【作者】 张璐瑶；

【导师】 朱小磊； 朱守诚；

【作者基本信息】 合肥大学 ， 环境工程（专业学位）， 2024， 硕士

【摘要】 由于湖泊富营养化导致蓝藻爆发是最常见的环境灾害。蓝藻在水体表面堆积,不光损害水质,导致饮用水危机,同时危害人类和水体动植物安全。对蓝藻资源化处理,不仅解决了环境污染问题,还能变废为宝产生经济效益。热化学是蓝藻资源化处理常见的方法,但是传统的热化学方法如燃烧,虽然可以产生热能,用于供热和发电,但燃烧的热效率有限,且产生的氮氧化物,硫氧化物等会对空气质量产生影响。使用热解技术处理蓝藻,可以将其转化为生物炭,生物油,生物气等可再生能源,不仅解决了环境污染问题,还实现资源的高效利用。论文针对蓝藻热解研究不足,以巢湖蓝藻为例进行了热解研究。首先通过热重实验比较分析了蓝藻在N₂与CO₂下热解行为的差异,运用补偿效应、主图法分、四种无模型方法计算了热解动力学参数,人工神经网络对蓝藻TG曲线进行建模预测,多重响应联合优化分析最佳热解条件。计算结果显示N₂和CO₂下蓝藻的活化能估计范围分别为98.7-205.6和95.1-176.5 k Jmol^-1,指前因子范围分别为2.643E+06-7.379E+16和9.770E+05-6.885E+13,反应机制分别为F4、F3反应顺序模型。人工神经网络模型预测的蓝藻TG曲线拟合度较高,均大于0.99。蓝藻最佳的热解条件为温度669.4℃,加热速率10℃/min和N₂氛围。其次研究了基于CaO/HY两种催化剂下蓝藻与PE共热解行为,计算了热解活化能与反应机制,热重红外联用和热解气相色谱质谱联用分析了催化共热解挥发性产物。结果显示加入HY分子筛后,蓝藻与PE共热解的活化能降低了23.1%,而加入CaO活化能只降低了5.7%;反应机制中,加入HY分子筛后由共热解开始的n级反应顺序模型转为扩散模型,CaO则是转变为随机成核模型。热重红外联用显示CaO在脱氧方面表现突出,而HY有利于去除氮化物。热解气相色谱质谱联用表明HY更有利于芳香烃的选择性,而CaO有利于芳烃向脂肪烃的转化。最后从65篇研究文章中收集了来自在无氧条件下热解的多种类型的微藻生物质数据,总共438组数据,4380个数据点。以工业分析、元素分析、热解转化率为输入变量,输出变量为活化能。构建的三层人工神经网络精度良好（R²=0.835,RMSE=30.311）。随机森林算法特征重要性程度占比显示转化率为最重要的因素,且模型的精度较高（R²=0.938,RMSE=22.008）。多元线性回归分析了C元素含量反向影响活化能大小,而固定碳和转化率正向影响活化能的大小（R²=0.209,RMSE=59.142）。更多还原

【Abstract】 Due to eutrophication of lakes,cyanobacterial blooms are the most common environmental disaster.Cyanobacteria accumulate on the surface of water bodies,impairing water quality,leading to drinking water crises,and posing threats to human and aquatic safety.Utilizing cyanobacterial biomass for resource recovery not only addresses environmental pollution but also generates economic benefits by turning waste into valuable resources.Thermochemical treatment is a common approach for cyanobacterial resource recovery.However,traditional thermochemical methods such as combustion,while capable of producing heat energy for heating and power generation,suffer from limited thermal efficiency and emit nitrogen oxides,sulfur oxides,and other pollutants that impact air quality.Utilizing pyrolysis technology for cyanobacterial treatment can convert cyanobacteria into biochar,bio-oil,biogas,and other renewable energy sources,thereby not only addressing environmental pollution but also achieving efficient resource utilization.This thesis addresses the insufficient research on cyanobacterial pyrolysis and conducts pyrolysis studies using cyanobacteria from Lake Chaohu as an example.Firstly,the differential pyrolysis behavior of cyanobacteria under nitrogen and carbon dioxide atmospheres was comparatively analyzed via thermogravimetric analysis.Kinetic parameters of pyrolysis were computed utilizing compensation effect,master plot method,and four model-free approaches.Artificial neural network was employed for modeling and predicting the TG curve of cyanobacteria,followed by a multi-response optimization analysis to determine the optimal pyrolysis conditions.The calculated results indicated that the activation energy ranges for cyanobacteria under N2 and CO2 atmospheres were 98.7-205.6 and 95.1-176.5 k Jmol-1,respectively,with pre-exponential factor ranges of 2.643E+06-7.379E+16 and 9.770E+05-6.885E+13,and reaction mechanisms were identified as F4 and F3 reaction order models,respectively.The neural network model exhibited high fitting accuracy for predicting cyanobacteria TG curves,all exceeding 0.99.The optimal pyrolysis conditions for cyanobacteria were determined to be a temperature of 669.4℃,a heating rate of 10℃/min,and a nitrogen atmosphere.(2)Furthermore,the co-pyrolysis behavior of cyanobacteria and polyethylene(PE)under the catalysis of two catalysts,CaO/HY,was investigated.The activation energy and reaction mechanism of pyrolysis were calculated.Thermal gravimetric analysis coupled with infrared spectroscopy(TGA-IR)and pyrolysis gas chromatography-mass spectrometry(Py-GC/MS)were employed to analyze the volatile products during catalytic co-pyrolysis.The results revealed that the addition of HY zeolite reduced the activation energy of cyanobacteria and PE co-pyrolysis by 23.1%,while the addition of CaO only reduced it by 5.7%.Regarding the reaction mechanism,the addition of HY zeolite shifted the reaction sequence model from an n-step reaction initiated by co-pyrolysis to a diffusion model,whereas CaO shifted to a random nucleation model.TGA-IR analysis demonstrated the outstanding performance of CaO in deoxygenation,while HY facilitated the removal of nitrogen compounds.Py-GC/MS analysis indicated that HY was more conducive to the selectivity of aromatic hydrocarbons,whereas CaO favored the conversion of aromatic hydrocarbons to aliphatic hydrocarbons.(3)From a pool of 65 research articles,data on various types of microalgae biomass under anaerobic conditions were collected,resulting in a total of 438 data sets comprising 4380 data points.Industrial analysis,elemental analysis,and pyrolysis conversion rates were employed as input variables,while activation energy served as the output variable.A three-layer artificial neural network was constructed,demonstrating good accuracy(R~2=0.835,RMSE=30.311).The random forest algorithm indicated that conversion rate was the most crucial factor based on feature importance analysis,with the model exhibiting high accuracy(R~2=0.938,RMSE=22.008).Multiple linear regression analysis revealed that carbon(C)content inversely influenced the magnitude of activation energy,while fixed carbon and conversion rate had a positive impact on activation energy(R~2=0.209,RMSE=59.142).更多还原