Fig. 2.2.3 some recommended improvements in the interim guidelines
The following possible causes, and their combinations, were found to have contributed to tile connection failures:
~ Inadequate workmanship in the field welds.
~ Insufficient notch-toughness of the weld metal.
~ Stress raisers caused by the backing bars.
~ Lack of complete fusion near the backing bar.
~ Weld bead sizes were too big.
~ Slag inclusion in the welds.
While many of the failures can be directly attributed to the welding and the material of the joints, there are more serious questions relative to the structural system that had evolved over the years mainly based on economic considerations.' The structural system used relatively few rigid-frames of heavy members that were designed to absorb the seismic forces for large parts of the structure. These few lateral-force resistant frames provide insufficient redundancy. More rigid-frames with smaller members could have provided a tougher and more ductile structural system. There is a question of size effect: Test results from joints of smaller members were extrapolated to joints with larger members without adequate test verification. The effect of a large initial pulse may have triggered dynamic forces that could have caused brittle fracture in joints with fracture critical details and materials. Furthermore, the yield stress of the beams was about 30% to 40% larger than the minimum specified values assumed in design, and so the connection failed before the beams, which were supposed to form plastic hinges.
As can be seen, there are many possible reasons for this massive failure rate, and there is blame to go around for everyone. No doubt, the discussion about why and how the joints failed will go on for many more years. The structural system just did not measure up to demands that were more severe than expected. What should be kept in mind, however, is that no structure collapsed or caused even superficial nonstructural damage, and no person was injured or killed. In the strictest sense the structure sacrificed itself so that no physical harm was done to its users. The economic harm, of course, was enormous.
7. Future directions of structural steel research and conclusion
The future holds many challenges for structural steel research. The ongoing work necessitated by the two recent earthquakes that most affected conventional design methods, namely, the Northridge earthquake in the US and the Kobe earthquake in Japan, will continue well into the first decade of the next Century. It is very likely that future disasters of this type will bring yet other problems to the steel research community. There is a profound change in the philosophy of design for disasters: We can no longer be content with saving lives only, but we must also design structures which will not be so damaged as to require extensive repairs.
Another major challenge will be the emergence of many new materials such as high-performance concrete and plastic composite structures. Steel structures will continually have to face the problem of having to demonstrate viability in the marketplace. This can only be accomplished by more innovative research. Furthermore, the new comprehensive limit-states design codes which are being implemented worldwide, need research to back up the assumptions used in the theories.
Specifically, the following list highlights some of the needed research in steel structures:
Systems reliability tools have been developed to a high degree of sophistication. These tools should be applied to the studies of bridge and building structures to define the optimal locations of monitoring instruments, to assess the condition and the remaining life of structures, and to intelligently design economic repair and retrofit operations.New developments in instrumentation, data transfer and large-scale computation will enable researchers to know more about the response of structures under severe actions, so that a better understanding of "real-life" behavior can be achieved. 钢筋混凝土结构研究英文文献和翻译(5):http://www.youerw.com/fanyi/lunwen_2833.html