Prediction of Converter Valve Cooling Capacity Based on Multidimensional Time Series Data

The cooling capacity of the converter valve is an important index to measure the reliability of the converter valve operation. Accurate prediction of the cooling capacity of the converter valve can make it take corresponding measures in advance to reduce the probability of failure. However, the current prediction methods for the cooling capacity of the converter valve core are all based on the analysis of data in a single time dimension and a single voltage dimension, and there are few studies on multi-dimensional data fusion methods. To this end, this paper adopts a method for predicting the cooling capacity of converter valves based on multi-dimensional time series data. This paper first analyzes the factors affecting the cooling capacity of the converter valve spool under different voltage levels; then analyzes based on multi-dimensional data such as voltage level, current level, cooling state, temperature, etc.; finally uses the least squares support vector machine (LSSVM) method to predict the cooling capacity of the converter valve. The experimental results show that the prediction of the cooling capacity of the converter valve based on the multi-dimensional time series data can increase the accuracy rate to the highest 95.9%.