P, 1}([p],
f), L^* _{P, 2}([p], f) is calculated by the following equation (19).

[0054]

[Equation 19]

In the present invention, the median noise level _LA50,
When calculating an evaluation value of road traffic noise such as the equivalent noise level L _Aeq, it is considered necessary to comprehensively evaluate frequency bands in which the degree of influence of vehicle noise is different, and target frequency bands are divided by the same frequency interval Δf. , The overall level of the sound sources S _{1 and} S ₂ at the prediction point P ([p]) (sum of energy of all frequency components)
_{[L P, 1] ([} p]), calculated by _{[L P, 2] ([} p]) the formula (20).

[0056]

(Equation 20)

Next, with respect to the directly propagating model as shown in FIG. 1B, the overall level (all the frequency components) at the prediction point P ([p]) is obtained by the completely same procedure using the boundary element method. Sum of energy) [L' _{P, 1} ] ([p]),
[L' _{P, 2} ] ([p]) is calculated.

As described above, the insertion loss ΔL _{P, k of the} noise propagation model with respect to the direct propagation model at the same predicted position
([p]) (the insertion loss of the noise propagation model in consideration of the effect of the sound insulation wall and the back sound absorbing plate in the two-layer road structure with respect to the direct propagation model) is calculated by the following equation (21).

[0059]

(Equation 21)

[Method of Calculating Noise Level of Two-Layer Road Structure by Boundary Element Method] The current "environmental standard relating to noise" and the evaluation quantity used in the Society of Acoustical Sciences are the median L of noise levels.
_The transition from _A50 to the equivalent noise level L _Aeq has already been made. However, the accumulation of past research results in this field is often based on the median _LA50 of the noise level. In particular, since the Monte Carlo method is initially formulated based on the median _LA50 of the noise level, the noise prediction calculation method using the boundary element method shown in this example is also used for comparison with the Monte Carlo method. It shall be performed for both the median _{L A50} and noise levels _{L Aeq.}

<Method of calculating median value _LA50 of noise level>
When calculating the median L _A50 of noise level in this calculation method, insertion loss [Delta] L P at the predicted point positions shown in the above equation (21) to ASJ _{Model1975, k} calculated by considering the _([p]) Do.

FIG. 2 is a model diagram schematically showing a noise propagation route of a two-layer road structure. S as sound source of general part
_1, it sets the _{S 2,} S as a sound source of a dedicated unit
_3, set the _{S 4.} The noise level LF _{, k ′} (k ′ = 1 to 4) in the one-row equal-space model directly propagated only at the predicted point P ([p]) only by the distance attenuation is calculated by the following equation using ASJ Model 1975.
Calculated in (22).

[0063]

(Equation 22)

Where l_{k '}Is sound source S_{k '}From the prediction point P
distance to ([p]), d_{k '}Is sound source S _{k '}Vehicles traveling
Is the average headway. Also L_{w, k '}Is each sound source S
_{k '}For example, in the second stage regulation type,
It is given by the following equation (23).

[0065]

(Equation 23)

Where V _{k ′} is the average vehicle speed, a
_{1, k '} is the mixing ratio of small vehicles, and a2 _{, k'} is the mixing ratio of large vehicles. Noise level L _R at the predicted point positions in consideration of reflection by the source S _1, S ₂ of the general _{section, 1 ([p]),} L R, 2 ([p])
Is the insertion loss ΔLP _{, 1} ([p]), ΔLP _{, 2} ([p]) calculated by the boundary element method shown in the previous section.
Noise level ΔLF _{, 1} ([p]), Δ
L _{F, 2} ([p]) is taken into account, and is calculated by the following equation (24).

[0067]

(Equation 24)

In the model considered this time, the setting is such that the cross section does not change in the traveling direction of the road. The above equation (22) according to ASJ Model 1975 agrees on the assumption of a cylindrical sound source (line sound source). Therefore, the model consistency between the insertion loss of the cylindrical sound source obtained by the boundary element method and the noise level given by the above equation (22) is obtained.

On the other hand, the noise propagating from the dedicated portion to the prediction point via the sound insulation wall is calculated by the following equation (25) from ASJ Model 1975.

[0070]

(Equation 25)

Where α _{D, k ″} and α _{I, k ″} are ASJ Mode
l Diffraction attenuation according to 1975 and correction values due to various factors. As described above, the median value _LA50 of the noise level at the prediction point considering all the sound sources can be obtained by the following equation (26).

[0072]

(Equation 26)

<Calculation method of equivalent noise level L _Aeq > In this calculation method, the insertion loss ΔLP _{, k} ([p]) at the predicted point position shown in the above equation (21) is considered in ASJ Model 1993, and the equivalent A method for predicting the noise level L _Aeq will be described. In this example, for the sake of simplicity, it is assumed that the road cross-sectional shape and the traffic conditions do not change and are constant with respect to the traveling direction of the road. M sections with a length Δl and a constant interval are set to the left and right in the road direction centering on the cross section including the prediction point. The distance from the center of the i ^* -th section to the prediction point P is represented by the following equation (28). Therefore, the length in the road direction to be calculated is l = (2m + 1) · Δl. The noise level L ′ _{F, k ′, i} ^* ([p]) (k ′ = 1 to 4) per vehicle directly propagated only by the distance attenuation to the prediction point P is ASJ Mod
In el 1993, it is defined by the following equation (27).

[0074]

[Equation 27]

[0075]

[Equation 28]

Here, L ′ _{w and k ′} are A-weighted noise power levels per vehicle of each sound source S _{k ′} according to ASJ Model 1993 (noise power level after human audibility correction).
In the two-vehicle classification (an evaluation method of classifying passing vehicles into two types of large vehicles and small vehicles by the Acoustic Society), the following equation (29) is used.

[0077]

(Equation 29)

Predicted point position P based on general-part sound sources S _{1 and} S ₂
The unit patterns _{LR, 1, i} ^* ([p]) and _{LR, 2, i} ^* ([p]) of the noise level considering the multiple reflection sound at ([p])
Is the insertion loss ΔL shown in the previous section.
_{p, 1} ([p]), ΔL _{p, 2} ([p]) directly propagated noise level L ′ _{F, 1, i} ^* ([p]), L ′ _{F, 2, i} ^* ([p]) Is calculated by the following equation (30).

[0079]

[Equation 30]

Note that a point sound source is assumed in ASJ Model 1993. However, assuming that the sound source is a cylindrical sound source under the condition that the cross section does not change, the insertion loss obtained by the boundary element method is not applied. Relevance has been obtained.

[0081] On the other hand, only part sound source S _3, the unit pattern of the noise level in consideration of diffraction and ground surface effect due sound insulating wall in a predictable location in relation to _{S 4 L 'R, k "} , i * ([p])
(k "= 3,4) is obtained by the following equation (31).

[0082]

(Equation 31)

Here, ΔL _{D, k ″} ([p]), ΔL
_{G, k ″} ([p]), ΔL _{M, k ″} ([p]) are ASJModel 1993
Is the correction value due to the factors of diffraction attenuation, ground surface effect, and weather conditions, and ΔLM _{, k ", i} ^* is set to 0 in this case. In this example, it is assumed that the road cross section and the traffic condition do not change in the road traveling direction. since it is, and the prediction point P ([p]) noise levels in L _{a eq ([p])} is given by the following equation (32).

[0084]

(Equation 32)

Here, N is the traffic volume per 3600 sec, Τ
(= 3600 sec) is the integration time.

[0086]

[Embodiment 1] Hereinafter, noise prediction of a two-story road structure was specifically performed by the method of the present invention described above, and a comparison was made with the Monte Carlo method previously proposed by the present inventors. The comparison was made based on the median value _LA50 of the noise level.

The present inventors have disclosed in Japanese Patent Laid-Open No. Hei 10-2062.
In Japanese Patent Publication No. 27, characteristics relating to the position of a sound insulating wall in a standard two-layered road structure having a sound insulating wall in a general part were clarified by the Monte Carlo method. As a result, contrary to the conventional concept, there was an extremely interesting result that the noise level decreased when the sound insulation wall was set back to the private area. In this embodiment, the median value _LA50 of the noise level is calculated using the boundary element method according to the above-described calculation method using the same model as the model set in the Monte Carlo method, and is compared with the result of the Monte Carlo method.

FIG. 3 shows a noise analysis model of the same two-layer road structure as that set by the analysis by the Monte Carlo method. A PC box girder is set as the girder structure, and the height of the general sound insulation wall is 8.0 m. Is set to The gap from the top of the sound insulating wall to the lower flange is D. General part sound sources S _{1 and} S ₂ are set as sound sources at the center of the road in the general part upper and lower lanes,
The dedicated part sound source S ₃ ,
Consider the S _4. The number of vehicles passing through the vehicle per hour N _G = 1,500 and the traveling speed V = 6 as the general-part sound sources S _{1 and} S ₂
0 km / h and a large-vehicle mixing ratio a ₂ = 0.2 are set. Similarly, as the dedicated section sound sources S _{3 and} S ₄ , the number of vehicles passing through the vehicle per hour N _H = 2000 and the traveling speed V = 100 km
/ H, a large-vehicle mixing ratio a ₂ = 0.2 is set. Also, P
C digits and road Omega sound absorption coefficient _R is alpha = 0, the sound insulating wall Omega _A was set to alpha = 0.8 in the specifications of the reverberation chamber sound absorption rate of the sound absorbing plate. Regarding the position of the general sound insulation wall, B = 0 m immediately below the position of the dedicated sound insulation wall, the negative side was set for the private side, and the road side was set to positive. The predicted position P is a public boundary.

Further, since the structure of this model is axisymmetric with respect to the center of the road, a half model is set by a mirror image. With this setting, the number of boundary elements can be reduced by half. The maximum element length was set to 5.6 cm, which is 1/6 of the wavelength (34 cm) of the highest frequency (1 kHz) to be analyzed.

In the calculation example, since the mirror image is used, the left and right sound sources have the same phase. However, if the phases of the two sound sources are changed, the interference characteristics change. Therefore, the frequency characteristics of the sound pressure shown in FIG. It is thought to change somewhat. But,
In the case of road traffic noise, there is no phase correlation between the left and right sound sources, and in principle, the phase of each sound source is given irregularly at each frequency. Ideally, it is desired to set a noise source having an irregular phase (uncorrelated phase characteristic), but it is difficult with the current solution. As a method, there is a method of setting a plurality of calculation cases in which the phase difference between the left and right sound sources is different and using the average value as a representative value, but performing this calculation involves a great deal of labor and is not realistic. . Since the reverberating sound in the two-story road structure is transmitted through a very complicated reflection path, interference waves between reflected sounds having very various phases actually occur, and equivalently, internal reverberation. The sound is almost uncorrelated. Therefore, it is considered that the effect of the phase condition of the left and right sound sources has little effect on the sound propagating to the private side.

First, FIG. 4 shows an analysis result by the method of the present invention.
Analysis frequency on the horizontal axis, the sound source S ₁ on the vertical _axis, the sound pressure level after correction in consideration of the S ₂ simultaneously ^{_{L P * ([p],}} f) is obtained by plotting. From this frequency characteristic, it can be seen that the sound pressure level fluctuates greatly near each frequency due to sound interference. From this, the validity of analyzing the target frequency band by dividing it into frequency steps Δf has been verified. That is, if a particular frequency is selected and analyzed only for that frequency, a large difference occurs in the noise level depending on the selected frequency. As in the method of the present invention, analysis is performed for each frequency step width Δf,
By integrating the sound pressure levels for each frequency step width Δf, the energy sum of all frequency components is obtained, and by evaluating with the numerical value based on this energy sum, it is possible to evaluate in a form that covers the target frequency band. .

Next, FIG. 5 shows the median value _LA50 of the noise level at the point 1.2 m above the civic boundary which is the predicted point P on the vertical axis.
And the horizontal axis represents the sound insulation wall position B, and the boundary element method (BE
M) and the value calculated by the Monte Carlo method are plotted. In the calculation range shown here, the deviation between the boundary element method and the Monte Carlo method is -1.5 dB or more.
It is about +2 dB. In general, the results obtained by the boundary element method and the results obtained by the Monte Carlo method show the same tendency. If the position of the general sound insulation wall is moved to the outside (private area), the noise level will be much reduced. Was also confirmed by the calculation by.

A detailed study and comparison shows that B = −3.0 to −7.0.
In the case of m, the boundary element method has a slightly higher value. This is because the Monte Carlo method counts the amount of energy of a vector touching the boundary as a calculation means, and the detection boundary is composed of a line segment having a finite length, and the size is usually about 3 to 10 m. The reflection noise level is evaluated as the average value of the detection boundary having a finite size because it is set, whereas the boundary element method obtains the value at the spot point in principle, so it is compared with the Monte Carlo method. It is considered that this is because the concentration of sound near the predicted point due to the reflection of the concave surface at the end is more finely evaluated. To confirm this,
In FIG. 5, a total of nine points are taken in a grid pattern of 1.0 m vertically, horizontally, and centered on the predicted point P, and the average value (BEM) of these points is calculated and plotted. When the result (BEM) is compared with the Monte Carlo method, it is found that the result is better than the former (BEM).

Next, a comparative study will be made on the change in the noise level with respect to the change in the undersizing amount D. FIG. 6 shows the model of FIG. 3 in which the vertical axis represents the median value _LA50 of the noise level, the horizontal axis represents the opening under the girder D, and the values calculated by the boundary element method and the Monte Carlo method are plotted. is there.

In the case of this model, the result of calculation by the Monte Carlo method is constant at about 72 dB in the range of D ≧ 2.0 m and hardly changes, whereas the result of calculation by the boundary element method has an aperture even in the area of D ≧ 1.0 m. A decrease of the noise level by about 1 dB is recognized for a decrease amount of 1 m of the amount D, and the result is a decrease of about 3 dB in the range of D = 1.0 m to 0.0 m.

[0096]

Second Embodiment Next, in a second embodiment, an equivalent noise level L _Aeq is applied to an actual two-layered road structure using the method of the present invention.
Prediction calculations were performed on a base and comparisons were made with measured values.

The road cross section shown in FIG. 7 is actually in service.
This shows the cross-sectional structure of a two-story road in
Noise was measured along the roadside. As the digit structure,
Similarly, it is a PC box girder, and the height of the general sound insulation wall is 6.0 m.
You. A back sound absorbing plate is laid at the end of the girder. sound
As a source, the general part sound source S
_1,S₂Set the center of the road in the dedicated lane
Dedicated sound source S₃, S ₄It was set. General part sound source S_1,
S₂From the actual traffic volume survey results,
Number of vehicles passing N_G= 1120 cars, running speed V = 64k
m / h, large car mixing rate a₂= 0.33 is set. As well
, Dedicated sound source S_3,S₄As per hour vehicle
Number of passing N_H= 1800 cars, running speed V = 80km /
h, Large car mixing ratio a₂= 0.28 is set. Also, P
C digit Ω_RHas the sound absorption coefficient α = 0, the sound insulation wall Ω_AIs the reverberation of the sound insulation board
The standard value of the room method sound absorption coefficient was set to α = 0.8. This model
Because the structure is axisymmetric with respect to the center of the road,
Half the model is set by the image. Also book
In the model, sound absorption at the end of PC girder according to the actual road structure
A rear sound absorbing plate with a ratio α = 0.8 is applied. For general section
In consideration of the sound absorption of drainage pavement,
Average vertical value obtained by averaging the measured value of the sound absorption coefficient and the A-characteristic
The incident sound absorption coefficient α was set at 0.23. As a prediction point,
Set the civilian boundary to 0m, set the road side to +, and the private side to-
Then, each height h = 1.2 m was set.

FIG. 7 shows the results of the analysis. The horizontal axis represents the distance x from the civilian boundary, and the vertical axis represents the equivalent noise level L _Aeq.
And the equivalent noise level L at each point on the roadside and on the private side
The value of the calculated values and measured values of _{ae q} is a plot of the.

It can be seen that the calculated value and the measured value generally agree well. For the two points on the road side from the civilian boundary, the measured values are slightly lower than the calculated values,
It is considered that this is because the rear surface (part a) of the sound insulation wall is actually a green belt, and the damping effect of this part occurs.

[0100]

As described above, according to the present invention, since the wave acoustic simulation by the boundary element method is used, the viaduct, the sound insulation wall, the positional relation of the road surface, the girder shape, the surrounding terrain, the arrangement of the sound absorbing material are used. Comprehensive model evaluation is possible, and even if the aperture ratio becomes small, the effect of diffuse radiation due to the wave nature of sound is considered in principle,
There is no restriction on the application, and the application range can be expanded. Also, since the diffraction and interference at the time of multiple reflection are taken into account, the prediction accuracy can be improved. Also,
Calculated insertion loss ΔL P _(dB) of noise propagation model for the direct propagation model in noise prediction point calculated by the boundary element, which existing noise level calculation formula, in particular by combining with Acoustical Society equation, The evaluation of road traffic noise can be predicted as an absolute amount.

[Brief description of the drawings]

FIG. 1 (A) is a noise propagation model, and (B) is a direct propagation model used in the description of the analysis by the boundary element method.

FIG. 2 is a model diagram schematically showing a noise propagation route in a two-layer road structure.

FIG. 3 is a noise analysis model diagram of a two-layer road structure set in the first embodiment.

FIG. 4 is a frequency characteristic diagram at a noise prediction point.

FIG. 5 is a noise level comparison diagram between the method of the present invention and the Monte Carlo method.

FIG. 6 is a noise level comparison diagram for the aperture amount D between the method of the present invention and the Monte Carlo method.

FIG. 7 is a noise level comparison diagram between the method of the present invention and measured values.

[Explanation of symbols]

 1 ... Viaduct girder 2 ... Sound insulation wall 3 ... Road surface

Claims (4)

[Claims] 1. A method for predicting road noise when a structure serving as an acoustic reflection surface is provided around a road surface, wherein a noise propagation model considering the structure serving as the acoustic reflection surface and an acoustic reflection surface are provided. In each of the direct propagation models with the structure omitted, a predetermined reference sound source is set at the noise generation position, and a complex acoustic impedance is set as a boundary condition for each boundary element, and a predetermined frequency band between target frequency bands is set. The sound pressure level at the noise prediction point for the sound source is obtained by the boundary element method for each frequency step width Δf, and the sound pressure level for each frequency step width Δf is integrated to obtain the entire frequency step width. The energy sum of all frequency components in the noise propagation model and the energy sum of all frequency components in the direct propagation model Based on the insertion loss of noise propagation model for the direct propagation model in noise prediction point [Delta] L P _(d
A first step obtaining a B), together with the directly obtained propagation noise level L _{F (dB)} from considering attenuation only the road noise source is propagated directly to the noise prediction point, the direct propagation noise level L _F (DB) and
A second step obtaining a total noise level L _R Considering reflected sound in the noise prediction point based on said insertion loss of noise propagation model for the direct propagation model in noise prediction point ΔL P _(dB), consisting of A road noise prediction method taking into account reflected sound. 2. A method for predicting road noise when a structure serving as an acoustic reflection surface is provided around a road surface, wherein a noise propagation model considering the structure serving as the acoustic reflection surface and an acoustic reflection surface are provided. In each of the direct propagation models with the structure omitted, a predetermined reference sound source is set at the noise generation position, and a complex acoustic impedance is set as a boundary condition for each boundary element, and a predetermined frequency band between target frequency bands is set. The sound pressure level at the noise prediction point for the sound source is obtained by the boundary element method for each frequency step width Δf, and the obtained sound pressure level at the noise prediction point is representative of the vehicle noise. Correction is made based on the audibility correction of the spectrum and the A characteristic and the frequency characteristic value of the set sound source to obtain a corrected sound pressure level for each frequency step width Δf, and a corrected sound pressure level for each frequency step width Δf. The energy sum of all frequency components is obtained by integrating the levels.Then, based on the energy sum of all frequency components in the noise propagation model and the energy sum of all frequency components in the direct propagation model, Insertion loss ΔL _P (d
A first step obtaining a B), together with the directly obtained propagation noise level L _{F (dB)} from considering attenuation only the road noise source is propagated directly to the noise prediction point, the direct propagation noise level L _F (DB) and
The following formula total noise level L _R Considering reflected sound in the noise prediction point based on said insertion loss of noise propagation model for the direct propagation model in noise prediction point ΔL P _{(dB)
(I) L R = L F} + ΔL P (dB) ... second step and the prediction method of road noise considering the reflected sound, characterized in that it consists of obtaining from (I). Wherein said direct propagation noise level L _F on the basis of the Acoustical Society of formula _{(ASJ Model 1975), L W} (dB) noise power levels of the sound source, the distance l between the sound source and the noise prediction point
(m), as the average headway distance d (m) of the vehicles running the sound source,
The following formula _{(II) L F = L W} -8-20 · log 10 l + 10 · log 10 (π (1 / d) tanh2π (1 / d)) (dB) ... claims 1 obtained from (II) one A method for predicting road noise in consideration of reflected sound described in Crab. Wherein said direct propagation noise level L _F is the Acoustical Society equation based on (ASJ Model 1993), per vehicle, the A-weighted noise power level after audibility correction human L '
_{W (dB),} the sound source position of m set at intervals of Δl the road direction, here the following formula _{(III) LF = L 'W} -8-20 · logl i (dB) ... (III), l i is The distance from the sound source i ^* th to the noise prediction point, and the following formula (IV) l _i = √ ((l ₀ ² + (i ^* · Δl) ² )) (i ^* = 0, 1, 2,... M) The road noise prediction method according to any one of claims 1 and 2, which is obtained from (IV).