基于人工神经网络的联合厌氧消化产气预测模型

有机固体废物厌氧消化为非均相体系，预警调控对系统的稳定运行非常重要。为了准确预测负荷跨越式提升的联合厌氧消化体系的日产气量，利用灰色关联分析方法量化了厌氧消化体系中常见因素间的关联度，评估了厌氧消化体系的复杂程度，构建了反向传播（back propagation，BP）神经网络对日产气量进行预测。结果表明，各指标间的灰色关联度均大于0.7，表明厌氧消化体系中各参数之间关系的高度的关联性和复杂性。利用试验获取的148组数据作为BP神经网络的训练集和测试集，构建的BP神经网络精度较高，对148组样本数据的平均预测准确率达到99.17%，对日产气量具有较好的预测能力。

A predictive model for combined anaerobic digestion gas production based on an artificial neural network

Anaerobic digestion of organic solid waste is a heterogeneous system, and early warning and control are very important factors in ensuring stable operation of the system. In order to accurately predict the daily gas production of a combined anaerobic digestion system with a large increase in load, the gray correlation analysis method was used to quantify the correlation between common factors in the anaerobic digestion system, and to evaluate the complexity of the anaerobic digestion system. A back propagation (BP) neural network was constructed to predict daily gas production. The results show that the gray correlation degree between each index is greater than 0.7, indicating a high degree of correlation and the complexity of the relationship between the parameters in the anaerobic digestion system. By using the 148 sets of data obtained in the experiments as the training set and testing set for the BP neural network, the resulting BP neural network has high accuracy, with the average prediction accuracy for the 148 sets of sample data being 99.17%, confirming its excellent ability to predict daily gas production.