基于机器学习算法的水煤浆提浓废水絮凝处理加药预测

Prediction of Dosing for Coal-water Slurry Concentrated Wastewater Flocculation Treatment Based on Machine Learning Algorithms

当前水煤浆提浓废水絮凝处理加药控制过程中只是根据现场废水检测仪器仪表的获取结果进行对应的加药,缺乏对加药后水体变化进行预测,加药控制结果不理想。因此本文提出基于机器学习算法的水煤浆提浓废水絮凝处理加药预测模型以及对加药量进行精准预测的加药控制方法。本文选择了机器学习算法中的随机森林(Random Forest)算法模型作为加药控制模型设计的核心,建立给药预测模型,对加药后的结果进行预测,当预测结果不佳时,则对加药参数进行再次修正,最终实现加药量精准控制。为了验证所提加药控制方法的可行性,进行实际应用实验,在陕西神渭管运公司蒲城终端站水煤浆提浓废水处理絮凝车间,对水煤浆提浓废水的指标进行实时监测。实验结果显示,水煤浆提浓废水絮凝处理的各项指标符合要求,证明该方法具有较高的可行性。

Currently,the dosing control process for the flocculation treatment of coal water slurry concentrated wastewater is only based on the results obtained from on-site wastewater detection instruments and meters for corresponding dosing,lacking prediction of water changes after dosing,and the dosing control results are not ideal.Therefore,a machine learning algorithm based dosing prediction model for coal water slurry concentration wastewater flocculation treatment is proposed,and a dosing control method for precise prediction of dosing amount is proposed.The Random forest algorithm model in the machine learning algorithm is selected as the core of the dosing control model design,and the dosing prediction model is established to predict the results after dosing.When the prediction results are poor,the dosing parameters are modified again,and finally the dosing amount is accurately controlled.In order to verify the feasibility of the proposed dosing control method,practical application experiments were conducted.Real time monitoring was conducted on the indicators of coal water slurry concentration wastewater treatment in the flocculation workshop of Pucheng Terminal Station of Shaanxi Shenwei Pipeline Transportation Company.The experimental results showed that the various indicators of coal water slurry concentration wastewater flocculation treatment met the requirements,proving that the method has high feasibility.