Abnormal data cleaning in thermal power plant based on self-organizing maps

This paper constructs a self-organizing maps (SOM) neural network model for the anomaly data cleaning in thermal power plant detection. The test data is trained 2000 times so that the vector of each weight is located at the center of the input vector cluster, and the 6*6 competitive network is constructed. The network classifies or eliminates the screening of data, and obtains a healthy sample library that can be used to predict the running state of the machine in the future, achieving a good data cleaning effect.