【摘要】 针对水性油墨黏度测量方法存在测量繁琐且无法实现在线检测的问题，利用超声波测量技术设计并搭建了试验装置，建立水性油墨黏度预测模型，实现对水性油墨黏度的无损检测。采用单片机、时间数字转换器、超声波换能器、温度传感器、上位机等搭建了水性油墨黏度超声检测装置，进行实验数据采集，分别构建多元线性回归、BP神经网络以及PSO-BP神经网络预测模型，PSO-BP神经网络模型的预测决定系数为0.935 3，预测标准差为0.037 6，均优于另外两个模型。并进行验证实验，检测装置的测量精度为86.2%，基本满足印刷车间的生产需要，表明所提出的基于超声波的水性油墨黏度检测方法具备一定可行性，可提高生产效率，为液体黏度实时检测系统研发提供一定参考。更多还原

【Abstract】 Aiming to address the issue of cumbersome measurement and the lack of online detection in water-based ink viscosity measurement methods,ultrasonic measurement technology is employed to design and construct an experimental apparatus. The primary objective is to establish a predictive model for water-based ink viscosity,enabling non-destructive viscosity testing. The experimental setup consists of a microcontroller,time-to-digital converter,ultrasonic transducer,temperature sensor,and host computer. Through data collection and analysis,prediction models including multiple linear regression,BP neural network,and PSO-BP neural network are constructed. The PSO-BP neural network model exhibits a superior prediction coefficient of 0.9353 and a lower prediction standard deviation of 0.037 6 compared to other models. Validation experiments demonstrate a measurement accuracy of 86.2%,meeting the requirements of printing workshop production. These findings suggest the feasibility of ultrasonic-based viscosity detection for water-based ink,which can improve production efficiency. Furthermore,this study contributes to the development of real-time liquid viscosity detection systems by providing valuable insights and reference.更多还原