Sổ tay tiêu chuẩn thiết kế máy P52 potx

CHAPTER 45 NOISE AND ITS CONTROL David A. Towers, RE. Senior Consultant Harris Miller Miller & Hanson Inc. Lexington, Massachusetts Erich K. Bender, Ph.D. Divisional Vice President Paul J. Remington, Ph.D. Principal Engineer Eric E. Ungar, P.E., Ph.D. Chief Consulting Engineer Bolt Beranek and Newman Inc. Cambridge, Massachusetts 45.1 INTRODUCTION / 45.1 45.2 NOISE MEASUREMENT AND ANALYSIS / 45.2 45.3 NOISE EFFECTS AND STANDARDS / 45.15 45.4 NOISE CONTROL/45.18 REFERENCES / 45.37 45.7 INTRODUCTION Noise is unwanted audible sound. Sound is essentially a fluctuating pressure distur- bance that may act locally or propagate away from its source. Extremely intense sound can cause structural damage or equipment malfunc- tions, but usually its effect on humans is the primary concern. Noise can annoy peo- ple, can lead to interference with speech communication, can interfere with the performance of mental and delicate manual tasks, and—if it is intense enough—can cause discomfort, pain, and temporary or permanent hearing damage. Fortunately, many aspects of noise and its effects are well enough understood to permit their con- sideration in the design process. Sound can occur and propagate in any gas, liquid, or solid medium, but sound in air is usually of primary interest. Sound may be produced by any phenomenon that can lead to fluctuating pressure disturbances. TTiese phenomena include (1) rapid expansion of gases or injection of fluid volumes, such as from explosions and engine exhausts; (2) repetitive interruptions or modulation of airflows, such as by siren disks or fluctuating valves; (3) turbulence, as present in fluid streams emerging from nozzles or duct grillages; and (4) vibrating solid surfaces. In many practical situa- tions, several noise-generating phenomena may occur simultaneously; for example, an impact press may generate noise not only because of the structural vibrations it produces but also because of the air it expels from between the impacting surfaces. Sound, being a pressure disturbance, can propagate in the medium in which it is generated. This propagation need not involve flow or net displacement of the medium; only the disturbance and the energy associated with it move away from the source. Pressure fluctuations in air can induce fluctuations in other media in contact with the air, and vice versa. Therefore, often sound from a given source reaches an observer not only via a direct air path but also via paths that may involve several media. For example, sound radiated from vibrating gears in a housing may propa- gate from the air in the housing through an oil layer and through the housing wall into the ambient air. In many practical situations, several parallel paths of sound transmission from a given source to a given observer—including some relatively tor- tuous paths along complex structures—may be similarly important. It usually is convenient to consider a noise problem from the "source-path- receiver" viewpoint. This approach facilitates accounting for all significant sources (noise generators), receivers (items or persons affected by noise), and paths along which the noise from the sources reaches the receivers. This approach thus encour- ages evaluation of all relevant facets of the problem. The remainder of this chapter introduces noise measurement and analysis, noise effects and standards, and noise control techniques relevant to machine design. For treatment of these subjects in greater depth, texts and handbooks on acoustics should be consulted (for example, Refs. [45.1] through [45.5]), as well as the specific references given throughout this chapter. 45.2 NOISE MEASUREMENT AND ANALYSIS 45.2.1 Noise Measures Sound or noise can be sensed by measurement of sound pressure, the variation in air pressure above and below its equilibrium value. The measure most commonly used is the root-mean-square (rms) sound pressure /? rms . The rms sound pressure is obtained by squaring the value of the sound pressure disturbance at each instant of time, averaging the squared values over the sample time, and taking the square root of the result. Because the range of sound pressure amplitude variations that the human ear can detect extends over several factors of 10, a compressed scale based on the loga- rithm of the mean square pressure is used. The decibel, abbreviated dB, is a measure of this scale. The corresponding noise descriptor is called the sound pressure level L p , defined as Lj = IOlOg(^y dB (45.1) \ PQ I where p Q is a reference pressure, standardized as 20 micropascals (uPa) [2.90 x 10~ 9 pounds per square inch (lb/in 2 )]. This very small reference pressure corresponds to O dB and represents approximately the weakest sound that can be heard by an aver- age young, alert person with an undamaged hearing mechanism. Since decibels are logarithmic measures, sound pressure levels cannot be added by ordinary arithmetic. The sound pressure level L p (total) corresponding to the combination of n sound pressure levels L p (i) is calculated from 1 / n \ L p (total) - 10 log X 10 L ' (0/1 ° (45.2) \/ = i / To describe noise adequately, one must measure not only its amplitude, which determines the magnitude of the pressure, but also its frequency, which determines its pitch. In any sound, the air pressure alternately rises and falls; for repetitive sounds, each time the pressure rises from its minimum value and returns to that value, it completes one cycle. The number of cycles occurring per second is called the frequency of the sound; the unit of cycles per second is hertz (Hz). Frequency is observed subjectively as the tone, or pitch, of a sound. The low frequencies (20 to 500 Hz) have a low-pitch, or bass, sound. The midfrequency range, from about 500 to 3000 Hz, is where most speech information is carried. High frequencies, from about 3000 to 20 000 Hz, tend to be prevalent in whistles, jets, and high-speed machines. The wavelength of a sound wave is defined as the distance the wave travels in a stationary medium during one cycle. Wavelength and frequency are related by ^ = J (45.3) where c = speed of sound, ft/s (m/s) /= frequency, Hz X = wavelength, ft (m). The speed of sound in gases depends on the temperature, but not on pressure. At 7O 0 F (21 0 C), for example, the speed of sound in air is 1128 ft/s (344 m/s), and the wavelength of a 1000-Hz sound wave is 1.128 ft (0.3438 m). The basic properties of a pure-tone (that is, single-frequency) sound wave are summarized in Fig. 45.!.This figure illustrates a time-history graph of the amplitude of a sound. Note that for this sinusoidal wave, the sound pressure amplitude rises from zero to a positive maximum, then falls through zero to a negative maximum, and then returns to zero during one complete cycle. For this type of wave, the rms value is 0.707 times the absolute value of the peak (positive or negative) amplitude. Noise from common sources, such as machinery, is usually more complex than the pure tone illustrated in Fig. 45.1. In general, noise consists of a combination of many sinusoidal components, all with different frequencies. Description of such noise requires a noise spectrum, which is a graph of sound pressure level versus frequency. Frequency analysis (or spectrum analysis) is essential for any comprehensive study of a noise problem for three reasons: (1) people have different hearing sensitivity and dif- ferent reactions to the various frequency ranges of noise, (2) different noise sources emit differing amounts of noise at different frequencies, and (3) engineering solutions for reducing or controlling noise are different for low- and high-frequency noise. Although a noise spectrum is useful for purposes of analysis, it is often conve- nient to use a single-number measure to describe a noise. The most commonly used measure of this type is the A-weighted sound level, expressed in units of dBA. From f This corresponds toprms (total) = ]T" Pnm(i), where the individual signals are at different frequencies and/or are uncorrelated. T - PERIOD - TIME FOR 1 CYCLE FIGURE 45.1 Basic properties of a sinusoidal (pure- tone) sound wave. many experiments with human listeners, it was found that human hearing is more sensitive to midrange frequencies than to either low or very high frequencies. This characteristic is taken into account by adjusting, or weighting, the various frequency components of a sound in accordance with the sensitivity of human hearing and then combining all the weighted components. The result is a single-number measure of sound level that corresponds approximately to the human subjective perception of the severity of the noise, as well as to its annoyance and hearing damage potential. Table 45.1 compares representative noise levels for common indoor and outdoor noise sources and environments. The extremes of noise range from O dBA (approxi- mate threshold of hearing) to 120 dBA (jet aircraft at 500 ft), although most com- monly encountered noise levels fall within the 40- to 100-dBA range. An understanding of the following subjective perceptions of changes in the A-weighted sound level is useful: • Changes of 1 dB or less cannot be perceived, except in carefully controlled labo- ratory experiments. • A 3-dB increase in A-weighted level generally is just noticeable. • A 10-dB increase in A-weighted level is perceived as approximately a doubling in loudness, independent of the initial noise level. All the discussion thus far has been related to sound pressure, since this is the property to which human hearing and microphones respond. However, as discussed later, the magnitude of sound pressure level resulting at a given location and due to a given source depends on the "strength" of the source, on the environment in which the noise source is located, on the distance of the observation location from the source, and sometimes on the direction. Therefore, it is useful in many cases to use a noise measure that describes the intrinsic strength of a given source, that is, its sound power. Sound power represents the total sound energy radiated by a source per unit of time and is proportional to the square of the sound pressure at any given location. SOUND PRESSURE (RELATIVE TO ATMOSPHERIC PRESSURE) ±A -PEAKAMPLITUDE A n -nm AMPLITUDE rim -0.707 I A| f- FREQUENCY - NUMBER OF CYCLES PER SECOND -VTHz A-WAVELENGTH _ SPEED OF SOUND TIME (MC) As is the case for sound pressure, the range of sound power encountered in acoustics is very large. Thus, a logarithmic (decibel) scale is also used to describe sound power. The sound power level L w is defined as W L w = Wlog^- (45.4) W 0 where W = source sound power in watts (W) and W Q = reference sound power, stan- dardized as 10~ 12 W. Sound power level is typically expressed in terms of dB with respect to ID' 12 W. 45.2.2 Sound Fields Meaningful measurements must take into account the variation of sound pressure level with position in the vicinity of a noise source. Figure 45.2 illustrates this general relationship and indicates the various sound field regions. For an ideal nondirectional "point source" in open space, the sound pressure level decreases at the rate of 6 dB per doubling of distance because of spherical spreading of the sound energy. This relation is usually called the inverse-square law, because it corresponds to the sound pressure's varying inversely as the square of dis- tance. However, the point-source approximation breaks down at distances very close to the source. At such distances, sound variation is more complex; in this near field, the sound pressure level may be either more or less than predicted by the inverse-square law, as shown in Fig. 45.2. The extent of the near field depends on the TABLE 45.1 Comparison of Various Noise Levels NOISE LEVEL (dBA) — 120 — —no— — 100 — — 90 — — 80 — — 70 — — 60 — — 50 — — 40 — — 30 — — 20 — — 10 — — O — EXTREMES JET AIRCRAFT ' AT 500ft ~ THRESHOLD OF HEARING HOME APPLIANCES IN ROOMS SHOPTOOLS BLENDER DISHWASHER AIR CONDITIONER REFRIGERATOR SPEECH AT 3 ft SHOUT LOUD VOICE NORMAL VOICE NORMAL VOICE (BACKTO LISTENER) MOTOR VEHICLES AT 50 ft DIESELTRUCK (NOTMUFFLED) DIESELTRUCK (MUFFLED) AUTOMOBILE AT 70 mph AUTOMOBILE AT 40 mph AUTOMOBILE AT 20 mph GENERALTYPE OF OUTDOOR ENVIRONMENT MAJOR METROPOLIS (DAYTIME) URBAN (DAYTIME) SUBURBAN (DAYTIME) RURAL (DAYTIME) GENERAL TYPE OF INDOOR ENVIRONMENT HEAVY INDUSTRY LIGHT INDUSTRY OFFICE DISTANCE FROM NOISE SOURCE (Logarithmic Scale) FIGURE 45.2 Sound fields in the vicinity of a noise source. frequency of the sound, the dimensions of the source, and the phase relations of the various radiating parts of the source. As a rule of thumb, the near field may be assumed to end at a distance about twice the largest dimension of the source (or at 4 times the largest dimension, for sources resting on an acoustically reflective floor). Note that sound pressure levels measured within the near field cannot be used to predict the sound pressure levels at other distances or to evaluate the source sound power level; for these purposes, one must take care to perform measurements in the acoustic far field (that is, at distances beyond the near field). In the acoustic far field, sound pressure levels decrease at a rate of 6 dB per dis- tance doubling, as long as there exists a free field, which is, for all practical purposes, a field in which the effects of any air volume boundaries are negligible. Such a free field can be obtained outdoors, in a large room at locations away from the walls, or in an anechoic chamber. (In the latter, the walls, floors, and ceiling absorb nearly all sound incident on them.) The extent of the free-field region is characterized in Fig. 45.2 by a line with constant slope. Sound from a source in any room—but most pronouncedly in a small room with "hard" (i.e., acoustically nonabsorptive) wall, floor, and ceiling surfaces—is reflected many times, so that the total sound at any location is composed of the sound radiated directly from the source (free-field sound) plus all the reflected components. If many reflected sound waves are arriving at an observation point from all directions, the sound field is called reverberant. In the reverberant field, the sound pressure level decreases less rapidly with distance than indicated by the inverse-square law, as shown in Fig. 45.2. Reverberant rooms, in which sound is uniform throughout, are often used to perform sound measurements that, in effect, average over all direc- tions (for example, for the purpose of evaluating sound power levels of sources). In practice, noise measurements often must be made in semireverberant fields, that is, where the sound propagation characteristics lie somewhere between free- field and reverberant conditions, as indicated by the transition zone in Fig. 45.2. The characteristics of a semireverberant environment are controlled largely by the amount of sound absorption in the room. These characteristics generally need to be evaluated and taken into account in analysis of the measured results. 45.2.3 Measurement Instrumentation Sound Level Meters. A sound level meter consists of (1) a transducer (micro- phone) to convert air pressure fluctuations to an electric signal, (2) an amplifier to SOUND PRESSURE LEVEL (dB) NEAR FIELD FAR FIELD SdBREDUCTIONPER DOUBLING OF DISTANCE REVERBERANT FIELD SEMI- REVERBERANT FIELD FREE FIELD raise the electric signal to a usable level, (3) weighting networks to modify the fre- quency characteristics of the instrument's response, and (4) an indicating device (meter) to display the measured level. Sound level meters are designated by class, depending on measurement accuracy and tolerances. International Electrotechnical Commission (IEC) standard IEC 651 defines four classes: type O, laboratory reference; type 1, precision; type 2, general purpose; and type 3, survey. Type O sets the most stringent accuracy and tolerance limits, followed by types 1, 2, and 3. Type 1 meters provide sufficient accuracy for field measurements in most cases and are usually selected when cost is not a major consideration. Standards for types 0,1, and 2 sound level meters are also provided in American National Standards Institute (ANSI) standard Sl.4-1983. The weighting network most commonly used in sound level meters is the A-weighting network. Its response, shown in Fig. 45.3, represents the average behav- ior of human hearing. Measurements made using this network are expressed in A-weighted decibels, abbreviated dBA. Other common weighting networks include the B, C, and D types, used for special purposes. Some sound level meters also include a "linear," or "flat," response, commonly employed when a sound level meter supplies an electrical signal to other instruments. The indicating meter on a sound level meter displays the sound level in decibels, relative to a standard reference sound pressure (20 uPa, or 2.90 x 10~ 9 lb/in 2 ). The speed with which the meter electronics and indicator respond also has been stan- dardized. Most meters include two choices for averaging time: fast, which has a time constant of about 1 A s, and slow, which has a time constant of about 1 s. The slow response is particularly useful for estimating visually the average value of a sound that fluctuates rapidly. Some sound level meters also have peak-hold and impulse- hold features, which are useful for measuring unsteady or impulsive noises. Microphones. A microphone is a transducer used to convert air pressure fluctua- tions to an electric signal. Of the different types currently available, the most com- monly used are the condenser, electret, and piezoelectric types. The choice of a particular microphone depends on its intended application and required perfor- FREQUENCY (Hz) FIGURE 45.3 Frequency response specified for the A-weighting filter of sound level meters (From ANSI Sl.4- 1983.) RELATIVE RESPONSE (dB) mance in terms of stability, precision, directivity, and frequency-response character- istics. Condenser microphones have excellent long-term stability and are insensitive to changes in temperature. However, they are sensitive to moisture. Electret micro- phones vary considerably in their long-term stability and sensitivity to temperature, and so are not as well suited as condenser microphones to measurement environ- ments with large temperature variations. However, they are less sensitive to mois- ture. Piezoelectric microphones are generally more rugged than condenser or electret microphones. Acoustical Calibrators. An acoustical calibrator is a device that produces a known, stable sound pressure level at the diaphragm of a microphone. The most common calibrators are the pistonphone and loudspeaker. A pistonphone calibrator produces a known sound pressure level within a closed cavity by means of moving pistons. Calibration is usually restricted to a single fre- quency (typically 250 Hz), and corrections for atmospheric pressure must be applied. Loudspeaker-type calibrators consist of a battery-operated oscillator and small loudspeaker. In contrast to the pistonphone, some loudspeaker-type calibrators operate over a wide frequency range (125 to 2000 Hz), and the sound pressure level developed is less sensitive to the atmospheric pressure. Spectrum Analyzers. A spectrum analyzer essentially produces a plot of sound pressure level versus frequency. Spectrum analyzers employ electronic filters to sep- arate the frequency components of a sound signal. The range of frequencies covered by an individual filter is called its bandwidth. Two basic types of filter sets are used in spectrum analyzers: those that use bands of constant bandwidth (that is, a fixed number of hertz) and those that use bands in which the upper frequency limit of the band is a fixed multiple of the lower frequency limit. Of the latter type, the band- width most commonly used in acoustic analysis covers a frequency range of one octave (that is, a 2-to-l frequency range); an analyzer having filters with this band- width is called an octave-band analyzer. Other analyzers use half octaves (A/2-to-l frequency range), one-third octaves (V2-to-l range), or even narrower bands. Nar- rowband filters are often required to determine pure-tone components, such as those resulting from operation of cyclic (reciprocating or rotating) machinery. For narrowband analysis, digital computer-aided real-time analyzers are widely used. The preferred center frequencies and band limits for spectrum analyzer filters are given in ANSI standard Sl.6-1984 and in International Organization for Stan- dardization (ISO) standard 266-1975. Values for octave- and one-third-octave-band filters covering the audio frequency range are given in Table 45.2. Filters that are incorporated in octave-, half-octave-, and one-third-octave-band analyzers have been standardized by ANSI (standard Sl.11-1966) and by the IEC (standard 225- 1966). 45.2.4 Measurement Procedures Once the purpose and required accuracy of a measurement are defined, one must select the proper measurement, recording, and analysis equipment. Microphone positions should be selected to yield a useful sample of the sound field in the area of interest, and the microphone orientations should be chosen on the basis of the frequency-response characteristics of the microphone and of the measurement environment (see microphone manufacturer's instructions). For out- door measurements or for other locations where the air is not calm, the microphone SOURCE: ANSI standard Sl.6-1984. should be fitted with a windscreen to avoid extraneous noise generated by air tur- bulence at the microphone. Before each set of measurements is made, all equipment should be calibrated according to the manufacturer's instructions. It is also a good idea to measure the electric noise floor (the lower measurement limit) of the instrumentation by replac- ing the microphone with an equivalent electric impedance (such as a capacitor) or by shielding the microphone from the acoustic background noise. It is good practice to monitor the output of the sound level meter during the mea- surements by listening with the aid of a high-quality set of headphones; this permits one to detect electromagnetic pickup, signals due to wind or humidity, or other inter- ference. TABLE 45.2 Center and Approximate Cutoff Frequencies for Octave and One-Third-Octave Frequency Bands Covering the Audio-Frequency Range Octave, Hz Lower band limit 11 22 44 88 177 355 710 1420 2840 5680 11360 Center frequency 16 31.5 63 125 250 500 1000 2000 4000 8000 16000 Upper band limit 22 44 88 177 355 710 1420 2840 5680 11360 22720 One-third octave, Hz Lower band limit 14.1 17.8 22.4 28.2 35.5 44.7 56.2 70.8 89.1 112 141 178 224 282 355 447 562 708 891 1 122 1413 1 778 2239 2818 3548 4467 5623 7079 8913 11220 14130 17780 Center frequency 16 20 25 31.5 40 50 63 80 100 125 160 200 250 315 400 500 630 800 1000 1250 1 600 2000 2500 3150 4000 5000 6300 8000 10000 12500 16000 20000 Upper band limit 17.8 22.4 28.2 35.5 44.7 56.2 70.8 89.1 112 141 178 224 282 355 447 562 708 891 1 122 1 413 1 778 2239 2818 3548 4467 5623 7079 8913 11 220 14130 17780 22390 For source noise measurements, it is desirable to measure the background noise level (by turning off the noise source) to determine whether the background noise has a significant effect on the measurements. The background noise level should be at least 10 dB below the source noise level, if it is not to affect measured results sig- nificantly; otherwise, the measured noise levels must be corrected to obtain the level of the source. Table 45.3 may be used to obtain the appropriate correction. At the conclusion of each set of measurements, the proper operation and cali- bration of all equipment should be rechecked, and all pertinent data should be recorded. 45.2.5 Data Evaluation A set of measured acoustic data usually must be evaluated with regard to the prob- lem of interest. This evaluation often requires conversion or extrapolation of the results. For example, sound pressure level measurements obtained for a machine in an anechoic chamber may need to be used to estimate the sound pressure level of the same machine at a different distance inside an industrial building. Or, one may want to use sound power level data acquired for a noise source in a reverberant room to estimate the sound pressure level at a given distance from the same source located outdoors. Such evaluations may be based on the relation between sound pressure and sound power level, as described below. For any sound source, the sound pressure level and sound power level are related by L p = L w + 10 log (-2^ + ±\ + 10 .5 (45.5) where L p = sound pressure level, dB re 20 uPa L w = sound power level, dB re 10~ 12 W Q = directivity factor (dimensionless) r = distance to observation point from acoustic center of source, ft R = room constant, ft 2 In mks units, this equation converts to L p = L ff + 101og (^ + 1) (45.6) where r is in meters and R is in square meters. TABLE 45.3 Correction Factors for Background Noise Difference between total noise level and background noise level, dB 8-10 6-8 4.5-6 4-4.5 3.5 3 Correction to be subtracted from total noise level to obtain source noise level, dB 0.5 1.0 1.5 2.0 2.5 3.0