Turbine Machine Working Condition Short-term Prediction Using Intelligent Analysis of Online Lubricant Monitoring

Aiming at the short-term load forecasting of steam turbines, a method combining the state trend prediction of equipment with oil monitoring technology is proposed. Firstly, the gray correlation analysis is used to screen out the factors that have a great influence on the turbine load. Then the main impact factors are used as input sample data, and the load forecasting model of support vector regression (SVR) is constructed by Python language. Taking a steam turbine as an example, the data from April 5 to Apri1 12, 2014 is used as a training sample, and the data from Apri1 13 to Apri1 14, 2014 is a test sample. The results show that the predicted and measured values of the load forecasting model constructed by SVR are in good agreement. The MSE of the training set and the test set are 0.0235 and 0.0827, respectively. This research method can be used as a short-term method. Load forecasting method.