Smoldering has broad prospects for application in the treatment of sewage sludge with high moisture content, but it faces the problem of high CO/NOx emission concentrations. To improve smoldering velocity and reduce emission concentrations of gas pollutants, intelligent control and refined treatment of sewage sludge smoldering need to be achieved. In this paper, the digital twin-driven sewage sludge smoldering treatment system is proposed, and the overall framework, operational process and key technologies of the system are described in detail. A digital twin system based on the back propagation neural network model is constructed, which achieves the accurate prediction of the variation trends and average values of CO/NOx emission concentrations as well as smoldering temperature and velocity. Nondominated Sorting Genetic Algorithm II is used for multiobjective optimization, providing effective control strategies for sewage sludge with distinctive characteristics. Smoldering features, emission concentrations of gas pollutions and equipment operating status are visualized using WebGL technology. Results show the maximum increase in smoldering velocity is 49 %, whilst CO can be reduced by 8–60 % and the maximum reduction in NOx is 51 %. This system can assist in applications such as monitoring state of sewage sludge smoldering, timely warnings and intelligent control.

中文翻译：

---

基于反向传播神经网络的污水污泥阴燃过程和 CO/NOx 排放数字孪生技术：一项室内实验研究


阴燃在处理高含水率的污水污泥方面具有广阔的应用前景，但面临 CO/NOx 排放浓度高的问题。为了提高阴燃速度，降低气体污染物的排放浓度，需要实现污水污泥阴燃的智能控制和精细化处理。本文提出了数字孪生驱动的污水污泥阴燃处理系统，并详细介绍了该系统的整体框架、运行流程和关键技术。构建了基于反向传播神经网络模型的数字孪生系统，实现了对 CO/NOx 排放浓度变化趋势和平均值以及阴燃温度和速度的准确预测。非支配排序遗传算法 II 用于多目标优化，为具有独特特征的污水污泥提供有效的控制策略。使用 WebGL 技术将阴燃特征、气体污染的排放浓度和设备运行状态可视化。结果表明，阴燃速度的最大增加为 49%，而 CO 可以减少 8-60%，NOx 的最大减少为 51%。该系统可协助监测污水污泥阴燃状态、及时预警和智能控制等应用。