厌氧消化论文不同种类生物质废物混合厌氧消化产气预测模型的研究.docVIP

厌氧消化论文：不同种类生物质废物混合厌氧消化产气预测模型的研究 【中文摘要】随着我国城市发展速度的加快,人民生活水平的逐步提高,固体废弃物的产生量也在高速地增长,特别是生物质废弃物所占的比重很大。而如果对生物质废弃物进行回收利用,不仅缓解了环境压力,也能够节约能源。本文对餐厨垃圾、畜禽养殖废物等生物质废弃物与城市污水厂剩余污泥混合,在不同试验参数条件下进行混合厌氧消化试验；以产生的大量实验室数据为原始数据,基于多元回归原理和BP人工神经网络原理,对其建立预测模型。在基于肥肉基质和剩余污泥混合厌氧试验的建模中,两个阶段多元回归模型的预测平均准确率分别为75.69%和79.29%；BP神经网络模型的预测平均准确率为79.05%。总体看来,两种模型的预测准确率相近,都有较好的预测结果,BP模型的结果较优。在基于高有机负荷的餐厨垃圾填料、餐厨垃圾与消化污泥混合厌氧试验的建模中,两个阶段多元回归模型的预测平均准确率分别为60.05%和22.21%；BP神经网络模型的预测平均准确率为72.40%。BP模型的结果明显优于多元回归模型。在分别以黄豆和芹菜为基质,和剩余污泥混合的控制pH值的厌氧消化过程的试验数据的BP建模,预测平均准确率分别为52.19%、69.13%；基于低有机负荷的餐厨垃圾填料、餐厨垃圾与消化污泥混合厌氧试验的建模中,预测平均准确率为86.30%。可以看出,BP模型的预测准确率一直处在一个较高的水平,并且比较稳定。对比所建立的两种模型的预测结果,BP人工神经网络模型用于混合厌氧消化的产气预测是可行的,并且优于多元回归模型的预测。本文建立的基于BP人工神经网络的混合厌氧消化产气预测模型为今后的生产实践提供了一种具有操作性强,效率高的预测方法。 【英文摘要】With the accelerated pace of urban development in China, the gradual improvement of living standards, solid waste generation is also growing fast, especially in the proportion of biomass waste. Biomass waste recycling relieves pressure on the environment and conserves energy.In this paper, for the digestion of different biomass gas production forecast, established multiple regression model and BP-ANN model.Based on fat matrix and mixed anaerobic sludge experiment, established multiple regression model,the average prediction accuracy is75.69%,79.29%; established BP-ANN models the average prediction accuracy is 79.05%. Prediction accuracy of two models are similar, the results of prediction is good.Based on high organic loading mixed food waste and anaerobic sludge, established multiple regression model,the average prediction accuracy is60.05%,22.21%; established BP-ANN models the average prediction accuracy is 72.40%. Obviously, BP-ANN models are better than multiple regression models.Based on soybean matrix and mixed anaerobic sludge experiment, established BP-ANN model, the average prediction accuracy is52.19%; Based on celery matrix and mixed anaerobic sludge experiment, established BP-ANN model, the average prediction accuracy is69.13