Heat Effects on Dynamic Response of 52m-Steel Chimney via in-situ Modal Testing

R. Livaolu
Gmhane University, Department of Civil Engineering, Gmhane, Turkiye

ABSTRACT: In this study, heat effects on dynamic behaviour of a 52m-steel chimney in Yesilyurt township of Samsun City in Turkey was studied via in-situ modal testing,. For this purpose before commissioning of the chimney, a series of test was realized and after the fabric get to work same tests were repeated for the same sensor locations to realize how the heat affect on the dynamic response of chimney. The ambient and forced vibration tests are proven to be fast and practical procedures to identify the dynamic characteristics of such kind of structures. Dynamic testing of the towers promises a widespread use as the identification of seismic vulnerability of lifeline structures becomes increasingly important. The data presented in this study is considered to be useful for the researchers and engineers, for whom the heat effect on steel chimney is a concern. The present study also aims to provide the designer with the material examples about the influence of heat on the seismic performance of steel chimney by means of reflecting the changes in dynamic behavior with the consideration of heat. Keywords: Steel chimney, Dynamic response, Heat effect, ambient vibration testing, In-Situ Modal Testing

1. INTRODUCTION Chimney structures are special-type structures which have different features than many other types of structures and are subjected to different loading conditions during their service life. The design of these slender structures becomes even harder considering the combined effects of local site conditions, the type of foundation, seismicity of the region, the magnitude and characteristics of wind loading, machine loading and loading which arise due to internal and external surface temperature difference. All the above factors necessitate a detailed work in the design of these type of structures. In particular, due to the fact that the dimensional proportion of the slender chimney structure in the longitudinal direction is high as compared to other dimensions, the soil-structure interaction effects in chimneytype structures should be carefully considered. In general design approach, the base of the structure is assumed to be fully fixed and due to the reasons stated in this assumption may therefore lead to unjustified errors in design. Besides the peculiarities regarding the slender geometry and the foundation behavior possible adverse local site conditions may further complicate the interaction and may completely alter the dynamic behavior of the structure. One other important factor which affects the behavior is the high levels of temperatures the chimney structure is subjected to during its service life. Particularly, chimneys built-up using steel plate assemblies are more prone to such temperature effects. As well known, steel suffers a progressive loss of strength and stiffness at increased temperatures. The change can be seen at temperatures as low as 300C. Although melting does not happen until about 1500C, only 23% of the ambient-temperature strength remains at 700C. It is considered that the dynamic behavior of the structure would be further altered under the effects of both soil-structure interaction and loading due to temperature change. Tall and slender steel structures are susceptible to wind action which is, in general, critical to their design (Cheng and Li, 2009). Besides common failure causes for these types of structures, such as

wind load, foundation settlement and earthquakes, chimneys are subject to high chemical loads, temperature loads, vortex shedding and ring oscillation ovalling (Simonovi et all, 2008; Tranvik, P. and Alpsten, 2002). Diverse loading of the chimney makes it prone to numerous different modes of failure, such as mechanical overload, force or temperature induced elastic yielding, fatigue, corrosion, stress concentration, buckling, wear, vibration etc. (Simonovi et all, 2008). In accordance with the above mentioned considerations, an existing 52 m long steel chimney structure was selected in Samsun Yeilyurt Ironworks to examine its dynamic characteristics. On-site modal test methods were employed both under operating and non-operating conditions. It is noted that under operational conditions outside temperature of the chimney is around 150-170 oC while the inside temperature may reach around 300 oC. Detailed field investigations were carried out to identify the material and structural properties of the chimney. Using the data obtained numerical models were produced and analyzed assuming both fixed-base boundary condition. So it is aimed to understand effect of the temperature loading on the dynamic behavior of the steel chimney structure, using the test results and numerical analysis estimations. 2. DETAILS OF TEST AND CHIMNEY This section provides a brief description of testing equipment, gives the details of the 52m-chimney and the soil deposit underlying it, and summarizes the arrangement of instrumentation. The details of the vibration tests and corresponding analytical solutions are also presented in this section. 2.1. Testing Equipment The testing equipment consisted of an inertial shaker with force capacity of 250 N, a GW-300 W power amplifier and a four-channel data analyzer (Data Physics DP Quattro). Inertial shakers are permanent magnet devices, which may be sealed for short test operations. Cooling is necessary for prolonged use of these devices, which can be achieved using a small cooling blower that circulates the air through the inertial shaker. The power amplifier used can produce a sine signal source to provide simple excitation for shaker. Alternatively, externally generated signals may be applied to the amplifier. Because of the type of structure as slenderness etc, however there was no reason to use shaker system, so machine load working all around the chimney and the wind load exposed was enough for the test. Two accelerometers located on perpendicular direction to each other (MMF KB12VD 0.5g range, seismic) were used for data collection. Figure 1 shows the test setup and accessories along with instrumentation equipment used in testing.

(a)

(b)

Figure 1. (a) Test setup and accessories (b) accelerometer.

2.2. Soil Profile A detailed geotechnical investigation of the test location was conducted. A total of nine boreholes were drilled up to a depth of 20 m. Standard Penetration Tests (SPT) (ASTM D1586) were conducted with 1.5 m intervals during drilling. A series of laboratory tests, such as Moisture Contest Test (ASTM D4643 08), Atterberg Limits Test (ASTM D4318 05), Triaxial Test (UU) (ASTM D2850 - 03a) and Direct Shear Test (ASTM D3080 04), were conducted on a variety of disturbed and undisturbed samples recovered from the boreholes. According to the results of the geotechnical investigation, the first 1.5 m of the soil profile was classified as organic soil with very low bearing capacity. A simple relationship between N60 and low strain shear modulus (G = 0.83 N60 ) was adopted in order to derive a reasonable approximation of shear wave velocity profile (Imai and Yoshimura 1970). It is recognized that SPT correlations are reliable only for granular material. However, using this correlation estimate dynamic stiffness of clayey soil layers herein would have a small effect given for the limited thickness of the clay layer. 2.3. 52m-Chimney and Instrumentation Details The chimney investigated in this study is located in the Yesilyurt township of Samsun City, Turkey (See Figure 2). The location of the structure is classified as seismic zone 2, according to Turkish Seismic Zoning. The structure is a steel chimney with a height 52 m from base. Chimney construction consists of windshield (outer wall of the chimney) and separate flue duct inside windshield (inner wall) that carries flue gases to the atmosphere and this layer is constructed to prevent heat effect on steel as 70 mm of refractor concrete (see Fig.3).

(a)

(b)

Figure 2. (a) Image and (b) schematics of chimney.

The chimney is supported on reinforced concrete (R/C) boot as seen Fig.2 and has been built as 1.25 m of inner radius with changing thicknesses material along to height between 20 mm to 8 mm. Chimney walls are made of bolted connection steel part (6m-tubes) strengthen with welded 100 NPU at every

1.5 m of level. Windshield sections are flanged together, while flue duct sections lean onto each other. The R/C boot fixed the R/C foundation on 9-piles. Due to realized test in test piece taken from the chimney, steel yielding strength was measured as 370 MPa. The steel Youngs modulus and unit weight were interpreted as 2.1x105 MPa and 78kN/m3, respectively. 2. NUMERICAL MODEL OF 52M CHIMNEY A series of numerical analyses was conducted to simulate the behavior extracted from the field tests. In order to analyze both SSI system and the fixed base system, the finite element software ANSYS (2006) was used. Modeling the chimney-foundation/soil system was performed using the finite element model depicted in Figure 3. The main body of the structure and connection elements are modelled using shell elements (six degree-of-freedom per node) for the supporting system and R/c boot and foundation with solid element (eight node, three degree-of-freedom per node). Soil was modeled using eight-node solid elements with three degrees of freedom in each node. Linear elastic material model was assumed for soil and structural components of the tower.

Z X

Figure 3. Considered finite element model of the 52 m chimney-soil/foundations in this study

The simulation of the infinite medium in the numerical modeling of dynamic soil-structure interaction problems is crucial. A general method of treating this problem is to divide the infinite medium into the near field (truncated layer), which includes the irregularity as well as the non-homogeneity of the foundation, and the far field, which is simplified as an isotropic homogeneous elastic medium (Wolf and Song 1996). More appropriate approximations can be achieved using the artificial and/or transmitting boundaries for preventing reflection and radiation effects of the propagating waves from the structure-foundation interface. There are different types of boundaries for frequency and time domains, including the viscous boundary (Lysmer and Kuhlemeyer 1969) and the damping-solvent extraction (Wolf and Song 1996). In this study, viscous boundary is used for three dimensions problems (see Fig 3). The 3D viscous boundaries were situated at two times of Rayleigh wavelength from the centre of the structure. Maximum element size within the FE mesh was between 1/6th and 1/8th of the minimum Rayleigh wavelength in order to allow the higher frequency components of the input motion to travel within the soil medium (Kramer 1996). 3. DISCUSSION OF RESULTS The results of the numerical analyses performed with fixed and flexible base assumptions are presented in this section along with the measurements of vibration tests. The results of the simulations

performed using large strain excitations are then presented. The comparisons of the results of numerical simulations with the field measurements are presented and a detailed discussion is provided in order to reflect the importance of the consideration of foundation flexibility in seismic design of these structures. 3.1. Fixed Base Model The mode shapes of the fixed base model are shown in Figure 4. The first three vibration modes were identified based on mass participation ratios. The first mode frequency was calculated as 1.372 Hz. The modal frequencies for 2nd and 3rd modes were 6.142 Hz and 15.125 Hz, respectively. Mass participation ratios for the first three modes were calculated as %87, %10 and %0.5, which constituted over %98 of the total mass of the structure. It is worth to say here that in calculation of mass participation ratios only steel chimney are taken into consideration and R/C boot mass are neglected. So this part of structure behaves quite rigidly.

Z Y X

Z Y X

Z Y X

Mod_1(f=1,372Hz)

Mod_2(f=6.142 Hz)

Mod_3(f=15,125Hz)

Figure 4. First three mode shapes and frequencies of fixed base model.

3.2. 52m-Chimney-Foundation/Soil Model As mentioned earlier, the foundation soil stratum is composed of high plasticity soil. Therefore, soilstructure interaction effects are expected to be substantial. Figure 5 gives the first three mode shapes of the interacting soil-structure model. The results show that there are significant deviations of the modal frequencies due to soil-structure interaction. The first mode frequency of soil-structure system reduced from 1.372 Hz to 1.016 Hz. The typical period lengthening effect due to the flexible foundation reduced the second and third mode frequencies to 4.553 and 15.695 Hz, respectively.

Z Y X
Y

Z X

Z Y X

Mod_1(f=1.016 Hz)

Mod_2(f=4.553 Hz)

Mod_3(f=15.695Hz)

Figure 5. First three mode shapes and frequencies of chimney on flexible foundation.

3.3. Results of Vibration Tests Figures 5 shows the normalized transfer functions of the system under ambient vibrations for three different measurements from two perpendicular channels and their power spectrums. In the Fig 5.a the result obtained before servicing of the chimney are given or ,in other worlds, dynamic behaviour for chimney is illustrated for ignoring heat effect. Due to taken measurement after beginning to the service of the chimney the power spectrum calculated is given in Fig 5b. 0.94, 4.595 and 1.14 Hz modal frequencies were measured for the first three modes using ambient vibration for non-heat condition. Similarly modal frequencies are occurred 0.90, 4.548 and 11.79 Hz under heat effect, respectively Identical modal frequencies were similarly obtained under all loading conditions.
Mod_1(f=0.94Hz)

(a)

Poweramplitude

Mod_2(f=4.595Hz) Mod_3(f=11.14Hz)

f(Hz)

(b)

Mod_2(f=4.548Hz)

Poweramplitude

Mod_1(f=0.90Hz)

Mod_3(f=11.79Hz)

f(Hz)

Figure 8. Power spectrum of the measured response to ambient vibration (a) for non-heat condition and (b) under heat effect..

3.4. Comparison of the Numerical Simulations and Field Measurements As mentioned before, steel suffers a progressive loss of strength and stiffness at increased temperatures. The change can be seen at temperatures as low as 300C. Although melting does not

happen until about 1500C, only 23% of the ambient-temperature strength remains at 700C. So when tests were realized in structure after servicing of the chimney, the measurement of heat at the bottom surface of windshield (outer wall of the chimney) was about 150~200 C due to wind direction. So estimated response given in the table 3.1 prove that heat reduce the material strength and accordingly the flexibility of the chimney getting higher. In this example, increases on heat was calculated as 200 o C approximately and this cause 9% loss on stiffness of all structure. The results presented earlier in this study showed that the soil-structure interaction effects on the studied structure is significant. It is well known that the tall and slender structures situated on soft soil deposits are exposed to such critical period lengthening effects (Aviles and Perez-Rocha 1998). The agreement between the results of numerical model and field tests shows that the numerical model is representative of the in-situ soil-structure conditions. Table 3.1 summarizes the results of numerical simulations and field tests. As can be seen from the table, the measured and simulated values differ by only about 7% for the first mode. The second mode field behaviour differed by 1% only. One can clearly see that first and second mode are enough to represent all dynamic behaviour of the structure. So these variations are considered to be minor. Thus, the numerical model is representative of the actual structure in small strain ranges.
Table 3.1. Comparison of Field Test Results and Numerical Simulations. Fixedbase model(Hz) 1.372 6.142 15.125 SSImodel (Hz) 1.016 4.553 15.695 Fieldtest Nonheat (Hz) 0.940 4.595 11.140 Fieldtest Underheat (Hz) 0.900 4.548 11.790 Error BetweenFEM andTest(%) 7 1 29 Differences Betweenheat condition(%) 4 1 6

Mode 1 2 3

5. SUMMARY AND CONCLUSIONS The dynamic response of 52m-chimney situated on a soft soil deposit was studied by means of field vibration tests and numerical simulations. Geotechnical investigations were carried out to define soil strata. The ambient vibration tests were conducted to identify the SSI effects and the heat effect on the small strain dynamic behavior of the structure. The following is the summary of the results and conclusions arising from this study. A total of three mode frequencies were obtained from the FEM of the systems under investigation. First and second modes which are extracted from among a great number of modes can be evaluated as sufficient because their contributions to total response are approximately 97% or over this value in all analyses. Therefore, one may say that only four modes can be adequate to estimate the total response of such a system investigated in this study. As well known, steel suffers a progressive loss of strength at increased temperatures. The change can be seen at temperatures as low as 300C. Although the temperature reaches 200C maximally in this study, nearly 9% loss of stiffness were observed. As mentioned in well-known literature, getting lost of steel elastic characteristic starts at almost 300C, by all means, this study showed that lower temperature than this value, the deterioration on dynamic characteristic of the investigated structure taken place. So the effect of the heat through this temperature value must be consider by the designer who must provide such a structure to heat effect. The results of the vibration tests provide strong support for the finite element models presented in this paper. Thus, it can be easily stated that the proposed finite element models themselves are the meritorious approximations to the real problem, and this makes the models appealing for use in comprehensive investigations.

ACKNOWLEDGEMENT

I would like to give my very special thanks to Yeilyurt fabric director M.E.etin LVAOLU (M.Sc) providing endless support and resources during test and I also want to thank Yeilyurt fabric employee for their support on preparation of the test setup.
REFERENCES ANSYS. (2006). Theory Manuel; Edited by Peter Kohnke, Twelfth Edition . SAS IP, Inc, 1266. Aviles, J. and Perez-Rocha, E.L. (1998). Effect of Foundation Embedment During Building-Soil Structure Interaction. Earthquake Engineering and Structural Dynamics, 27: 15231540. Cheng, J. and Li, Q.S. (2009). Reliability analysis of a long span steel arch bridge against wind-induced stability failure during construction, Journal of Constructional Steel Research, v.97:3-4, 132-139 Imai, T. and Yoshimura, M. (1970). Elastic shear wave velocity and mechanical characteristics of soft soil deposits. Tsuchi to Kiso, 18:1, 17-22. Kramer, S.L. (1996). Geotechnical earthquake engineering. Prentice-Hall Inc., Englewood Cliffs, N.J. Lysmer, J. and Kuhlmeyer R.L. (1969) Finite dynamic model for infinite media, ASCE. Engineering Mechanic Division Journal, 95, 859877. Simonovi A.M., Stupar, S.N and Pekovi, A.M., (2008). Stress Distribution as a Cause of Industrial Steel Chimney Root Section Failure. FME Transactions, v.36, 119-125 Tranvik, P. and Alpsten G.: Dynamic Behaviour Under Wind Loading of a 90 m Steel Chimney, Alstom Power Sweden AB, Vxj, 2002. Wolf J.P. and Song C.H. (1996). Finite element modeling unbounded media, The 11th World Conference on Earthquake Engineering, San Francisco 70, 1-9.