Final Report Summary - RERCSGM (Risk Estimation of Rocking Components Subjected to Ground Motions.)
A wide range of structural and non-structural elements are not tied or bolted to their support medium and thus are free to experience large displacements and rotations, when subjected to ground motions, which may overall be characterized as rocking. Unlike typical civil engineering problems, where failure criteria are associated to fracture, yield, or maximum allowable levels of stress and strain, failure of rocking bodies may occur either due to excessive displacements that may lead to collisions, or large rotations that may lead to the body overturning.
The structural performance of rocking bodies subjected to ground motions is of great importance due to a number of reasons: Some rocking objects may play a significant role in the infrastructure, as in the case of: trains, electrical transformers, medical equipment and carts in hospitals or even supercomputers and data storage centres. There are also rocking objects whose failure might have dangerous implications for human societies and the environment as in the case of: containers of nuclear waste and water tanks, or might lead to direct human losses as in the case of seismic isolated building. Moreover, there are cases where the rocking objects may be priceless and irreplaceable as in the case of museum exhibits and historic monuments. Finally, the recent recognition of the fact that damage to non-structural components is a major fraction of direct property losses , has increased the interest in simulating the behavior of these bodies, a behavior dominated by sliding and rocking.
It is therefore important to have models that accurately describe the behavior of rocking objects when subjected to ground motions, understand the mechanisms that may lead to failure, quantify the associated risk and suggest solutions to minimize it. The project aimed at the following objectives:
1. Further development of the rocking models and expansion to additional applications. This involves understanding the limitation of the assumptions of existing models and relaxing them to increase their accuracy and the range of their applicability, as well as creating novel models for simulating the response of more general rocking systems.
2. Estimating the risk of structural and non-structural components subjected to 3D ground motions.
This objective includes the development of new - and the comparison of existing- methods for generating 3D ground records corresponding to various intensity measures and using them with deterministic models to obtain suitable curves that capture the probability of failure of the objects.
Different intensity measures will be considered and evaluated in terms of their ability to capture the risk of failure of rocking objects.
3. System Identification of Rocking Bodies Using Heterogeneous Data Fusion and novel system
Identification Algorithms. This requires the development of suitable algorithms for identifying the properties of rocking models based on experimental data. The goals here are related to: devising suitable experiments that capture the dynamics of the problem, creating new system identification algorithms capable of handling the discontinuous behavior of rocking systems and improving data, using data fusion techniques from heterogeneous sensors.
The project achieved several important milestones within the directions of the objectives. There have been several achievements beyond the state of art. Those can be summarized in the following bullet points:
• The work has shown both theoretically that there is an uncertainty of Inverted Pendulum models in terms of the energy lost during impacts. The work further showed that this energy depends on the contact interface and that this can be exploited for design purposes in both single and stacked rocking bodies.
• A series of experiments were conducted in the Carleton lab of Columbia University illustrating the previous points experimentally.
• The rocking of stacked bodies was studied. Conclusions were drawn towards the use of pedestals in museums and their design to minimize the response of the exhibits.
• A robust framework for estimating the risk of rocking bodies was developed. The framework allows for obtaining the probability of failure or a rocking body for a given intensity measure or versus a design period. Those can be used by engineers for comparisons versus non rocking systems.
• A method based on the KL decomposition was developed for generating ground motions compatible with seismic intensity measures. The obtained records are not confined by a model and share the properties of the list of records that was used to produce them.
• A framework for Palaioseismology was developed. This will lead towards a methodology for the assessment of seismicity of regions given historic evidence.
• A method for identifying non-smooth systems was developed: The Discontinuous modification allows for improving online Bayesian algorithms in identifying non-smooth Systems. Such systems often occur when modeling damage.
• Two robust algorithms were developed: the DEKF and DUKF that are both more robust than their standard counterparts and allow for identifying problems involving impacts, sliding, plasticity or fracture (and generally non-smooth behavior)
• The Discontinuous modification was used to successfully identify in online manner a rocking body.
The project achieved the tasks described in the proposal and often exceeded them. The direction of the Identification of Rocking bodies was extended to the broader problem of non-smooth systems thus allowing applicability in several different fields of engineering.
As a result of this project 3 papers have been published in international journals(with a fourth accepted at the day of submitting this report), 4 papers have been published in proceedings of international conferences and 12 presentations were given in international conferences with peer review of abstracts.
A PhD student, Maria Garcia Espinosa, has worked in the project since October of 2014. The student is working towards the completion of her Phd. She has co-authored (together with the PI M Chatzis) 2 published papers in international journals and 1 paper in the proceedings of international conferences. Maria has presented in the 16th WCEE in Chile and has contributed to two more conference presentations.
Thanks to the support of the Marie Curie CIG, the PI Chatzis, acquired a new biaxial shake table for the Dynamics Lab of the University of Oxford, ensuring continuation of the project and in-house testing capabilities for his group.
At the end of the project the PI Chatzis had achieved his full integration in the University of Oxford, having returned to Europe from the US, and wholeheartedly thanks the EC for this Marie Curie Career Integration Grant that substantially facilitated the process.
The structural performance of rocking bodies subjected to ground motions is of great importance due to a number of reasons: Some rocking objects may play a significant role in the infrastructure, as in the case of: trains, electrical transformers, medical equipment and carts in hospitals or even supercomputers and data storage centres. There are also rocking objects whose failure might have dangerous implications for human societies and the environment as in the case of: containers of nuclear waste and water tanks, or might lead to direct human losses as in the case of seismic isolated building. Moreover, there are cases where the rocking objects may be priceless and irreplaceable as in the case of museum exhibits and historic monuments. Finally, the recent recognition of the fact that damage to non-structural components is a major fraction of direct property losses , has increased the interest in simulating the behavior of these bodies, a behavior dominated by sliding and rocking.
It is therefore important to have models that accurately describe the behavior of rocking objects when subjected to ground motions, understand the mechanisms that may lead to failure, quantify the associated risk and suggest solutions to minimize it. The project aimed at the following objectives:
1. Further development of the rocking models and expansion to additional applications. This involves understanding the limitation of the assumptions of existing models and relaxing them to increase their accuracy and the range of their applicability, as well as creating novel models for simulating the response of more general rocking systems.
2. Estimating the risk of structural and non-structural components subjected to 3D ground motions.
This objective includes the development of new - and the comparison of existing- methods for generating 3D ground records corresponding to various intensity measures and using them with deterministic models to obtain suitable curves that capture the probability of failure of the objects.
Different intensity measures will be considered and evaluated in terms of their ability to capture the risk of failure of rocking objects.
3. System Identification of Rocking Bodies Using Heterogeneous Data Fusion and novel system
Identification Algorithms. This requires the development of suitable algorithms for identifying the properties of rocking models based on experimental data. The goals here are related to: devising suitable experiments that capture the dynamics of the problem, creating new system identification algorithms capable of handling the discontinuous behavior of rocking systems and improving data, using data fusion techniques from heterogeneous sensors.
The project achieved several important milestones within the directions of the objectives. There have been several achievements beyond the state of art. Those can be summarized in the following bullet points:
• The work has shown both theoretically that there is an uncertainty of Inverted Pendulum models in terms of the energy lost during impacts. The work further showed that this energy depends on the contact interface and that this can be exploited for design purposes in both single and stacked rocking bodies.
• A series of experiments were conducted in the Carleton lab of Columbia University illustrating the previous points experimentally.
• The rocking of stacked bodies was studied. Conclusions were drawn towards the use of pedestals in museums and their design to minimize the response of the exhibits.
• A robust framework for estimating the risk of rocking bodies was developed. The framework allows for obtaining the probability of failure or a rocking body for a given intensity measure or versus a design period. Those can be used by engineers for comparisons versus non rocking systems.
• A method based on the KL decomposition was developed for generating ground motions compatible with seismic intensity measures. The obtained records are not confined by a model and share the properties of the list of records that was used to produce them.
• A framework for Palaioseismology was developed. This will lead towards a methodology for the assessment of seismicity of regions given historic evidence.
• A method for identifying non-smooth systems was developed: The Discontinuous modification allows for improving online Bayesian algorithms in identifying non-smooth Systems. Such systems often occur when modeling damage.
• Two robust algorithms were developed: the DEKF and DUKF that are both more robust than their standard counterparts and allow for identifying problems involving impacts, sliding, plasticity or fracture (and generally non-smooth behavior)
• The Discontinuous modification was used to successfully identify in online manner a rocking body.
The project achieved the tasks described in the proposal and often exceeded them. The direction of the Identification of Rocking bodies was extended to the broader problem of non-smooth systems thus allowing applicability in several different fields of engineering.
As a result of this project 3 papers have been published in international journals(with a fourth accepted at the day of submitting this report), 4 papers have been published in proceedings of international conferences and 12 presentations were given in international conferences with peer review of abstracts.
A PhD student, Maria Garcia Espinosa, has worked in the project since October of 2014. The student is working towards the completion of her Phd. She has co-authored (together with the PI M Chatzis) 2 published papers in international journals and 1 paper in the proceedings of international conferences. Maria has presented in the 16th WCEE in Chile and has contributed to two more conference presentations.
Thanks to the support of the Marie Curie CIG, the PI Chatzis, acquired a new biaxial shake table for the Dynamics Lab of the University of Oxford, ensuring continuation of the project and in-house testing capabilities for his group.
At the end of the project the PI Chatzis had achieved his full integration in the University of Oxford, having returned to Europe from the US, and wholeheartedly thanks the EC for this Marie Curie Career Integration Grant that substantially facilitated the process.