Experimental in Structural Dynamics (Base Isolation System: Modelling)

Copyright © 2019 by author(s) and International Journal of Trend in Scientific Research and Development Journal. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0) (http://creativecommons.org/licenses/ by/4.0) ABSTRACT This project is to understand and have the ability to perform dynamic test. Furthermore in this project, the author can investigate how the dynamics of a multistory building is modified by base isolation. The dynamic properties are also very important for the dynamics analysis of the structure. The main task in this project is to determine the dynamic behavior of a 5-storey steel structure model with base isolation system and without base isolation system by experimentally and numerically.


INTRODUCTION
Base isolation, also known as seismic or base isolation system, is a collection of structural elements which should substantially decouple a superstructure from its substructure resting on a shaking ground thus protecting a building or non-building structure's integrity. Base isolation system is one of the effective tools for earthquake building design and enables to reduce and prevent a potentially devastating seismic impact. The aim of base isolation is to reduce the forces impacted to a structure to such a level that no damage to the structure or nonstructural elements occur. Base isolation lengthens the fundamental vibration period of the structure, and thus reduces acceleration and hence the earthquake-induced forces in the structure. Fig (1.2) Base isolation demonstration at The Field Museum in Chicago Base isolation system consists of isolation units with or without isolation components, where: 1. Isolation units are the basic elements of base isolation system which are intended to provide the mentioned decoupling effect to a building or non-building structure. 2. Isolation components are the connections between isolation units and their parts having no decoupling effect of their own.
By their response to an earthquake impact, all isolation units may be divided into two basic categories: shear units and sliding units. Fig (1.3) Comparison between fixed-base system and base-isolation system mechanism during an earthquake

Numerical Modelling
The numerical modeling of the structure was done using the finite element program SAP 2000.

Parameters of Numerical Model
The model parameters used for the analysis in numerical analysis are taken as following:

Modeling in SAP2000
The numerical modeling for the base isolated system was attempted with the used of spring elements. The structure was supported on spring in three directions. In this modeling, the model is limited so that it can only reach nonlinear condition in U2 and U3 direction, which is means that in U1 direction the system is fixed and can only response in linear range, and in U2 direction, the system can slide on the axis. The stiffness value for the lateral springs was taken from the value obtained for the springs actually used in the experimental structure.

Spring parameter used:
Vertical Spring = 1,000,000 N/m Lateral Spring = 1,000 N/m

Fig (2.2) Base isolated model in SAP2000
The From the graph above we can see that Braced model system stiffen the structure, which is conclude by an increased in Eigen frequency. On the other hand, base isolated model make the structure more flexible. We can see that the Eigen frequency decreases, indicating that the structure become more flexible.

Experimental Test 3.1. Equipments
The dynamic response measurement of the experimental model was measured using 8 acceleration sensor and one force sensor. Seven acceleration sensors were used to measure the response of the structure and one sensor was used to measure the base excitation. The force sensor was used for application of impulse force on the structure. The response of the system was measured through a data collection system.
The sensor set up was arranged in combination of four set up in order to capture the response of every node of the structures. Two sensors in node 21 (top floor) in X direction and Y direction were fixed as the reference sensor and other sensors were changed in respective set up. The experimental test was performed for three structural systems, the base model, base model with bracing in the second and fourth floor and base model with base isolation. Same combination of the sensor arrangement was used for the three structural systems.

Modeling of Base Isolated System on the Workshop
The project objective required a design of a base isolated system to be tested. The 1st design attempt was made using two copper pipes intended to slide over each other. The set up consisted of two long copper pipes of smaller diameter and four small pieces of pipes of larger diameter. The load of the structure was intended to be taken by the small pieces of pipes and required to slide over the length of the long pipes. The long pipes were supported on two strip of wood rigidly fixed in the base plate of the shaker. The larger diameter pipe was inserted in the small diameter pipe and the structure was placed on top of the larger diameter pipe. During the experiment, the author realized that the acceleration values required to overcome the sliding friction was too high for the shaker. The load from the structure was too high on the copper pipes it was not possible to slide the structure over the copper pipe. In addition, the long copper pipes supported on the wooden strips were acting as beam supported on the support with the cantilevers and the structure load was acted on the cantilever part. This induced additional load on the long copper pipes thus making it more difficult to slide over it. The 2 nd design attempt to model a base isolated system was attempted with 8 pieces if GI pipes, like shown in the picture below. The attempt was to roll the structure over the pipes between the two bolts used to fix the wooden plank in the base plate. Four wooden strips were fixed in the base plate of the shaker with the help of bolts. The head of the bolts were intended to limit the movement of the rolling pipes. The structure was placed over the pipes and made to roll over the pipes. The length of the pipes used was not long enough and the whole system was not stable. In the end the author realize that, it is needed to design a base isolated system not only can slide but also stable, so it is needed something which is able to pull the structure back to its initial position.
The 3rd or final attempt was finally a success. The used of GI pipes were then replaced by longer copper pipes. The pipes were placed in between the head of the bolt which control the movement of the structure. Also four springs were placed in the middle of the base plate to control the lateral movement of the structure. The spring assured that the system moved back to its original state after excitation. Fig (3.7) 3 rd attempt to model base isolated system (success) The stiffness of the spring used in the base isolation system was computed experimentally. Since there was no laboratory set up available for the test, the test was performed with only the available materials like wooden clamp and glass for putting the additional load.
The test was performed by adding weight to the glass container and thus measuring the elongation of the spring.  Fig (3.9) Calculation of spring stiffness The stiffness of the spring was found to be 374.7 N/m.

Data Processing 4.1. Data Processing with MACEC
The data obtained from the experimental testing was processed using MACEC for the determination of model frequencies and modal vectors. MACEC is an analysis tool box of Matlab. MACEC requires definition of the structure with grid with the coordinates of nodes and elements. It also requires assignment of master nodes and slave nodes definition in order to have a complete visualization of the structures and mode shapes.
The data procession steps consist of: 1. Making the geometry of the structure In order to do this, first the nodes were defined and then connected each node by making a link between nodes to build the structural model. The model consists of 5 storey and 24 nodes. 2. Define the master nodes and the slave nodes 3. Conversion of the digital data obtained into signals 4. Processing the data Sampling frequency of 2048 was used on the experimental test. This step also includes dissemination of the signal, and removal of redundant channels and also assigning degrees of freedom for the nodes. Sensors numbers are assigned to corresponding nodes with direction of measurement.

Analyzing the Data using Stochastic Subspace
The singular values of the system can be got from this step. These values indicate maximum possible system order equal to 1608.

Perform Modal Analysis
In this step the dots were selected. The dots represent the Eigen frequency from stabilization plot in order to get the mode shape of the system. 7. Combine the different sensor setup In order to get the full mode shapes of the system, all of the results from different sensor setup were needed to combine.

Eigen frequencies from MACEC
The analysis of experimental data showed that there were some jumps in the frequencies values for all the three experimental set up. It was noted that frequency range of Model 10 for the braced system obtained from numerical simulation was not available. Similar was the case with the torsion mode for the base isolated system. The frequencies ranges obtained from data processing largely depend on the quality of the experimental data obtained and also on the structural system under observation. The experimental data was obtained for noise excitation of frequency ranging from 0 to 100 Hz and also from impulse excitation. There were some measurement done with sine waves for base isolated structure and the base model. The sine wave data is used for the comparison of the top displacement between the base model and the isolated structural systems. The data was sampling frequency of 2048 was chosen and the time of observation was 200 seconds. In this case, it was able to get 4 Eigen frequencies in the range from 1 to 20 Hz, but from braced model it was only able to get 3 Eigen frequencies in the same range. On the other hand, it was also unable to get the Eigen frequencies for torsion mode. To compensate with this problem, the Eigen frequency was extracted from Impulse excitation result, to have the Eigen frequency for torsion mode.

Fig (4.11) Comparison of Eigen frequencies of each system from MACEC results
From the graph above, it can be seen that the results are more or less similar with the numerical simulation with SAP 2000. Braced model stiffen the structure, so that the Eigen frequency a bit higher. On the other hand, base isolated model make the structure more flexible, indicating by the lower Eigen frequency compare to base model or braced model.

Data Processing with LABVIEW
LABVIEW software is used as a tool to extract the portion of the data in order to get the signal from the digital data obtained from the experimental analysis. The *.lvm files are the time history data with the one column being the time and other column the response acceleration measured. Every lvm file contained 2 second time interval data. Since we had the measurement for 200 sec for every measurement, this means that we have 100 *.lvm file that we need in order to get the time history signal. Since a periodic sine function input was applied to excite the structure, we only took 4 measurement data, which meant that we have a time history signal with duration for 8 sec.
Both the structure systems (Base model and Base isolated model) were excited with sine function excitation with same amplitude and frequency. For measurement data without base isolation, we used measurement 2 nd with the frequency 1.7 Hz and amplitude of 0.2 m/s 2 and for measurement data with base isolation we used measurement 18 th with the frequency with the same frequency and amplitude.
The experimental data was processed using the Labview. The acceleration data was two times integrated to get the displacement history and relative displacement between the top nodal displacement and the support displacement was computed. The obtained result shows that there is significant reduction in the relative displacement of a base isolated system.

Numerical Simulation in SAP 2000 Using Experimental Data
Numerical simulation for base isolated model also performed. This time SAP2000 is used to analyze the base isolated structure with time history function. It was taken the data in the range of 5 sec only, because SAP2000 took a lot of time to load the measurement file, if all the data are used. In the end, the measured response and the simulated response are compared in the graph below:

Discussions and Conclusion
The project aimed at experimental analysis of the base isolated system in comparison to a base model. The base isolated system was designed using available material and equipments in the lab. The experimental result showed that the system used for base isolation reduced the time period of the structure and the relative displacement of the top with respect to the support. The base isolated system introduced in the structure make the structure more flexible thus reduced the affect of the earthquake loads on the structure. The use of bracing elements stiffens the structure. It provides additional stiffness to the structure and decreases in time period.
The model itself needs some improvement regarding the stability of the structure and also robustness to have similar result in repeated measurements. The base isolated method largely depends on the behaviors of the springs attached which provided the stability to the structure. The base isolation system can be improved by providing stable arrangement of spring system and also introducing dampers to reflect the real time model for base isolation.
The analysis of the experiment data showed some deviation from the results obtained from the numerical results. Model frequency value for the second mode for base isolated system show deviation from the numerical result. The spring analogy was used for modeling in SAP as isolators while the real structure was isolated using steel pipes and making roll over the pipes. The deviation may have resulted due to some minor difference in the principles applied for modeling of base isolation.
The experimental results are also affected by the distribution of the sensor. The mass of the sensor also contributes to the response of the structure. The sensor arrangement had a mass concentration in every floor. The sensor arrangement should be arranged that there is no mass concentration and thus result minimum contribution to the structural response.