Visual recognition processing of power monitoring data based on big data computing

The operation control of power units is usually carried out by the control personnel with the help of distributed control system. Although it can ensure the safety of unit operation and meet the requirements of power generation loads, the economy of unit operation and the accuracy of control process still need to be further improved. Therefore, by designing multiple view mapping and association, it provides interactive visualization support for relevant experts in the key links of model establishment and evaluation. In the exploration stage of estimating model parameter, the user can get the delay range by line chart and focus + context technology, while in the model screening stage, the user can provide the combination of screening views, selecting the model by its accuracy on different data sets, and finding the model anomalies by the model structure view. Besides, in the model evaluation stage, the user can get the delay range by predicting line chart and model accuracy radar chart. In addition, the method in this paper keeps between 4.2–7.2 in most distributions, and the maximum value is 18. The time series trend of the data segment is consistent, and the absolute value of the weight coefficient is basically 0 after being superimposed, which has great advantages compared with other methods, proving the effective results of the research content in this paper.