Research And Application Of Improved Gray Bernoulli Model For Water Consumption Prediction

Water,one of basic natural resources,impacts on human life and socio-economic development.In recent years,with the social and economic development,the industry,agriculture and other needs for water continues to increase,leading to inadequate water resources capacity.Therefore,it is necessary to strengthen the water resources management and realize the effective planning and scientific dispatching of water resources in water supply systems.In the water supply system planning and water resources scheduling study,the primary goal is to forecast water consumption.Water consumption has complexity and non-linear characteristics,need to seek the complex scientific theory and nonlinear prediction method to characterize the change and development trend of water consumption.The predicted value of water consumption will directly affect the reliability of practicality and water dispatching decision.Compared with the widely used multivariate statistical method,the non-linear water consumption forecasting method has great space to explore and research value.For the purpose of improve the forecasting accuracy of water consumption,based on the water supply system's large-area planning and small area short-term optimization and controlling needs,this paper took the water consumption of different regions,different categories and different periods as the research object.With the support of gray theory and gray prediction method,the model suitable for water consumption prediction was constructed by improving the gray Bernoulli model,and the potential law and development trend of historical water consumption data were excavated.The research work not only realized the high precision prediction of water consumption,but also provided technical support for the optimal regulation of water supply system.The main work and conclusions are as follows:?1?Based on the data of annually water consumption in life and agricultural in Hubei Province,the data of monthly water consumption in Wuhan.The trend of water consumption in different regional and different time periods was analyzed.Compared and analyzed several commonly water consumption prediction methods.?2?Experiments proved the limitation of classical gray model?GM?1,1??model in water consumption forecast.This paper introduced the nonlinear gray Bernoulli model?NGBM?1,1?model?for nonlinear problems.Based on the problem that the model could not identify the fluctuation of the water consumption data and the accuracy of the predicting results can not reach the requirement of certain water management accuracy level.This paper put forward the improvement idea of"external"and"internal".By the metabolism of the original data,the external improvement model was established.By the improvement of the gray sequence accumulation with the fractional operator and replacement of a new dynamic gray action b₁cos??k?+b₂k+b₃ to the original solid state action,the improved gray Bernoulli model for water demand prediction?FNGBM?1,1|cos,t?water consumption prediction model?is established,Strengthen the model's ability of water consumption volatility identification and development trend prediction.At the same time,taking into account the importance of model parameters selection,this paper optimized the model parameters with the adaptive step size artificial fish swarm algorithm.?3?Based on the requirement of future water use plan in the large area and the short-term water control in small areas,annually water consumption of life and agriculture in Hubei Province and the monthly total water consumption in Wuhan in2013 and 2014 were taken as examples to evaluate the water consumption prediction of the model.The results showed that FNGBM?1,1|cos,t?water consumption prediction model had the characteristics of less data demand,simple construction process,strong operability and wide application range.Compared with NGBM?1,1?model,the prediction accuracy was improved,and the degree of reality was reduced,reflecting the strong excavation capacity of the potential information,the ability to predict the future water consumption and so on.Provided a convenient and common new technical means For the water consumption prediction.