Fault Prediction of Phased Array Radar T/R Components Based on Data Mining -- Method Study

Based on the processing flow and key techniques of data mining, a fault quantity prediction method of phased array radar T/R components is studied. The method mainly consists of two parts: fault data pre-processing and prediction model building, in which the fault data pre-processing mainly performs missing value processing, outlier analysis and data smoothing on T/ R component fault count data to obtain high-quality training data; while the prediction model building is to build a prediction model for T/ R component fault count data and analyse the model performance to meet the prediction requirements. Finally, the effectiveness of the prediction method is verified by using the actual T/ R component failure count data of a certain type of radar.