Study On The Performance Evaluation And Forecast In Pavement Based On GANN

Highway transportation is a nationally public foundation facility which provides service for our society, and represents the modernization level of the whole social development and the real national economy strength. Along with the social economic development, as the increase of volume traffic and heavy load, highway has widely appeared damaged phenomenon. Therefore, in order to prolong its service life, we should take some measures such as maintain and improve its quality based on technique to improve its grade. The pavement performance is defined as an ability to guarantee all vehicles to transport normally and safely. The driving comfort and vehicles' engine performance are affected directly by its capability, and the road transportation quality is also determined by its quality. At the same time, we could find out the inside shortage existed in the management of the road engineering, it also benefits to take into improvement or control in engineering construct and perfect the theory of road construction in initial or using stage.This text introduced various breaks and damages that may appear in highway and then analyzed the results of the diseases, studied the index of the pavement and discussed the measurement of various index and its rating standard. Based on the complementation of advantages of neural network and genetic algorithm, we established the GANN(genetic algorithm and neural network) model and then used the model to evaluate and forecast the pavement performance of the road.Through the state highway road, provincial highway and the county road (excluding township road and dedicated road) in Shanghai, we analyzed the factors that impact the pavement performance of the road and chose the appropriate indicators to establish a genetic neural network models. According to the relationship of the pavement condition index (PCI), road quality index (RQI), surface strength index (SSI), sideway-force coefficient (SFC) and the pavement quality index (PQI) from 2002 to 2004, we used this model to evaluate the pavement performance of the road in 2005, the test results and the actual results are very consistent. This model shows very good simulation between the four indicators and the quality of road;moreover, according to the relationship between the cost of road maintenance and pavement conditions index, road quality index, traffic conditions, the length and area of maintenance in road every year, we used the tested training model to forecast the conservation of funds in each district of highway department of shanghai in 2006, the results of the forecast indicated that this ideas and methods were feasible, this model could provide effective guidance and help to the maintenance of road.