Research On Dynamic Monitoring Method Of Rare Earth Element Content Based On Time Series Characteristics

China is a country worthy of the name of rare earth.The country has been working hard for the efficient separation and extraction of rare earth metal resources.However,most rare earth companies currently have difficulty in real-time control of the component content distribution in the rare earth extraction and separation process,which leads to the inability to adjust the process parameters in time when the component content distribution deviates from the optimal working conditions and control the quality of export products.Aiming at the current situation that it is difficult to monitor the component content in the rare earth extraction process in real time,and the current component content detection methods are time-consuming and memory-consuming,a dynamic monitoring system for the content of rare earth element components based on time series characteristics is designed.The main research contents are as follows:1.A dynamic monitoring method of element content based on time sequence characteristics is proposed.The image acquisition device is used to obtain the time series image of the extraction tank solution,and the principal component analysis method is used to extract the time sequence characteristics of the image in the mixed color space.Combined with the production index,the time series characteristic threshold of condition judgment is obtained by the least square method;when the working condition is in the non-optimal state,the image retrieval algorithm based on the HSV color histogram feature is used to obtain the component content value.2.Aiming at the low prediction accuracy of the least squares method and the image retrieval algorithm based on HSV color histogram feature,the least squares support vector machine(LSSVM)classifier based on whale optimization algorithm(WOA)and the HSV color space hybrid feature retrieval algorithm are introduced to improve the monitoring system,and a dynamic monitoring model of element component content based on WOA-LSSVM is proposed.Through the praseodymium/neodymium extraction tank mixed solution test,and compared with the least squares curve fitting,support vector machine(SVM),least squares support vector machine based on particle swarm optimization(PSO-LSSVM)working condition classification model,the experimental results show that the prediction performance of the model is the best,and the model improvement effect is good.3.Based on the image acquisition device developed by the laboratory and the method proposed in this thesis,a dynamic monitoring system of rare earth elements content based on time series characteristics and a friendly man-machine interface are developed with MATLAB GUI as the software platform.The experimental results show that the developed system has high prediction accuracy and strong real-time performance.The simulation experiments and system test results show that the dynamic monitoring system of rare earth element component content based on time series characteristics,which combines WOA-LSSVM classifier and image retrieval algorithm based on mixed features,can meet the requirements of on-site detection of rare earth extraction process and realize the dynamic monitoring of component content.