Analysis And State Prediction Of Biological Nitrogen Removal Process In Piggery Wastewater Microaerobic Treatment System

Manure-free piggery wastewater is a kind of organic wastewater with high am-monia nitrogen（NH₄⁺-N）and low carbon nitrogen ratio（C/N）.Using the traditional nitrification-denitrification process to remove nitrogen not only consumes high energy,but also needs to supplement a large amount of organics,which will emit more carbon dioxide.The microaerobic system can realize the symbiosis of microorganisms with different oxygen utilization phenotypes（anaerobic,aerobic,etc.）and nutritional types（chemoheterotrophic,chemoautotrophic,etc.）.It not only has the advantages of low energy consumption and less excess sludge discharge,but also can achieve efficient si-multaneous carbon and nitrogen removal through the coordinated metabolism of various functional flora.Based on the start-up and regulation operation of the internal aeration microaerobic reactor（IAMR）,this study investigated and analyzed the efficiency and main influencing factors of the IAMR in treating manure-free piggery wastewater.Based on the construction and analysis of a mathematical model,the reaction kinetics characteristics,quantitative nitrogen metabolism network and the operating state prediction of the IAMR were studied.This research can provide theoretical guidance and technical support for the optimal control and practical engineering application of microaerobic treatment.With activated sludge from the secondary sedimentation tank of the municipal sewage treatment plant as the inoculum,under the conditions of HRT 8 h,DO 0.30 mg·L^-1and 32^°C,the IAMR can be started up and run stably within 140 days.When the influent COD,NH₄⁺-N and total nitrogen（TN）were about 255.0,237.9 and 260.3 mg·L^-1,respectively,the removal rates of IAMR were 77.6%,94.6%and 82.6%,and the effluent concentration were only 57.3,13.1 and 58.2 mg·L^-1,which fully meet the discharge requirements of"Pollutant Discharge Standard for Livestock and Poultry Breeding Industry"（GB 18596-2001）.Compared with the previously developed upflow microaerobic activated sludge bed reactor（UMSR）,IAMR can save reflux energy consumption and 35%of aeration energy consumption.The results of regulating the influent C/N,sludge load and oxygen supply of IAMR showed that the nitrogen removal efficiency of the microaerobic system was not affected by the influent C/N;the COD sludge removal load should be lower than200 mg COD·g MLVSS^-1·d^-1,otherwise easy to induce filamentous bulking;the optimal sludge retention time（SRT）should be 42-62 d for the stable and efficient operation of the system.The NO₃^--N accumulation in the IAMT had a strong correlation with the oxygen supply,and the correlation coefficient（pearson coefficient）was as high as0.93.Therefore,by controlling the oxygen supply,the issue of nitrogen removal efficiency decreased with NO₃^--Naccumulation can be avoided.The research results showed that the oxygen supply of the IAMR was controlled to be less than 4243.2 mg·L^-1·d^-1,the NO₃^--N accumulation rate of the system can stable below 14.0%.The results of metagenomic sequencing and characteristic substrate metabolism showed that,there were nitrification-denitrification,partial nitrification-denitrification,partial nitrification-Anammox and partial denitrification-Anammox in the microaerobic activated sludge system.These biological denitrification pathways were intertwined and constitute the complex nitrogen metabolism network in the microaerobic activated sludge system.Kinetic-based mathematical model showed that simultaneous partial nitrification-Anammox and partial denitrification-Anammox were the most important mechanisms in the microaerobic activated sludge system,and the contribution to TN removal reached82.7%-87.9%.Partial denitrification was an important NO₃^--N reduced pathway in the IAMR,and the its contribution to the total NO₃^--N reduction is as high as 84.6%.To reasonably control the excess sludge discharge and sludge load of the microaerobic activated sludge system,a recurrent-gated neural network model for predicting the MLVSS at the end of the sludge discharge cycle was established.The cross-check results showed that the prediction error of the model was only about 4.9%.On the other hand,multivariate statistics,joint probability distribution and correlation analysis showed that COD removal load,NH₄⁺-N removal load and NO₃^--N accumulation load all had strong correlation with the oxygen supply of the system.And the appropriate oxygen supply can control the NO₃^--N accumulation at a low level,which plays an important role in ensuring system’s TN removal efficiency.Aiming at the problem of NO₃^--N accumulation,this paper further established a deep learning model combined with a recurrent-gated neural network and a fully connected neural network to predict it.The model can accurately predict effluent NO₃^--N of the IAMR in the next 5 days,and the prediction error was only about 10.1%.The deep learning model proposed and established in this study can provide technical guidance and mathematical prediction tools for the load design of the microaerobic biological treatment system,the control of excess sludge discharge and bad NO₃^--N accumulation.To sum up,based on the start-up and regulation operation of IAMR,this paper evaluated the efficiency of the microaerobic activated sludge system in treating manure-free piggery wastewater,analyzed the influence of the main factors,and clarified the reaction kinetics,nitrogen metabolism network and main biological nitrogen removal mechanisms of the microaerobic treatment system.Moreover,neural network prediction models for predicting biomass growth and NO₃^--N accumulation were constructed,and an operation control strategy with sludge load and oxygen supply was also determined.This study can provide theoretical guidance and technical support for the optimal control and engineering application of microaerobic biological treatment for manure-free piggery wastewater.