Proceedings of the Fourth National Conference on Wind Engineering 2007, SERC, Chennai, 30th October to 1st November 2007, pp.477484 ! " 1 Ph.D. Student, 2 Director Wind Engineering Research Centre, Tokyo Polytechnic University, 1583, Iiyama, Kanagawa 2430297, Japan Email: d0484201@st.tkougei.ac.jp, yukio@arch.tkougei.ac.jp # In this paper, a method for calculating the peak factor for any stationary random process which is Gaussian or nonGaussian and narrow band or nonnarrow band is shown. Current estimation of calculating peak values for wind loads and load effects in ASCE 702, AIJ is based on the assumption that they are of Gaussian distribution. However this assumption is erroneous when the random process is nonGaussian. Also in the major codes the bandwidth parameter in the frequency domain is ignored. In the wind engineering problems we come across various random processes which show nonGaussian and the nonnarrow band processes. It is important to consider these parameters while calculating the peak value of any random processes in the wind engineering applications. The generalized peak factor shown in this paper has a wide application irrespective of the Gaussian, non Gaussian, narrow band, nonnarrow band processes. Wind tunnel tests were conducted for low and mediumrise building models with simultaneous pressure measurements on their surface. There are different pressure points considered for the validation of peak pressures. The normal stress acting on the corner columns are also considered for validating the derived peak factor. The generalized peak factor validated with experimental results, shows good agreement when compared to the existing methods. The probability distribution function, power spectral density etc., are also shown for the various random processes. $%&’($% ! ! " # # $ ! ! %&$’ () & *+((+, %-. *+((, &/0 *12), # $ 3 *12)4,