سال انتشار: ۱۳۹۱

محل انتشار: نهمین کنگره بین المللی مهندسی عمران

تعداد صفحات: ۸

نویسنده(ها):

Arash Adib – Assistant Professor, Shahid Chamran University, Engineering Faculty, Civil Engineering Department
Mohammad Vaghefi – Assistant Professor, Persian Gulf University, Engineering Faculty, Civil Engineering Department
Soroosh Alahdin – MSc, Khuzestan Water &Power Authority (KWPA Co.), Ahvaz,Iran

چکیده:

Drought and shortage of water are very important problems. For overcoming on these problems, attention to water resources management is essential. For water resources management, it needs to sufficient and confident hygrometry data. The most of Iran’s hydrometric stations have not sufficient data. For preparation of needing data, synthetic data must be produced. The method of producing of synthetic data has to make used of probability concepts and saves main characteristics of data too. The Markov chain method is a suitable method for generation of synthetic data. In this research, synthetic hydrometric data are generated by the monthly Markov chain method and the annual Markov chain method in five hydrometric stations of the upstream of the Dez River. The constructed dams do not regulate discharge of the Dez River in this region. Among of these stations, the Telezang station has the most exact hydrometric data. The Telezang station was selected as base station. It was evaluated relation between data of other stations to data of the Telezang stationby the multi sites Markov chain method. Linear regression relations were extracted by this method. These relations show discharge of other stations as function of discharge of the Telezang station. By using of discharge of the driest day and the wettest day of each month and the generated monthly hydrometric data of each station, it is calculated the probable highest daily discharge and the probable lowest daily discharge in each station and each month. At the end, artificial neural network was trained by a number of observed hydrometric data and generated hydrometric data. The results of artificial neural network were compared to a number of observed hydrometric data that they did not apply to training of network. Training of artificial neural network by generated hydrometric data improved results of network. For more improvement of results of network, genetic algorithm was applied for training of network and optimization of parameters of network. Artificial neural network showed correctness of generated hydrometric data