AMBR工艺畜禽养殖污水处理及BP神经网络水质预测研究

发布时间：2018-03-12 09:05

本文选题：AMBR　切入点：养殖污水　出处：《宁夏大学》2017年硕士论文　论文类型：学位论文

【摘要】：畜禽养殖业规模化发展,造成养殖粪尿和冲洗水急剧增多,生态环境保护形势十分严峻,人工检测水质具有滞后性且耗时长,存在不达标污水偷排、误排等现象,因此,采用折流厌氧反应器+膜生物反应器组合(AMBR)工艺处理养殖污水,分析运行过程中污染物去除特性和效率,并利用BP人工神经网络预测模型,对AMBR工艺处理养殖污水进行模拟,在输入参数溶解氧(DO)、悬浮物污泥浓度(MLSS)、酸碱度(pH)改变的情况下,对出水水质作出动态预测。为畜禽养殖污水高效处理提供行之有效的方法,弥补了人工检测产生的滞后性,也为AMBR在高浓度养殖污水处理领域的技术优化及推广应用提供相关技术参考,降低畜禽污染程度,保护生态文明。主要研究结果如下:(1)AMBR系统对化学需氧量(CODcr)、氨氮(NH3-N)、总氮(TN)、总磷(TP)和悬浮物(SS)的平均去除率分别为94.17%、84.17%、83.37%、41.63%和98.33%,按去除率的大小排列为:SSCODcrNH3-NTNTP。AMBR系统对于进水浓度波动较大的指标(CODcr:1298.29mg/L～3372.42mg/L,NH3-N:79.35mg/L～207.6mg/L,SS:809.9mg/L～32396.4mg/L)处理效果依然稳定良好,体现了 AMBR系统超强的抗冲击能力和自适应调整恢复能力。(2)AMBR系统对CODcr、NH3-N和TN的去除主要依靠生物降解作用,其生物降解贡献率分别在60%～85%、82%～90%和69.96%～90.44%,由于NH3-N和TN分子较小,能透过微孔膜组件随出水排出。系统对NH3-N的去除主要靠硝化菌硝化作用,以及厌氧氨氧化菌将反硝化产物亚硝酸盐作为电子受体,将氨氮转化为氮气实现的。对SS的去除依靠膜系统截留过滤作用,其贡献率平均为61.26%,主要由于AMBR膜组件采用膜孔径为0.1～0.2μm的超滤膜,养殖污水中的胶体、悬浮物SS和高分子有机物等粒径大于0.2μm则均被膜的物理筛滤作用截留。(3)对于TP的去除率明显低于其他水质指标,可能存在的原因有两方面:①系统内部环境被累积的高浓度有机物破坏,聚磷菌的碳源供给受到竞争抑制,气化成磷量减小。②缺氧区DO浓度约为0.45～0.6mg/L,高于厌氧除磷所需溶解氧浓度(DO=0.2mg/L),非聚磷菌繁衍增长抑制聚磷菌反硝化作用,不利于TP的去除。(4)BP神经网络出水CODcr、和NH3-N、TN、TP和SS水质预测的相对误差平均值分别为 0.99%、1.67%、0.08%、0.02%和 0.03%,平均绝对误差率和分别为 1.01%、2%、8.1%、2.12%和4.99%。综上可知,BP神经预测模型对于上述指标都具有良好的适应性和准确性,其中性能最佳的是CODcr。(5)当DO、pH(不超过pH=8)、MLSS增加时有利于出水TN、NH3-N、和SS的去除;运行条件调整到DO=3mg/L,MLSS=16000 mg/L,pH=7时,畜禽养殖污水处理系统处理净化程度最大化,处理效果达到最佳;由于CODcr、NH3-N和TN的去除率在MLSS从10000增加到16000过程中增长缓慢,本着节约污水处理厂投入资金且保证污水处理效果的原则,MLSS控制到10000～12000能够达到CODcrCODcr、NH3-N和TN的最佳去除效果。
[Abstract]:The large-scale development of livestock and poultry breeding industry has resulted in a sharp increase in feces, urine and washing water, the ecological environment protection situation is very severe, the artificial detection of water quality has lag and consuming time, there are some phenomena, such as substandard sewage stealing discharge, wrong discharge, etc., therefore, Culture wastewater was treated with baffled anaerobic reactor membrane bioreactor (MBR) process. The pollutant removal characteristics and efficiency were analyzed. BP artificial neural network model was used to simulate the treatment of aquaculture wastewater by AMBR process. Under the condition of changing the input parameters of do, MLSS, pH and pH of suspended sludge, the dynamic prediction of effluent quality is made, which provides an effective method for efficient treatment of livestock and poultry wastewater, and makes up for the lag produced by manual detection. It also provides relevant technical reference for the technical optimization and application of AMBR in the field of high concentration aquaculture wastewater treatment, and reduces the pollution degree of livestock and poultry. Conservation of ecological civilization. The main results of the study are as follows: the average removal rates of the chemical oxygen demand (COD), NH _ 3-N (NH _ 3-N), total nitrogen (TNN), total phosphorus (TP) and suspended solids (SSs) of the 1 / 1 / AMBR system are 94.1717 / 84.171.37% and 98.33%, respectively, and the removal rates are in the order of: SSCcrNH _ 3-NTNTP.AMBR system for influent concentration. A highly volatile indicator, CODCR: 1298.29 mg / L, 3372.42 mg / L, NH3-N: 79.35 mg / L, 207.6 mg / L, SS: 809.9 mg / L 32396.4 mg / L) treatment is still stable and good. The results show that the removal of NH3-N and TN by AMBR system mainly depends on biodegradation, and the contribution rate of biodegradation is in the range of 600.8585% and 69.96% / 90.440.44, respectively, because the NH3-N and TN molecules are small, and the effect of biodegradation on the removal of NH3-N and TN of the system is mainly dependent on the biodegradation of NH3-N and TN, and the biodegradability of the system is 69.96% and 90.44%, respectively, because of the small NH3-N and TN molecules. The removal of NH3-N by the system mainly depends on nitrification by nitrifying bacteria, and nitrite, the denitrification product, is used as the electron receptor by anaerobic ammonia-oxidizing bacteria. Ammonia nitrogen was converted to nitrogen gas. The removal of SS depended on the retention and filtration of membrane system. The contribution rate of SS was 61.26 on average. The main reason was that the AMBR membrane module adopted ultrafiltration membrane with a membrane pore diameter of 0.1 渭 m and 0.2 渭 m, and the colloid in the aquiculture wastewater. The removal rate of TP was obviously lower than that of other water quality indexes when the suspended solids SS and macromolecule organic particles were larger than 0.2 渭 m in the physical sieve and filtration action of membrane, the removal rate of TP was obviously lower than that of other water quality indexes. There are two possible reasons for this: the internal environment of the system has been destroyed by the accumulation of high concentrations of organic matter, and the supply of carbon sources of phosphorus accumulating bacteria has been inhibited by competition. The concentration of do in anoxic region was about 0.45 ~ 0.6mg / L, which was higher than that of dissolved oxygen concentration (do) 0.2mg / L ~ (-1) for anaerobic phosphorus removal. The proliferation and growth of non-phosphorus accumulating bacteria inhibited denitrification of phosphorus accumulating bacteria. The average relative error of water quality prediction of TP and NH _ 3-N _ N _ N _ N _ N _ N _ T _ N _ TP and SS was 0.99 ~ 1.67 ~ 0.08% and 0.03%, respectively. The average absolute error rate and the average absolute error rate were 1.01 ~ 2% and 8.12% and 4.99%, respectively. In summary, it can be seen that the BP neural prediction model is suitable for the above mentioned fingers, and the average absolute error rate is 0.02% and 0.03% respectively. In summary, we can see that the BP neural prediction model is suitable for the above mentioned fingers. All of them have good adaptability and accuracy. The best performance was CODcr.5) when the pH of DOA (no more than pH 8) was increased, the removal of TNNH _ 3-N and SS was favorable, and the operating conditions were adjusted to 3 mg / L MLSS = 16000 mg 路L ~ (-1) pH = 7:00, the purification degree of livestock and poultry wastewater treatment system was maximized and the treatment effect was the best. Because the removal rate of COD _ (Cr) NH _ 3-N and TN increased slowly in the process of increasing MLSS from 10000 to 16000, the best removal efficiency of COD _ (Cr) _ (Cr) NH _ 3-N and TN could be achieved by using the principle of saving the investment of wastewater treatment plant and ensuring the sewage treatment effect by controlling the COD _ (Cr) NH _ 3-N and TN to 10 000 ~ 12000.
【学位授予单位】：宁夏大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：X713

【参考文献】