A Case Study to Report the Advantage

of Using Signed Response Quantities

in Response Spectrum Analysis

Sanjib Das and Santanu Bhanja

Abstract Static method of seismic analysis is based on the assumption that fun-
damental mode of vibration is predominant and there are no mass and stiffness
irregularities in the structure. These assumptions of static analysis do not hold good
for the irregular and/or tall structures. In such structures, Response Spectrum
analysis is preferred as it provides an easy solution compared to Time History
analysis. The results of Response Spectrum analysis are computed from modal
responses obtained by performing modal analysis. Individual modal responses are
combined using modal combination methods (SRSS, CQC etc.) to get maximum
absolute responses. Modal combination process generates absolute response
quantities. Unsigned member forces at the both ends of the members do not satisfy
static equilibrium for the Response Spectrum load cases. Support Reaction of the
supported nodes in Response Spectrum load case is also unsigned quantity. It does
not allow understanding whether the supported node is subjected to uplift or not.
Hence load combinations with these unsigned response quantities may lead to
untenable and unrealistic support reaction. It may be observed that footing designed
with such load combination is being subjected to huge uplift and requires unrealistic
thickness. This problem can be easily eliminated if the sign of the responses of the
dominant mode of vibration is considered. This has been clearly demonstrated in
the present case study of a watch-tower constructed in seismic zone four in West
Bengal where unsigned response quantities resulted in a huge uplift force in support
reaction which resulted in unrealistic thickness of the foundation. It was observed
that on using sign of the dominant mode, the supported nodes were not actually
subjected to such huge uplift. It can be inferred from the present case study that the

S. Das (&)
Department of Civil Engineering, National Institute of Technical Teachers’
Training and Research, Kolkata, West Bengal, India
S. Das
Structural Group, Bentley Systems India Pvt. Ltd, 3rd Floor, Tower-A,
DLF IT Park, 8 Major Arterial Road, Rajarhat, Kolkata 700 156, West Bengal, India
S. Bhanja
Department of Civil Engineering, National Institute of Technical Teachers’ Training &
Research, Block—FC, Sector—III, Kolkata, Salt Lake 700 106, West Bengal, India

© Springer India 2015 831

V. Matsagar (ed.), Advances in Structural Engineering,
DOI 10.1007/978-81-322-2193-7_65
832 S. Das and S. Bhanja

responses obtained from modal combinations should be used for load combinations
only with proper sign which is often not considered by the designers.

Keywords Response spectrum
Signed
Watch tower

1 Introduction

In the static method of seismic analysis, it is assumed that the fundamental mode of
vibration dominates the response and mass and stiffness of the structure are evenly
distributed thus giving a regular mode shape. In tall and/or irregular buildings, these
assumptions are invalid and dynamic analysis needs to be performed. Dynamic
analysis using Response Spectrum method is preferred as it is easy to use. Response
spectrum analysis is a technique for performing an equivalent static lateral load
analysis of structures for earthquake forces. It is useful in the approximate evalu-
ation of the reliability and safety of structures under earthquake forces [1]. The
major problem lies in the fact that the response quantities obtained from Response
Spectrum analysis are absolute quantities as they are computed using modal
combination methods like SRSS, CQC etc. The response quantities can be axial
forces acting on a column, moment about its major axis or storey shear in a frame
etc. The nature of the responses cannot be determined. The unsigned responses
derived from a load combination involving a Response Spectrum load case may
provide un-realistic results [2]. If they are used in the design of structural elements
specifically columns, the results will be erroneous. Some special consideration is
required in the Response Spectrum analysis by which the nature of the response
quantities can be determined. A structural element is designed on the basis of the
worst algebraic load combination that leads to maximum stress resultants [3]. Since
member forces are expressed in local axes, unsigned member forces at the both ends
of the members do not satisfy static equilibrium for the Response Spectrum load
cases. Hence load combinations with these unsigned response quantities may lead
to untenable and unrealistic member forces at the start and end of the member. The
support reaction obtained in a Response Spectrum load also poses the same
problem as the nature of the forces cannot be ascertained, only the magnitude is
obtained. It becomes difficult to ascertain whether the reaction at a node is causing
uplift or compression [4]. This problem can be easily eliminated if the sign of the
responses of the dominant mode of vibration is considered. This has been clearly
demonstrated in the present case study of a watch-tower situated in seismic zone IV
in West Bengal where unsigned response quantities resulted in a huge uplift force at
the foundation level. As a result of that the foundation for the structure had to be
designed for high tensile force. The depth required for the foundation to stabilize
this huge uplift was unrealistic. Considering the sign of the dominant mode, it was
observed that the support reaction was not actually exhibiting that huge tension and
a realistic depth of foundation could be obtained. It can be inferred from the present
A Case Study to Report the Advantage of Using Signed Response … 833

case study that the responses obtained from modal combinations should be used for
load combinations only with proper sign which is often not considered by the
designers.

2 Objective of the Present Research Work

The objective of the present research work may be outlined as follows:

• To report in brief the analysis and design considerations adopted for the present
case study
• To highlight how the unsigned response quantities in the in the response
spectrum load cases may lead to unrealistic support reactions
• To discuss the methodology of using the sign of the dominant mode in the
response quantities
• To demonstrate how the use of signed response quantities can resolve the
problem

3 Introduction to the Case Study

The present case study consists of a watch-tower constructed at Jalpaiguri district of

West Bengal, India. The typical administrative building plan is shown in Fig. 1.
Height of the tower was 15.05 m. The circular platform of the tower was at a
height of 11.8 m from the ground level with a radius of 3.5 m. The entire model was
supported by four circular columns which were tied at different levels. There were
intermediate platforms at different levels. The main tower was made accessible by a
series of stairs resting on a series of columns. These columns were not connected to
the main tower model of the watch tower has been provided in Fig. 2.
The watch-tower was constructed in seismic zone IV and response spectrum
analysis was proposed to capture the dynamic response of the watch-tower [5]. The
watch-tower model was prepared in STAAD.Pro V8i software. Figure 3 shows the
3D structural model prepared using the software.
The model was analysed for dead load, live load and seismic forces. Response
spectrum analysis was performed. The structural elements were designed with the
maximum forces obtained with load combinations as per the relevant Indian
standards code.
834 S. Das and S. Bhanja

Fig. 1 Plan of watch tower

Fig. 2 3D Model of watch tower

A Case Study to Report the Advantage of Using Signed Response … 835

Fig. 3 3D Structural model

of watch tower prepared in
software

4 Statement of the Actual Problem

The model was analysed for the primary load cases and load combinations created
as per relevant Indian standards. Support reactions from the superstructure were
obtained and they were adopted for the design of the foundation. A rectangular mat
foundation having a size of 4.6 m × 4.5 m was adopted. It was observed that 12 ms
depth of foundation was required to stabilize this model for the uplift forces coming
from the superstructure which was absurd and technically not feasible. It was
observed that two supported nodes marked as node number 18 and 19 were
experiencing very high uplift forces. The support reactions should be equal in
magnitude but opposite in sign of the forces acting at the column connected to that
node. Figure 4 indicates that if unsigned values of response spectrum quantities are
used, the axial forces developing at the ends of the member for a load combination
involving the gravity and Response spectrum loads fails to satisfy static equilibrium
for that element.
836 S. Das and S. Bhanja

Fig. 4 Support reactions

5 Modifications Needed in the Analysis and Design

Considering the response spectrum load case i.e. Load case 3, the support reaction
at node 19 was obtained as + 386.18 kN (i.e. acting upwards) and the axial force in
the member 128 at that node was +386.16 kN (i.e. acting upwards) and at node 32
was 386.18 kN (i.e. acting upwards) as shown in Fig. 4. This clearly demonstrates
that unsigned responses obtained in Response Spectrum load cases violates the
conditions of static equilibrium. It was observed that load combination 18 [0.9
times dead load (Load case 1) +1.5 Response spectrum load case in X direction
(Load case 3)] was the critical one for the foundation design. It was a tough task for
the designers to find out a suitable depth for the mat foundation. Figure 5 shows the
support reaction for the critical load combination 18.
At the supported node number 18 and 19 the support reactions in FY direction
for load case 1 and 3 were 153.90 kN and 386.18 kN respectively. The support
reaction in FY direction for the load combination 18 was computed as
(0.9 × 153.910 −1.5 × 386.19) = −440.76 kN. This has been shown in Fig. 6.
It can be inferred that the results of the Response spectrum load case being
unsigned when added with other gravity load cases may lead to huge uplift forces in
the foundation which is actually non-existent. This situation can be avoided by
using signed Response Spectrum results. The mode having maximum mass par-
ticipation is considered as the dominant mode and all the response quantities are
assigned the sign of dominant mode [6]. On using the sign of the dominant mode,
A Case Study to Report the Advantage of Using Signed Response … 837

Fig. 5 Critical design forces

Fig. 6 Support reactions for different load cases

the member forces/support reactions gets modified which also satisfies statical
equilibrium for the different elements. The program is capable of finding out the
dominant mode and use the sign of the responses from the dominant mode in the
final result of response spectrum analysis.
It was found that support reaction for the supported node number 19 changed
drastically on using the sign of dominant mode. Proper sign of the support reaction
obtained from the dominant mode of vibration was assigned (Fig. 7). It was
observed that the foundation was to be designed for uplift only for support number
18 and support 19 was not experiencing any uplift. Figure 8 shows the support
reactions obtained after using the sign of dominant mode.
The mat foundation designed with these modified forces resulted in 0.9 m
thickness which was quite a reasonable value. Figure 9 shows the finite element
model created for mat foundation.

6 Findings from Present Case Study

Instead of using the unsigned numerical values of the responses, designers should
use the signed response spectrum results. Designers can use the relevant command
to instruct the program to use the sign of the member forces in Response Spectrum
838 S. Das and S. Bhanja

Fig. 7 Dominant mode of vibration in X direction

Fig. 8 Support reactions after using the sign of dominant mode

results for the most dominant mode—where the mass participation is the highest
among all the considered modes in the analysis. If engineers use simple algebraic
summation of the signed response spectrum quantities with the other load cases, the
final result used for the design will have the proper sign. This can be easily
understood from the output results of the present case study.
A Case Study to Report the Advantage of Using Signed Response … 839

Fig. 9 Support reactions acting on the mat

7 Conclusion

Direct use of unsigned response spectrum analysis results in design sometimes may
lead to an irrational/untenable situation. For example support reactions may yield
very high tension arising from a load combination where the response spectrum
load case is involved. It can be concluded from this case study that use of signed
response spectrum results, where the sign of the response spectrum quantities
obtained from the dominant mode, provides a realistic analysis result which can
directly be used for proper design of structural elements.

References

1. Sen TK (2009) Fundamentals of seismic loading on structures, 2nd edn. Wiley, New York,
pp 64–65
2. Wilson EL (2002) Three-dimensional static and dynamic analysis of structures, 1st edn
Computers and Structures, Inc, pp 233−234
3. Dutta, TK (2010) Seismic analysis of structures, 1st edn. Wiley, Mississauga, pp 115−116
4. Housner GW, Jennings PC (1982) Earthquake design criteria. Earthquake Engineering Research
Institute, Berkeley, pp 33−34
5. IS 1893(Part-I) (2002) Indian Standard. Criteria for earthquake resistant design of structures.
General Provisions Buildings, Bureau of Indian Standards
6. Das S (2014) Use of signed response quantities in response spectrum analysis—a case study.
National Institute of Technology, Durgapur, Innovation in Civil Engineering Structures