پديد آورندگان :
نجفي، اسماعيل نويسنده دانشجوي دكتري ژيومورفولوژي دانشكده علوم جغرافيايي، دانشگاه خوارزمي تهران , , صفاري، امير نويسنده دانشيار گروه ژيومورفولوژي دانشكده علوم جغرافيايي، دانشگاه خوارزمي تهران , , قنواتي، عزت‌الله نويسنده دانشيار گروه ژيومورفولوژي دانشكده علوم جغرافيايي، دانشگاه خوارزمي تهران , , كرم، امير نويسنده دانشگاه تربيت معلم تهران ,
كليدواژه :
دبي هاي حداكثر لحظه اي , شبكه عصبي مصنوعي , شبيه سازي , كلان شهر تهران , ايستگاههاي هيدرومتري
چكيده لاتين :
Introduction
Numerical analysis and advanced statistical has turned into hydrology as an exact science. Today this knowledge in the design of hydraulic structures such as reservoirs and diversion, constructions of irrigation and drainage canals and bridges, river engineering and flood control, watershed management, road construction, municipal and industrial wastewater and environmental and physical geography widely is used. Thus, the role of hydrology in the correct and exact assessment of water projects and conservation measures will be very sensitive. Artificial neural network is a simplified model of the human brain is capable of showing non-linear relationship between the input and output components of each system. The network is trained to predict the learning process used in the future. The wide spread use of artificial neural networks as an experimental model and efficient in different sciences such as hydrology represents the the high value of these models. Important factor in the decision and discuss the construction of river structures, location and construction of bridges is be aware of the maximum discharge instantaneous, simulation and analysis in hydrometric stations. Aim of research is analysis and simulate of peak discharge through artificial neural network techniques in order to location, construction and optimal maintenance of bridges.
Methodology
The research method is descriptive -analytical. For simulation and analysis of maximum instantaneous discharges using an artificial neural network, the data rate of hydrometric stations in Haft howse(Darakeh), sooleghan(Kan), Qlak (Darabad), Maqsud Beyk (Darband) that itʹs data from the Water Resources Management of Iran (Tamab) and Tehran regional Water Authority was received. Then Using with Excel software flow diagram depicting hydrometric stations were analyzed at each moment. The following data for the month of peak discharge, month after month peak discharge, months discharge before the peak of discharge, maximum discharge day and peak discharge Code months as network input and output of the network was set up as instantaneous peak discharge.
Results and discussion
For making of neural network, default of NeuroSOlution5 Multilayer Perceptron (MLPS) was used, and after fitting the various functions and rule learning BiasAxon Momentum transfer function in the hidden layer and output layer of the network as the best mode was used. In neural networks, data is generally divided into three, training, validation and testing can be divided. in this study, the data were divided into three sections like above, on the basis of the results graphically in Figure 10 Shows that the predicted values to the actual (measured) and the correlation coefficient has a value of approximately optimal fit in (R=0.66). The mean squared error is (MSE=1.59) lower amounts of test data to show the utility of the training data. In contrast to this, the mean squared error (MSE=0.02) for the training data to the test and the lower value indicates better performance and indicates that the network is properly and well-trained. considering the hydrometric stations are located in the watershed outlet, resulting in a number of different rates (maximum instantaneous flow rate of a maximum one-day, monthly and annual) are recorded, as well as a sensitivity analysis with respect to the input data and the output from the neural network, the effect of month peak discharge, the greatest impact on output or maximum had a peak moment, It can be stated that these factors and environmental factors feeder programs and structures associated with the watershed cross the bridge and the construction and communication played a crucial role the optimal position. Simulation diagram between the measured data and estimated data using neural networks, also indicate that a high correlation are exist between actual and predicted values. Finally we can say that, in the study area and according to the discharge Statistics hydrometric stations haft howse, sooleghan, Qlak, Maqsud Beyk in Tehran metropolitan The highest frequency peak discharge rates for the month of maximum moment in the spring months (April and march) occurred, in addition to rainfall in the spring, because of its mountains and snowy slopes, snow melt water in basins and the area under study, the flow rate is also influenced and therefore requires that any plan for the basin and floodways and the construction of structures and construction of a bridge over the river and positioning of bridge they must consider discharge values in this months and great interest to periods return of them.
Conclusion
The results show that in all stations that studied using artificial neural network to simulate peak discharge at the moment of maximum daily discharge data have a high efficiency. It can be said that in the Haft howse stations and Qlak discharge fluctuation is large but sooleghan and Maqsud Beyk has little fluctuation is around the mean. Given the high fluctuation of peak discharge of Haft howse (Darakeh) and Qlak (Darabad) It can be deduced that the bridges that are built on the river or stream have a flood risk and possibly unstable And to location and construction of such structures must be thinks necessary to measured and Hydro geomorphological characteristics of the upstream basins and flood return periods are considered. Also, due to the numerous rivers and floodways in Tehran metropolis, North-South gradient, changes in land use and manipulation of space and floodways channels, creating a fusion basins, impermeable surfaces and paved city in recent decades and as mentioned above, the maximum instantaneous flow rate fluctuations in different years and record flooding, flood event in Tehran is permanent. To prevent loss of life and property caused by floods and increase the welfare and safety of the citizens, require a comprehensive management and interdisciplinary with systematic (basin) approach.
