ZhangJun,RaoJingang,JianSangui,et al.Prediction of PAC Dosage in Coagulation Process Based on Multimodal Large Model[J].China Water & Wastewater,2026,42(17):20-25.
基于多模态大模型的混凝PAC投加量预测
- Title:
- Prediction of PAC Dosage in Coagulation Process Based on Multimodal Large Model
- 关键词:
- 污水处理厂; 混凝; 聚合氯化铝投加量预测; 时间序列; 多模态大模型
- Keywords:
- wastewater treatment plant; coagulation; prediction of polyaluminium chloride(PAC) dosage; time series; multimodal large model
- 摘要:
- 精准预测混凝过程的聚合氯化铝（PAC）投加量是智慧污水厂高效运行的关键环节。传统方法多依赖经验公式或单一数据源，而如长短期记忆网络（LSTM）等典型的深度学习模型虽然能够捕捉时序动态规律，但普遍存在对多源信息利用不足、对工艺知识表达能力有限的问题，难以满足复杂工况下的预测需求。为此，构建了一种面向智慧污水厂的多模态大模型加药预测方法。该方法在利用污水厂多变量时序数据的同时，引入基于规则生成的文本模态，通过将进水COD、SS、NH3-N、TP、TN等关键指标映射为语义短语，并拼接成自然语言描述。该规则化文本与时序数据在统一特征空间中对齐，实现了跨模态特征融合。在模型方面，以大语言模型LLaMA为基座模型，使用LLaMA编码器对规则化文本进行编码，并与对齐后的时序序列联合建模以实现学习与预测。结果表明，该方法在预测精度上优于传统统计方法和单模态深度学习模型，能够为智慧污水厂提供高效的PAC投加量预测工具。
- Abstract:
- Accurate prediction of polyaluminium chloride (PAC) dosage during the coagulation process is a key step for the efficient operation of smart wastewater treatment plants（WWTPs）. Traditional methods mostly rely on empirical formulas or single data sources, and typical deep learning models such as long short-term memory (LSTM) can capture temporal dynamics but often make insufficient use of multi-source information and have limited capability to express process knowledge, making it difficult to meet prediction requirements under complex operating conditions. To address these issues, a multimodal large model-based dosing prediction method for WWTPs is proposed. The method utilizes multivariable time-series data from the plant and introduces a rule-based text modality: key influent indicators such as COD, SS, NH3-N, TP and TN are mapped into semantic phrases and concatenated into natural language descriptions. The rule-based text and time-series data are aligned in a unified feature space to achieve cross-modal feature fusion. In terms of the model architecture, large language model Meta AI (LLaMA) is utilized as the foundation model. Specifically, the LLaMA encoder encodes the regularized text, which is then jointly modeled with aligned temporal sequences to facilitate learning and prediction. The results showed that this method outperformed traditional statistical and single-modal deep learning models in prediction accuracy, providing an efficient PAC dosing prediction tool for smart WWTPs.
相似文献/References:
[1]王 亮.马来西亚Pantai地埋式污水厂环网供配电结构设计[J].中国给水排水,2018,34(22):63.
WANG Liang.Power Supply and Distribution Structure Design of Ring Network for Pantai Underground Wastewater Treatment Plantin Malaysia[J].China Water & Wastewater,2018,34(17):63.
[2]侯晓庆,邓 磊,高海英,等.MBR工艺在神定河污水处理厂升级改造工程中的应用[J].中国给水排水,2018,34(22):66.
HOU Xiao-qing,DENG Lei,GAO Hai-ying,et al.Application of MBR Process in the Upgrading and Reconstruction Project of Shending River WastewaterTreatment Plant[J].China Water & Wastewater,2018,34(17):66.
[3]邱明海.北京市垡头污水处理厂改扩建工程设计技术方案[J].中国给水排水,2018,34(20):13.
QIU Ming hai.Reconstruction and Expansion Design Technical Plan of Beijing Fatou Wastewater Treatment Plant[J].China Water & Wastewater,2018,34(17):13.
[4]郝二成,郭毅,刘伟岩,等.基于数学模拟的污水厂运行分析——建模与体检[J].中国给水排水,2020,36(15):23.
HAO Er-cheng,GUO Yi,LIU Wei-yan,et al.Operation Analysis of Wastewater Treatment Plant Based on Mathematical Simulation: Modeling and Examination[J].China Water & Wastewater,2020,36(17):23.
[5]张月,王阳,张宏伟,等.阳泉市污水处理二期工程BARDENPHO工艺设计和运行[J].中国给水排水,2020,36(16):64.
ZHANG Yue,WANG Yang,ZHANG Hong-wei,et al.Design and Operation of BARDENPHO Process in Phase Ⅱ Project of Yangquan Wastewater Treatment Plant[J].China Water & Wastewater,2020,36(17):64.
[6]祝新军,蔡芝斌,姚斌,等.绍兴污水处理厂气浮设备的优化改造[J].中国给水排水,2020,36(16):101.
ZHU Xin-jun,CAI Zhi-bin,YAO Bin,et al.Optimization and Modification of Air Floatation Equipment in Shaoxing Wastewater Treatment Plant[J].China Water & Wastewater,2020,36(17):101.
[7]王文明,杨淇椋,蔡依廷,等.MSBR工艺在高排放标准污水处理厂的应用[J].中国给水排水,2020,36(16):111.
WANG Wen-ming,YANG Qi-liang,CAI Yi-ting,et al.Application of MSBR Process in Wastewater Treatment Plant with Stringent Discharge Standard[J].China Water & Wastewater,2020,36(17):111.
[8]郝二成,郭毅,刘伟岩,等.基于数学模拟的污水厂运行分析——控制与优化[J].中国给水排水,2020,36(17):23.
HAO Er-cheng,GUO Yi,LIU Wei-yan,et al.Operation Analysis of Wastewater Treatment Plant Based on Mathematical Simulation: Control and Optimization[J].China Water & Wastewater,2020,36(17):23.
[9]王阳,张月,王晓康,等.高排放标准下的改良AAO+深度处理工程案例[J].中国给水排水,2020,36(18):56.
WANG Yang,ZHANG Yue,WANG Xiao-kang,et al.Project Case of Modified AAO and Advanced Treatment Process under High Emission Standards[J].China Water & Wastewater,2020,36(17):56.
[10]郑枫,慕杨,孙逊.MBR工艺用于山东省某污水处理厂扩建工程[J].中国给水排水,2020,36(18):81.
ZHENG Feng,MU Yang,SUN Xun.MBR Process Used in Expansion Project of a Sewage Treatment Plant in Shandong Province[J].China Water & Wastewater,2020,36(17):81.