Keywords: seismic micro-zoning, ANN, earthquake affecting coefficient, response spectrum, Tangshan.

China is one of the countries with frequent continental earthquakes. The earthquakes caused the loss of more than 400,000 lives in the 20th century. In Tangshan City, the calamitous earthquake in 1976 resulted in the loss of 243,000 lives and there have been many micro-earthquakes in recent years [1]. Investigation of natural disasters in cities is the key of city planning and disaster reduction. Therefore, more attention has been paid to seismic micro-zoning and earthquake resistance [2,3]. The main work of seismic micro-zoning is the design of earthquake response spectra. Earthquake affecting coefficient and characteristic period (T_g) are the main design parameters of earthquake response spectra, other parameters can be calculated based on the structure of building to be designed. In the former research, the influence of site on seismic micro-zoning has been investigated [4], therefore, precise calculation of earthquake affecting coefficient is the key to seismic micro-zoning. Distribution of earthquake affecting coefficient in cities cannot be described precisely without consideration of the influence of the basement rock condition. It is very difficult to consider the influence of the underground rock condition in seismic micro-zoning, because the data of the basement rock condition cannot be obtained from experiments. From the relationship analysis between the earthquake and the stress field, we know that the basement rock condition can be represented by the numerical simulation of stress field in earthquake zones.

Tangshan, a heavy industrial city with six million people, and located in the northeast of Hebei Province, has frequently suffered from natural disasters such as earthquake, karst collapse, and ground collapse of mined-out areas. Earthquake is the primary disaster in Tangshan City and the calamitous earthquake in 1976 destroyed most of the buildings and infrastructure [5]. By analyzing the geological data of Tangshan City, structural stress field of the basement rock is calculated based on elastic theory and the finite element method, also the distribution of the basement fracture probability is predicted. It was found that the distribution of earthquake affecting coefficient is controlled by the basement displacement, the basement fracture probability, site category, physiognomy and special geological condition, etc. The relationship between the earthquake affecting coefficient and influencing factors is complicated, so the prediction model is constructed on the basis of an artificial neural network (ANN). During the calculation, 4 elements are decided for the input layer and 1 for the output layer. There are 2 concealed layers, 6 elements for No.1 concealed layer, and 3 for No.2 concealed layer. The learning coefficient is 1.2 and the inertia coefficient is 0.5. After repeated calculation of 197012 times, the distribution of the earthquake affecting coefficient in Tangshan City is calculated precisely.

In the seismic micro-zoning of Tangshan City, the earthquake affecting coefficient is obtained from calculating the result based on the ANN described in this article, and T_g of different sites can be obtained from site investigation. All parameters for the design of earthquake response spectra are obtained, and different response spectrum for every kind of site is designed. There are three kinds of sites in Tangshan City, each with three characteristic periods (T_g) based on the result of earthquake risk evaluation. So nine kinds of earthquake response spectra were obtained, in which the parameter of earthquake affecting coefficient () becomes a variable. Therefore, earthquake response spectrum is precisely designed. Earthquake loads can be calculated carefully for building design. For example, if the exceeding probability is larger than 30%, T_g is 0.3 and is between 0.2 and 0.28 in Area I, T_g is 0.4 and is between 0.28 and 0.56 in Area II, T_g is 0.55 and is between 0.56 and 0.64 in Area III.

The distribution of the earthquake affecting coefficient can be predicted precisely by predictive model based on the ANN, which makes the parameter of the earthquake affecting coefficient become a variable. Therefore, seismic micro-zoning can be designed carefully. According to the results, some advice on the design of earthquake-proof buildings and disaster reduction planning in Tangshan City is proposed.

1

Q.J. Zhu, Y.P. Su, "Influence of Base Rock Condition on Earthquake Affecting Coefficient", Chinese Journal of Geotechnical Engineering, 26, 198-201, 2004.

2

T. Topal, V. Doyuran, N. Karahanoglu, etc, "Microzonation for Earthquake Hazards: Yenis_ehir Ssettlement, Bursa, Turkey", Engineering Geology, 70, 93-108, 2003. doi:10.1016/S0013-7952(03)00085-1

3

Z. Ding, Y.T. Chen, G.F. Panza, "Estimation of Site Effects in Beijing City", Pure and Applied Geophysics, 161, 1107-1123, 2004. doi:10.1007/s00024-003-2495-9

4

I.A. Parvez, F. Vaccari, G.F. Panza, "Site-specfic Microzonation Study in Delhi Metropolitan City by 2-D Modelling of SH and P-SV Waves", Pure and Applied Geophysics,161, 1165-1184, 2004. doi:10.1007/s00024-003-2501-2

5

Q.J. Zhu, Y.P. Su, T.Q. Liu, "Neural Network Predictive Model of Disaster Risk Evaluation in Tangshan City", Journal of Electronics and Information Technology, 24(supp1), 10-16, 2003.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £80 +P&P)