Windows Application For Computing Spectral Shapes

The Spectral Shapes Windows Application

Applications Opening Window

Earthquake response spectrum 

Single Degree Of Freedom System 

Design Response Spectra

Damping in spectral shapes

Frequencies of Spectral ordinates

 

Earthquake response spectrum 

 

            Earthquake response spectrum has its origin in earthquake engineering, and determines the influence of the earthquake vibrations on an engineering structure. Under the influence of earthquake vibrations the base of a structure, partially embedded in the ground, tends to move with the ground. The rapid ground motion produces stresses and deformations in the structure. These deformations vary in different parts of the structure. A rigid structure tends to move with its base, so that the dynamic forces on the base and elsewhere in the structure are nearly equal. To prevent damage from differential movements in the structure, it must be designed to withstand the forces of deformation. It should be able to absorb energy and dissipate it through vibrations, damping or inelastic deformations. Response spectra are the tools used to describe the effect of the structure to the earthquake vibrations. These are defined in terms of the maximum responses of a set of single degree of freedom (SDOF) systems characterized by their natural frequencies and damping ratios, when each oscillator is subjected, independently, to the vibration at its base. (The number of independent coordinates required for specifying the configuration or position of a system is referred to as the number of degrees of freedom. A single degree of freedom system can, thus, be completely defined using   only one spatial coordinate.)  With suitable modifications to the response spectra to account for the effects of local geology in modifying the earthquake vibration these are used for earthquake resistant (or aseismic) design of structures . For understanding  earthquake response spectrum the response of the SDOF to the earthquake ground motion must be understood.   

 Single Degree Of Freedom System 

The figure given below shows a very simple structure. Here mass less elastic columns CE and DF support a slab ABCD of mass m.  The elastic columns are characterized by a spring constant k.  When subjected the earthquake ground motion the structure deviates from its original position, as shown in the figure, and dissipates energy due to internal damping proportional to velocity of the mass. The constant c represents the damping in the structure. In the treatment given below the displacement of base of the structure in the horizontal direction is represented by ug(t) and the horizontal displacement of the slab by u(t) .The slab is in equilibrium under three different forces, namely:

(a)    The inertial force fI = m (d2u/dt2 +d2ug/dt2),

(b)   The spring force fS =k u, and

(c)    The damping force fD = c (d2u/dt2).

  In the equilibrium state fI + fD+ fS=0. This condition gives the equation of motion during the earthquake vibration.

                 m  (d2u/dt2) +c (du/dt) +k u =-m(d2ug/dt2)

or              d2u/dt2 +2wx (du/dt). +w2u = -d2ug/dt2… (1)

Here w= Ö(k/m) and x = c/2Ö(km)

This structure is a single degree of freedom  (SDOF) system of natural frequency w and damping c. For any given earthquake displacement time history ug(t) the equation of motion can be solved for u(t) at any specified frequency and damping. For each time history, and for a specified frequency and damping value, a maximum value umax of u(t) can be determined, where umax represents the value of the maximum displacement of the slab with respect to its original position. A plot of umax against frequency is called displacement response spectrum of the earthquake. The ordinates of this plot are called spectral displacement, and are represented by SD(w,x). Thus for any frequency and damping value

SD =max |u(t)| … (3)

The maximum spring force is the given by:

fS = k umax … (4)

The spring force can also be visualized as a pseudo inertial force on the mass m, so that

fS = ma … (5)

Here `a’ is pseudo absolute acceleration. From equations (4) and (5) one gets a = umax. At high frequencies the maximum value of `a’ during the vibration may be obtained from Equation (1). This turns out to be w2u. The pseudo absolute acceleration (PAA) is represented by the symbol SA.

SA= uamax  =w2u  … (6)

SA is called the spectral acceleration. The maximum strain energy in the spring equal to (1/2) k SD2. This can be associated with a velocity SV, such that the kinetic energy associated with it is equal to the maximum strain energy, i.e.

(1/2) m SV2 = (1/2)k SD2…(7)

Thus SV = w SD … (8)

SV is also a pseudo velocity because it is not related to the actual velocity of the structure. Thus the three different spectral ordinates are related with each other as follows:

SA = w SV = w2SD…. (9)    

   

Damping is Specified by the user.

Response spectrum is computed at the following frequencies (Hrtz)

0.10 

0.15 

0.20 

0.25 

0.30 

0.35 

0.40 

0.45 

0.50 

0.60 

0.70 

0.80 

0.90 

1.00 

1.10

1.20 

1.30

1.40 

1.50 

1.60 

1.70 

1.80 

1.90 

2.00 

2.10 

2.20 

2.30 

2.40 

2.50 

2.60 

2.70 

2.80 

2.90 

3.00 

 

3.15 

3.30 

3.45 

3.60 

3.80 

4.00 

4.20 

4.40 

4.60 

4.80 

5.00  

5.25 

5.50 

5.75 

6.00 

6.25 

6.50 

 

6.75 

7.00 

7.25 

7.50 

7.75 

8.00 

8.50 

9.00 

9.50  

10.00

10.50

11.00

11.50

12.00

12.50

13.00

13.50

 

14.00

14.50

15.00

16.00

17.00

18.00

20.00

22.00

25.00

28.00

31.00

34.00

37.00

40.00

43.50

45.50

50.00

 

The Spectral Shapes Windows Application

    The application may be used for generating mean and mean + standard deviation design response spectral shapes. The values of mean and the mean squares are stored in the applications directory, and are used to update the spectral shapes when more ground motion time histories become available. The application has the provision of generating spectrum compatible accelerograms, which are basically useful for understanding the application. The opening window of the application has a self explanatory menu.

Applications Opening Window

Clicking on the 'Generate time history' menu item opens the following three dialog boxes in succession. The time histories, which are generated, use a trapezium window for strong motion defined by a rise time, a duration of constant signal amplitude followed by a signal decay time. The generated time histories are compatible with a specified spectral shape given in an ASCII file ( extension .spc). The time history file to be used by the application for generating or updating spectral shapes must have the extension .ths (it is again an ASCII file, and the extension .ths has been chosen only to indicate that it is a time history file).

 

 

After the menu item 'Fresh spectral shapes is clicked on the windows menu bar the dialog box for initializing the parameters of the spectral shapes appears as shown on the left hand side.  of the figure given below: The generated mean and mean + standard deviation shapes then appear as shown on the right hand side of this figure. The generated mean and mean + standard deviation shapes are listed in ASCII files SPShapeMean.SPC and SPShapeMeanPlusSigma.SPC in the applications' directory. The application makes use of a file named Shape_0.spf, which contains the spectral frequencies and the sums and the sums of the squares for each frequency. After the spectral shapes are generated the Shape_0.spf file is modified. In subsequent updates of the spectral shapes fresh .spf files are created.

 

 

Updating The Spectral Shapes

     On clicking on the menu item Update shapes update initialization dialog box appears as shown on the right hand side of the following figure. In the update Input file section the .spf file corresponding to the the last update is selected, and in the time history file name the file which contains the new set of the time histories is selected. A number is added before the extension .spf in the spectrum shape output file to produce a new update file (which will be used in the next update). As the various update files (with .spf extension) are stored (until deleted manually) the updates can be effected from any stage. On clicking OK in this dialog box the updated mean and mean + standard deviation spectral shapes appear as shown on the right hand side of this figure with the details as shown.

 

Design Response Spectra 

 Design Response Spectra (DRS) specify the upper limit within which the response of the structure should be restricted during an earthquake in order to protect it from damage. DRS are, generally, specified in terms of a spectral shape and a normalization factor.  Peak ground acceleration (PGA) has been widely used as the normalization factor.  In the early days of aseismic design use of standard response spectra was proposed [1]. These were derived from eight accelerograms - two horizontal components each from four earthquakes recorded at sites within 8 to 56 km distance from the earthquake source. These earthquakes are listed below:

Table-   : Earthquakes used for deriving Standard Response Spectra.

S.No.

EARTHQUAKE

DATE

MAGNI-TUDE

1.

Imperial valley, California

December 12, 1934

6.5

2

El Centro, Imperial Valley, California

May19,1940

7.1

3.

Olympia, Puget Sound, Wash.

April 13, 1949

7.1

4.

Taft, Kern County, California

July 21, 1952

7.7

Spectral shapes computed from the accelerograms of these earthquakes were normalized to have equal area under the zero damped spectral intensity curve for a predetermined frequency range to produce a standard spectral shape. The spectral intensity, SI, was defined as follows:


                               2.5

              SI(b) = ň SV(b,t)dt    

                          0.1

Where SV represents the velocity spectrum and b is the coefficient of damping. Spectral intensity is not, any more, considered a good descriptor of earthquake damage, and hence is not used for normalizing response spectra. Current practices are based on normalization of accelerograms with respect to PGA before computing spectral shapes. Though, it is not considered ideal, PGA has found wider acceptance because:

§         Acceleration, being a measure of force, is easy to comprehend.

§         The procedures of computing spectral shapes are better defined and understood.

§         The computations deliver, directly, the ordinates of the response spectral shape (called the dynamic amplification factors - DAF, because when multiplied by the design value of PGA these provide the ordinates of the design response spectrum).

 In 1969 Newmark and Hall proposed that spectral acceleration at different frequencies are related to the PGA through amplification factors[2]. These amplification factors are he dynamic amplification factors (DAF). They determined the DAFs using statistical analysis of the observed response spectra at certain control points, which were chosen to define the response spectra. These DAF values are given below in Table 9.  Here, the response is defined at four discrete points, called control points. Between the controls points the spectra have been assumed to be smooth, and are defined by straight-line segments. These spectral shapes have been in use as standard or site-independent spectra shapes. Standard response spectral shapes from 33 accelerograms from twelve earthquakes recorded under varying site conditions were derived by Blume and his associates [3] .   In derivation of these spectra the accelerograms were grouped according to several different criteria (e.g. PGA values, site soil characteristics, epicentral distance, geographical locations etc.) to determine the influence of various factors on the shape of the response spectral shapes. On the basis of statistical analysis of various groups, response spectral shapes corresponding to three different levels of exceedance probability were estimated. These three levels were called the `large’, ‘small’ and `negligible’ exceedance probability levels, and correspond to the mean, M (50% probability of exceedance), M+s (15% probability of exceedance) and M+2s (2.3% probability of exceedance) s being the standard deviation . It was suggested that risk variability due to factors like regional seismicity, geotectonics etc. be considered before choosing the spectral shapes for a site.  A comparison of these shapes with the corresponding shapes of Housners spectra showed that the Housners shapes were lower below the mean shapes for periods shorter than 0.4 seconds, and higher for longer periods.

 

  TABLE-1 : DAF VALUES FOR RESPONSE SPECTRUM FOR HORIZONTAL DESIGN RESPONSE SPECTRA: RELATIVE VALUES OF AMPLIFICATION FACTORS AT CONTROL POINTS.

 

Percent of critical damping

Acceleration

Displacement

 D

(.25Hz )

 A

(33 Hz)

B

(9 Hz)

C

(2.5 Hz)

0.5

1.0

4.96

5.95

3.20

2.0

1.0

3.54

4.25

2.50

5.0

1.0

2.61

3.13

2.05

7.0

1.0

2.27

2.72

1.88

10.0

1.0

2.28

2.28

1.70

  These studies also inferred, from the analysis, that the effect of signal duration on the response spectral shapes for periods smaller than 0.5 seconds was small. For longer periods the response tends to increase with increase in signal duration. The International Atomic Energy Agency (IAEA) and the United States Atomic Energy Commission (USAEC) adopted the M + s shapes for design of nuclear power plants in this manner [4],[5].

 



[1] Housener,G.W.(1959). Behavior of structures during Earthquakes. Proc. Am. Soc. Engrs. Vol. 85.

[2] Newmark, N.M. and Hall, W.J.(1969). Seismic Design Criteria For Nuclear Power Facilities. 4 WCEE, Santiago, Chile, Proc. pp 37-50.

[3] Blume, J.A. and Associates (1973). Recommendations for Shapes of Earthquake Response Spectra. WASH-1254, Government Printing Press, Washington D.C.

[4] IAEA(1979). Earthquakes and Associated Topics in Relation to Nuclear Power Plant Siting. Safety Guide  50-SG-S1, International Atomic Energy Agency, Vienna.

[5] USAEC(1973). Design Response Spectra for Seismic Design of Nuclear Power Plants. Regulatory Guide 1.60