| doi:10.3850/978-981-08-6218-3_SS-Th040 |
 Final Paper PDF
|
PREDICTION OF BUCKLING LOAD OF STEEL RACKING FRAME USING NON DESTRUCTIVE METHOD
Kalaikumar Vallyuthama, V. J. Kurianb, S. P. Narayananc, Mohd Shahir Liewd and Nabilah Abu Bakare
Civil Engineering Department, Universiti Teknologi PETRONAS, Perak, Malaysia.
akalaikumar@petronas.com.my
bkurian_john@petronas.com.my
cnarayanan_sambu@petronas.com.my
dshahir_liew@petronas.com.my
enabilah@petronas.com.my
EXTENDED ABSTRACT
In storage industries, steel racking frames are widely used to store goods. Due to the limitations of the storage space, the capacities of the frames are stretched to the ultimate limit. To maximise the use of racking frame, its capacity needs to be determined, ideally by a non destructive method. Predicting buckling load of steel racking frame by using the measured natural frequency is proposed in this work. An ideal equation developed earlier was adopted for predicting the buckling load. The ideal equation is determined from the plot of non-dimensional axial load versus nondimensional squared frequency. In the dimensionless plot, when the ratio of the axial load is 1 the corresponding squared frequency is zero and when the ratio of the squared frequency is 1, the ratio of axial load is zero. To validate the ideal equation, simple experiments and numerical analysis were carried out and the results were compared with the analytical values. The experiments works were carried out using three different struts in the laboratory. Incremental loads were applied on the struts until failure and the readings were recorded. The identical struts were modelled in the LUSAS to be analysed using one-dimensional analysis. In addition to that, the buckling load and natural frequency for the struts were computed theoretically. The buckling load and natural frequency obtained from the experiment were compared with the corresponding values from the numerical and analytical analysis. It was found that, there is a large difference between the values obtained from experiment compared to the numerical and analytical analysis. The difference might be due to the fatigue effect on the struts due to repetition experiments, varying cross sectional area along the length and the imperfections of the section. The numerical and analytical analysis showed only a slight difference of less than 1%. This has validated the one dimensional numerical model analysis. The study was extended using the strut but with different type of support conditions using the numerical and analytical analysis only. Once again a good agreement was found between the numerical and the theoretical analysis. The one-dimensional analysis could always satisfy the ideal equation regardless of any end support condition. Further to that, the two-dimensional model was studied. A simple frame was adopted from a literature review and modelled. This was also analysed using numerical and analytical methods. To simplify the study, the bracing effect and imperfection of the member were ignored. The connections in the frame were assumed as rigid. It has been found that the numerical and theoretical result agree closely. The percentage difference in value between the numerical and the theoretical buckling load is 5.17 and numerical and theoretical natural frequency is 2.4. When plotted, a linear relation is obtained between axial load and the squared frequency. However, it could not satisfy the ideal equation. Further study is required to explain this outcome. For the three-dimensional study, the simple frame is duplicated and converted into two three-dimensional models and analysed using LUSAS. The theoretical work for the three-dimensional frame was not conducted as it involves lengthy and complex mathematical method. A good agreement was achieved when experimental result is compared with the ideal equation. This finding indicates the possibility of adopting this non-destructive method to determine the capacity of the steel racking frame.