ABSTRACT
This section contains a brief review of experimental research in the field of monotonic behaviour of semirigid connections. This is essentially aimed at identifying the main features of the moment M - rotation <]) curves that are of interest for the purpose of developing analytical methods of prediction. In particular, the significant phases of behaviour that are common to most test curves are evidenced, even if it is to be pointed out that the relative importance of each phase varies from case to case. In past years, a large number of test programs has been conducted in order to provide a better understanding of the behaviour both in USA and in the European countries. Among the former ones, reference is made to tests performed by Azizinamini et al. (1987) on top-and-seat angle connections with web angles and to those by Lewittet al (1966), reported by Kishi and Chen (1990), on double web angle connections. For what concerns the tests conducted in Europe, curves obtained by Jaspart (1991) and by Bernuzzi et al. (1992) for top-and-seat angle connections are considered. Samples of all the above test programs are reported in Figure 1. It can be seen that significant slip phenomena affected the moment-rotation response of top-and-seat angles tested by Jaspart and Bernuzzi et al., whereas joints tested by Azizinamini et al. and by Lewitt were able to develop continually increasing moments through the full range of imposed rotations. Effect of bolt slippage is examined in a companion paper (Bernuzzi et al. 1994) with reference to tests made by Bernuzzi et al. (1992). Except for this aspect, all curves denote similar behaviour, characterized by three phases: 1) elastic phase, with rotational stiffness almost constant and approximately equal to its initial value SJiinil; 2) yielding phase, within which the moment-rotation curves exhibit the largest nonhnearity; 3) hardening phase, within which the response can be considered linear, as in the first phase, with a rotational stiffness Sjh far smaller than S]Mt. As far as the hardening phase, De Stefano and Astaneh (1991) and Jaspart (1991) have investigated the physical phenomena that are involved in this state, leading to the conclusion that for monotonic behaviour the ratio S]h/S]lmt is related to the ratio of the steel hardening modulus Eh to the Young's modulus E. Namely, in case of top-and-seat angle connections, Jaspart (1991) suggested the following relationship:
A precise evaluation of the ratio EJE is rather complex, since there is no well-established definition for Eh. Furthermore, Ek is dependent on several factors. Massonet and Save (1980), based on a wide collection of experimental data, showed that EJE is affected by the strength of the steel. However, for low and medium strength (mild) steels, that are the most widely used, values of EJE were always close to 1/50. Kato et al. (1990) conducted a statistical analysis on mechanical properties of steels usually used in Japan. The mean values of EJE reported in their study was close to 1/60. It has to be said, though, that if one wants to assume a reasonable bilinear approximation for the entire stress-strain curve, this value grows up to approximately 1/30. In any case, it has to be said that there is no
570 de Stefano, Bernuzii, d 'Amore, de Luca and Zandonini general agreement about values of Eh/E to be considered for different types of steels. At the connection level, experimental information on the ratio Slh/SJim[can be inferred from the results obtained by Attiogbe and Morris (1991), when fitting moment-rotation curves of five types of semirigid connections with analytical expressions depending on both stiffness parameters. Referring to top-and-seat angle connections, the ratio SjtJSj:ml was ranging between 1/15, for low stiffness connections, and 1/43 for high stiffness.
