Tools › Mechanical Design › Acoustics › Construction Noise and Vibration › Method

Construction Noise and Vibration: Method

Open the Construction Noise and Vibration tool

Construction Noise and Vibration

What the tool does

Before a construction job starts, the engineer or the environmental consultant is asked two questions about each nearby building: how loud will the work be there, and will the ground vibration annoy the occupants or damage the structure. This tool makes the screening estimate that two U.S. government documents describe: Section 7 of the Federal Transit Administration's Transit Noise and Vibration Impact Assessment Manual (FTA Report No. 0123, 2018) and the Federal Highway Administration's Roadway Construction Noise Model (RCNM, user's guide FHWA-HEP-05-054, 2006). It is a first estimate for one receiver. Blasting is outside its scope: neither blast vibration nor air overpressure is estimated.

Noise of one machine

Each machine is described by its maximum A-weighted sound level at 50 ft (15.24 m), L50. RCNM carries it to the receiver at distance D as a point source in the open (its Eq. 1) and then weights it by the acoustical usage factor, the fraction of the time the machine works at full power (Eq. 2):

Lmax = L50 − 20 log(D / 50 ft) − shielding,   Leq = Lmax + 10 log(usage factor),   L10 = Leq + 3 dBA (Eq. 3).

The FTA manual writes the same relation with one more term for the ground (its Eq. 7-1):

Leq = Lemission + 10 log(usage factor) − 20 log(D / 50) − 10 G log(D / 50).

G is the ground factor of FTA Table 4-26: 0 for hard ground, and for soft ground 0.66 when the effective path height is 5 ft or less, 0.75 (1 − H / 42) between 5 and 42 ft and 0 above. The tool opens with G = 0, which gives the higher level. The shielding term is RCNM's; the FTA equation has none, and the tool subtracts the entered value in every assessment. Several identical machines in one row add 10 log(number) to the Leq and leave Lmax unchanged.

Levels are combined by energy: L = 10 log Σ 10Li/10. Two equal sources give 3 dB more than one. The total Lmax is the level of the loudest machine (RCNM Section 5.3).

Equipment levels

RCNM Table 1 lists 57 kinds of equipment with an impact flag, a usage factor, the Lmax limit of the Central Artery/Tunnel specification 721.560 and the average of the levels measured on that project. The tool ships 56 of them (the blasting row is left out). RCNM opens on the measured ("Actual") level; where the table has none, the specification level is used. The Advanced option switches to the specification level or to the higher of the two. FTA Table 7-1 lists 33 kinds of equipment, several of them railroad machines, with one typical level at 50 ft and no usage factor: outside the General Assessment you enter the usage factor of such a row. A custom row takes a measured or certified level.

Three assessments

FTA General Assessment (Section 7.1, Option A): for a project whose equipment roster is not yet known. Every machine runs at full power for one hour (usage factor 1), the ground effect is ignored, and only the two noisiest pieces are added. The result is a 1-hour Leq, set against Table 7-2: 90 dBA by day and 80 by night for residential land, 100 and 100 for commercial and for industrial land.

FTA Detailed Analysis (Option B): all equipment, with usage factors, hours and the ground factor. Table 7-3 gives an 8-hour Leq criterion (residential 80 by day and 70 by night, commercial 85 and 85, industrial 90 and 90) and a 30-day average criterion: Ldn 75 for residential land, and a 24-hour Leq of 80 for commercial and 85 for industrial land. The manual describes these metrics in words and prints no formula for them in Section 7. The tool computes them as follows. Each machine contributes its full-power energy times its usage factor times its hours in the period. The 8-hour Leq divides the hours, up to 8, by 8, for the day (7 a.m. to 10 p.m.) and for the night (10 p.m. to 7 a.m.) separately. The Ldn is 10 log[(Σ Ei (hday + 10 hnight)) / 24], the day-night level with the 10 dB night adjustment of the manual's Appendix B, multiplied by the share of working days in 30 days. The 24-hour Leq is the same without the factor 10.

RCNM calculation: Lmax, Leq and L10 of each machine and in total against limits you enter. RCNM's own default limits are those of the Central Artery/Tunnel specification and depend on the measured baseline (for example the larger of 75 dBA and baseline + 5 for the daytime residential L10). The tool does not model baselines: it opens with 75 for L10 and 85 for Lmax, and you replace them with the limits of your project.

The criteria fill in when the assessment is chosen and stay editable. The manual is explicit that they are guidelines: "No standardized criteria have been developed for assessing construction noise impact. Consequently, criteria must be developed on a project-specific basis unless local ordinances apply."

Distance at which a criterion is met

For each comparison the tool also moves every machine to one common distance and solves for the distance at which the combined level equals the criterion: D = 50 ft × 10(L at 50 ft − criterion) / (20 + 10 G). The chart shows the same curve. It answers "how far away must the work be", not "how far away must each machine be".

Vibration

FTA Table 7-4 gives, for 14 kinds of construction activity, the peak particle velocity (PPV) at 25 ft and an approximate rms velocity level Lv in VdB re 1 micro-inch per second. Each source is carried to the receiver by itself (Eqs. 7-2 and 7-3):

PPV = PPVref (25 / D)1.5,   Lv = Lv,ref − 30 log(D / 25).

The exponent 1.5 is the manual's value for "normal propagation conditions". In Advanced mode a site exponent n from measurements replaces it (30 becomes 20 n). The manual's Lv values come from the PPV with a crest factor of 4, that is Lv = 20 log(PPV / 4 / 10−6), rounded; a custom source uses the same rule.

Damage is judged on PPV against Table 7-5: 0.5 in/sec for reinforced-concrete, steel or timber buildings without plaster, 0.3 for engineered concrete and masonry without plaster, 0.2 for non-engineered timber and masonry buildings and 0.12 for buildings extremely susceptible to vibration damage. Annoyance is judged on Lv against the General Vibration Assessment criteria of Table 6-3, as Section 7.2, Step 4 directs: 65 VdB for buildings where vibration would interfere with interior operations; 72, 75 and 80 VdB for residences with frequent, occasional and infrequent events; 75, 78 and 83 VdB for institutional land uses. The distance at which each criterion is met is D = 25 ft × (PPVref / criterion)1/n and D = 25 ft × 10(Lv,ref − criterion) / (20 n).

PPV is entered in in/sec in both unit systems, as the manual prints it; in SI the results are also shown in mm/s (1 in/sec = 25.4 mm/s). The VdB reference stays 1 micro-inch per second, which is not the reference used in most metric standards.

What it assumes

Point sources in the open with 6 dB per doubling of distance; one shielding value per machine, taken from a rule of thumb (RCNM Appendix A) or a separate barrier calculation; usage factors that describe the whole period; no reflections from buildings, no wind or temperature effects, no background noise added; all machines of a row at one distance. For the Detailed Analysis: the same working day repeated on each working day of the 30. For vibration: one source at a time, the manual's average source levels, one attenuation law for all soils, and no adjustment between the ground and the inside of the building.

When it misleads

The FTA manual states that its distance relation assumes about 300 ft or less; beyond that, ground and atmospheric effects need a more detailed method, and the straight-line curve of the chart is an extrapolation. Close to a large machine the point-source assumption fails. The RCNM levels are averages of measurements on one project in Boston, some from a single sample (the sample counts are in Table 1 of the guide), and a given machine can differ by several decibels. Table 7-4 gives one vibration level per kind of equipment while, in the manual's words, "there is considerable variation in reported ground vibration levels from construction activities"; soil, pile type and hammer energy can move the result by a large factor, and the manual warns that sources working together could raise the level without saying by how much. A vibration source closer than 25 ft is extrapolated upward. The damage criteria are screening thresholds for identifying locations to study, not a prediction of cracking, and the annoyance criteria were written for repeated transit events inside buildings. Local ordinances and contract specifications, where they exist, govern.

What the tool does not do

Blasting; barrier insertion loss by path-length difference; the baseline-relative limits of RCNM; L10 statistics other than the fixed adjustment; several receivers or moving equipment along an alignment; octave bands and low-frequency noise; ground-borne noise; vibration-sensitive equipment criteria in one-third octave bands; the manual's Detailed Vibration Analysis with measured transfer mobility; combined vibration of several sources.

Checks

Every row of RCNM Table 1 and of FTA Tables 7-1, 7-2, 7-3, 7-4, 7-5 and 6-3 is compared by the test script with values read by a separate program from the text of the two documents. The printed example reproduced is Figure 19 of the RCNM guide: five machines at 50 ft with total Lmax 89.6 dBA, total L10 88.3 dBA and every printed row value and limit exceedance. The decibel-addition Examples B-1 and B-2 of the FTA manual are reproduced as well. Section 7 of the FTA manual prints no worked construction example: the 8-hour Leq, the 30-day Ldn and all vibration results are checked against closed forms and an independent numerical calculation only. One inconsistency in the source is kept as printed: Table 7-4 lists a hydromill in rock at 0.017 in/sec and 75 VdB, where the crest-factor rule gives 73 VdB.