【摘要】 水文水井钻探是开发地下水资源的一项重要技术手段，钻探过程中产生的泥浆排放会对生态环境造成一定程度的破坏。为对泥浆进行有效净化并循环利用，开展了泥浆在线监测方案的研究，首先利用管道黏度计对泥浆密度、表观黏度、塑性黏度、动切力进行实时监测，然后通过预先训练基于泥浆密度、流变性参数和含砂量的神经网络模型，预测经过固相控制之后的泥浆中的含砂量。结果表明：通过管道黏度计监测的泥浆密度、表观黏度、塑性黏度、动切力的平均误差分别为0.2%、1.7%、3.4%、3.7%；通过神经网络模型预测的含砂量的平均误差为15.9%，预测模型不依赖模型训练中使用的泥浆配方，能够适用于相同体系的其他泥浆配方，表明预测模型具有一定的泛化性。该泥浆在线监测系统能够满足绿色勘查的要求，适合水文地质及水井钻探的现场应用。更多还原

【Abstract】 Hydrological well drilling is an important technical means for developing groundwater resources, and the discharge of mud generated during the drilling process can cause a certain degree of damage to the ecological environment. In order to purify the mud and recycle it effectively, the online monitoring scheme for the mud is studied. Firstly, the pipeline viscometer was used to measure the mud density, apparent viscosity, plastic viscosity, dynamic shear force in real-time. Then, through pre-trained neural network models based on mud density, rheology parameters and sand content, the sand content in the mud after solid phase control was predicted.The test results show that the average measurement error of mud density, apparent viscosity, plastic viscosity, and dynamic shear force through the monitoring of pipeline viscometer is 0. 2%, 1. 7%, 3. 4%, and 3. 7% respectively. The average prediction error of sand content through the neural network model is 15. 9%. The prediction model does not rely on the mud formulations used for model training and can be applied to any other mud formulation with the same system, which shows that the model has a certain degree of generalization. This online mud monitoring system meets the requirements of green exploration and is suitable for on-site applications in hydrogeology and water well drilling.更多还原