Deliverables
The conventional pseudodynamic (PsD) process, involving intermittent application of displacements, has been replaced by the Continuous Pseudodynamic (CPsD) method for testing large structures via a reaction-wall or reaction frame. This reduced to a large degree the major disadvantages of reaction-wall testing, strain-rate effects and stress relaxation. A new continuous PsD algorithm has been proven and dynamic substructuring has been incorporated into the test programme. Implementation aspects of the PsD/CPsD method have been studied, particularly the need to achieve a high-speed algorithm so as to improve on the capabilities of the CPsD method.
The use of the MCS algorithm within the CPsD method has been studied, the objective being to overcome the problems association with fixed gain control when the dynamics are not defined. Another successful research result has been the implementation of the CPsD algorithm on single board computers.
The extensive theoretical research described above has been accompanied by three series of experimental tests:
- CPsD tests on large-scale structures.
- Low speed CPsD tests on large structures with non-linear substructuring. They have been specially designed to assess the behaviour of long-span irregular bridges subjected to asynchronous input motion.
- High speed CPsD tests with substructuring in which the critical sections are tested at full-scale, with the remainder modelled numerically. The purpose of these tests is to study the behaviour of strain-rate sensitive devices used to protect structures against earthquakes.
Project Reference: FMGE980103
Inverse and forward kinematic models have been developed for multi-DOF shaking tables. These are essential for the translation of kinematic data (displacement, velocity and acceleration) between table and actuator axes. The analytical background to the MCS algorithm has been developed for shaking table applications, involving two types of retrofit strategies. These have allowed MCS to be implemented on a shaking table controller facility without the need to remove the existing controller. Two schemes were identified as the 'inner-loop' and 'outer-loop' strategies.
The concept of composite filters has been developed with details of each type of filter. The purpose of composite filters is to combine kinematic data measurements into a single 'optimal' estimate of the required signal. In particular, displacement and acceleration responses have been combined into a better estimate of displacement over a 0-30Hz frequency range. This composite filtering technique has also been shown to be valuable in the generation of demand data, e.g. for the generation of displacement demand from pre-recorded seismic acceleration data.
A software package TABCTRL (Table Control) has been developed for the adaptive control of multi-axis shaking tables. Using MCS, new research areas have been developed in non-linear behaviour of test pieces, multi-support input, sub-structuring on shaking tables and the effect of spurious motions.
Project Reference: FMGE980103
A completely new approach to the interpretation of seismic test results has been implemented at LNEC Lisbon, using research performed there at JRC Ispra. The research was divided into three parts:
- Improvement of data transfer protocols between data acquisition systems and post-processing computers.
- Computer implementation of numerical techniques for system identification.
- Trial applications to ongoing seismic tests.
Through the first research part, an almost online assessment of data acquired during shaking table tests is possible; this increases the accessibility of the facility to external users. The software developments for the second research part - system identification techniques were in the object-orientated language LabView, which is compatible with the new LNEC DACE system of the first part above. A feature of the new procedures is the ability to use existing software libraries for structural analysis, signal processing and system ID. Based on the Ibrahim Time Domain Method, one of the features of the new procedure is the ability to obtain the dynamic characterisation of structural properties using the final decay motions at the end of the input. The trial applications referred to in the third research part were conducted in collaboration with EERC Bristol on a specially designed test-piece, which exhibited pronounced non-linear behaviour. Two principal results emerged:
- It is possible to obtain a seismic vulnerability assessment by testing a single test-piece, rather than the expensive alternative of using virgin test-pieces for each level of excitation.
- Bayesian inference strategies can be used to update fragility, on the basis of prior knowledge, plus data from experimental tests.
Project Reference: FMGE980103
Searching for OpenAIRE data...
There was an error trying to search data from OpenAIRE
No results available