Study On Soft-measurement Model Of Ammonia-nitrogen And Operational Energy Consumption In Urban Sewage Treatment Plant

Sewage treatment plant is the basic infrastructure in city.It's an important means to control water pollution and improve the quality of urban water environment.The inlet water quality condition is an important basis for the design of sewage treatment plant.The fluctuation of the inflow water quality will make it difficult for the sewage treatment process to stably operate in the design conditions.This will not only affect the treatment results,but also make the equipment deviate from the efficient working state for a long time and increase the energy loss.According to the statistics,the average power consumption of sewage treatment in China is 0.29 kWh/m³ and the total electricity consumption is about100×10⁸kW·h.In the process of operation,the energy consumption expense accounts for60%to 80%of the operation and maintenance cost.High power consumption restricts the development of the industry of sewage treatment.Understanding the treatment effect of the sewage treatment plant systematically,according to the condition of inflow water quality,adopting the real-time measurement,optimizing the process parameters to control the key water quality indicators,which is important to improve the efficiency of sewage treatment and reduce the operating power consumption.At the meantime,it is helpful to promote the energy conservation and sustainable development of the sewage treatment plant.This passage take Qingyuan sewage treatment plant as a study,through investigating and collecting the relevant date of water quality indicators,running energy consumption.,using MTALAB software to analyze the relation between water quality indicators,adopting soft measurement model to predict ammonia nitrogen value,knowing the water quality situation quickly,accurately and accurately,which is helpful to up to effluent standard and to make management efficient for the sewage treatment plant.According to the characteristics of the sewage treatment plant,it use genetic algorithm to built ammonia nitrogen and operational energy consumption forecasting model,by using genetic algorithm to optimize it.The main conclusions of this paper are as follows:?1?The main energy consumption of the plant in Qingyuan is the pump and Power consumption of fan.In summer,the plant consumes lots of electricity,but it is low consumed in the spring and winter.Tons of water consumption changes marginally in the meantime.The closer the water inflow of sewage treatment plant water to the designed water inflow was,the lower the tons of water power consumed and the higher the treatment efficiency was.In 2016,the waste water treatment plant consumes 0.243kW·h/m³ of water and electricity,the annual electricity consumption is 1.9?10⁶kW·h,and the unit COD power consumption is3.9 kW·h/kg.The pump and fan are affected by the water load,studying the energy model of the waste water treatment plant and the variables of the selected model,which is needed to consider the operating status of the pumps and fans,operating time,and hydraulic load.?2?Based on the inflow water quality parameters of a sewage treatment plant,the soft-sensing model of NH₃-N is carried out.Analyzing the relevance among NH₃-N?pH?DO?COD,in order to use multivariate linear modeling method to fit NH₃-N in MATLAB software.The results show that NH₃-N is calculated by the regression equation,and there is no significant difference between the measured values and the soft measurements,which indicates that the established regression model is helpful to the management of production and operation in sewage treatment plant.?3?To break the low dimensional variable,this paper break the similar research institute,trying to adopt the high-dimensional variable idea,putting the 33 dimensional variable into the input variable.Input variable include lift pump,reflux pump and blower are included,as well as ORP,MLSS,aerobic pool DO,anoxic pool DO,water inlet flow Q,COD,ammonia nitrogen,etc.The results show that the high-dimensional variable can effectively fit the energy consumption.?4?This article adopts BP neural network to built the prediction model of power consumption,and it is optimized by using algorithm.The result shows that the neural network is suitable for fitting the calculation of electricity consumption,and the average error is reduced to 0.0016 by combining with the genetic algorithm,which can effectively improve the accuracy and reliability of the simulation prediction.