An online identification method of flooding state of packed tower based on deep learning

Abstract

An online identification method for the flooding state of packed towers based on deep learning. First, run the packed tower equipment to obtain video data and label data during its operation; then preprocess the data, including video capture and image timing processing, And use the processed data and label data to make a data set, divide the data set into three parts: training set, verification set and test set; then establish a convolutional long-term short-term memory neural network model and use adaptive moment estimation to train the model; finally use The test set evaluates whether the trained network model meets the recognition requirements. Compared with the traditional method, the present invention is more accurate and intuitive, uses deep learning technology to process and analyze the video data during the operation of the packed tower, and finally realizes the online identification of its operating state.

Description

Technical field [0001] The present invention relates to a method identification method based on video analysis, and the present invention belongs to the field of filler tower, and the processing and analysis method involving the video data of the filler tower. Background technique [0002] The filler tower is an important separation equipment. Due to its characteristics of its large flux, low pressure drop, high efficiency, there is a wide range of applications in various fields such as chemical, environmental protection, food and medicine. It is the key to determining production benefits and product quality. . Filler investment generally accounts for 20% of total investment, and its energy consumption accounts for 50% of all unit operations. When the filler is in a normal operating state, the efficiency increases with the increase of operating speed, and the energy consumption and carbon emissions are also reduced; when it is in the liquid, the efficiency will fall rapidly, and ...

AI Technical Summary

Problems solved by technology

The visual observation method is only effective for transparent towers, and there is a delay phenomenon. When liquid flooding is observed, the damage may have occurred, and there are uncertain subjective factors through human visual observation; the liquid holdup measurement method needs to stop the gas-liquid two-phase It is not suitable for on-line monitoring in actual industrial production; the monitoring method of key variables in the tower is an indirect method, which cannot directly reflect the occurrence of liquid flooding, and there are certain errors and lags

Images