A password will be e-mailed to you.

As a new development containing an underground parking level was being planned next to a soaring and almost century-old multi-wythe brick chimney, the owner decided to preserve the chimney and use it as an icon to the development.

Constructed in 1927, this brick chimney originally belonged to Nashville’s first full-service hospital. It is 117 feet tall, with a 12-ft. overall diameter at the base and a 6-ft. 6-inch OD at the top; the gross thickness tapers from 24 in. at the base to 7 in. at the top.

When the building design team realized that rock blasting may take place as close as 30 ft. from the chimney at a depth of 12 ft., they contacted Genesis Engineering Group to determine the chimney’s dynamic response to ground motion induced by blasting and reinforce the chimney as required.

Vibration velocity, and not displacement, is typically used to evaluate damage potential, because the majority of studies done on the subject focuses on damage thresholds based on peak particle velocity (PPV).

The Tennessee Blasting Standards Act, like most laws that govern the use of explosives and their potential effects on structures, evaluates the potential effects on a structure in terms of ground vibration primarily and air concussion secondarily. The accepted practice for this evaluation is to measure the energy transmitted through the ground and air and evaluate its harmful effect in terms of PPV for the ground vibration and decibels for the air sound pressure level. Typically, the alternate PPV threshold graph as enacted in 2008 is used, in which PPV is limited with different blast frequencies, up to a 2 in./second maximum threshold for blast frequencies greater than 30 hertz.

Most blasting damage studies have historically evaluated the effects on a structure with various types of materials or various levels of damage based on the vibration level the structure is subjected to, and not particularly the vibration response of the structure due to energy transmitted in the form of vibration traveling through the ground that can cause a secondary vibration response in the structure itself.

Because of the age of the chimney and concerns related to the specific condition of chimney materials, it was determined that for this project, PPV, acceleration and displacement would be examined. The typical concerns were expanded to include the response of the chimney, so acceleration data was computed from the velocity data in order to evaluate the stress to the chimney’s members that developed along the length of the chimney as a result of the blasting.

Before the blasting operation started, pre-blasting inspections, utilizing diagrams and photographs of the surrounding structures in the area, were performed in order to document pre-existing conditions and problems with structures prior to the start of any construction or blasting activities. These inspections are helpful in evaluating and addressing unwarranted claims of construction or blasting damage that might occur during the course of the project.

Figure 2
Seismograph Geophone Sensor and Simplified System Diagram.

The ground vibration data collecting sensor in the seismograph, also called the geophone, is shown in Figure 2. The EXAD-8 seismographs used on the project are innovative, cost-competitive, high-accuracy seismograph units that combine analog and digital technology to consistently capture the peak particle velocity at three orthogonal directions. Three geophones were rigidly attached to the east face of the chimney along the height, at bottom, mid-height, and top (approximately 1 ft. below the top of the chimney). The attachment bracket was fitted with a geophone retention plate that clamped the geophone to the bracket so it would move with the chimney’s motion. These three seismograph geophones were manually triggered at the same time in order to get comparative time history data.

Figure 3
Seismograph Geophone Sensor attached to chimney.

Pre-blast analysis
From a structural dynamic analysis standpoint, it is not uncommon to approximate blast vibration as a sinusoidal vibration using methods developed from studies of building response to earthquake (see "Report of Investigation 8507 – Structure Response and Damage Produced by Ground Vibration from Surface Mine Blasting," United States Bureau of Mines; "Blast Vibration Monitoring and Control," by Charles H. Dowding). The analysis team decided to employ two methods: 1) time history analysis, and 2) response spectrum analysis using statistical data from Dowding. Both methods require engineering judgment and experience in selection of analysis parameters, such as damping ratios of multi-wythe brick chimneys and wave propagation velocities through the rock at the site.

Knowing that vibration propagation and attenuation in the ground is influenced significantly by the specific rock or soil properties at the site, the analysis team pulled together time history data from past blasting records performed at a site near downtown Nashville with similar limestone rock formation, similar blast hole diameter size, lift depth, blast pattern layout and closest exposure distance and initiation pattern. A total of seven time-history records from previous blasting projects were used in the analysis. It is important to note that this site has less rock per cubic foot of excavated material at the closest exposure distance to the chimney as compared to the seismic data from the records used in pre-blast dynamic analysis. Hence, the analysis team anticipated the actual blast for this site would be less significant from a vibration standpoint. Geometrical non-linear dynamic analyses were performed using a homogenous, multi-degree-of-freedom finite element model.

Additional analyses using the response spectrum method modified from earthquake engineering with amplification factors based on statistical data from Dowding were performed, based on multiple charge weight per delay and scaled blast distance to obtain envelope forces.

Figure 4
Analysis team setting up the geophones.

Using simple elastic stress calculations with forces produced from the analysis, the mortar joint tensile stresses were determined and compared to the code’s prescribed values. Calculations also showed that the existing foundation on rock is capable of resisting all code-prescribed lateral forces, as well as the dynamic reactions from the analysis. Additionally, with the high frequency nature of the blast operation and low natural frequency of the chimney (the calculated fundamental natural frequency is 0.86 Hz), the analysis determined that it was unlikely that the chimney would be impacted by this blast operation. The analysis team concluded that the chimney would not be negatively affected by the planned blasting and no additional reinforcing would be required other than repointing of all mortar joints.

Blasts and post-blast analysis
With the relatively high forces shown in the preliminary dynamic analysis, line drilling of the perimeter of the excavation and blasting site was conducted as the rock strata allowed. Line drilling consisted of drilling closely spaced holes 4 in. in diameter spaced 10 to 12 in. on center. The holes were drilled to a depth that was 2 ft. below the finished grade of the site. This line drilling was designed to serve as an interceptor to provide damping of the ground vibration and a containment line for any physical rock and material shifting that would typically occur in the crater zone. Line drilling typically replaces pre-split blasting and can provide even better rock perimeter walls around a site in addition to the vibration damping that can accompany the line drilling. Unlike the vibration damping from pre-split blasting, line drilling can influence the first blast on a project since the drilling can be performed before any blasting takes place. See Figure 1 for the location of the line drilling.

Figure 1
Overall site layout.

A total of three blasts were in the proximity of 30 to 42 ft. from the chimney (refer to Figure 1 for locations). Blast number is organized incrementally in the order of time when it was performed, at 8:47 a.m., 9:20 a.m., and 1:10 p.m., respectively. The vibration from Blast #3 was considerably lower, likely due to reduction of confinement of the material from Blast #1 (therefore not shown). In general, the blasts were comprised of 24 to 26 holes, drilled 3 in. in diameter and spaced on a 4 ft. burden and 4 ft. spacing pattern. Hole depths varied from 9.5 to 14 ft. and were loaded with four to 5 decks of explosive charge. Each deck had 2 x 16 Unimax dynamite for a total of 82.5 to 277.5 pounds of explosives in the blast. Charge weight per delay varied from 1.25-lb. (Blasts #1 and #3) to 2.5-lb. (Blast #2).

Figure 6
Time-displacement plot for Blast #1 – predicted displacement by Time History Analysis and Actual Measured Displacement.

Time history analysis was performed for comparison with predicted values using the recorded time history data from the three geophones mounted on the chimney. A summary of base shears and base moments is shown here:

Base Shears and Base Moments Summary
Base Shear (kip)
Base Moment (kip-ft)
Blact #1 (actual)
Response Spectrum Method using similar site parameters
Blact #2 (actual)
Response Spectrum Method
Time History Analysis using time history data with similar site parameters
60 (average)
816 (average)
Seismic (QE)
Winde (W)

In general, the forces produced from pre-blast analyses were conservative. Factors that may have swayed the results include the location of the line drilling, variance in rock formation, accuracy of the data due to weather impact, and damping effect assumed in the analysis versus actual, which is difficult to be determined without further testing.

Figure 5
Comparison of response spectra.

The analysis team was also able to generate the response spectra for the actual blasts and compare them to the predicted response spectra (see Figure 5). It appears that the predicted principal frequencies using the response spectrum analysis approach are approximately seven to 16 percent lower than the actual blast frequencies. For a low natural frequency structure like this chimney, the response spectrum analysis approach appears to be relatively conservative in predicting internal forces when compared to the results from the time history analysis using actual time history data. For Blast #2, the pseudo spectra velocity calculated from the actual time history peaked at frequencies between 30 and 75 Hz, and appears to be 67 percent higher than predicted using the response spectrum method.

During this study, the analysis team found that both methods present room for interpretation due to unknown parameters that have to be assumed in the dynamic analysis, and the prediction requires extensive experience in blasting operations and geotechnical knowledge of the local rock formation. The success of the project is most vividly demonstrated by the fact that the chimney is still standing and undamaged.

C.W. Yong, P.E., S.E., LEED AP, is a structural engineer and principal (managing member) with Genesis Engineering Group, LLC, a full service engineering firm located in Tennessee, Florida, and Texas (www.geneng.net). Contact him at cwyong@geneng.net. Ernest R. Grigoryan, P.G., is a geophysicist and vice president of Instrumentation and Quality Control with VCE, Inc. Wade Hutchison, P.E., is a civil engineer and the president of VCE Inc.